The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
FOREWORD
-On October 29 and 30, 1979, the Federal Reserve Bank of St. Louis
and the Center for the Study of American Business at Washington University co-sponsored their fourth annual conference.
This volume presents
the papers and comments delivered at the conference, entitled
11
Stabil i za ti on Po 1i ci es:
Lessons from the ' ?Os and Implications for
the '80s."
The conference was divided into three sessions.
The first of
these considered recent developments in the theory of stabilization
policy and empirical evidence on the effects of stabilization policies.
The second session focused on evaluations of the monetary and fiscal
policies pursued in the 1970s.
The third and final session discussed
international aspects of stabilization policies.
This foreword pre-
sents summaries of the three sessions.
Stabilization Policies:
Theoretical and Empirical Issues
In his paper "Recent Developments in the Theory of Stabilization
Policy," John Taylor focuses on current theoretical work on the response
of output and employment to changes in aggregate demand.
guishes between two approaches:
He distin-
information-based theories in which
uncertainties about economy-wide disturbances are emphasized, and
contract-based theories in which temporary riqidities in wages and
prices are emphasized.
The former set of theories, combined with
rational expectations, are the foundation of the "new classical
microeconomics."
In these models only unanticipated disturbances
affect real variables, and systematic policy has no effect on real
variables.
The contract models, on the other hand, allow for temporary
rigidities in wages and prices and therefore yield more traditional
conclusions about the short-run response of output and employment
to demand disturbances and policy actions.
Taylor contrasts the implications of these two approaches for two
important issues of stabilization policy:
the possibility of improving
economic performance of output and employment through systematic variation in policy instruments, and the cumulative output loss associated
with anti-inflationary monetary restraint.
The information-based
models suggest an absence of any gains associated with policy activism
and an ability to decelerate inflation without a prolonged or serious
rise in unemployment.
The contract-based models suggest that there may
be gains to policy activism and that there may be sizable costs in
terms of foregone output associated with policies aimed at reducing
inflation.
In his comments on the Taylor paper, Hyman P. Minsky rejects both
the new classical microeconomics and other theories based on "neoclassical" economics as meaningful frameworks for understanding the
role and effects of stabilization policies.
models:
Minsky believes that these
(l) lack the potential for economic instability that makes
policy actions potentially useful; (2) ignore important developments,
beginning in the mid-60s, that radically changed the environment in
which stabilization policy must operate; (3) abstract from essential
aspects of economic institutions, particularly the evolution of the
financial system and of financial practices which have made the economy
increasingly susceptible to financial instability with the accompanying
threat of a serious debt-deflation process.
ii
The second paper in the first session, "Empirical Evidence on the
Effects of Stabilization Policies" by Laurence H. Meyer and Robert H.
Rasche, begins with a survey of monetary and fiscal multipliers.
These
are examined both across various large scale macroeconometric models
and simple reduced-forms, and over time, to assess the degree of consensus and the nature of the evolution in policy multipliers as the
various macroeconometric models have been refined.
The authors give
special attention to the difference in estimated fiscal policy multipliers between the large scale income-expenditure econometric models,
on the one hand, and the St. Louis reduced-form equation on the other
hand.
Meyer and Rasche then develop the implications of both the largescale income-expenditure models and smaller monetarist models for the
two issues highlighted in Taylor's presentation:
the cumulative output
loss associated with anti-inflation policies and the gains from policy
activism.
They contrast the large cumulative output losses implicit
in both conventional estimated Phillips curve equations and monetarist
models with the implications of rational expectation macro models.
Meyer and Rasche note, however, the importance of balancing the gains
from reducing inflation against the transitional costs associated with
reducing inflation.
They conclude with a survey of empirical evidence
on the gains to policy activism, based on model simulations which
compare the simulated performance of the economy under fixed rules,
ad hoc rules with feedback, and optimal control.
In his comments on the Meyer and Rasche paper, Neil Wallace
rejects as useless any results based on the current generation of
large-scale econometric models and reduced forms.
ii;
According to Wallace,
these models are not "coherent'' in the sense that their conclusions are
not derived from a mutually consistent and defensible set of assumptions.
However, he admits that the same criticism can also be applied
to almost all the recent rational expectation macro models.
The
important contribution of these models, according to Wallace, is not so
much the policy ineffectiveness conclusion which has attracted so much
attention, but the demonstration that the assumptions made about how
economic agents forecast future values of variables have great influence
on the response of real variables to macroeconomic policies.
Stabilization Policies:
Critique of the '?Os and Preview of the '80s
The second session focused on evaluations of the monetary and
fiscal policies pursued in the '?Os.
In "The Case for Gradualism in
Policies to Reduce Inflation," Allan H. Meltzer rejects as myth the
view that the current inflation has its roots in the Vietnam War era
deficits.
Instead, Meltzer states that the proximate source of the
current inflation is the monetary policy of the early 1960s, and that
inflation persists because monetary policy continues to sustain anticipations of future inflation.
Meltzer then develops the rationale for a policy of "gradualism'' -pre-announced, gradual, sustained declines in the rate of growth of
money.
Meltzer emphasizes the importance of conducting monetary policy
in a way that permits individuals to quickly recognize permanent shifts
in the rate of monetary growth.
If monetary growth is volatile, indivi-
duals have difficulty in inferring from observed money supply figures
what direction the Federal Reserve is likely to take in the future.
This situation results in a slow adjustment of expectations about
iv
future monetary growth and inflation to a permanent decline in the rate
of monetary growth -- and, as a consequence, a serious cumulative output loss.
By announcing its target and reducing the variance of actual
monetary growth around its target, the Fed promotes more rapid revision
in inflation expectations and minimizes the cumulative output loss associated with anti-inflation policy.
In "Federal Budget Policies of the 1970s:
Some Lessons for the
1980s," Michael E. Levy is critical of monetarist explanations of the
persistent inflation of the last decade and a half.
While recognizing
the important role of monetary change in the inflation process, Levy
argues that monetarist explanations, such as that provided in the
previous papers by Allan Meltzer, fail to take the analysis far enough.
Although they identify the Federal Reserve as the ultimate source of
inflation, monetarists do not give the Fed's inflationary behavior
adequate explanation.
The fundamental source of the inflation of the last decade and a
half, according to Levy, lies in the drastic changes in social attitudes
and in economic policies that got underway in the mid-60s and persisted
throughout the '70s.
This new social activism resulted in large and
rapidly growing federal programs designed to transfer real after-tax
income from the productive sectors to nonproducers, a dramatic increase
in both the size and scope of civilian programs, increased reliance on
deficit spending, and a new wave of socially-oriented regulation.
The
dominant forces behind the persistent inflation were the following:
the increased expansionary thrust of the budget; the acceleration in
monetary growth in order to accommodate the deficit financing; the
acceleration in wage demands as workers attempted to reverse the
V
decline in after-tax rea1 income associated with tax-transfer programs;
the increased reliance on "inflationary" social security taxes; the
increased business costs associated with regulation; and the slowdown
in productivity and real growth resulting from disincentives, both to
work and invest.
In his comments on the Levy paper, William Poole suggests that
the slowdown in productivity could have raised the inflation rate
associated with a given rate of monetary growth by only one or two
percentage points.
The remainder of the rise would have to be asso-
ciated with increases in monetary growth to accommodate the inflation
initiated by the other factors cited by Levy.
Poole states, however,
that Levy does ·not provide any evidence that the factors he cited were
quantitatively important sources of inflation pressure.
Moreover,
Albert Burger, in his discussion paper, notes that Levy gives relatively
little attention to the behavior of the Fed and hence leaves unanswered
the question that motivates Levy's objection to the monetarist explanation of inflation:
"Why did the Fed accommodate these i nfl ati onary
forces?"
In his luncheon address, Lawrence K. Roos, president of the
Federal Reserve Bank of St. Louis, described what he termed the
"shortcomings" of the past monetary policy actions and announced his
enthusiastic support for the Fed's recently announced change in the
method by which future monetary policy will be conducted.
Although
the new policy approach, which places ·primary emphasis on the growth
of reserves and monetary aggregates, holds the promise of avoiding the
policy errors of the past, Mr. Roos cautioned that there are several
steps which must be taken if the policy change is to fully achieve its
vi
desired results.
Among the necessary steps are increased focus on the
growth in the monetary base, the avoidance of monetary policy surprises,
and a commitment to a long-run policy viewpoint so that neither
political pressures nor false expectations force abandonment of the new
policy.
He emphasized that the new policy must be given at least a
year to prove its value and should not be expected to dissipate inflation in a matter of months.
Stabilization Policies:
International Aspects
Jacob Frenkel began the third session of the conference with a
thorough analysis of the experience with flexible exchange rates in the
1970s.
His paper, "Flexible Exchange Rates in the 1970s," sets forth
the way in which the asset market or monetary approach to exchange rate
determination helps to explain this experience, particularly the
observed volatility in exchange rates and the relation between exchange
rates and both domestic and foreign interest rates and price levels.
Within this framework Dr. Frenkel highlights the central role of
expectations, particularly expectations about future inflation, in
determining exchange rates.
An explanation of the volatility of ex-
change rates is aided by the view that these rates are a financial
variable whose value is sensitive to expectations about future developments and is capable of quickly incorporating new information about
these developments.
The central role of inflation expectations
suggests, .ccording to Frenkel, an "intimate connection between monetary
policy and exchange rate policy" and imposes a "unique responsibility
on the monetary authorities in affecting the rate of exchange."
vii
H. Robert Heller outlines in his paper, "International Stabilization Policy Under Flexible Exchange Rates" the adverse effects that
the move to flexible exchange rates has had on international trade,
international capital movements, and foreign investment.
Heller takes
the position that the increased uncertainty about exchange rate fluctuations has resulted in a significant increase in costs to the business sector and that the adverse effect of this uncertainty has been
particularly evident in the decreased willingness of investors to
undertake direct investment and long-term construction activity abroad.
He also suggests that speculative capital flows may have accentuated
rather than reduced the fluctuation in exchange rates.
These increased
costs, moreover, were not offset by any benefits associated with
flexible exchange rates, such as greater freedom for domestic stabilization policies.
Heller notes that it will be impossible to return to fixed exchange rates as long as national inflation rates differ so widely.
He
concludes his paper with a series of recommendations for improving the
functioning of the international monetary system under flexible exchange
rates.
To preserve the dollar standard, the United States must act to
maintain the real purchasing power of the dollar.
This, in turn, will
require better control of monetary aggregates and will be facilitated
by adoption of longer-term monetary aggregate targets.
In his comments on the Frenkel and Heller papers, David Laidler
emphasized the implications of flexible exchange rates for the response
of inflation and output to deceleration in monetary growth by a single
country.
Flexible exchange rates, according to Laidler, impart an
added degree of price flexibility; hence they permit both a more rapid
viii
deceleration in inflation and a reduction in the cumulative output loss
associated with anti-inflationary policies.
This fact suggests that
empirical approaches which do not explicitly allow for the effect of an
open economy under flexible exchange rates may seriously overestimate
the cumulative output loss.
In his comments on the Heller paper, Geoffrey Wood remarks upon
the lack of evidence to support Heller's contention that flexible
exchange rates have had a harmful effect on international trade.
Wood
also objects to Heller's contention that destabilization of capital
movements has been an important source of volatility of exchange rates.
In Wood's view, the volatility of exchange rates simply reflects the
underlying volatility of national monetary policies.
Washington University
Laurence H. Meyer
Associate Professor
of Economics
ix
RECENT DEVELOPMENTS IN THE THEORY OF STABILIZATION POLICY
John B. Taylor
During the past decade the theoretical framework underlying
macroeconomic stabilization analysis has undergone a number of significant developments.
Theories designed to explain the crucial linkage
between aggregate demand policy and real economic variables have been
revised following the research on the "new microfoundations" of employment and inflation.
Critical expectations effects of stabilization
policy have been incorporated into the theoretical framework through
the use of rational expectations.
Optimal control techniques have
become sophisticated enough to be used on large nonlinear econometric
models, and more recently have been adapted for use in models with
endogenous expectations.
Supply considerations have been recognized as
having important policy implications and, when necessary, have been
incorporated into policy analyses.
Theories underlying the choice
between rules and discretionary policy have been altered and refined.
These developments are likely to play an important role in the practical evaluation of economic policy in the years ahead.
This paper reviews these developments in the theory of stabilization policy and outlines some of their implications for macroeconomic
policy evaluation.
The first section reviews the theories which have
John B. Taylor is Professor of Economics, Columbia University. The
author wishes to thank Robert Barro, Jerry Green, Dale Henderson, and
Laurence Meyer for helpful comments on an earlier draft, and the
National Science Foundation for financial support.
been developed to explain the effect of policy variables on the real
economy.
As there is still little consensus here, a number of alter-
native representative models are presented and compared.
The second
section examines the implications of these different theories for the
problem of reducing the rate of inflation, which is likely to be one of
the more important policy issues in the years ahead.
The third section
discusses a number of issues which have arisen in recent policy analyses
and which are closely related to the changes in the theoretical framework:
The Lucas critique of traditional policy evaluation procedures,
the applicability of optimal control, the choice of rules versus discretion, and the applicability of the new equilibrium approach to
stabilization policy.
With few exceptions this review focuses on theoretical research
on domestic stabilization policies.
International considerations and
empirical results are reviewed in other papers prepared for this conference.
Some of the topics reviewed here have recently been the sub-
ject of a large number of survey and expositional works.
The variety
of survey papers by Barro (1979), Buiter (1979), Fishcer (1979),
McCallum (1979), Phelps (1979), Prescott (1977), Santomero and Seater
{1978), and Shiller (1978) and the books by the Ball committee (1978),
and Sargent (1979) provide further detail and alternative perspectives
on the topics reviewed here.
Expectations play a predominant role in any discussion of stabilization analysis.
For the discussion that follows, the benchmark assump-
tion will be that expectations are formed rationally.
Variations from
this benchmark -- due perhaps to the necessity of people gradually learning about whether the economy has undergone a structural change -- are
-2-
considered in the course of the discussion along with variations in the
model underlying the policy analysis.
THEORIES OF AGGREGATE DEMAND EFFECTS ON REAL OUTPUT AND EMPLOYMENT
In the idealized world of complete markets with perfect information about opportunities in all markets, changes in the money supply
or more generally, changes in aggregate demand -- do not affect real
economic variables such as real GNP and employment.
Apart from distri-
bution effects, aggregate demand fluctuations are translated point-forpoint into price fluctuations.
Money is neutral.
Many of the theoret-
ical developments in macroeconomics in t~e 1970s have been concerned
with explaining, in more detail and with more rigor than earlier
theories, why this neutrality is not observed in the real world.
A
reasonably firm understanding of the mechanism generating this nonneutrality is certainly necessary for evaluating stabilization policy
because aggregate demand management tools, such as money growth and
government expenditure plans, are the primary instruments of stabiliza t 1. on po 11. cy. 1
1The effects of government policies which impact directly on
relative prices can be evaluated in principle using the standard allocative theories of microeconomics. Some examples: a relative lowering of tax rates on capital would be expected to stimulate investment
by raising the desired capital-labor ratio; a higher steady rate of
inflation has allocation effects by acting as a tax on real money balances; and unemployment insurance can raise the equilibrium unemployment rate by driving a wedge into the work-leisure tradeoff. Apart
from disagreement over the magnitude of the relevant elasticities for
measuring these effects, there has been a general consensus among
economists that such policies have real effects. However, because
these policies are used for allocative or distributional purposes, they
are not generally flexible enough to be considered seriously in stabilization analysis. Nevertheless, their importance cannot be overlooked
in analyzing macroeconomic trends. See Feldstein (1978) for a summary
of such effects on unemployment.
-3-
Recent theories of the observed link between aggregate demand and
real variables can be grouped into two types -- information-based
theories in which the uncertainties about economy-wide disturbances are
emphasized, and contract-based theories in which temporary rigidities
in prices and wages are emphasized.
At the risk of becoming too taxon-
omic, it will be useful to further classify each of these theories.
Among the information-based theories it is important to distinguish
between those in which the uncertainty is whether an observed economic
change is local or economy-wide, and those in which the uncertainty is
whether an economic change is temporary or permanent.
Similarly, among
the contract-based theories it is important to distinguish between
those that emphasize relative price shifts due to asymmetrical rigidities (for example, wages are rigid while prices are flexible), and
those that emphasize the general persistence of all prices due to nonsynchronous price {or wage) setting relative to a prevailing trend in
prices (or wages).
Uncertainty about Local Versus Aggregate Economic Conditions
Perhaps the most significant finding of the research 2 on the "new
microeconomics" is that imperfect information about economic conditions
outside an individual's own market or industry can have profound implications for the behavior of inflation and employment.
Suppose aggre-
gate demand increases because of a higher rate of money growth.
Then
individual firms will find an increased demand for their products, and
will respond by increasing their production (and perhaps running down
2see Phelps et~ (1970).
-4-
:heir inventories of finished goods).
But much of this higher real
Jroduction may be due to the misperception on the part of each firm that
the increased demand is a relative shift toward the product it sells.
3ecause there is always imperfect information about whether an increase
in sales is a local phenomenon, this misperception and the consequent
rea 1 output response is una voidable.
If, on the contrary, each firm
,new that the increase in demand was common to all firms in the economy,
3nd was due to the purely nominal increase in the money supply, then its
1roduction response would be much smaller.
If prices and wages were
generally flexible, then firms would know that prices and wages should
➔ uickly
rise to offset the increase in the money supply, and therefore
that an increase in output would not be warranted.
In the limiting case
of perfectly flexible prices, good information about what is going on
elsewhere in the economy enables fil11ls to respond just as they would be
predicted to do in the money-neutral world of general equilibrium theory.
But even with perfectly flexible prices, imperfect information creates
a non-neutrality in which firms respond to aggregate demand stimulus by
increasing real output.
The link between aggregate demand and real vari-
ables, according to this theory, depends in no essential way on price or
wage rigidities.
As long as there is imperfect information about the
source of aggregate demand shifts, the correlation between aggregate demand
and real output will exist.
Of course, the possibility of a coincidence
of perfectly flexible prices and wages with these well-known empirical
correlations means that policy implications will be much different.
Simple descriptions of this theory are found in Phelps et al.
(1970) and Lucas (1973).
The algebra of the Lucas presentation is
-5-
convenient for our purposes and can be represented in terms of a simple
quantity theory of aggregate demand.
(l)
y + p = m+ v
combined with an "aggregate supply" equation
(2)
y = u(p - p).
All variables are measured in logarithms and should be thought of
as deviations from secular trends: y is real GNP, pis the aggregate
A
price index, mis the money supply, and vis velocity.
The p term
represents a forecast of the price level before the information about m
~
and v becomes available.
The difference between p and p represents the
average difference between each firm's observation of demand conditions
during the period and its guess about economy-wide demand conditions.
This difference represents the misperception or mistake discussed above
which causes firms to increase their production.
production responses is y.
The sum of all firms'
(It turns out that it is convenient alge-
braically to use prices to index demand conditions.)
A
Substituting from (l) into (2) and noting that from (2) that y=O,
we find 3
(3)
y
u(m - m + v - v).
3we take p to be a rational (unbiased} forecast of p; hence
E(p-p) = 0. "Biased" forecasts are treated in Section 1.2 below and
arise because of information confusion about what is the actual model
underlying policy or the structure of the economy. These ~"biased"
forecasts have forms which resemble adaptive expectations, but unlike
adaptive expectations are closely related to the structure of the model.
-6-
Hence real output responds positively to unanticipated money m-m and
h
unanticipated velocity v-v.
This is the critical link between real
variables and aggregate demand which the theory explains.
However, because only unanticipated changes in aggregate demand
affect real output, the policy implications of this linkage theory are
striking:
if the monetary authorities change their policy instrument m
in a way which can be predicted by individuals in the economy, then in
our notation m~m and the change in m does not affect real output at all.
And from equation (1) the change in mis translated entirely into a
point-for-point change in p, apart from any unanticipated shifts in
velocity.
This famous "policy-ineffectiveness" result, emphasized by
Lucas (1973), Sargent and Wallace (1975) and Barro (1976), has understandably stimulated a large volume of research.
The significance of this theory for practical stabilization analysis is not simply the neutrality result -- the idealized general
equilibrium model has long been known to yield neutrality as discussed
Rather the significance is due to the appearance of neutrality
above.
in a model which explains the empirically observed correlation between
aggregate demand policy and real output.
The theory would be of little
practical importance if it did not generate this important empirical
result.
The econometric work of Sargent (1976) and Barro (1977, 1978),
has been aimed at making this empirical connection more formal and
rigorous.
I think it is fair to say that this empirical work has demonstrated that the theory is consistent with these correlations.
facts have been more difficult to reconcile with the theory.
Other
The per-
sistence of unemployment is one regularity which does not emerge from
-7-
the simple theory, and was used as a critique of the theory by Hall
(1975) and Modigliani (1977).
A number of modifications of the theory
to account for this persistence have been suggested.
Lucas (1975) em-
phasized that unanticipated shocks could cause firms' capital stock to
get out of line, and this would have repercussions on production in
later periods as the capital stock is adjusted.
sized adjustment costs in changing employment.
Sargent (1979) emphaBlinder and Fischer
(1978) have placed more emphasis on finished-goods inventory being
drawn down or accumulated.
Optimal inventory adjustments in later
periods will then require production changes and thereby cause a correlation between output changes at different dates.
All these theoret-
ical modifications of the basic information-based model with perfectly
flexible prices can in principle explain persistence, but it has yet to
be demonstrated whether actual inventory behavior or costs of employment
adjustment are sufficient to explain the persistence.
There is, of course, much other evidence which the theory can be
tested against.
Two pieces of evidence which seemingly run counter to
the theory are procyclical productivity changes, and a slight tendency
for real wages to vary procyclically, though the latter is much less
pronounced.
Sargent (1979), extending the work of Lucas (1970), has
shown, however, that these observations are consistent with the limitedinformation flexible price models.
His proof involves disaggregating
employment into straight-time and over-time, and assuming that straighttime employment is more costly to adjust, but that over-time workers
must be paid more on average.
Under these conditions firms will find
it optimal to employ more straight-time workers than over-time workers
on average, but to make larger changes in employment among over-time
workers than straight-time workers, when demand conditions change
across the business cycle.
This behavior implies that real average
hourly earning will tend to increase during booms, because of the shift
of the mix of workers toward higher paid overtime employment, even
though real wages may fall for both groups of workers.
Moreover, since
fewer over-time workers are employed on average than strai~ht-time
workers, their marginal productivity is higher.
Hence, the shift to-
ward more over-time employment causes average productivity in the economy to increase.
Sargent (1978) has attempted to see if this intricate
theory is sufficient to explain the phenomena quantitatively, and finds
that, although there are some discrepancies, the theory generally conforms to the facts.
Another explanation for the procyclical behavior
of real wages is given in Phelps (1969) using a model of inventory
behavior.
New data now becoming available on real inventories may per-
mit a check of this explanation.
From the point of view of stabilization theory a number of extensions of the basic information-based model represented in equation (2)
should be mentioned.
Cukierman (1979) has shown that the limited-
information assumptions can be generalized to permit firms to change
their expenditures in order to better determine the source of economywide events.
This makes the information structure endogenous to the
rest of the economy, including policy, and thereby removes the
criticism that the theory unrealistically places an exogenous information structure on economic agents.
He finds that the general results
of the theory are robust with respect to this modification.
Mccallum and Whitaker (1979) have shown that the policy neutrality result does not apply to such aggregate demand tools as automatic
-9-
stabilizers because these react simultaneously to changes in economic
conditions, rather than with a lag as in the feedback monetary policy
discussed above.
For example, with progressive taxes, after-tax income
immediately changes as a fraction of total income when nominal income
fluctuates.
This can have direct real stabilizing effects.
It should
be emphasized, however, that in principle monetary policy could be made
to operate just as simultaneously as the automatic stabilizers.
This
has not been the case in practice, however, except for extreme interest
rate pegging where the central banks' supply of reserves responds
instantaneously to changes in demand.
Uncertainty about Temporary Versus Permanent Changes in Economic
Conditions
The theory discussed above emphasizes lack of information about
whether demand changes are local or economy-wide.
From the viewpoint
of stabilization policy, an equally important type of uncertainty is
the lack of information about whether an observed economic change is
temporary or permanent.
Theories which emphasize temporary versus per-
manent effects are, of course, not new to macroeconomics, as exemplified
by Friedman's (1956) original permanent income theory of consumption.
Muth (1960, 1961) also emphasized the distinction in his original work
on rational expectations.
Here we are concerned with the importance of
this uncertainty for the link between aggregate demand and real output.
The general point is that a shift in nominal aggregate demand, which is
expected to be permanent will have a much smaller effect on real output
and a correspondingly larger effect on prices, than a shift which is
expected to be temporary.
-10-
Suppose, for example, that in an attempt to reduce the rate of
inflation the central bank reduces the growth rate of the money supply.
The information problem which economic agents face is whether this
change is a permanent one, or whether the central bank will soon give
up on its resolve to lower the growth rate of the money supply.
In
reality, this information problem is not trivial, and cannot be eliminated simply by announcing that today's start at monetary restraint is
the beginning of a permanent shift in policy.
Lack of credibility
about whether the shift is indeed permanent may be cured only by the
public observing the results of the new policy.
During the transition period when people learn whether the shift
is temporary or permanent, the policy of restraint can have real output
effects, even if prices are perfectly flexible. This can be illustrated
using the algebra introduced above. 4 Equation (2) can be written in
terms of inflation rates rather than price levels by subtracting the
lagged price from p and p.
This gives
when rrt is the expected rate of inflation.
Suppose that rrt
=
rrt so
that there is initially no uncertainty, but that starting in period t+l
the central bank reduces the rate of growth of the money supply to a
level that will generate an inflation rate of rrs
'
<
rrt for s
>
t.
If
the new policy is not fully credible, then people will not immediately
adjust their expectations to rrs.
A reasonable assumption would be that
they expect a level of inflation which incorporates the new information
4The following discussion is based on Taylor (1975).
-11-
about rr as well as the previously expected rate of inflation.
In
simple terms:
s
(5)
=
t+ l , t+2, ...
Formula (5) can be derived more formally using Bayesian techiques which
incorporate the uncertainty about whether the new inflation rate is
permanent or whether the observed change is a temporary occurrence.
The
parameter\ will be time dependent in general, however, and this should
be taken into account if one is interested in quantitative policy
evaluation.
To see the effects of the new monetary policy on real output
assume for simplicity that ns is equal to a constant n* for s
>
t + 1.
Then from (5) we have
(6)
for s
rr
>
5
= \
t + l.
S-1
E
i
S'
(1-\) rr* + (l-\) rrt
i=O
Hence, rrs converges torr*, but will be greater than rr*,
if rr* is less than rrt (if the new monetary policy is to aim for a lower
rate of inflation).
The gap between the expected rate of inflation rrs
and the actual rate of inflation rr* will be larger, the smaller is\.
Hence, the less credibility there is about the new policy, the larger
the inflation gap and the larger the reduction in real output.
will be no reduction in real output if \=l.
There
In this way the uncertainty
about permanent versus temporary effects has an important influence on
the way policy is linked to real economic variables.
The type of model represented here in very simple terms has been
emphasized in stabilization policy analyses by Fellner (1976),
-12-
B. Friedman (1979), and Taylor (1975).
A full macroeconomic model
developed by Brunner, Cukierman, and Meltzer (1979) uses the distinction between permanent and temporary effects to examine the influence
of supply shocks as well as demand shocks on production.
Flood and
Garber (1979) have provided estimates of similar credibility parameters
in the case of monetary reform in the German hyper-inflation.
These types of models have been criticized, especially when used
for policy analyses of the type discussed here, because they appear to
depend on policy deception (see Barro (1978)).
While the potential for
deception is clearly present in these models they are equally applica~e
to situations where all parties disclose their intentions.
nately, disclosure does not generate immediate credibility.
UnfortuIt is the
problem associated with this lack of credibility which these models
emphasize.
Contracts and Relative Price Effects
Imperfect information is not the only reason that aggregate
demand would be expected to influence real output.
Temporary rigid-
ities in prices or wages might force some of the change in nominal
demand into changes in real production.
Since casual observation
suggests that such rigidities are pervasive either in the form of explicit contracts or less formal implicit contracts, economists have
been willing to take these rigidities as given.
The main theoretical
development in this area during the past several years has been to
recognize that the form W'lich these rigidities takes is important for
stabilization analysis.
Attempts have been made to model these rigid-
ities with more detail than was previously available, and to trace out
-13-
the implications for policy.
Two different forms of this type
Of
analysis can be usefully distinguished.
The most common form of this type of model assumes that wages are
at least temporarily rigid, but that prices are perfectly flexible in
the sense that firms cannot directly influence profit margins by
marking up their prices relative to wage costs.
Firms simply adjust
their demand for labor when the real wage shifts against them.
Recent
examples of this type of model are found in Fischer (1977), Phelps
(1978), and Calvo (1980).
Letting wt represent the nominal wage and
keeping the notation introduced earlier, the most rudimentary form of
this model is
When the real wage rises firms reduce output and employment, until the
marginal productivity of labor is increased.
If wt is partially pre-
determined, perhaps because of multiperiod contracts which were set in
previous periods, then the link between aggregate demand and real output fo 11 ows directly.
If aggregate demand is determined according to
equation (1) then
(8)
a
a(vt- wt)
Yt = ~ t +
l+a
and clearly changes in nominal mt get translated into real output.
The
mechanism is simply that a higher money supply raises prices which
lowers the real wage and stimulates employment and production.
The major advance in using this type of model has been to develop
the mechanism determining the nominal wage.
-14-
Fischer assumes, for
ixample, that there are overlapping contracts with a fraction of the
:ontracts set in each period so as to keep the expected real wage
:onstant.
A consequence of this assumption is that aggregate demand
!ffects do not persist for longer than the length of the longest con:ract.
Another consequence is that wage or price trends have no
:endency to persist.
In these two respects this type of model has many
°eatures which are similar to the results of the information-based models.
"his has led Gramlich (1979), for example, to conclude that wage-rigidities
fo not add much in the way of policy implications to a rati ona 1 expecta-
:i ons mode 1s.
In pri nci p1e, of course, announced monetary po 1icy affects
·eal variables in such models, even with rational expectations.
>een emphasized by Fischer (1977).
This has
The question is whether they describe
;he wage and price dynamics in an empirically accurate way that is rele1ant for policy analysis.
The main feature of these models is their dependence on real wage
:hanges for all employment effects.
As discussed above, it has been
Hfficult to find much variation in the real wage over the business
:ycle.
Empirical checks of this model along the lines of Sargent (1978)
Jsing the distinction between straight-time and over-time workers would
:herefore be very useful.
On the other hand, there are important policy problems where
:hanges in real wages are the central issue.
For example, a supply
;hock could shift the marginal productivity downward requiring a reduc:ion in the real wage.
With sticky wages, this reduction might be
lifficult without monetary intervention.
In effect the monetary
ruthorities can use monetary policy to shift the price level to a
msition such that the real wage is equal to the level which workers
-15-
would have aimed for, if they had known about the shock when they
signed the contract.
This is the conclusion of Phelps (1978) who bases
his analysis on such a model.
Gordon's (1975) analysis of agricultural
supply shocks reaches a similar conclusion if farm prices shift up
while industrial prices are assumed to be relatively riqid.
Blinder
(1979) also emphasizes these relative price rigidities in examining the
appropriate response of policy to an oil price shock.
One difficulty
with all these analyses is the possibility that the assumed rigid wage
(or price) eventually adjusts to offset the policy-induced shift in
relative prices.
In the Phelps analysis, this is not much of a diffi-
culty in principle because the real wage is pushed toward what workers
and firms would have negotiated otherwise.
Another difficulty, already
alluded to, is that the models do not capture much of the persistence
effects of inflation and unemployment which now seem to present important policy problems.
In this respect they are similar to the informa-
tion-based models reviewed above.
Staggered Contracts and Inflation Persistence
By most measures the variability of the general price level in
recent years has been larger than the variability of all but a small
number of relative prices.
For example, the real wage has been rela-
tively stable compared with the sharp rise in nomina1 wages and prices.
Moreover, changes in both nominal wages and prices are more highly correlated with business cycle fluctuation than changes in the relative
wage.
For these reasons, one might suspect that analyses which focus
on real wage changes as the sole cause of employment shifts might be
omitting at.her factors.
-16-
Another class of models which are based on rigidities in wages
nd prices deemphasize the aggregate effects of relative price shifts
nd focus on the problems of general price movements.
These models
mphasize the fact that all prices and wages are not set in unison
cross the economy but are generally staggered, and that a primary
eterminant of the price decision is the prevailing price outstanding
n the market.
Hall (1979) has recently developed a microeconomic
odel which gives an explanation for the importance of setting prices
elative to the prevailing price.
An example of this type of model is given in Taylor (1979).
irms and workers decide on a wage xt in period t which is to last for
wo periods.
The contract wage xt is set according to the expected
revailing wage during the contract period with suitable adjustments to
eflect demand conditions.
Hence
9)
here wt~ 1/2(xt + xt_ 1 ) is the average wage at time t.
ions of Yt represent demand pressure on wage decisions.
The expectaIf we make
he additional assumption that profit margins are relatively stable
hen pt~ wt+
y
where
y
is a constant parameter which we can set to
ero without loss of generality.
By holding the relative wage constant,
he model purposely abstracts from relative price changes and focuses
n general price movements.
In this model, as with the previous model based on price rigidties, aggregate demand policy has a direct effect on real output.
If
quation (1) is the aggregate-demand relationship, then the mechanism
-17-
works as follows:
the price level is predetermined since the wage is
predetermined and profit margins do not adjust.
Hence, an increase in
the money supply increases real balances, which tends to increase the
real demand for goods.
This results in an increase in production and
hence an increase in employment.
Eventually wages and prices will ad-
just because the favorable demand conditions will give firms the incentive to pay increased wage demands.
and reduce real money balances.
This in turn tends to raise prices
Eventually a new equilibrium is
reached at a higher price level but with the same level of production.
Money is neutral in the long run.
What is different about this model compared with those discussed
in the previous section is that convergence to the new equilibrium
takes time, and there is never any important shift in relative wages
(there is a period during which the workers who had settled their contracts when the money supply was changed tend to fall behind other
workers but this is not necessarily integral to the workings of the
model).
The inertia in wage movements following the shift in money
supply can be demonstrated by solving the model to obtain 5
where Band o depend on the parameter
a.
Hence, a change in the money
supply sets off a series of changes in the contract wage xt and hence
in the average wage wt.
This series of changes in wt is matched by the
price level Pt and, if the money supply is held fixed at the new level,
5The derivation requires the use of rational expectations to
solve out for the expectation variables.
-18-
is reflected in a similar pattern of changes in real output.
Because
of these persistence effects this tyoe of model would seem to be more
useful for examining stabilization problems associated with reducing
inflation, or more generally achieving price stability, than the models
discussed in the previous section.
If changes in real wages are also
thought to be important, then they can easily be incorporated into the
analysis.
Theoretical frameworks of this kind have been used for
policy analysis by Phelps (1978a), Gertler (1977), Modigliani and
Papademos (1978), Papademos (1979), and Taylor (1980).
These models have some similarities to the "disequilibrium"
models developed by Clower (1965) and Barro and Grossman (1976).
Im-
portant differences not generally found in "di seq uil i bri um" models are
the use of rational expectations, a reasonably explicit description of
the contract mechanism, and a reliance on the more traditional aggregate demand framework without the development of market spillover
effects or of binding supply constraints.
These differences largely
reflect empirical considerations or modelling strategies.
It is not
yet clear what is to be gained empirically or theoretically from incorporating disequilibrium spillover effects.
A recent paper by Green and
Honkapohja (1979) has attempted to bring rational expectations into a
framework which corresponds more closely with the disequilibrium
models.
However, their approach is designed to avoid explicit treat-
ment of the nonlinearities caused by setting market transactions equal
to the minimum of supply and demand.
Rational expectations are much
easier to deal with in linear models, and this is one reason the
"demand is determining" assumption is used.
Another reason is that the
assumption seems to be empirically realistic in many situations.
-19-
~omparison of the Alternative Theories
What sets the contracting models off from the information-based
models is of course the use of "sticky" prices, and the corresponding
disuse of the market-clearing assumptions.
In the contract models,
markets "clear" in the short run in the sense that supply adjusts to
meet the demand; in the long run, prices eventually adjust to clear
markets.
In the information models, on t~e other hand, prices instan-
taneously adjust to clear markets in the short run.
better?
Which approach is
I have used the contracting approach because it corresponds
more closely with my interpretation of the market mechanisms in the
real world.
It is not just the widely discussed long-term labor con-
tracts which suggest this interpretation, but also the much more common
(at least in the U.S.) implicit contracts, which are much shorter and
a re usually not called contracts.
In fact, long-term labor contracts
have so many indexing provisions that they probably correspond more
closely with shorter contracts.
Research in this area has shown that
"contracts" do not have to be very long to generate a very lengthy
persistence of wage and price inflation.
ample.)
(See Taylor (1980), for ex-
But in using these contracting models, one has to be aware
that without an explicit utility maximization framework, there is a
possibility that the models are not robust to changes in policy.
Again
my preference has been to make the most of these models in situations
where the contracting mechanisms are relatively robust.
At the same time, it is difficult not to appreciate the theoretical elegance of the information models, and the potential to use the
traditional tools of microeconomics to conduct policy analysis with
these models.
But even the information-based models have some ad hoc
-20-
assumptions, especially when they need to be modified for empirical
work.
One of the major recent developments in the literature on market-
clearing rational expectations has been to pursue a more theoretically
rigorous approach with the aim of omitting the remaining ad hoc
features, in particular the money demand equation or quantity theory
equations (such as equation (l} in this paper).
Cass and Shell ( 1979) . 6
See Wallace (1977) and
The work by Azariadis (1975}, Baily (1974), and D. F. Gordon
(1976) does not provide as much of a foundation for contract models as
one might have originally thought.
These theories do not suggest why
contracts are set in nominal terms without contingencies.
In fact,
Barro (1979) has suggested that these microeconomic theories are more
useful in showing that the market-clearing models are useful "as if"
devices.
Calvo and Phelps (1978) and Hall and Lilien (1979) have pro-
vided alternative theories of contracts which emphasize the practical
and theoretical difficulties of making contracts contingent on
everything.
Most of the policy discussions associated with the theories reviewed above have been about the effectiveness of policy or whether
policy activism is useful or not.
In the market-clearing setting, only
6A useful appraisal of the overlapping generations model approach
advocated by Wall ace is contained in Cass and She 11 ( l 979). The major
appeal of this approach is the enormous theoretical mileage one gets
from the disaggregation of generations. At an abstract level this disaggregation is very similar to the disaggregation of contracts according
to when they are negotiated -- a feature of the contracting models
discussed in Section 1.4. More generally one suspects that different
types of disaggregation are likely to yield additional theoretical insights. Another example is the two-sector model explored by Sargent
and Wallace (1971), Henderson and Sargent (1973), and Foley and
Sidrauski (1970).
-21-
unanticipated changes in aggregate-demand policy matter, so announced
policies do affect output.
In contracting models aggregate-demand pol-
icy has effect whether it is anticipated or not.
Hence, in these models,
policy is effective and, in certain cases, policy activism is desirable.
Some examples of the optimal reaction to supply shocks were discussed above.
Mccallum (1977) has argued that price rigidities are not really
the source of the policy effectiveness in the contracting models.
In
criticizing the contract model used by Phelps and Taylor (1977) he shows
that monetary policy is ineffective if one removes inventory effects on
production, but uses the supply equation in the form of equation (2).
However, inventory effects on production are an important part of
models where prices do not adjust to clear markets.
Firms will want to
increase production, for example, if inventories are drawn down below
optimal levels because price adjustments are not quick enough.
This is
the rationale behind the inventory effects on production in the PhelpsTaylor model.
Omitting the term attributes suboptimal inventory
management to rational firms.
This point has been demonstrated by
Frydman (1979) in a critique of McCallum's results.
The main outcome of the policy-effectiveness debate is a general
consensus that rational expectations per se does not rule out effective
aggregate-demand management.
It is the flexible-price market-clearing
assumption that makes po 1icy ineffective for short run s tabi1 i zat ion policy .7
7Fischer (1978) and Lucas (1975) mention the nonneutrality that
comes even in market-clearing models from the substitution out of money
into real capital when the expected rate of inflation rises. However,
this mechanism is not seriously considered as a tool of aggregate demand-management. Moreover it is likely to be offset by tax effects. A
useful discussion of the relationship between rational expectations and
policy effectiveness is found in Lucas (1980).
-22-
POLICIES TO STABILIZE PRICES
The practical policy implications of these models can be alternatively stated from the viewpoint of price stabilization rather than
from the viewpoint of policy intervention to affect output.
Suppose,
for example, that the rate of inflation is generally agreed to have
become too high, either because of past policy mistakes or unavoidable
velocity shifts, and that tne monetary authorities want to reduce the
rate of inflation.
The important question is whether the monetary re-
straint necessary to achieve this goal of price stabilization will
cause a recession and how large that recession will be.
The answer to
that question will obviously influence the policymakers' choice of how
11uch restraint to apply.
If we take literally the information-based models, which emphasize
the uncertainty between aggregate and local shocks, then if this policy
of restraint is announced it will not have any effect on real output.
There will be no recession since inflation will match the reduction in
monetary growth point for point.
This striking conclusion is, of
course, contrary to the views of many economists and policymakers, and
I think for this reason the model is still rejected by many economists
as a practical guide to policy.
On the other hand, if there is uncertainty about whether the
changes in policy are permanent or temporary {as discussed above), then
the real effects of policy will exist, and a recession would be expected to occur.
The size and duration of the recession would depend
on the speed with which people begin to believe that the central bank
is firm in its resolve to restrain money growth.
If the credibility is
high or increases quickly, then the recession could be very mild.
-23-
Fellner (7979) indicates why he thinks that credibility is likely to
increase quickly, if a clear announced policy of restraint is undertaken, and that people's expectations of inflation would be swiftly
revised downwards.
The contract-based models yield different conclusions.
The
models which emphasize real wage shifts because of asymmetric rigidities do not suggest any reason for a recession to last longer than the
length of the average contract.
The inflation rate could be put on its
new target path in the first period; in the second period wages would
adjust.
In fact, if the restraining policy was announced and believed
one period (year?) in advance, there would be no decline in output.
In
this case, this type of contract model does not give results that are
much different from the market clearing models.
The general staggered contract models suggest, on the other hand,
that the recession would be somewhat longer because the adjustment process is passed on gradually from one contract to the next.
However,
because there are some forward-looking features to these models (see
equation (9), the recession would not be expected to be as severe as
would be implied by the simple reduced forms (see equation (10)).
The
policy of restraint (if it is believed) would change the parameters of
(10), so as to reduce the size of the recession.
Accurate quantitative
estimates of how much the parameters would be expected to change have
yet to be obtained, though simulation results in Taylor (1980) suggest
that it is likely to be significant.
In sum, each of the models reviewed here has implications about
the real effects of a policy of price stabilization.
(These models,
ignore, of course, any direct positive real effects that a more certain
-24-
irice level might bring; see Fischer and Modigliani (1979) for a dis:ussion of these direct effects.)
In the cases where the real effect
s likely to be significant, it would be interesting and useful to
:ompare empirically its magnitude with the estimates provided by con1entional econometric techniques as summarized by Okun (1978).
l
This is
feasible and well-defined estimation problem as the discussion above
1akes clear.
ALTERNATIVE TECHNIQUES FOR THE ANALYSIS OF STABILIZATION POLICY
This section gives an overview of several recent developments
:oncerning the choice of alternative techniques to analyze stabili~ation policy.
Some of these issues are intimately connected with the
~heoretical developments summarized in the first section.
rhe Lucas Critique of Econometric Policymaking
Econometric models have played a large role in policy formulation
in recent years.
It is rare that the staff members of policymaking
igencies do not run alternative policies through the major large scale
~conometri c models before meeting with their
jo not have formal models of their own.
II
pri nc i pals," even if they
Whether this heavy use of
~conometric models actually influences the decisions of policymakers is
~nother question.
Political or other noneconomic considerations are
frequently a factor.
But when "pure" economic advice is sought, the
results of the econometric models are certainly taken into account.
For
~xample, the property of almost all econometric models that nonaccomnodative monetary policy has small effects on prices and large effects
)n output, undoubtedly influences policymakers to choose more accomnodative policies than they otherwise would.
-25-
Lucas (1976) has criticized this type of econometric policymaking.
He argues convincingly that the parameters of these models are not invariant to changes in policy, so that the policy experiments performed
on these models (which treat the parameters as fixed) give misleading
results.
R. J. Gordon (1976) suggests that suitable modifications of
econometric policy evaluation procedures could deal with the Lucas
criticism.
The parameters could, in principle, be made endogenous.
The parameters of econometric models can shift for many reasons,
but the one Lucas emphasized was that rational economic agents would
forecast the future effects of policy, and accordingly, modify their
behavior in a way not described in the econometric models.
To deal
with this problem it is necessary at least to reestimate the econometric
models taking these expectation effects into account.
The most prac-
tical way to do this with existing econometric techniques is to use the
rational expectations assumption.
Having specified and estimated an
econometric model with rational expectations it is then possible to
perform a policy analysis to take account of the expectations effects.
This is the approach taken by Taylor (1979a}.
A simple quarterly
econometric model of the U. S. economy was estimated during the 19541976 period, imposing rational expectations on economic agents.
Using
the estimated parameters of this model, alternative policies were compared, and for a given set of policy preferences, optimal policies were
calculated.
Because the model incorporated contracts of the kind dis-
cussed above, a policy tradeoff between inflation and unemployment was
implied by the model and this was calculated using the estimated parameters.
The tradeoff was characterized by a "best" relationship
-26-
)etween output stability and price stability. 8 This optimal relation;hip apparent·ly dominated actual policy during the period as well as
~he policy of a constant growth rate for the money supply.
Constant
noney growth would have given better results than actual policy, how~ver, according to these estimates.
Anderson (1979) and Fair (1979) have tried to estimate the quantitative significance of the Lucas critique by simulating conventionilly estimated econometric models, with rational expectations inserted.
fhey both find the effects to be quantitatively significant, but their
"esults are difficult to interpret because the conventional models were
,ot formulated as rational expectations models.
For example, Anderson
(1979) finds that the Phillips curve is much steeper when he imposes
0
ational expectations on the model.
But clearly the specifiers of his
nodel would have altered its specifications if they knew rational exJectations would be imposed.
It is likely that the adaptive expecta-
tions distributed lags used in such models are designed to caoture
}ther dynamic properties than pure extrapolative forecasting.
Quantitative work of this kind with rational expectations is only
just beginning.
More experience with these techniques will be
1ecessary before they can be accurately appraised as significant im)rovements over conventional econometric policy evaluation procedures.
rhe results available thus far are promising, are already giving rough
8
Flemming (1976) p. 73 suggests that a tradeoff between output
;tability and price stability might be a good way to characterize the
)olicy problem. Phelps and Taylor (1977), Taylor (1980), and Green and
,onkapohja (1979) have calculated theoretical tradeoffs of this kind.
\n international comparison of such tradeoffs is given in Taylor
:1980a).
-27-
empirical estimates of the effect of policy, and indicate that further
research is fruitful.
Two objections can be raised against these attempts to account
for the Lucas critique.
One is that the rational expectations assump-
tion is not accurate because it does not incorporate learning on the
part of individuals about the economy.
If this learning problem is
significant, then these techniques will have to be modified.
Learning
effects are likely to be a serious empirical problem immediately
following a major economic reform.
This was illustrated above for the
case where the monetary authorities change their policy and people do
not know whether it is a permanent or temporary change.
However, even
if learning problems are significant, these techniques will be useful
for evaluating alternative policy procedures over a long period of time.
For example, it is useful to know if a less accommodative monetary
policy during the 1960s and 1970s would have increased the amplitude of
business cycle fluctuations as much as conventional econometric models
would imply.
If the use of rational expectations gave results much
different from other models over long enough periods for the rational
expectations assumption to be realistic, then the results would be
taken into consideration in recommending how accommodative policy
should be in the 1980s.
Another objection to the quantitative use of rational expectations as described here is that there are other reasons that parameters
of a model could change.
For example, even if rational expectations
were used, behavioral relations for contract-wage determination might
shift with policy as workers and firms change contract lengths.
While
expectations are probably a significant source of parameter drift, this
-28-
does not mean that models can ignore other behavioral shifts.
Success-
ful policy evaluation requires careful modelling of all behavioral
relations.
The New Equilibrium Approach to Policy Evaluation
Lucas and Sargent (1978) have suggested that the pervasiveness of
these other sources of parameter shifts means that minor modifications
of econometric models are not sufficient.
They recommend a "new equi-
librium" approach to modelling in which all economic relations are
based on explicit utility maximization analysis.
If tastes and tech-
nology remain relatively constant -- or can be modelled as exogenous
factors -- then this approach, in principle, will avoid these other
types of parameter shifts.
The approach is attractive because once one
has developed a model based on sound utility maximization principles,
macroeconomic policy analysis is conducted like any other welfare
analysis in microeconomics.
Explicit externalities can be located and
offset by optimal policies, and no approximate aggregate welfare
criteria such as output and price stability are necessary.
One would
design policy to maximize the welfare of the representative individual.
Attempts to design business cycle or econometric models along these
lines include the work by Barro (1976), Lucas (1975), Hansen and
Sargent (1980) and Kydland and Prescott (1980).
This approach represents a fundamental change in macroeconomic
policy evaluation and its full practical implementation will take a
long time as emphasized by Lucas and Sargent (1978).
As an alternative
to the approach outlined in the previous section, several reservations
about this new equilibrium approach might be mentioned.
-29-
Does utility
maximization provide any additional constraints on an economic model
which do not already come from a set of explicit decision rules and
rational expectations?
If it does not, then the gains from beginning
each analysis with explicit utility maximization are not clear.
For
example, one of the major~ hoc features of decision rules designed
for empirical work is that they include lags to capture the gradual
adjustment of firms to new economic conditions.
With utility maximi-
zation, these lags are "exp 1ai ned" by adjustment cos ts which tend to
make it optimal for firms to adjust slowly.
But one has almost as much
freedom to choose adjustment costs in a utility framework as one does to
choose lag length when writing down decision rules.
Unless good micro-
economic or technological information is available to measure these
adjustment costs, the utility maximization approach does not seem to
provide additional information in this case.
Another reservation concerns the practical use of the welfare of
the representative individual as the criterion for stabilization policy.
In principle this approach is better than the alternative approach of
postulating an aggregate measure of welfare, which might include
measures of inflation or aggregate employment stability.
gate welfare approach has advantages in practice.
But the aggre-
It is very difficult
to incorporate some of the welfare gains of price stability into individual utility functions.
The gains from a relatively stable aggregate price
level involve such considerations as providing a more certain framework
for private decision making.
Unti1 one finds a way to incorporate
these complex effects into individual utility functions, the use
-30-
of aggregate criteria may serve as satisfactory and workable
alternatives.
Rules Versus Discretion
The debate between those favoring rules versus discretion has
not diminished in recent years but the arguments have been modified.
A
definitional change is that rules are now rarely taken to mean holding
policy instruments constant.
Feedback rules, in which the money supply
responds in a systematic way to economic developments, are rules as
much as constant money growth.
Kydland and Prescott (1977) have suggested that the problem of
time inconsistency (see also Calvo (1979)), implies that rules
should be used rather than discretion.
Time inconsistency can arise
because of taste change or because people forecast future behavior
of policymakers.
In both cases policymakers may be tempted to change
plans after they have announced the optimal path.
Time inconsistency
does not imply that optimization techniques cannot be used (see
Fischer (1980) for a discussion of this issue), but it does raise
questions of how policy should be implemented.
Kydland and Prescott
(1977) argued that rules would be a way to reduce the incentive for
policymakers to change plans.
Rules do not generally exploit the
initial conditions of a maximization problem as much as fully optimal
policies.
If policymakers do not exploit initial conditions today,
then people might expect that they will not exploit initial conditions
in the future.
But of course there is no logical guarantee.
-31-
This
preference for rules over discretion is a practical, rather than a
logical, implication of time inconsistency problems. 9
Another practical reason to prefer rules over discretion is that,
especially with rational expectations, it is difficult to estimate the
impact of alternative discretionary paths with great accuracy.
The
rational-expectations assumption is not accurate unless one can assume
people are familiar with how policy works; this might require that they
have experience with one type of rule for a long period of time. 10
Fischer (1979) has suggested a compromise resolution to the rules
versus discretion debate:
rules should be used in normal times, but in
the case of an unanticipated disaster (such as a financial panic) discretion should come into play.
It is difficult to disagree with this
eclectic solution to the problem, but practical implementation might
prove difficult.
Objective measures of what is normal and what is ab-
normal are difficult to obtain in economics.
A less constructive, but perhaps more realistic resolution to the
rules versus discretion debate comes from deemphasizing the distinction
between the two.
If policymakers make the same policy decision when-
ever their staffs' econometric forecasts are the same, then in effect
9Monetarists who advocate the use of a fixed money growth rule,
suggest that, because of initial conditions (a high inflation rate inherited from the past), the growth rate be diminished to the target
path slowly when starting out on such a plan. There is a time inconsistency argument here. If higher rates of money growth are advocated
because of initial condition, then what is to keep people from expecting a return to high money growth when similar conditions arise
again in the future?
10 Another practical reason is that statistical estimates of
policy effects are considerably less complex if one can focus on rules.
-32-
they are using rules.
The rules might be difficult to describe and
even more difficult to estimate, but they are rules nonetheless.
If
this is a good description of the way policy works, then research which
focuses on alternative rules rather than discretionary paths might turn
out to be the more practically useful type of policy research.
Such
research might suggest ways in which the policymaking process (rule)
should be modified in order to improve the performance of the economic
system.
CONCLUDING REMARKS
This overview has been aimed at recent theoretical research in
stabilization theory.
Earlier research on such issues as the choice of
intermediate targets, problems of lags in the effect of policy, and the
effect of parameter uncertainty on the choice of policy instrument has
been omitted largely because theoretical developments in these areas
have been relatively minor in recent years.
It should be emphasized
that these older problems continue to be of practical importance.
The
continuing efforts to persuade the Fed to switch to a reserve targeting
procedure in their short-run operating strategy is a case in point.
The practical interpretation of these earlier stabilization issues
has been changed in some cases, however, by the theoretical developments reviewed in this paper.
For example, Poole's (1970) analysis of
the choice of policy instrument loses most of its practical relevance
in the mark~t-clearing models where monetary policy is ineffective.
But in the contracting models , where monetary policy effects on real
output are significant, Poole's analysis needs only slight modifications
to account for the rational expectations effects.
-33-
Interest rate
pegging frequently leads to instability in rational expectations models,
whether prices are flexible or temporarily rigid.
This policy impli-
cation, which was emphasized by Sargent and Wallace (1975), appears to
be robust to change in the theory which is used. 11 That many other
important policy implications are not robust to changes in alternative
theories -- as was emphasized here for the policy objective of price
stabilization -- suggests that additional theoretical and empirical
work to sort out and test these theories should be high on any agenda
for future research on stabilization policy.
11 such instability can occur in the model used by Phelps and
Taylor (1977) for example. Because prices are set at levels which
clear markets on average, market-clearing conditions are used to determine expected future prices which in turn are used to determine the
current price setting. Extreme interest rate pegging can make future
prices and hence the current price level undetermined.
-34-
REFERENCES
Anderson, P. A. ( 1979) "Ra ti ona 1 Expectations Forecasts for NonRa ti ona l Models, Journal of Monetary Economics, 5, 67-80.
Azariadis, C. (1975) "Implicit Contracts and Underemployment Equilitrium," Journal of Political Economy, 83, 1183-1202.
Baily, M. N. (1974} "Wages and Employment under Uncertain Demand,"
Review of Economic Studies, 41, 37-50.
Ball, R. J. et al. (1978) Report of the Committee on Policy Optimization, Her Majesty's Stationery Office, London.
Barro, R. J. (1976) "Rational Expectations and the Role of Monetary
Policy," Journal of Monetary Economics, 2, l-32 .
. (1977) "Unanticipated Money Growth and Unemployment in the
United States , " American Economic Review, 67, 101- 15 .
--~-.--,-
. (1978) "Unanticipated Money, Output and the Price Level in
---t~h-e-·united States/ Journal of Political Economy, 86, 549-80 .
1
. (1979) "Developments in the Equilibrium Approach to Business
Cycles," unpublished, University of Rochester.
---=---.-
Barro, R. J. and H. J. Grossman (1976) Money, Employment and Inflation,
Cambridge, Cambridge University Press.
Blinder, A. (1979) "Supply-Shock Inflation: Money Expectations, and
J1.ccommodation," M. J. Flanders and A. Razin (Eds.), Development
in an·Inflationary World, Academic Press, Forthcoming.
Blinder, A. and S. Fischer (1978) "Inventories, Rational Expectations
and the Business Cycle," unpublished.
Brunner, K. , A. Cu ki erman, and A. Meltzer ( 1979} "Stagflation, Macroeconomics and the Permanence of Economic Shocks," CarnegieMellon Working Paper, February.
Suiter, W. (1979) "The Macroeconomics of Dr. Pangloss, A Critical
Survey of the New Classical Macroeconomics," unpublished,
Princeton University.
Ca1vo, G. (1979), "On the Time Consistency of Optimal Policy in a Monetary Economy," Econometrica, 46, 1411-1428 .
• (1980), "Tax Financed Government Spending in a Neoclassical
--~M~o~d-el with Sticky Wages and Rati ona 1 Ex pee ti ons, n Journal of
Economic Dynamics and Control, 2, Forthcoming.
-35-
Calvo, G. and Phelps, E. S. (1978), "Employment and Contingent Wage
Contracts," in Stabilization of the Domestic and International
Economy, K. Brunner and A. Meltzer (Eds.), North-Holland,
Amsterdam.
Cass, D. and K. Shell (1979), "In Defense of a Basic Approach,"
California Institute of Technology, Social Sciences Working Paper
No. 282.
Clower, R. (1965), "The Keynesian Counterrevolution: A Theoretical
Appraisal," in Theory of Interest Rates, Conference on the Theor
of Interest and Money, F. H. Hahn and F. P. Brechling, ,eds. ,
London, Macmillan.
Cukierman, A. (1979) "Rational Expectations and the Role of Monetary
Policy: a Generalization," Journal of Monetary Economics, 5,
213-30.
Fair, R. C. (1979), "An Analysis of a Macroeconomic Model with Rational
Expectations in the Stock and Bond Markets," American Economic
Review, 67, 539-552.
Feldstein, M. (1978) "The Private and Social Costs of Unemployment,"
American Economic Review Papers and Proceedings, 68, 155-158.
Fellner, W. (1976) Towards a Reconstruction of Macroeconomics: Problems
of Theory andPolTcy, American Enterprise Institute, Washington .
~ - ~ - - ,. (1979) "The Credibility Effect and Rational Expectati ans,"
Brookings Papers on Economic Activity, l, 167-178.
Fischer, S. (1977) "Long Term Contracts, Rational Expectations, and the
Optimal Policy Rule" Journal of Political Economy, 85, 191-206.
~ - - - - · (1979) "On Activist Monetary Policy with Rational Expectations," in S. Fischer (ed.) Rational Expectations and Economic
Policy, University of Chicago Press, Chicago (forthcoming) .
. (1980) "Dynamic Inconsistency, Co-operation, and the
--~B-e-ne-vo lent Dissembling Government, Journa 1 of Economic Dynarni cs
and Control, 2, forthcoming.
11
Fischer, S. and F. Modigliani ( 1978) "Towards an Linders tan ding of the
Real Effects and Costs of Inflation," Weltwirtschaftliches
Archiv, 114, 810-833.
Flemming, J. (1976), Inflation, Oxford University Press.
Flood, R. P. and P. M. Garber ( 1979} "An Economic T!1eory of Monetary
Reform," Jou ma l of Political Economy, Forthcomi nq.
-36-
Foley, D. K. and M. Sidrauski (7970), "Portfolio Choice, Investment and
Growth," American Economic Review, 70, (March) 44-63.
Friedman, 8. (1979) "Optimal Expectations and the Extreme Information
Assumptions of 'Rational Expectations' Macromodels, 11 Journal of
Monetary Economics, 5, 23-42.
Friedman, M. (1956) A Theory of the Consumption Function, New York,
National Bureau of Economic Research .
. (1968) "The Role of Monetary Policy," American Economic
--~R_e_v~iew, 58, 1-17.
Frydman, R. (1979) "Sluggish Price Adjustments and the Effectiveness of
Monetary Policy under Rational Expectations," unpublished paper,
New York University.
Gertler, M. (1977) "Rational Inflationary Expectations and the Disequilibrium Versus Equilibrium Methods in Macroeconomics,"
Stan ford Workshop on I nfl a ti on, Discuss ion Paper No. 15.
Gordon, D. F. (1976} "A Neo-Classical Theory of Keynesian Unemployment;'
in K. Brunner and A. H. Meltzer (Eds.) The Phillips Curves and
Labor Market, North-Holland, Amsterdam.
Gordon, R. J. ( 1975) "A lterna ti ve Responses of Policy to External
Supply Shock," Brookings Papers on Economic Activity, 183-204 .
"Can Econometric Policy Evaluation be Salvaged?: A
The Phillips Curve and Labor Markets, eds. K. Brunner
and A. H. Meltzer, North-Holland, Amsterdam.
. (1976)
---~Comment" in
Gramlich, E. M. (1979} "Macro Policy Responses to Price Shocks,"
Brookings Papers on Economic Activity, 125-167.
Green, J. and S. Honkapohj a ( 1979) , "Variance Mini mi zing Monetary
Policies with Lagged Price Adjustment and Rational Expectations,"
Harvard Institute of Economic Research, Discussion Paper Number
721.
Hal1, R. E. (1975) "The Rigidity of Wages and the Persistence of Unemployment," _Brookings Papers on Economic Activity, 2, 301-35.
- - - - ~ .( 1979) "The Role of Preva i1 i ng Prices and Wages in the
Efficient Organization of Markets," unpublished paper, Stanford
University.
Hall, R. and D. Lilien (1980) "Efficient Wage Bargains under Uncertain
Supply and Demand, 11 American Econor1i c Review, forthcoming.
Hansen, L. P. and T. J. Sargent (1980) "Formulating and Estimating
Dynamic Linear Rational Expectations Models," Journal of Economic
Dynamics and Control_, 2, forthcoming.
-37-
Henderson, D. w. and T. J. Sargent (1973), "Monetary and Fiscal Policy
in a Two-Sector Aggregative Model," American Economic Review, 63,
345-365.
Kydland, F. and E. C. Prescott (1977), "Rules Rather than Discretion:
The Inconsistency of Optimal Plans," Journal of Political Economy,
85, 473-492.
.(1980), "Dynamic Optimal Taxation,
--~R-at~,.-.o-n-a'-"'l~E~x_p_e_c~ta-t~i_o_n_s_,-and Control Theory" H Journa 1 of Economic
Dynamics and Control, 2, l, forthcoming.
Lucas, R. E. (1970) "Capacity, Overtime, and Empirical Production
Functions," American Economic Review Paper and Proceedings, 60,
23-27 .
. (1973} "Some International Evidence on Output-Inflation
---T-r-a-deoffs, t American Economic Review, 63, 326-34~
1
.(1975) "An Equilibrium Model of the Business Cycle,"
---o-f--=Political Economy, 83, 1113-44.
Journal
·
.(1976) "Econometric Policy Evaluation: a Critique," Journal
_ _ _o_f,........,_Monetary Economics, Supplement, 19-46 .
. (1980} "Rules, Discretion, and the Role of the Economic
---A.-dr-"v~isor" in S. Fischer (Ed.), Rational Expectations and Economic
Policy, forthcoming.
Lucas, R. E. and T. J. Sargent (1978) "After Keynesian Macroeconomics,"
Federal Reserve Bank of Boston, After the Phillips Curve: Persistence of High Inflation and H1gh Unemployment.
Mccallum, B. T. (1977) "Price-Level Stickiness and the Feasibility of
Monetary Stabilization Policy with Rational Expectations,"
Journal of Political Economy, 85, 627-34 .
---. (1979) "The Current State of the Policy-Ineffectiveness
Debate," American Economic Review, Papers and Proceedings, 69,
240-245.
Mccallum, B. T. and J. K. Whitaker (1979) "The Effectiveness of Fiscal
Feedback Rules and Automatic Stabilizers under Rational Expectations," Journal of Monetary Economics, 5, 171-86.
Modigliani , F. ( 1977) "The Monetarist Controversy or, Should We Forsake
Stabilization Policies?'',American Economic Review, 67, 1-19.
Modigliani, F. and L. Papademos (1978) "Optimal Demand Policies Against
Stag fl a ti on," We ltwi rtschaft l i ches Arch iv, 114, 736- 78 1 •
-38-
Muth, J. F. (1960) "Optimal Properties of Exponentially Weighted Forecasts," Journal of the American Statistical Association, 55,
299-306 .
- - - ~ . (1961) "Rational Expectations and the Theory of Price Movernents," Econometrica 29, 315-335.
Okun, A. (1978) "Efficient Disinflationary Policies," American Economic
Review, Papers and Proceedings, 68, 348-352.
Papa demos, L. ( 1979) "Stagflation and Monetary Policy," unpublished
paper, Columbia University.
Phelps, E. S. (1969) "Short-Run Employment and Real Wage in Competitive
Markets," International Economic Review, 10, 220-232.
_ _ _ _ et al. (1970) Microeconomic Foundation of Employment and
Inflat.lorlTheory, Norton, New York .
. (1978) "Commodity Supply Shocks and Full-Employment Monetary
---Policy," Jou ma l of Money, Credit and Banking, 10, 2, 206-221 .
---. (l978a)
and Monetary
"Disinflation without Recession: Adaptive Guideposts
Policy," Weltwirtshaftliches Archiv 114 .
. (1979) "Introduction: Developments in Non-Walrasian Theory"
---in E. S. Phelps Studies in Macroeconomic Theory, Academic Press,
New York.
Phelps, E. S. and J. B. Taylor (1977) "Stabilizing Powers of Monetary
Policy under Rational Expectations," Journal of Political Economy,
85, 163-190.
Poole, rJ. (1970) "Optimal Choice of Monetary Policy Instruments in a
Simple Stochastic Macromodel," Quarterly Journal of Economics,
197-216.
Prescott, E. S. ( 1977) "Shaul ct Contra 1 Theory Be Used for Economic
Stabilization," in K. Brunner and A. H. Meltzer (Eds.) Optimal
Policies, Control Theory and Technology Exports, North-Holland,
Amsterdam.
Santomero, A. M. and J. J. Seater (1978) "The Inflation-Unemployment
Trade-Off: A Critique of the Literature," Journal of Economic
Literature, 16, 499-544.
Sargent, T. J. (1976) "A Classical Macroeconometric Model for the United
States," Journal of Political Economy, 84, 207-238.
_ _ _ _. (1977) "The Persistence of Aggregate Employment and the
Neutrality of Money," unpublished.
-39-
Sargent, T. J. (1978), "Estimation of Dynamic Labor Demand Schedules
under Rational Expectations, Journal of Political Economy,
86, l 009- l 044 .
- - -• ( 1979) Macroeconomic Theory, Academic Press·, New York.
Sargent, T. J. and N. Wallace (1971) "Market Transactions, Costs,
Asset Demand Functions, and the Relative Potency of Monetary and
Fiscal Policy," Journal of Money, Credit and Banking, 3, 469-505 .
. (1975) '"Rational Expectations' the
Optimal Monetary Instrument and the Optimal Money Supply Rule,"
Journal of Political Economy April 241-254.
--~~~~~--,----;----,-~·
Shiller, R. (1978) "Rational Expectations and the Dynamic Structure of
Macroeconomic Models: A Critical Review, Journal of Monetary
Economics, 4, 1 (January), 1-44.
Taylor, J. B. (1975) "Monetary Policy during a Transition to Rational
Expectations," Journal of Political Economy 83, 1009-1021.
• ( 1979) "Staggered Wage Setting in a Macro Model , " American
---E~c-o-nomic Review, Papers and Proceedings~ 108-113~
_____ . (1979a) "Estimation and Control of a Macroeconomic Model
with Rational Expectations," Econometrica, September, forthcoming .
• ( 1980) "Aggregate Dynamics and Staggered Contracts," Journal
---o~f-Political Economy, forthcoming~
• (1980a) "Output and Price Stability: An International
Comparison," Journal of Economic Dynamics and Control, 2,
February, forthcoming.
-----c,c---
Wallace, N. (1977) "Why the Fed Should Consider Holding M0 f"'.onstant,"
Federal Reserve Bank of Minneapolis Quarterly Review, Summer
Issue, 2-10.
-40-
ON THE EFFECTS OF STABILIZATION
STABILIZATION POLICY
EMPIRICAL EVIDENCE ON
H. Meyer and Robert H. Rasche
Laurence H.
Macroeconometric
Macroeconometric research in the
the 1970s
197Os has been
been dominated
dominated by
by the
the
refinement of large-scale income-expenditure macroeconometric
macroeconometric models,
models,
the
the attenpt
attempt to
to reconcile the
the policy multipliers derived from these
models with
yielded by
simple reduced-forms,
models
with those
those yielded
by simple
reduced-forms, the
the refinement
refinement and
and
estimation of the relation between inflation
inflation and unemployment, and the
the
application of optimal control techniques to macroeconometric models.
These four themes provide the focus for this
this paper.
The
The first section reviews the implications of
of various nacroecomacroecononetric
nometric models for monetary and fiscal multipliers.
We are particuparticu-
larly concerned here with
with the degree of consensus
consensus across models and the
evolution of
of estimated
estimated models over time.
The second section
section discusses
discusses
attempts to reconcile
reconcile the divergent implications of
of income-expenditure
income-expenditure
structural nodels
models and the St.
St. Louis
Louis reduced-form for fiscal policy
multipliers.
In the third section
estisection we develop the implications of esti-
mated
mated Phillips
Phillips curve equations and monetarist
monetarist models for the response
of unemployment, output, and inflation to
to traditional demand management
management
policies.
And in the fourth section we consider
consider the accumulated evievi-
on the results of
dence on
on the gains from policy activism, drawing on
optimal
sinulations with
optimal control
control simulations
with aa variety
variety of
of nacroeconometric
macroeconometric models.
models.
Laurence H. Meyer is Associate
Professor of Economics at Washington
Associate Professor
Scholar at
at the Federal
Federal Reserve
Reserve Bank of St.
St.
University and Visiting Scholar
Louis.
Louis. Robert H. Rasche is Professor of Economics at Michigan
Michioan State
University.
University.
-41-41-
During the last half of the ‘70s
'70s increased attention has been
focused on the way in which economic agents
agents form expectations, particuoarticu11
1
arly inflation
i nfl ati on expectations,
ex pee tat i ans, and on
on “equilibrium”
equil i bri um!! macroeconomic
macroeconomic mode
1s
larly
models
embodying “rational
"rational expectations.”
expectations."
These models yield dramatic
dramatic concon-
clusions about both
both the costs of eradicating inflation and the gains
from activism.
We therefore consider
consider the implications of
of rational
rational exex-
pectation models in both
both the third and fourth sections, although there
is as yet only a small literature
literature on empirical applications of these
upon.
models to draw upon.
AA COMPARISON OF POLICY MULTIPLIERS ACROSS
ACROSS MODELS AND TIME
In this
this section we review the evidence from structural models
models and
reduced-forms about the size and time
time pattern of policy multipliers.
We are interested in the average size
multipliers, the
the consensus
consensus
size of multipliers,
estimated multipliers.
across models, and the evolution over
over time
time in
in the estimated
AA Comparison
Comparison of Multipliers_Across_Models
Multipliers Across Models
(1975) has surrgnarized
summarized the
the consensus across models rather
Christ (1975)
pessimistically:
" ... though models forecast reasonably well over
“.
. .
horizons of four to
to six quarters, they disagree
disagree so
so strongly
strongly about the
effects of
of important
important monetary
and fiscal
fiscal policies
effects
monetary and
policies that
that they
they cannot
cannot be
be
considered reliable guides to such
such policy
policy effects, until it can be dedeof them are wrong in this
this respect and which
which (if any)
any) are
termined which of
right.”
right." (p. 54)
54)
1, 2, and 33 present
Tables l,
present policy multipliers from seven econoeconometric models
models (Bureau of Economic Analysis (BEA), Brookings (B),
(BJ, UniverUniversity
sity of Michigan (MQEM), Data Resources, Inc. (DRI),
(ORI), Federal
Federal Reserve
Bank of
of St.
Louis (St.L),
(St.L), MIT-Pennsylvania-SSRC
MIT-Pennsylvania—SSRC (MPS),
and Wharton
Bank
St. Louis
(MPS), and
Wharton (W))
(W))
—42—
-42-
as reported in Fronin
Fromm and Klein (1976).
The multipliers
multipliers are reported
for the
the first quarter
quarter and fourth, eighth, twelfth, sixteenth, and
twentieth quarters and
and for three policy changes -- an increase
increase in
in real
real
——
government expenditures on
on goods and services, aa decline in personal
money supply or nonborrowed
nonborrowed reretaxes, and an increase in either the money
serves.
The mean
mean and coefficients of
of variation for the
the various multimultipliers are also reported.11
TABLE
TABLE 11
Fiscal
Policy
Fiscal Po
1icy - Increase in
in Government
Government Expenditures
Expenditures
-
Model
IC*
BEA
BB
MQEM
DRI 74
ORI
74
St.L
St. L
MPS
W
w
62
62
561
56I
621
62]
611
61I
621
62I
651
65I
RMSE(4Q)*
6.94
6.94
5.13
13
5.
6.20
4.60
4.98
4.23
4.64
(w/o St.L)
5t.L)
Mean (w/o
St.
St. dev.
dev. (w/o St.L)
s.d./mean
s.d./mean
Mean (w/St.L)
St.
St. dev. (w/St.L)
(w/St.L)
s.d./mean
s.d./mean
Multiplier
lQ
lQ
4Q
40
SQ
l2Q
l6Q
l 6Q
20Q
1.1
l. 1
1.8
7.8
1.4
l. 4
1.3
l. 3
0.5
l. 2
1.2
1.3
l. 3
2.2
2.2
2.8
1.77
l.
2.1
2. l
0.5
2.2
2.0
2.0
2.2
2.2
2.7
1.4
l. 4
2.2
—0.2
-0.2
2.2
2.2
2.4
1.88
l.
1.6
l. 6
2.0
1.0
7.0
1.7
l. 7
-0.2
-0.2
-0.5
2.4
1.33
l.
1.5
7.5
1.1l
l.
1.77
l.
—0.2
-0.2
1.35
l.
35
2.17
2.
17
0.36
0.17
1.93
1.
93
..71
77
0.37
0.24
0.18
1.23
l.
23
.39
0,32
0.32
2.18
2. l 8
0.43
0.20
1.84
l.84
.98
. 98
0.53
2.4
1.0
7.0
2.0
2.0
—0.2
-0.2
0.7
2.6
2.6
1.75
l.
75
0.76
0. 76
0.43
0.43
1.47
7.47
1.01
l. 07
0.69
0.69
IC
initial
*IC=
initial conditions
conditions for
for policy
policy simulation;
simulation;
square error for four quarter forecast of
of
of dollars at 1958 prices) over 1961-1967
*
=
1.37
l. 37
1.03
l. 03
0.75
1.14
7.14
1.11
l.ll
0.97
l.
1.99
1.17
l. 17
0.86
0.74
.97
.97
.94
.94
0.97
RMSE = root
root mean
mean
RMSE
real GNP (billions
period.
=
1The multipliers are reported with
with and without the St.
St. Louis
model multipliers. The latter are based
based on aa reduced-form income equaequation
than on
particularly in
the case
tion rather
rather than
on aa structural
structural model
model and,
and, particularly
in the
case of
of
the fiscal
substantially from the
multipliers based
fiscal multipliers, differ substantially
the multipliers
on
on the structural models.
—43—
-43-
The mean
mean fiscal
fiscal expenditure
expenditure multiplier
multiplier is
is just
just over
1-1/4 in
in the
the
The
over 1—1/4
first quarter and builds to 2—1/4
2-1/4 by the end of year
year two; however,
however, the
cumulative multiplier is still over one after five years.
there
While there
is
about the
multipliers through
three
is considerable
considerable consensus
consensus about
the multipliers
through the
the first
first three
deteriorates sharply.
years, the agreement deteriorates
Note that in all cases the
multiplier
multiplier peaks within
within three years, generally
generally within four to
to eight
eight
quarters; and cumulative fiscal multipliers fall to
to zero or below
below by
the
16th quarter
the fifth quarter for the St. Louis model, by the 12th
12th to 16th
for the MPS model and by
by the 24th quarter for the BEA model.
But it
it
TABLE 22
Fiscal Policy - Tax Cut
—
Model
Multiplier
Multiplier
________
1Q
lQ
4Q
SQ
8Q
12Q
120
l6Q
16Q
MQEM
DRI
DR! 74
St.L*
St.
L*
MPS
Ww
0.4
1.0
0.6
0.9
00
0.4
0.5
1.22
1.
1.6
1.6
1.2
1. 2
1.3
1. 3
00
1.3
1. 3
1.22
1.
1.44
1.
1.6
1.11
1.
1.2
1.2
00
2.1
2. 1
1.7
1. 7
1.1
1.66
1.
1.1
1. l
0.9
00
2.2
1.9
1. 9
0.8
1.5
1. 5
1.22
1.
0.6
Mean (w/o St.
5t.L)
L)
St.
St. dev.
dev. (w/o St.L)
St. L)
s.d./mean
0.63
0.63
0.26
0.
41
0.41
1.30
0.16
0 .16
0.12
0. 12
1.52
1. 52
0.37
0.37
0.24
0.24
1.47
0.52
0.52
0.35
1.25
1. 25
0.47
0.38
Mean (w/St.L)
(w/St. L)
St. dev. (w/St.L)
s.d./nean
s.d./mean
0.54
0.34
0.63
1.11
0.
51
0.51
0.46
1.30
1.
30
0.66
0.51
1.26
1.
26
0.73
0.58
1.07
1.07
0.64
0.64
0.60
0.60
BEA
B
B
1.8
1.6
*
Multipliers reported for St. Louis model are based on
* Multipliers reported for St. Louis model are based on
absence of aa tax variable in
in the model’s
model's reduced-form
equation for income.
income.
-44-
multiplier to reach zero in
takes eight to ten years for the cumulative multiplier
in the Brookings and
and
the Wharton and Michigan models and still longer in
ORI models.22
DRI
The tax multipliers
multipliers are smaller than the expenditure
expenditure multipliers;
multipliers;
they build from an
an initial
initial mean value of
peak of 1.5 at
at the
of 0.63 to aa peak
end of the
the second year.
In the case of aa tax change,
change, there is less
consensus in
deterioration in
in the first quarter, but no
no deterioration
in later
later quarters.
quarters.
The
The tax
tax multipliers tend to peak aa bit later than the expenditure
expenditure
multipliers,
multipliers, generally between
between the 8th and 12th
12th quarters, and then
then
decline.
TABLE 33
Monetary Policy
Policy
Monetary
Model
BEA
ORI
DRI
St.L
MPS
Ww
Multiplier
MV*
RU
RU
Ml
RU
RU
RU
Mean (w/o
(vito St.L)
St.
St. dev. (w/o)
lQ
lQ
4Q
4Q
SQ
SQ
l2Q
l2Q
16Q
l6Q
00
D.3
0.3
1.1
l. l
0.3
0.3
1.44
l.
0.2
4.1
4. l
4.4
3.2
4.5
4.5
0.4
8.3
2.8
2.8
8.4
7.2
7.2
0.7
6.5
1.2
l. 2
12
.4
12.4
8.6
8.6
0.7
2.8
2.8
—0.4
-0.4
14.
14.55
8.0
8.0
0.5
1.24
l.
24
3.0
0.65
0.65
6.08
0.63
0.63
7.05
0.69
6.50
6.50
0.95
0.95
(Ml == narrow money supply;
* MV
MV == monetary
monetary variable
variable (Ml
supply; RU
RU
nonborrowed
as in
non
borrowed reserves; initial conditions same as
I.
Table l.
*
=
2Note also that the fact that the cumulative multiplier
multiplier turns
negative
negative does not guarantee
guarantee a negative long-run multiplier since
sinte these
models are subject to oscillatory convergence
convergence to their
their long—run
long-run values.
—45—
-45-
~
There are only four comparable multipliers for monetary policy
(those using nonborrowed reserves).
The initial quarter
quarter mean
mean multiThe
multi-
plier is small
small and the mean multiplier peaks at the end of
of the third
year at
at aa value of 7.
7.
There is less consensus about monetary comcom-
pared to
to fiscal policy; the coefficient
coefficient of variation is larger in all
but
but one quarter
quarter for monetary policy multipliers.
While the St.
St. Louis
cumulative multiplier peaks in the
the fourth quarter and goes to
to zero by
the 16th
16th quarter, large scale model multipliers generally
generally peak
peak after 88
to
to 12
12 quarters
quarters and the MPS multiplier reported by
by Fromm and Klein is
still rising from the 12th
still
12th to
to 16th quarters.
quarters.
The
The large
large scale models
thus suggest that monetary
monetary policy has aa more persistent effect on outoutput than is the case in
in the St.
St. Louis model.
The exception is
is the DRI
DR!
model in which
which the cumulative monetary policy multiplier
multiplier falls to
to zero
by the 20th quarter.
While the multiplier results do differ across models there
there is
is
clearly considerable
considerable consensus particularly over
over the first two years in
the
fiscal policy
policy when
when we
we exclude the
the St.
St. Louis
Louis results.
results.
the case
case of
of fiscal
The
problem is evaluating how much divergence
divergence in the multipliers is conconsistent with using the models for policy recommendations.
recommendations.
Later we
we
will discuss the use of
of stochastic simulations which
which allow
allow for multimultiplier
plier uncertainty within
within aa particular model,
model.
valuable
approach suggested
by Chow
Chow (1977).
(1977).
valuable approach
suggested by
Here
Here we want
want to note
note the
the
Chow
notes that
Chow notes
that while
while
policy recommendations
from alternative
alternative structural models
models
recommendations derived
derived fron
differ from each other, they may
may nevertheless be closer to each other
than to a passive policy of constant growth rates in the policy instruinstruments.
The comparison Chow
Chow suggests and implements
implements is the improvement
policy derived from
in economic
economic performance in one model using optimal policy
-46-46—
a second model relative to
to the economic performance under
under passive
policy.
Chow uses the
the multiplier properties of
of the Wharton and MichiMichi-
gan models to construct reduced-form
reduced-form equations for real and nominal GNP
including government expenditures and nonborrowed reserves as the
policy instruments and
and employs aa conventional quadratic
quadratic loss function
involving deviations
deviations in
in real
and nominal
from their
targets (in
involving
real and
nominal GNP
GNP from
their targets
(in
each case
average historical
values over
each
case average
historical values
over the
the period
period in
in question).
question).
The results of this
this experiment are mixed.
mixed.
If
If the Michigan
Michigan model
structure and
were the
the true structure
and the policy
policy recommendations
recommendations were derived
from
the Wharton
Wharton model,
model, active
would improve
improve performance
from the
active policy
policy would
performance relarelative to aa passive policy;
policy; costs under the
the active policy would be
be under
25
25 percent of those under
under aa passive policy
policy although
although they would be 70
percent
greater than
if the
the policy
percent greater
than if
policy were
were derived
derived using
using the
the true
true strucstructure.
ture.
hand, if
On the
the other hand,
if the
the Wharton
Wharton model
model were the
the true structure
structure
and
the policy
policy recommendations
were derived
model,
and the
recommendations were
derived from
from the
the Michigan
Michigan model,
the
cost under
an active
policy would
three times
cost of
the cost
under an
active policy
would be
be three
times the
the cost
of aa
passive
policy and
and about
17 times
about 17
times the
the cost
cost when
when the
the true
true model
model was
was
passive policy
used.
used.
And, of course, the Michigan and Wharton
Wharton multipliers
multipliers are quite
close at least for fiscal policies, compared
compared to say the Brookings and
the St. Louis models.
Thus there are other comparisons that would
would lead
to
to even less
less favorable
favorable results
results for
for activism.
activism.
AA Comparison of Policy Multipliers Over Time
We expected
expected to
find aa secular
We
to find
secular decline
decline in
in the
the value
value of
of fiscal
fiscal
multipliers and aa secular
secular rise in
in monetary policy multipliers for large
scale econometric models from the
the late ‘SOs
'60s versions
versions to the versions of
of
the midmid- to late ‘70s.
'70s.
However, published information on such
—47-47-
multipliers
multipliers is relatively scarce and what is available
available is frequently
not constructed on aa comparable basis.
This, of course, increases
increases the
value of the NBER/NSF model comparison studies but makes multiplier
multiplier
comparisons pieced together
together from the literature hazardous.
Perhaps
Perhaps the
most serious problems for comparing multipliers
multipliers across nodels
models or over
time are differences in
specin initial conditions and differences in the spec-
ification of policy instruments, particularly
particularly for monetary policy.
The
large scale models are invariably nonlinear,
nonlinear, implying
implying that
that their multimultipliers are sensitive to initial conditions, particularly
particularly the
the degree of
of
economic
economic slack.
But there is painfully little reported evidence of the
the
degree of this sensitivity.
There are aa bewildering
bewildering number of possipossi-
bilities for aa change in tax rates and even differences in
nultipliers
in multipliers
for different
different government expenditure components.
The
The most serious
problem, however, may be differences in assumptions about the monetary
policy
policy instrument.
instrument.
Monetary
Monetary policy,
policy, particularly
particularly in
in the
the late SOs
60s verver-
sions, has been identified with
with changes in short—tern
short-term interest rates.
with either
In other
other cases, monetary
monetary policy
policy is
is identified ,1ith
either the
the money
money
supply
or some
often nonborrowed
reserves.
supply or
some reserve
reserve aggregate,
aggregate, most
most often
nonborrowed reserves.
The
The
choice affects both monetary
monetary and fiscal multipliers since fiscal
fiscal multimultipliers assume unchanged monetary
monetary policy;
policy; fiscal multipliers will, of
course, be much larger under fixed short-term
short-term interest rates than under
fixed values of the money
money supply or
or nonborrowed reserves.
In Tables 44 and
and 55 we have
have pieced together some policy multipliers
multipliers
for alternative versions of Michigan,
Michigan, Wharton,
Wharton, and MPS models.
The
The
Michigan ‘70
'70 and
and Wharton ‘68
'68 models
models assume
assume constant short—term
short-term interest
Michigan
rates while
while the others
others assume
assume constant unborrowed reserves.
reserves.
It is sursur-
prising (to us at least) that the
the fiscal multipliers
multipliers in
in the
the late ‘60s
'60s
-48-
TABLE 44
TABLE
Real Non Defense Government Expenditure Multipliers - Real GNP
-
QQ
Michigan 70
a Michigan 75b
70a
75b
1.55
1.4
l.
l. 4
c
Wharton 68
68c
2.0
Wharton 75
b
75b
1.3
l. 3
Wharton 79
d
79d
1.1l
l.
e
MPS 69
69e
1.3
l. 3
MPS 75
b
75b
1.22
l.
4
2.1l
2.
1.7
l .7
2.0
2.0
1.77
l.
1.8
l.8
2.2
8
1.99
l.
1.44
l.
2.0
2.3
1.8
l.8
1.6
l.6
2.2
12
12
n.
a.
n.a.
l.
1.00
2.
2.1l
2.6
1.77
l.
1.1
l. l
0.7
,,'
a
"''
aa
S. H. Hymans
Hynans and H. T. Shapiro, "The
“The DHL-III
OHL-III Quarterly Model of the U.S. Economy,”
Economy," Research
Seninar
Seminar in Quantitative Economics, University
University of
of Michigan, 1970, Table
Table 4, p. 22.
bb C.
R.
p.
c
Fromn and L. R. Klein, "The
“The NBER/NSF Model Comparison
Results,” in
Fromm
Comparison Seminar:
Seminar: An Analysis
Analysis of Results,"
in L.
L.
Klein and E. Burmeister (ex), Econometric Model Performance, Pennsylvania, 1975, Table 6,
402.
M. K. Evans and L. R. Klein, The Wharton Econometric_Forecastjj9j~o4~j,
Econometric Forecasting Model, Econonics
Economics Research Unit,
ed. , 1968,
1968, Table 5,
SB.
University of
of Pennsylvania, 2nd ed.,
5, p. 58.
dd Unpublished Wharton multiplier simulations kindly provided by R. M. Young, Wharton Econometrics
Forecasting Associates.
e
DeLeeuw and E. M. Granlich,
Policy,” Federal Reserve Bull
Bulletin,
June
F. Deleeuw
Gramlich, “The
"The Channels of Monetary Policy,"
et in, June
1969, Table 4, p. 489. Shock applied fully to
to federal real wage payments.
three models (including the two with constant short—
shortversions of the three
term rates) are so small; they peak at 2.0 or
or less.
less.
One important
difference in the later versions
versions of
of Michigan and
and MPS models is the
sharp decline
decline in the cumulative
cumulativ2 multiplier from its peak
peak value
value by the
the
12th quarter.
12th
There was aa. tendency in
in earlier versions for multipliers
multi p1i ers
to stabilize
stabilize at
to
at about 7.5-2.0
1.5—2.0 for aa longer period.
This contim:,es
This
continues to
be the case in the Wharton model;
model; in both
both the
the ‘75
'75 and ‘79
'79 versions the
be
fiscal multipliers are stable or rising during the first three years.
We have been able to find comparable unborrowed
multiunborrowed reserves multi-
pliers at different points in
in time for only two models:
pliers
model and the MPS model.
These
S.
These are reported in Table 5.
the Wharton
In
In these
models there is
is aa fairly dramatic evolution of the nonetary
monetary policy
policy
multiplier.
1968 Wharton model the unborrowed reserves multiIn the 7968
multi-
plier for real
real GNP reached aa fairly constant level in
in the 11.5
.5 to
to 2.0
after about one year.
range after
In the MPS model the multiplier
multiplier is stable
In
in the 10.0
70.0 range during the second and third years.
In the later
TABLE 55
Unborrowed Reserve Multipliers
(Real GNP/Nominal Reserves)
68c
Wharton 68
c
0.0
1.5
l. 5
2.1
2. l
1.7
l. 7
l1
44
88
12
72
75bb Wharton 79dd
l,harton 75
Wharton
79
l.4
l.
1.4
1.22
4.5
4.8
4.8
9.1
7.2
9. l
8.6
13.33
13.
Notes - See Table 4.
—
-50So
—
-
69ee
MPS 69
MPS
0.7
5.4
5.4
l10.0
0. 0
12.4
72.4
MPS 75b
b
75
0.3
3.2
8.4
9.4
1ersions
,ersions of both models, the multiplier is continually growing over the
First three years.
Note also the substantial
substantial increase in the size of
‘68 verthe monetary policy multipliers in the
the Wharton model from the '68
ver-
;sion
ion to
to the ‘75
75 and ‘79
79 versions.
versions.
1
1
We view the Wharton ‘68
68 multipliers as
1
fairly typical of the conventional wisdom
wisdom of the midmid- to
to late ‘SOs,
'60s,
prior to the development of the MPS model.
“ST. LOUIS”
COMMENTS ON THE "ST.
LOUIS" EQUATION
Since the original Andersen-Jordan
Andersen-Jordan article
article (1968)
(1968) (AJ) that proproposed aa single equation
equation test of the
the relative importance of
of monetary
monetary and
fiscal policies on nominal
nominal GNP, nunerous
numerous replications
replications have been perper-
formed, across time, across countries, and across functional forms and
and
aa number of
of criticisms, mostly statistical in
in nature, have
have been levied
against
against the equation.
The purpose of
of this section is to review
review the
criticisms that have been raised against the
the equation
equation and to
to evaluate
evaluate
criticisms.
how robust the
the equation
equation appears to be against these criticisms.
The conclusions of the
the Andersen-Jordan investigation are
are by
by now
now
almost universally
almost
universally known.
known.
The
conclusion that
The conclusion
that remains
remains most
most controvercontrover-
sial is the zero cumulative fiscal multiplier
multiplier for nominal
nominal GNP.
This
This
conclusion
conclusion did not conform well to the conventional wisdom of the late
1960s, nor was it
it consistent with other econometric results.
ConseConse-
quently, for the past decade there has been considerable
considerable skepticism of
of
that yields this conclusion.
conclusion.
the specification that
Functional Forms, and_Distributed_Lags
and Distributed Lags_
Time Periods, Functional_Forms,
The
Ad equation was estimated
The AJ
estimated over
over the period 52/1-68/Il
52/1-68/II and subsub-
over the
sequently reestimated by Andersen and Carlson (1970) (AC) over
53/1-69/TV period as part of
53/1-69/IV
of the St.
St. Louis model.
model.
-51—51—
In each case
case
monetary policy had aa powerful and significant effect
effect while the tax
variable (change in
in high employment receipts) was insignificant and exex-
government expenditure
cluded from their preferred regression and the government
variable had only a small
small and transitory
transitory effect.
Silber (1971) subsesubse-
into Republican
Republican (53/I-60/IV) and Democratic
quently split the period into
(61/I—69/IV)
sigthat fiscal variables
variables were sig(61/I-69/IV) administrations and found that
nificant in the latter but not in the former.
former.
Silber argued that these
results are
are consistent
consistent with
more systematic
systematic use
fiscal policy
results
with the
the more
use of
of fiscal
policy in
in
the latter period.
At aa minimum, these results suggest that the
the time
time
period used in
estimation can
can dramatically
dramatically affect the conclusions
in the estimation
and that the estimates
estimates may
may reflect the particular policies pursued over
the estimation period.
More recently
has extended
extended the
More
recently Friedman
Friedman (1977)
(1977) has
the sample
sample period
period
employed
employed by AC through 76/Il
76/II and
and concluded
concluded that “even
"even the St. Louis
equation now believes in fiscal policy.”
policy."
In Table
Table 66 we report the rere-
sults of the
the Ad
AJ and AC equations along with
with estimates
estimates over
over alternate
time periods including Silbers
Silber's two subperiods (Sl and S2),
S2), Friedman’s
Friedman's
extended period (F), and the period 1960/1-1976/Il
1960/1—1976/Il (MR).
The results
The
matter~ The size and sigsuggest that both money and the time period
period matter'.
significance
nificance of
of fiscal policy
policy multipliers
multipliers is
is not
not definitely
definitely settled by
by
these results.
In response to Friedman, Carlson (1978)
(1978) has
has pointed
pointed out that the
first difference form of
of the estimated equation,
equation, while appropriate over
the AC period, is not appropriate over
over the longer period
period because of
of
heteroskedasticity,
heteroskedasticity, implying that the tt values of coefficients
coefficients reported
by
by Friedman are unreliable.
When all variables
variables are defined
defined as
as rates of
two periods are
change, Carlson finds that the results of the two
—52-52-
TABLE 66
TABLE
Time Periods
Time
Periods
A-l
AJ
52/l-68/l
52/1—68/IlI
AC
53/l-69/IV
53/I—69/IV
Sl
51
53/l-60/IV
53/I—60/IV
M
M
5.83
(7.25)
5.57
(8.06)
5.58
(. 43)
(.43)
GG
0.17
0.
17
(0.54)
0.05
(0.17)
-1. 77
-1.77
(.
90)
(.90)
2.36
2.36
(.
67)
(.67)
Sample
'
w
"'
'
T
T
R2
.60
.66
Se
4.01
4. 01
3.84
.652
4.23
F
F
53/1-76/1
53/1-76/IlI
MR
MR
60/1-76/1 I
60/1-76/11
9.20
9.20
( 2. 35)
(2.35)
4.
94
4.94
(6.3)
5. 72
5.72
(1.07)
(1,07)
l. 75
1.75
(2.11)
(2.11)
-3.92
-3.92
(2.78)
(2.78)
l. 42
1.42
( 4. 3)
(4.3)
2.44
2.44
( 5. 57)
(5.57)
-1.67
- l.67
(2.90)
(2.90)
S2
52
61/1-69/1
61/1—69/I
.73
.73
.66
.69
.69
3.30
7.54
7.84
7.84
consistent with
with the
the hypothesis
hypothesis that
that the
the specification
specification is
is stable
stable and,
and,
consistent
like
like the original AC equation,
equation, indicate
indicate that any effect of government
expenditures
expenditures is small and temporary.
temporary.
Allen and Seaks (1979), using the
growth rate specification, find that the
the fiscal variable sums
sums to zero
in both
both Silber subperiods
(Eisenhower and Kennedy-Johnson)
Kennedy-Johnson) but
but is sigsigin
subperiods (Eisenhower
Nixon-Ford era (69/II-77/I).
nificant in the Nixon—Ford
(69/11-77/I).
Over the period 60/1-76/Il
we find that
that both expenditure and tax variables enter
enter significantly
into
both first
difference and
and rate
of change
change specifications.
specifications.
into both
first difference
rate of
In
In Table
Table
77 we
we report the results of the AC equation in
in difference
difference form over both
the original period (AC) and over
over Friedman’s
Friedman's extended
extended period (F) and
and in
rate of change form over Friedman's
Friedman’s extended period (C) along with
with the
Allen-Seaks
the Nixon-Ford
Allen-Seaks results
results over
over the
Nixon-Ford period
period (AS)
(AS) and
and both
both functional
functional
(MR1 and MR2).
forms over the 1960/1—76/Il
1960/1-76/II period (MRl
From
From these results
we
we can conclude
conclude that money,
money, time
time period, and functional form matter.
Ad type
The results of
of AJ
type equations are estimated using
using polynomial
distributed lags.
lags.
This technique requires selection of
of lag
lag length,
degree
degree of polynomial,
polynomial, and
and end point
point constraints.
constraints.
Schmidt
Schmidt and Waud
Waud
(1973) caution
caution that introduction of inappropriate constraints can
result in biased and inconsistent estimates and demonstrate
demonstrate how changes
in degree of polynomial
polynomial and
and end point constraints can
can substantially
substantially
alter the
the conclusions about policy multipliers.
Others
Others have
have found
of lag
length of
lag can affect conclusions also.
We can conclude,
conclude, therefore,
therefore, that
that the choice
choice of time period,
period, funcfunctional form, and
and lag constraints matters aa great deal.
deal.
money appear
very robust.
robust.
money
appear very
The results
results for
The results
results for
for fiscal
policy are
are dramatidramatiThe
fiscal policy
cally affected by these factors.
-54-54-
TABLE 77
Functional
Functional Form
Sample
Form*
'
en
u,
'
AC
F
C
AS
MR1
MR2
AC
53/T-69/IV
53/1-69/1 V
F
53/1-76/11
C
53/1-76/lI
53/1-76/]]
AS
69/11—77/I
69/11-77/1
MRl
60/1-76/Il
60/1-76/11
60/1-76/Il
60/1-76/11
Delta
Dot
Dot
Dot
Delta
Dot
Delta
MR2
M
M
5.57
(8.06)
4.94
4.
94
(6.3)
1.06
1.06
(5.59)
.90
(1.93)
( l . 93)
5.72
5. 72
(5.31)
..75
75
(3.08)
(3.08)
G
O
0.05
((0.17)
0. 77)
1.42
(4.3)
.03
(.40)
(.
40)
.36
(2.07)
( 2. 07)
2.44
(5.57)
..37
37
(2.82)
-1.67
- l . 67
(2.90)
-.29
-.29
(2.25)
( 2. 25)
T
T
FR22
.66
.66
.40
Se
Se
3.84
77,54
.54
3.75
3.75
Delta:
* Delta:
*
Dot:
Dot:
first difference
difference specification
specification
first
rate of change specification
..56
56
.69
.42
7.84
3.02
3.02
Biases Associated With Choices of Independent Variables
Biases
reduced-form multipliers and
The inconsistency between the AJ/AC
Ad/AC reduced-form
the multipliers in large-scale
large—scale econometric models
models generated
generated aa search
(on both sides of the controversy) for an
an explanation.
Monetarists
criticized large-scale econometric models for failing to
to capture the
crowding—out
money demand
crowding-out phenomenon through misspecification of the money
equation (e.g. excluding
excluding a wealth
wealth effect)
effect) and
and failure
failure to
to explicitly
explicitly ininclude aa government financing constraint.
The income expenditure
counterattack focused on
on the unreliability of
of reduced—forms
reduced-forms due
due to
to aa
variety of
of problems, some
some more
more easily correctable
correctable than
than others, associassociated with the choice of
of independent
independent variables.
variables.
been:
The key issues have
What are appropriate measures
measures of the policy instruments?
can
the possibility
possibility of
of reverse
be avoided?
avoided?
can the
reverse causation
causation be
How
How
What biases
are
What
biases are
introduced by omission of nonpolicy exogenous variables?
The Measurement
Measurement of Policy Instruments
There
with specifying the policy
There are two interrelated problems with
instruments.
The first is the problem of specifying the instrument
that the policy authority
authority directly controls.
For example, if
Fed
if the Fed
sets
sets policy by controlling the value of
of the monetary base, employing
employing a
monetary aggregate other
other than
than the monetary base as aa proxy for the
policy instrument may bias the policy multipliers if the other
aggreother aggregate varies endogenously relative to the base.
AA second problem arises
arises
even if the instruments themselves are included if policy itself syssys-
responds to economic developments.
developments.
tematically responds
this case,
case, the
In this
policy instruments themselves become endogenous
endogenous and reverse causation
again may bias the multiplier results.
results.
-56—56—
In this section we take
take up the
in the next the
the
problem of specifying the policy instruments and in
of endogeneity of
of policy.
problem of
noted in
in aa DeleeuwThe problem of reverse causation was noted
DeLeeuw-
Kalchbrenner (1969)
(1969) comment on
on the Ad
AJ paper.
Indeed it was the concern
over this issue that arose
arose out o~the
o• the Friedman-Meiselman debates that
motivated the
the choice of the high
high employment fiscal policy
policy measures
measures by
Andersen and Jordan.
Andersen
Deleeuw and Kalchbrenner’s
Kalchbrenner's main concern is with
with
DeLeeuw
monetary base
base or money supply as
as the variable
variable the Fed
the choice of the monetary
directly controls.
They point out that the choice
choice among the monetary
base, the nonborrowed base, total reserves, and nonborrowed reserves
depends on whether
whether the Fed
Fed offsets
offsets the effect of movements
movements in member
bank
on the base and of movements
bank borrowing on
movements in currency
currency holdings on
reserves.
They
They express no
no special preference among these
these alternate
alternate
measures suggesting only that results which hold for some
some measures and
not for others should be
be viewed with great caution.
caution.
Their
Their empirical
results indicate that fiscal multipliers are affected by the
the choice of
monetary instrument;
instrument; in particular, fiscal multipliers
multipliers of approximately
produced in the MPS model result when nonborrowed
nonborrowed reserves are
the size produced
substituted for the monetary base.
substituted
The treatment of fiscal instruments
instruments in the Ad/AC
AJ/AC equations
equations has
also drawn considerable
considerable comment.
In order to
to avoid the bias
bias associated
tax revenues and expenditures
expenditures
with the income induced movements in tax
transfer payments) under
under preexisting schedules of tax
tax and
(mostly transfer
Ad/AC equations use high employment expenditures.
transfer rates, the AJ/AC
High employment receipts were tried but dropped from the
the preferred
equation
equation due to lack
lack of significance.
The high employment
employment surplus
surplus was
also employed in
in an alternate
alternate specification.
—57—
-57-
is clearly an inappropriate measure of
of stimulus
stimulus assoassoThe latter is
actions because
because it groups components which
which are exexciated with fiscal actions
pected to have different multiplier responses.
The same problem arises
even
in the
high employment
employment expenditures
even in
the case
case of
of high
expenditures because
because this
this variable
variable
includes both
both expenditures on
on goods
goods and services and transfers while
while
economic theory suggests that transfers should be
be netted against taxes.
Suggestions for improved specification of
of fiscal variables have been
DeLeeuw-Kalchhrenner (OK), Gramlich (1977),
(1971), and Corrigan
made by Deleeuw-Kalchbrenner
(1970).
(1970).
Gramlich employs
purchases of
goods and
and services
of goods
services
Gramlich
employs government
government purchases
rather than high
high employment expenditures,
expenditures, and assumes
assumes no adjustment
adjustment is
necessary to purge it of effects of changes in income.
Government
Government exex-
penditures are employed in
in aa composite
composite variable including grants—in—aid
grants-in-aid
and exports
exports with
with an
an adjustment
adjustment introduced for defense inventory
accumulation.
Deleeuw
high employment
DeLeeuw and Kalchbrenner suggest adjusting high
receipts
changes in
in this
endogenous
receipts to
to purge
purge changes
this variable
variable of
of the
the effects
effects of
of endogenous
movements in prices.
Gramlich uses
uses high
high employment
employment net tax
tax revenues
(taxes minus transfers) also
also adjusted along lines suggested by
by DK.
DK.
The
difficulty
difficulty with
with all
all these series
series for
for tax
tax revenues
revenues is
is that
that the series
series
for changes include nonzero entries in periods
periods during
during which no changes
in tax rates or transfer
transfer programs occurred.
occurred.
Corrigan
Corrigan has
has suggested an
alternate
alternate tax variable, the initial stimulus measure,
measure, that indicates
the tax revenues released
released or absorbed by tax
tax rate changes.
This series
This
has
has plenty of zeros~
zeros: For each
each tax, the initial stimulus measure is the
the
change in tax rates times the lagged tax base.
An unweighted sum for
all taxes is the variable
variable Corrigan
Corrigan used and it continues to be
be used in
the New York Fed version of the St. Louis equation.
equation.
-58-
The discussion above suggests that the simple specification
specification of
of
both
equaboth monetary and fiscal instruments
instruments employed in the Ad
AJ and
and AC equa-
tions may be improved
improved upon
upon and that such improvements might alter the
the
relative importance of monetary and fiscal multipliers.
multipliers.
However, the
the
modifications
modifications suggested
suggested above
above have not generally
generally resulted in dramatic
dramatic
changes in the estimated
estimated multipliers in simple reduced—form
reduced-form equations.
While many of these suggestions seem valid, they have not helped to
resolve the differences between the St. Louis equation
equation and econometric
econometric
model
s.
models.
Endogeneity of Policy
Even if
direct policy
policy actions, our estiestiif we obtain measures of direct
mates of their
their effects will be biased if these actions themselves are
systematically
systematically related
related to economic developments.
This
This problem has
has
widely been noted in comments on the Ad
AJ equation,
equation, but most critics
critics inincluding
DeLeeuw and Kalchbrenner considered the problems in measuring
cluding Deleeuw
the instruments the more likely source of bias.
with
with endogenous policy are easy to illustrate.
illustrate.
The biases associated
If
If aa policy instrument
eliminate completely the
varies in response to
to disturbances so as to eliminate
instability in income, the regression of
variof the change in the policy varion changes
able on
changes in income (zero by assumption)
assumption) will yield aa zero coefcoefficient on the policy instrument.
Thus, endogeneity
endogeneity of policy may
result in aa downward
downward bias
bias in the policy multiplier, with the downward
bias
policy.
bi as a funucion
func i.:. ion of the effectiveness
effectiveness of po
1 icy.
We
He can, therefore,
interpret
the zero
multiplier on
instruments as
evidence of
interµret the
zero multiplier
on fiscal
fiscal instruments
as evidence
of
their effectiveness rather than of their insignificance~
insignificance'. While the
endogeneity of
of policy may introduce biases into the estimates of policy
policy
-59—59—
multipliers from both reduced-form equations and structural models,
on the bases of simulation
simulation results
Goldfeld and Blinder (1972) suggest on
that the bias is much more serious for reduced-forms.
reduced-forms.
If
If policy
responds to economic developments with aa lag, the bias
bias is
is reduced but
not eliminated.
Omitted
Variables
Omitted Exogenous Variables
The
The third major source of bias in the choice of
of independent
variables in the Ad/AC
non—
AJ/AC equation is alleged to be the omission of
of nonpolicy exogenous variables.
Andersen and Jordan explained
explained in
in an
an apap-
pendix to their original paper why they believed that the omission
omission of
other
other exogenous variables
variables did not bias
bias their measured impact of the
the
monetary and
and fiscal policy variables:
these variables are presumed to
be independent of monetary and fiscal policies and their average
average effect
effect
is registered in the constant term.
Modigliani
Modigliani (1971)
(1971) made the
the first
detailed
detailed critique of the St. Louis reduced-form model on
on the grounds of
of
(1976) reported aa more exexomitted variables and Modigliani and Ando (1976)
tensive
simulation results
tensive set
set of
of simulation
results supporting
supporting their
their view
view that
that omission
omission
of
of exogenous variables
variables may severely bias
bias the results
results of reduced forms.
The ingenious
ingenious simulation experiments
experiments involved estimation of an
an
Ad
AJ type equation on data generated
generated by
by non—stochastic
non-stochastic simulations
simulations of aa
model.
economy.
economy.
The model represents
represents the known structure
structure of aa hypothetical
The simulated values
values of
of nominal income from the
the model are the
“actual”
"actua 1" values
va 1ues of
of income
income in
in the hypothetical
hypothet i ca 1 economy.
AA reduced-form
is
is estimated
estimated using
using these
these simulated values
values for
for income,
income, and the
the resulting
resulting
estimated multipliers are compared
compared with their “true”
"true" values (the values
implied by
structural model).
by the structural
The comparison
comparison of the reduced—form
reduced-form
-60-60-
multipliers with their “true”
"true" (structural model) values tests the
ability of simple reduced-forms,
only aa couple of
of policy inability
reduced—forms, including only
instruments, to replicate the true
true value of the policy multipliers.
In the 1971 paper, Modigliani emphasized the
the finding that the
of the St. Louis
Louis equation on MPS simulated values yielded
yielded aa
estimate of
money multiplier in excess
excess of
of the “true”
"true" MPS
MPS multiplier
multiplier and reached
reached the
11
“unequivocal
conclusion” that reduced-form money multipliers
unequivocal conclusionu
multipliers are upward
biased.
biased.
This bias was attributed
attributed to positive correlation between the
money supply and omitted exogenous
exogenous variables.
For example,
example, if the Fed
Fed
attempts to stabilize interest
interest rates (as monetarists assert they often
attempts
do), then the money supply will be positively correlated
correlated with real
sector exogenous
exogenous demand variables
variables and the monetary
monetary policy multiplier
multiplier
can be
be expected
expected to
to be biased upward.
Modigliani and Ando (1976) turned their
their attention to
to biases in
in
the estimates
estimates of fiscal effects and suggested that correlation
correlation between
omitted exogenous variables
variables and
and fiscal instruments in this
this case might
account for the small
small size and transitory effects
effects of fiscal
fiscal instruments
Louis equation.
in the St. Louis
Estimates of the AJ
Ad type equation
equation on values
of
of the change in nominal income based on
on simulations
simulations with the MPS model
yield fiscal multipliers like
like the original Ad
AJ equation and contrary
contrary to
structure of the MPS model.
the structure
They concluded
concluded that the St.
St. Louis
approach is "a
“a severely biased and quite unreliable method of estiestimating the response
response of a complex
complex economy to fiscal and
and monetary
monetary policy
actions
actions” (p. 42).
11
To
To demonstrate the role
role of omitted variables in
in the bias
bias in the
Ad equation, they remove any correlation between policy instruments and
AJ
nonpolicy exogenous variables in the structural models by assuming all
-61—61—
nontrended
nontrended exogenous variables are constant
constant at their
their means and all
all
trended exogenous
exogenous variables grow along aa constant
constant trend.
trend.
trended
The predicted
The
value of nominal income for this adjusted structure
structure is
is computed and
used
used to reestimate the Ad
AJ equation.
Fiscal
approFiscal multipliers now
now of aopro-
priate size and magnitude
magnitude confirm the crucial role
role of omitted
omitted exogenous
variables in biasing the estimates of
of the policy multipliers in
in the
initial Ad
AJ equation.
In both papers,
papers, Modigliani
Modigliani and Modigliani
Modigliani and
and Ando (MA) are carecareful to note that the evidence they present does not permit them either
either
to accept the MPS multioliers
multipliers or reject the St.
St. Louis ones.
ones.
But their
results should make those who use St. Louis type
type reduced-form equations
uneasy about the validity
validity of
of the
the multiplier results,
results, particularly those
for fiscal instruments.
omitted variable
variable bias may
may be
While the analysis demonstrates that omitted
aa source of serious inferential
inferential error in the impact of policy actions,
the
to be
be nonconstructive
nonconstructive in
that it
the conclusion
conclusion appears
appears to
in the
the sense
sense that
it does
does
not provide any evidence on
on the particular source
source of
of the bias in the
experiments that were conducted
conducted and it
it suggests abandoning the entire
approach
approach without attempting to investigate the issue of biases
biases in
in the
St. Louis results directly.
St.
It would be useful to identify the
the sources
of bias
bias in the estimated multipliers by introducing the
the most important
directly into the reduced-form equation.
exogenous variables directly
AA number of studies
studies have
have attempted
attempted to address the alleged
alleged biases
in the
the St. Louis approach directly by including nonpolicy
nonpolicy exogenous
variables.
variables.
Gordon (1976), fur
for example,
example, added aa “shock
"shock proxy,’
proxy," concon-
sisting
exports, consumer
sisting of
of the
the sum of net exports,
consumer expenditures on automobiles
and non—residential
non-residential fixed investment
investment to the St.
St. Louis specification.
-62—
-62-
Although monetary multipliers
multipliers decline and fiscal multipliers increase
over his
his longer sample period, the
the multiplier results with and without
the shock proxy remain qualitatively alike; monetary
monetary multipliers are
significantly positive while the
the sum of
of the lag coefficients on the
the
government expenditure variable is not significantly different from
zero.
Recently, Dewald and Marchon
Marchan (1978)
(1978) have estimated
estimated expanded St.
six different
different countries, including
including the United
Louis equations for six
States.
States.
They included exports as aa separate independent variable,
variable, disdis-
missing the
the conglomerate variable constructed by Gordon as including
including
too
too many endogenous influences.
For the United States, the Gordon
result is replicated; the impact
imimpact of monetary policy is reduced, the impact of
of fiscal policy is left
left essentially unchanged, and the exports
exports
variable has aa significant contemporaneous impact.
AA major monetarist
contention
contention is that the influence of aa maintained change in
in the monetary
rate should
should be
be aa proportional change in the
the growth rate of nomnomgrowth rate
inal income.
This
This hypothesis is
is alleged
alleged to be aa universal phenomenon.
phenomenon.
However, while Dewald and Marchon
Marchan cannot reject this hypothesis
hypothesis for the
U.S. data, the monetary response for the
the U.S. is the strongest
strongest of
of any
of the six
six countries
countries investigated.
The long-run elasticities of nomnom-
inal GNP with respect to the money stock in
in the other five countries
never exceed
exceed .5.
In France they found this elasticity
only .07
elasticity to be
be only
.07
and in
in two countries (France and the U.K.) this
this estimated
estimated elasticity is
not
not significantly
significantly different from zero.
zero.
-63-63—
Resolvjflq
hePuzzleLReduced—Fonn
Resolving the
Puzzle: Reduced-Form Versus
Versus Structural
Structural Model MultiDl4~!_
Multipliers
Two further tests by
by Modigliani (1977)
(1977) attempt to
to resolve the
puzzle of conflicting multiplier results.
First of all, he suggests
that despite the apparent large differences in the
the AC and
and MPS multimultipliers, the two sets of multipliers
differenU
multipliers may not be ~ilinificantly different'.
To
To test
test for significance of the difference in multipliers,
multipliers, Modigliani
Modigliani
presumes
presumes that the MPS multipliers are the true ones and tests whether
the AC multipliers differ significantly from the MPS multipliers.
The
The
result
result is that
that they
they are not significantly different
different at the Si)
5.0 percent
level.
Modigliani
Modigliani concludes,
concludes, “This
"This test resolves the puzzle by
by showing
showing
that there is really no puzzle:
the
the two alternative estimates of
of the
expenditure multipliers are not inconsistent, given the margin of error
of the estimates.
It implies that one should accept whichever of two
estimates
more reliable
reliable and
estimates is
is produced
produced by
by aa more
and stable
stable method,
method, and
and is
is
generally more sensible.
To me, these
these criteria call, without question,
question,
for adopting the econometric model estimates."
estimates.” (p. 10)
10)
For
For those who would
would still
still opt for the reduced-form multipliers,
multipliers,
Modigliani compares the
the post-sample prediction performance of
of the AC
equation with one in which the coefficients of
of government expenditures
plus
plus exports were constrained to equal those based
based on multipliers
multipliers dederived
models.
rived from simulations with the MPS models.
begins
197011.
begins in
in 197011.
dominates:
The
The post sample simulation
For
the first
first four
equation
For the
four years,
years, the
the MPS
MPS based
based equation
“distinctly larger”
larger" errors
errors in eight
eight
the AC equation yields "distinctly
only three quarters,
quarters, and results in a
quarters, smaller errors in only
squared
1/3 larger than for the
the MPS based equation.
squared error l/3
Over the
next
“miserably” but
next two years, both
both equations
equations perform "miserably"
but the MPS based
equation is still
still 11“aa bit better.”
better. 11
-64-64-
Conclusion
The income
income expenditure counterattack on
on reduced-forms, particuparticularly the Modigliani-Ando results on
on the implications of
of omitted exogeexogenous variables, and the ability to
to dramatically
dramatically alter the fiscal policy
substanmultipliers by
by choice
choice of time period and functional form, have substan-
on reduced-form
equations for small
small and
tially weakened the case based on
reduced—form equations
transitory
transitory fiscal effects on
on nominal income.
income.
The implied monetary
policy multipliers, on the other hand, have proven robust, at least for
the United States.
ASSESSING THE CUMULATIVE
CUMULATIVE OUTPUT LOSS
LOSS OF
OF ERADTCATING
ERADICATING INFLATION
prominent policy issue of the
the ‘70s
'70s and one that seems certain
AA prominent
'80s is the appropriate policy
policy response
response
to dominate at least the early ‘SOs
to aa prevailing high rate of
of inflation.
inflation.
The view that there is
long—
is aa long-
run trade-off
trade—off between inflation and
and unemployment,
unemployment, widely
widely held
held at the
end of the ‘60s,
'60s, is now held by
by only aa small minority.
The key issues
are the nature of the short—run
short-run relation between inflation and unemunemployment
exployment and the process by
by which economic agents form inflation
inflation ex-
pectations.
Macroeconomic
Macroeconomic models, both
both income expenditure
expenditure and
and none—
mone-
tarist versions, suggest
suggest that while the
the traditional demand management
techniques remain quite capable
capable of
of reducing the rate of inflation,
inflation, the
cost of such
such a policy
policy in
in terms
terms of
of cumulative output
output loss
loss would be
be
great.
Despite
Despite the importance of
of the
the issues, there is substantial
substantial disdis-
agreement about the cost of eradicating inflation
inflation and
and little evidence
evidence
on the benefits
benefits derived
derived as aa consequence.
In this
this section we present
present evidence on
on the cumulative
cumulative output loss
estimated Phillips
associated with
with reducing inflation
inflation based on
on both
both estimated
-65-
curves and monetarist
monetarist models.
Then we discuss
discuss the most serious limitalimita-
tion of these results -- the failure to allow the results to be influinflu--
enced by the
the degree to
to which the public
public believes policy authorities are
committed to aa consistent anti—inflation
anti-inflation policy.
In the final analysis,
analysis,
the cost of anti-inflation policies in the
the form of output loss must be
balanced against
against the benefits associated with aa reduced rate of inflainflation.
tion.
Empirical evidence on the cost of
of inflation and hence the benebene-
fits
fits of
of reducing inflation
inflation is
is quite limited.
limited.
Our discussion
discussion of
of the
the
benefits of
of anti-inflation policies is therefore
therefore confined to
to deterdeter-
mining how large the per period gains
gains would have to be in
in order
order to
justify incurring the cumulative output loss which we
we calculated
calculated from
the Phillips curves
curves and monetarist models.
Econometric Evidence on the
the Size of the Cumulative Output Loss
Loss
Three alternative sources of evidence on
on the cumulative output
loss associated with the use of
of demand management policies to moderate
inflation are discussed
discussed below.
estimated Phillips curves.
curves.
The first is evidence
evidence directly
directly from
Here we
we calculate how long
long unemployment
unemployment
must be increased by either
either 11 percentage point or 33 percentage points
points
to reduce inflation by
by
above the rate consistent with steady inflation to
7.5
7.5 percentage points.
points.
The second and
and third sources
sources use monetarist
models whkh
which include either aa Phillips curve or aa reduced-form
reduced-form equation
relating inflation
inflation to monetary
monetary change.
Here we
we simulate the effects on
inflation
inflation and output of
of aa phased
phased deceleration in monetary growth.
Results Based on Estimated
Estimated Phillips
Phillips Curves
Three recent
recent studies
studies have considered the cost of reducing inflainflation in the context of
of traditional Phillips curve regressions (Perry
—66-66-
(1978), Okun (1978), and Cagan
Cagan (1978)).
Perry’s
Perry's results are based on aa
wage change equation using the inverse of his weiqhted
weiqhted unemployment
rate and lagged wage change estimated using annual observations
observations over
over
the
the 1954-77 period.
period.
‘nonaccelerating
His preferred equation yielded
yi e 1ded aa "nonacce
1era ting
employment (NAIRU) of 4.0 in
inflation rate of employment"
in terms of
of his
his weighted
unemployment
unemployment rate (corresponding to about 5.5 percent in the official
unemployment rate in ‘77):
'77):
unemployment
(1)
Mn W
W
-1.88
nln
-1.88 ++ 7.44 (1/Uw) ++ 0.79 A1nW
tlnW_11 ++ 0.21 A1nW
AlnW_22 ++ 1.07 ONIX
DNJX
(—2.2)
(-2.2) (3.5)
(4.6)
(1.1)
(2.9)
=
S.E. •= 0.70
where
where WW adjusted
adjusted hourly
hourly earnings in the
the private
private nonfarm sector and
and
=
DNIX is aa dummy for the controls equal to
to —1
-1 in 1972
DNTX
1972 and 1973
1973 and +l
+1 in
19744 and
and 1975.
197
Any unemployment rate in excess of
of the critical unemployment
rate, if maintained long
long enough,
enough, will permit aa cycling down
down of inflainflation.
tion.
To compute the
the cumulative
cumulative output loss of eradicating
eradicating inflation,
we
we begin with Mn
Aln PWset equal to
to 10.0
10.0 in the two
two lagged years
years and at
NAIRU.
NAIRU.
Our "moderate"
policy consists
consists of
of increasing the weighted
unemOur
moderate’ policy
weighted unem-
ployment rate 1.0 point above
above NAIRU in period 11 and holding it here
until ~1n
P declines
declines to
\ln W
to 2.5, the
the rate presumed equal to trend
trend growth in
labor productivity and,
and, therefore, consistent with price stability.
stability.
The wage inflation rate falls from 10.0 to 9.6 percent
percent in the first
year and declines about 0.3 percentage points per
per year thereafter
year
taking 23 years to reach aa 2.5 percent rate.
rate.
An alternative "radical"
An
radical
policy is modeled
modeled as aa 33 percent point increase in unemployment
unemployment beginbeginning in period one and again sustained until
until wage
wage change declines
declines to
—67—
-67-
2.5 percent.
percent.
This takes
~k 1111 years:
takes~
years!
Note
Note that the nonlinearity
nonlinearity in
in
Perry’s
Perry's wage equation ensures that
that the cumulative excess of person
years of unemployment and, hence, cumulative
cumulative output loss will
will be
greater in
in the more radical policy case.
Using Okun’s
perOkun's estimate of 3.2 as the impact on output of aa 1l percent
cent point increase in
in unemployment,
unemployment, we can convert
convert the excess unemunemployment
unemployployment into
into output loss. 3 One
One percentage
percentage point increase in unemploy~
ment
ment reduces output 3.2 percent or $45.6 billion dollars (calculated
(calculated at
1978 value for real potential GNP).
The
The 33 percent point increase in
an initial year output loss
unemployment involves an
loss of $136.7
$136.7 billion.
To
find the
undiscounted output
potenTo find
the cumulative,
cumulative, but
but undiscounted
output loss
loss we
we assume
assume potential output will rise at a 3.3
3.3 percent rate.
rate.
This yields
yields aa cumulative
loss of $1532.6 billion for the moderate policy and $1778.0
$1778.0 billion
billion for
4
4
the radical
policy.
The discounted output loss is essentially
radical policy.
essentially the
initial year
year loss and the number of years
years required to
product of the initial
complete the program (not accounting for the 3.3
3.3 percent rate of growth
growth
potential output is the same as discounting by aa 3.3
3.3 percent
percent rate);
in potential
the discounted losses are $1047.9 billion and
and $1503.6
$1503.6 billion in
in the
modest and radical cases, respectively.
Charts l1 and 2.
The results are depicted
depicted in
in
(Perry 1 refers to
to the
the moderate case and Perry 22 to
the
the radical
radical case.)
3Estimation of
sugEstimation of the Okun law relation over more recent data suggests that 3.2 may be an overestimate of the output loss associated
with aa one percentage point increase
increase in unemployment;
unemployment; the
the recent estiestimates
mates are about 2.5.
41f
1f the Okun’s
Okun's law coefficient is 2.5 instead of
outof 3.2, these outlosses should be reduced by
by about 20 percent.
put losses
-68-
PP or
or WW
CHART I1
11.0
n.o
CHART
MODERATION IN INFLATION VIA MONETARY DECELERATION
MODERATION
IN INFLATION VIA MONETARY DECELERATION
10.0
9.0
9.0
8.0
7.0
6.0
5,0
5.0
"''
<.O
4.0
PERRY I
'
3.0
PERRY 2
2.0
1 .0
1.0
0.0
CAGAN
cAGAN
-1
.0
-1.0
-2.0
-3.0
-3.0
1
2
3
3
4
55
6
77
88
99
10
11
12
13
14
15
16
17
10
19
20
21
21
22
23
23
2
IEAR
Okun finds that
that aa variety of estimated
estimated Phillips curves (PC5)
(PCs) in
in
the literature
quantitatively similar
similar conclusions.
the
literature yield
yield quantitatively
conclusions.
The six
six
The
equations considered by Okun
Okun yield aa first year reduction
reduction in
in inflation
of from 1/6
1/2 percentage point
l/6 to l/2
point and an average of
of 0.3 percentage
points for aal1 percentage point increase in unemployment.
Gramlich
Gramlich
(1979)
similar conclusion.
(1979) reached aa similar
There are
are two
two aspects
of the
which deserve
aspects of
the Perry
Perry specification
specification which
deserve
There
unemfurther discussion: expectations are formed adaptively and
and the unemployment
ployment rate enters nonlinearly.
The
The Phillips curve
curve is uniformly
uniformly
drawn as a nonlinear
nonlinear relation and
and there have been
been aa number of
of theorettheoret-
ical explanations (including
(including Lipsey and Tobin) and some empirical supsupport
port (Perry’s
(Perry's influential
i nfl uenti al l1966
966 study, for example).
However, nonlinear
nonlinear
and linear specifications
specifications seem
seem to do about as
as well over sample
sample through
existence of nonlinearity would provide
the mid-197Ds.
mid-l970s. 55 The existence
provide aa rationrationale for the gradual as opposed to radical policy approach; the
the greater
the nonlinearity, the
the greater
greater the cumulative output
output loss
loss under the
radical as opposed gradual policy.
The inflation inertia implicit in the Perry equation derives from
two sources:
sources: actual inflation is
is built into
into expected inflation
inflation with aa
lag and actual inflation responds gradually to unemployment in
in excess
of the critical
critical rate.
To the extent that the lag in incorporating
incorporating
actual
negotiations is
actual inflation
inflation into
into future
future wage
wage negotiations
is long,
long, indexation
indexation
might substantially
substantially reduce the
the inflationary inertia.
ation, there would be aa lag.
Even with
with index—
index-
Assuming that the full effect occurs
5Cagan
cagan (1977) has recently noted the surprising
surpr1s1ng lack of evidence
of nonlinearity
nonlinearity and this has been
been confirmed in aa careful examination
examination by
Papademos
Papademos (1977).
(7977).
—70—
-70-
rL
EL
2,200
CUMULATIVE OUTPUT LOSS BASED ON PERRY &
CURVES
& CAGAN PHILLIPS CuRVES
CHART 2
2,000
2,000
1 ,800
1,800
PERRY 2
1,600
1 ,600
1.400
1 ,400
1,200
1 .200
PERRY 11
1 ,000
1,000
'
'
___,
800
600
400
CAGAN
200
00
-200
-400'
—400
-600
1
22
3
4
5s
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
YEAR
24
cumulative out—
outwithin the first year would not dramatically reduce the cumulative
put costs.
The cumulative output loss would
would decline about 20 percent
percent
in each case.
Thus, the critical determinant
determinant of the gradual decline
decline in
in
inflation is the extremely
extremely small per period deceleration
deceleration in
in inflation
inflation
associated with labor market disequilibrium (excess unemployment) in
in the
conventional Phillips
Phillips curve, not with the slow response of inflation
inflation
expectations to
to changes in
in the actual inflation rate.
Cagan develops a PC equation beginning with the natural rate
specification and assuming adaptive expectations
expectations Cagan’s
Cagan's estimated PC
PC
is:
(2)
( 2)
Pt
=
u—u
ut - ut 22
Pt_i
0.95 (~
~ )
Pt-i - o.95
2 - l
-
u +u
(ut
+ ut-l
- 0.23 ( ~ t- 3
-
+ ut-2
t-2 --
ul
where PPis
is the quarterly rate of
of change in
in the CPI, uu is the unemployunemploy-
ment rate for prime age males and
ii
u is estimated
estimated from the constant
constant of
the regression
regression (3.7,
(3.7, for this
this regression)
regression) and
and the equation is estimated
estimated
1953—1977.
over the
the period 1953-1977.
using quarterly observations over
As is clear in Charts 1
1 and 2, the Cagan equation generates
generates aa
dramatically more
more rapid decline in inflation
inflation and smaller
smaller cumulative
cumulative
output loss.
0 at a 7.5 percent inflation rate (in
Beginning in period Oat
the current and last
last period) and at NAIRU, aa one percentage point
point inincrease in the unemployment rate reduces inflation by the full 7.5 perper-
centage points by
by the eighth year with cumulative output loss of $4.2.9
billion, about aa quarter
quarter of that associated with
with the Perry and Okun
results.
—72—
-72-
Evidence
Evidence Based on the St. Louis Model
To provide additional evidence on
on the output
output effects of using
using
stabilization
stabilization policy to reduce inflation, we ran
ran simulation
simulation experiments
with
with the St. Louis model.
model. 66 We begin
begin with a base run in
in which the rate
rate
of monetary growth is at aa steady 7.5 percent
percent rate beginning
beginning in 1968/111
of
1968/TI!
through 1978/IV.
1978/TV.
This builds in inflation inertia and provides the
debase against which we can evaluate the effects of gradual monetary de-
celeration.
Beginning in
in 1973/I
1973/l we gradually decelerate monetary
growth by l1 percentage point in the first quarter of each year.
We
then compare the policy runs with base
base run and compute the cumulative
output loss associated with the policy.
The first set of simulations with
with the St. Louis
Louis model employ the
l953/I-78/IV.
version of the model estimated
estimated over the
the sample
sample period 1953/1-78/IV.
The general practice
practice at the Sank
Bank is
is to
to employ
employ the estimates
estimates of the
model using all
all available data
data for forecasting and policy simulations.
78/TV, however, has aa very large
The version estimated
estimated through 78/IV,
large coefficoefficient on the
the demand slack variable in the model’s
model's Phillips curve,
curve, almost
three times the size of the coefficient estimated with data through
through
71/I! or 75/1,
75/I, for example.
71/ll
by the
the lines labeled StL1.
StLl.
The results are reported in
in Charts
Charts 33 and 44
There
There is a rapid deceleration in inflation
and aa low cumulative
cumulative output loss.
The inflation rate begins to decline
very slowly; it takes two years
years to reduce the inflation rate by Il perpercentage point.
Thereafter the deceleration speeds up so that after
66For a description of
of the St.
St. Louis model
model,, see Andersen
Andersen and
Carlson (1970). The model includes aa reduced—form
reduced-form equation for nominal
income and aa Phillips curve equation for price change; output
output is then
solved for via an identity.
-73—
-73-
5-1/2 years,
years, inflation has declined
declined by 7.5 percentage points.
5—1/2
unThe un-
employment rate
rate rises slowly at first and the maximum
maximum increase is
is only
only
1.8 percentage
percentage points, during the
the sixth
sixth year.
The cumulative
cumulative output
loss
loss is
is only about $200 billion.
The output loss is, of course, sensitive
sensitive to the coefficient
coefficient on
the demand variable in the Phillips
Phillips curve.
Using aa version of the
model estimated
estimated through
through 71/111, where the coefficient on
on the demand
smaller than in the first version discussed,
variable is substantially smaller
inflation decelerates much more gradually;
gradually; after six years the inflainfla-
points below that
tion rate in the policy run is only four percentage points
in
in the base run.
At this point unemployment
unemployment is
is four
four percentage points
points
higher than in the base run.
loss is $350
The cumulative output loss
billion at this point and escalating rapidly.
These
These results are dede-
Charts 33 and 44 by the lines labeled Stl2.
picted in Charts
StL2.
Evidence
Evidence Based on Reduced—Form
Reduced-Form Equations
7
Given reasonable doubt about the
the validity
validity of the
the Phillips
Phillips curve, 7
it
it is useful to consider the implications of
of reduced—form
reduced-form models that
are not
not tied directly to an explicit Phillips
Phillips curve.
curve.
We consider two
examples: Stein’s
Stein's (1978) two equation model of inflation and unemployunemployment and AJ
AJ type equations for nominal
nominal income
income and inflation.
The
The
results are depicted in
in Charts 33 and
and 44 by
by the lines labeled Stein
(Stein 11 for the moderate case and Stein 2
2 for the radical case) and
and
StL3.
77 See,
see, for example, Stein (1978).
—74-74-
VIA MONETARY DECELERATION
INFLATION VIA
IN INFLATION
MODERATION IN
MODERATION
MONETARV DECELERATION
CHART
CHART 3
Pp
10.0
9.0
9.0
8.0
77 .0
6.0
""''
'
STL
22
'~
5,0
5.0
ST L~ ~.
4.0
STL I
3.0
3.0
STEIN
STEIN II
2.0
STEIN
2
11 .0
00
1
22
33
44
5
66
77
88
9
10
11
12
12
YEAR
YEAR
Stein model -· In the Stein
Stein model
model,, both
both unemployment and
The Stein
--
inflation are driven by
by the rate of
of monetary growth.
Stein’s
Stein's two
equation model is:
is:
u(t)
(3)
A
6
(4)
At ~t)
dt)
=
33 · 0.6
0.6 u(t-1)
-
+
+
0.4
(t-i)
(t-1) · 0.4
0.4 ~
µ7 (t-i)
(t-ll
c
-
•~04~
Ul (t~i)
·0.4 TT (t-i)
(t-1) ++ 0.4 lll
(t-1)
is the unemployment rate,
where uu is
rate,
rate
rate of monetary growth.
w
·rr
is the
the inflation rate and
and ~
JJl is
is the
The critical unemployment
unemployment rate is 5.0 and
and the
equilibrium rate of inflation is the rate
rate of monetary
monetary growth.
ning at u
u
=
5,0 and ~(t)
+(t-l)
5.0
rr(t) == TT(t-1)
=
=
BeginBegin-
7.5 == ,,7(t)
p~(t) = ~~
we
µ7(t-l),
we deceldecel1(t-1),
erate the rate of monetary growth either (a)
(a) gradually by 11 percentage
point per
~
per year until µ7
=
=
0 or (b) immediately to
to 0.
In the gradual
policy, unemployment rises beginning
beginning in year 22 and peaks in year 8 at
6.6 percent returning to almost 55 percent
16.
percent by
by year 16.
The
The inflation
inflation
rate
rate begins to
to decelerate in year
year 22 initially at aa 0.4 percent point aa
year rate but ultimately reaches 1.0
1.0 point per year by year 7.
7.
The
The
inflation rate is
is down to 22 percent by year 88 and thereafter declines
gradually
gradually to about zero
zero by
by year 16.
$687.5 billion.
billion.
The
The cumulative output loss is
Interestingly, the gradual policy incurs aa smaller
smaller
cumulative
loss, $613
$613 billion.
cumulative output
output loss,
The St.
St. Louis reduc
reduced-form
equation for income
income with aa reduced—
reduced—for~euatjon
A second simulation
equaform for inflation -- A
simulation based on reduced-form
reduced-form equa—-
tions combined the reduced—form
reduced-form for nominal income
income in
in the St. Louis
model with aa reduced-form equation for inflation.
inflation.88 The inflation
88The reduced-form equation for
The
inflation used in
in this section was
developed
developed by
by Jack Tatom of the Federal Reserve Bank of St. Louis. An
earlier
“Does the Stage
earlier version of
of this equation was used by Tatom in
in "Does
of
Rate?” Federal Reserve Bank
of the
the Business Cycle
Cycle Affect the Inflation Rate?"
of
of St. Louis Review,
Review, September
September 1978, pp. 7-15.
—76—
-76-
700
CHART 4
CU*JLATIVE
CUMULATIVE OUTPUT LOSS ASSOCIATED WITH NONETARY
MONETARY DECELEMTION
DECELERATION
600
STEIN 11
STEIN 2
500
400
""'
'
STL 33
STL
300
200
STL 1
100
100
0
2
3
4
55
6
77
8
9
10
11
12
13
14
15
16
17
18
19
YEAR
20
reduced-form includes aa twenty period distributed lag on the rate of
of
reduced-form
change in the money supply and aa four quarter distributed
distributed lag
lag on
on the
differential in
in the rate of change in producer prices for energy and
the price index
index for the nonfarm business
business sector, and two dummies for
the effects of the freeze and
and Phase II and for the subsequent
subsequent catch up
effects.
The St. Louis equation yields
yields values for nominal income;
income; the
the
inflation reduced
reduced form is employed to generate price level
level predictions;
and the price level is used to
to deflate nominal
nominal income to
to yield real
output
output predictions.
The results in
in Charts 33 and 44 depicted
depicted by
by the line
labeled
labeled StL3, reflect the response to the same phased monetary decelerdeceleration
St. Louis
Louis model
described
ation employed
employed with
with the
the other
other St.
model simulations
simulations described
above.
Note the similarity with the St.
St. Louis
Louis results with
with aa Phillips
71/Il), StL2,
curve (based on the
the sample period through 71/11),
StL2, in
in Charts
Charts 33 and
4.
With the reduced-form equation
equation inflation
inflation declines more rapidly,
rapidly, by
about
.20 - .30
.30 percentage
over most
period; corabout .20
percentage points
points per
per year
year over
most of
of the
the period;
cor-
respondingly, the output loss is somewhat
somewhat smaller.
But the time
pattern and magnitude
magnitude of
of both the deceleration
deceleration in inflation and the
cumulative output loss are remarkably similar.
Again note that the
output loss per quarter has not peaked after six years of the phased
phased
deceleration
deceleration so
so that the cumulative output
output loss
loss is still rising raoidly
rapidly
at the end of six
six years.
Qualifications of the Empirical
Empirical Analysis
Analys_is_
Qualifications
The results reported above are derived both
both from explicit
explicit
Phillips curves, and from monetarist reduced-forms.
The existence
existence of aa
cumulative output loss associated with eradicating inflation is
-78-
therefore
therefore generally consistent with
with both income-expenditure
income-expenditure structural
models and monetarist reduced-forms.
reduced—forms.
The
emThe major deficiency of the em-
pirical analyses on which the
the results described
described above are based is the
failure to
to allow
allow the public’s
public's perception of current
current and future policy
policy
to affect expectations about future inflation.
inflation.
The Credibility Effect
The results reported above based on Phillips curves all related
inflation in the current period to aa distributed lag on past inflation
rates where the latter are intended to reflect the rate
rate of inflation
expectations
exexpectations (and/or direct the influence of past inflation as
as for example via catch—up
catch-up effects).
This specification does not allow
allow the
degree of
credibility associated
with announced
anti—inflation policies
degree
of credibility
associated with
announced anti-inflation
policies
or
or even the
the expected
expected influence
influence of recent
recent policy actions to influence
influence
inflation expectations.
The estimates
estimates of cumulative output loss gengen-
erated
overerated by
by such
such models
models are,
are, therefore,
therefore, almost
almost certain
certain to
to be
be over-
estimates.
Fellner (1979),
(1979), for example, maintains
maintains that
that" ... the
...
standard model coefficients
coefficients...
... would
would change significantly
significantly for the
the
better -- in
in the
the direction
direction of
more rapid
reduction of
better
of aa much
much more
rapid rate
rate of
of reduction
of
--
inflation
given slack
inflation for
for any
any given
slack -- if
if aa demand
demand management
management policy,..
policy ...
--
changed to aa credible policy of consistent demand disinflation."
disinflation.”
But
by
much does
does the
overestimate inflationary
by how
how much
the standard
standard model
model overestimate
inflationary inertia?
inertia?
By 10
10 percent, 50 percent?
We do not have any reliable quantitative
quantitative estimate of the degree
deoree
to which
the deceleration
deceleration of
of inflation
to
which policymakers
policymakers can
can speed
speed the
inflation by
by
clearly defining their anti—inflation
anti-inflation policies and convincing the
public that they intend
intend to
to follow through.
—79—
-79-
Nevertheless,
Nevertheless, there would
be nearly universal agreement that anti—inflation
anti-inflation policies ought to be
be
set out Clearly
clearly and supported by both the Treasury and the
the Federal
Federal
Reserve in such aa manner as to maximize
maximize the Credibility
credibility effect.
Rational
Rational Expectations and the Cumulative Output Loss
of rational expectations models advocated,
In the extreme form of
for example, by
by Sargent and Wallace (1976),
(1976), the cumulative output loss
associated with aa credible
credible policy of monetary deceleration should be
zero.
zero.
These models have two essential features:
1) they are equilibl)
equilib-
rium models in which prices respond immediately and fully to monetary
monetary
change and real variables
variables such
such as
as unemployment
unemployment and output
output respond
resoond only
only
to unanticipated inflation; and 2) inflation expectations are formed
rationally, taking into
into account knowledge both about the structure
structure of
the economy and the systematic features of policy.
In such aa model, inflation should
should moderate imediately
immediately in
in reresponse to
to the monetary
monetary deceleration,
deceleration, provided, of course,
course, that
that the
policy was announced in
in advance and believed (or otherwise
otherwise expected).
We had
had thought
thought of running
running simulations with an
an RE version of the
the St.
lines suggested
suggested by
by Andersen (1979).
Louis model along lines
On aa moment’s
moment's
sufficiently obvious that computer
computer
reflection, the implications were sufficiently
simulations could be dispensed ,dth.
with.
The St. Louis model has
has aa
Phillips curve in which inflation
inflation depends on aa demand variable (x) and
expected inflation (Pe)
an adaptive
adaptive
expected
(pe) where the latter is determined from an
expectations
expectations model with weights taken from aa regression of
of the nominal
interest rate on past inflation rates:
(5)
p
P
=
+
Sx
+
~P
-80-80-
Andersens RE version imposes the condition that Pe
Andersen's
=
E(P);
E(P); i.e.,
i.e., that
that
subjective inflation expectations equal
inflaequal the model’s
model's forecast
forecast for inflation.
In this case:
(6)
C
E(P)a+Sx+eE(P)
a + sx + sE (Pl
(P)
( 6' )
(6’)
E
E (P)
C
l
(a+ S x)
l~E(sx)
l -s
and Andersen substitutes
substiti~tes
(7)
p
=~(a+
S x)
1-c
for the St. Louis
Louis Phillips curve.
curve.
= .86,
its value in the St. Louis model.
Andersen sets c~ =
HowHow-
ever, if
if cc is meaningfully viewed in this case as
as the coefficient on
expected
expected inflation,
inflation, the value of .86 estimated in the St.
St. Louis
Louis model
should not be
be accepted as the magnitude
magnitude of that
that parameter
parameter in the
the RE
RE
version of the St.
St. Louis model because the value of
of cc was estimated
estimated
under
under the assumption that expectations were formed
fanned adaptively.
Taking
1, as seems essential to
to the RE model,
model, equation 77 no longer
longer is aa
c == l,
meaningful equation
equation for
for P.
P.
meaningful
(6’)
(6')
Instead we
obtain from
c = l
Instead
we obtain
from (6)
(6) where
wheres=
0 = ct + sx
O=a+sx
so that there is aa unique value of x*
=
- a/s
s/B corresponding, of course,
course,
to the natural
natural rate of unenployment.
unemployment.
xx can differ
differ from x* only on
on
—
account of random disturbances (with zero
zero mean).
In
In this
this case any
effect
effect of monetary deceleration on
on the rate
rate of growth of nominal income
transformed immediately and fully into
is transfonned
into aa decline in
in inflation
inflation
without any cumulative output loss.
This
This seems to us a more
-81-81-
meaningful RE version of the
the St.
St. Louis
Louis model than
than that employed by
Andersen. 9
Andersen
Balancing the Gains from Reducino Inflation~g~jpstthe Transitional
Balancing the Gains from Reducing Inflation Against the Transitional
Costsi0
Costs ~
The cumulative output loss is
is aa measure of the cost of
of antiinflation policies.
To
To evaluate the desirability of
of such
such policies we
we
also need to assess the gains from reducing inflation.
Unfortunately,
the costs of inflation
inflation (and hence the
the benefits
benefits of reducing inflation)
inflation)
are
not as
or easily
easily quantifiable
cost of
unemployment.
are not
as clearcut
clearcut or
quantifiable as
as the
the cost
of unemployment.
Fischer
Fischer and Modigliani (1978) provide aa careful
careful outline of the costs of
inflation.
The costs include the
the welfare loss
loss associated with the
incentive
reduction in
capital acincentive to
to economize
economize on
on cash
cash balances,
balances, the
the reduction
in capital
ac-
cumulation due to disincentives for saving
saving and investment that reflect
after—tax
the way in which the tax system
system permits inflation to affect after-tax
9There
is aa second
second and related objection
objection to
to Andersen’s
Andersen's approach.
In the
the St.
St. Louis model ca is not the sum of the coefficients
coefficients on
on lagged
inflation
the sum of the
inflation rates. Indeed thesum
the coefficients
coefficients is generally
generally about
1.0. The reason for this
curve does
1.0.
this is that the St.
St. Louis Phillips
Phillips.curve
does not
estimate the weights on
on lagged inflation
inflation directly
directly within the estimation
of
of the Phillips
Phillips curve itself. First,
First, an
an equation
equation for aa short-term
interest
interest rate
rate is estimated as
as aa function of
of the rate
rate of
of monetary
monetary growth
growth
and distributed lags on
on both
both the rate of change in
in output and on
on past
past
inflation
inflation rates divided by the ratio of unemployment
unemployment to the full—
fullemployment rate. The sum
sum of
of the
the coefficients on lagged prices from
from the
interest rate equation in the original Andersen/Carlson
Andersen/Carlson article was
1.27
1.27 so
so the sum of weights on
on lagged inflation
inflation rates in the Phillips
(1.27/(u/uf)), approximately 1.0. The sum
curve is .86
.86 (l.27/(u/uf)),
sum of the
the inflainflation coefficients from the interest rate
rate equation
equation vary
vary considerably
considerably
estimate of
over different sample periods
periods and the
the estimate
of ac always
always compensates
compensates
to yield
yield a sum
sum on past inflation rates of about
about 1l .0. This
This reinforces
our view that
that the value of as in equation (6) should be taken as 1.0.
1.0.
10This section was added to the original paper and was motivated
lOThis section was added to the original paper and was motivated
by
tzer at
by comments by Jerry Jordan and Allan Mel
Meltzer
at the conference.
conference.
-82-82-
redistriburates of return and the cost of capital, and the arbitrary redistribution of income and wea1th
due to unanticipated inflation.
wealth due
While Fischer and Modigliani
Modigliani do
do provide estimates of
of some compocompo-
nents of the costs of inflation,
inflation, neither their study nor others
others permit
permit
us to compute aa meaningful estimate
estimate of
of the benefits that would
would accrue
from reducing inflation
inflation which
which could in turn be compared
compared with the cost
cost in
in
terms of cumulative output loss.
What we
we can compute is the minimum
minimum
size of
of the permanent gain
gain in
in output per year due to
to eradicating
eradicating inflainfla-
tion which would just justify incurring the cumulative
cumulative output loss assoassociated
ciated with the transition to price stability.
per year.
benefits as a gain in real output per
We will refer to the
the
of the gain
Some components of
may, however, be
be welfare or
or utility gains that would
would not necessarily
necessarily
show up in computed measures
measures of real output.
While
While such welfare gains
gains
difficult to evaluate
evaluate than
than output gains, they are no
no less
are even more difficult
in developing
developing aa measure of the benefits of reducing inflation.
inflation.
important in
Figure l1 depicts
depicts the comparison we wish to make.
Figure
dashed XX
The dashed
line is the rate of growth of (potential) output if
if inflation
inflation remains
Figure
Figure 11
x
X
—
t
n
0on
-83—83-
indefinitely at
at 7.5 percent.
percent.
If anti-Inflation
anti-inflation policies are
are pursued,
If
output is assumed to follow the solid
solid line.
occur between
between t
=
0 and t
O
=
The transitional costs
nn as unemployment rises above the rate
associated with potential output.
However, if
if there are costs of
of inin-
flation, output will rise above the level that would have prevailed
prevailed if
the initial
initial steady inflation rate
rate had continued.
continued.
We define
as the
define GGas
present value of
of the permanent per period output gain,
gain, evaluated from
period nn to
(8)
(8)
oo
GG = r
i=n (l+r)’
This can be
be compared
compared to
to the
the present value of the cumulative
cumulative output
output loss
(L)
( L)
(9)
LL
=
n—l
L.
n-1
z
E
i=O (l+r)1
i0
where L~
Li is the output 1loss
ass in the ith
ith period ((i=0,
i =0, . . . n—l).
n-1).
. . .
Assuming
Assuming that the unemployment rate is maintained above the rate
consistent with potential output by aa fixed amount for nn periods,
periods, the
the
loss in period i can be
be expressed
expressed as
((10)
l O)
L~
E
(1+p)1
where [U is the loss in the first period and oa is
is the
the rate of growth in
potential
potential output.
((10’)
10' )
If r=p, the expression for LL simplifies to
L
L == n[
nTi
This is precisely
precisely the
the way
way we calculated the discounted value of
of the
cumulative output loss above for the Perry and Cagan equations.
-84-84-
To simplify further, we
~ for all ii
we assume g~
gi is
is aa constant g
>
>
n.
g which first equates the cost of unWe then
then solve for the value of g
unemployment
employment and the gain from eradicating inflation -- the minimum
minimum value
-—
of the permanent per period gain from eradicating
eradicating inflation
inflation that would
justify
justify incurring the transitional costs.
The
The value of ~
g for the
Perry, Stein, and Cagan results are presented
presented in
in Table
Table 8; we calculated
calculated
them under the assumption of aa 3.3 percent discount rate and for two
TABLE 88
The Minimum Value of the Per Period
Period Gain
that Justifies Eradicating aa 7.5 Percent Inflation Rate
Value of ~
g (billions of 72 $)
3.2
3.2
2.5
Equation/
Model
Perry
Perry
Cagan
Stein
Stein
Stein
11
22
73.0
70.9
70.9
16.66
16.
31.
31.00
25.4
25.4
11
22
57.0
57 .o
55.4
13.0
24.2
19.8
respecalternative values of the Okuns
0kun's Law coefficient
coefficient (3.2 and 2.5, respectively).
The minimum value of ~
g varies from $13
$13 billion per year
year based
on Cagan’s
Cagan's Phillips curve to $73 billion based on the Perry’s
Perry's Phillips
curve under aa moderate policy.
Note that this
this analysis
analysis provides an alternative perspective
perspective on
Mote
the case for gradualism.
Under gradualism, the costs may
may be
be reduced if
if
the Phillips curve is nonlinear.
But the benefits are
are also more
more
gradual (in our analysis, postponed until
until inflation is eradicated).
Thus, we find that although the costs are smaller under the gradual
gradual pol-~
pol1), the size
size of
of the per period
period gain
gain
icy using the Perry equation (Perry 1),
-85-
justify eradicating inflation
inflation is
is smaller under
under the more
required to justify
radical policy (Perry 2).
The radical policy also yields
yields aa smaller
minimum per period gain using the
the Stein model, although this result
result was
expected in this case because the cost turned out to be lower in
in the
using Stein’s
Stein s model.
radical case using
1
The calculations reported above presumed that the gains from rereducing inflation could be
be meaningfully represented
represented as
as aa fixed real
real sum
per period.
What if the gains are more meaningfully
meaningfully specified
specified as
as aa
real
real sum which grows at the
the same rate as potential output?
For
For example,
the cost of aa fully anticipated increase in
in inflation
inflation is
is generally
generally
measured by
by the reduction in the area under the demand curve for money
money
balances as
as wealth owners reduce their demand for money
money in response to
the associated rise in
in nominal interest rates.
rates.
The
The decline in
in demand
for real
generally viewed
real money due to
to aa rise in
in the interest rate
rate is generally
as proportional to
to the overall scale of money holdings which, in turn,
is determined by
by the level of transactions (e.g. real income).
The
cost of a given rate of inflation and
and hence the
the benefits
benefits of eliminating
eliminating
the inflation may therefore grow
grow at the
the rate
rate of
of increase of
of potential
output.
(8’)
(8')
In this case where
where ~
g is
is the value of the
the gain in
in period nn (the
G= ;
G=~
i=n
i=n
g (1+r_l2_
(l+r)i1
(l+r)
first period in
in which aa gain is
is registered).
For r~ .: :_ r, GG -+~ co.
>
00 •
corresponds to the result recently
recently derived by Feldstein (1979):
(1979):
This
if the
cost of inflation grows at aa rate
rate equal to or
or greater than the
the discount
rate, any positive initial gain (any ~
g
>
>
incurring
incurring any finite transitional cost~
cost'.
-86-
0) is sufficient to justify
resu1ts suggest that the case for anti—inflation
ctnti-inflation policies
These results
should not be dismissed lightly, even when there are large transitional
costs of eradicating Inflation.
inflation.
The range of the estimates of the
cumulative output loss, the uncertainty about the adjustment in those
results required to allow
allm,· for the
tr12 credibility effect, and the lack of aa
quantitative estimate of the cost of Inflation
inflation makes it extremely difdifficult
fic.1lt to make a meaningful comparison
:::omparison of the costs and benefits of
1
anti—inflation
anti-inflation policy.
It should not be surprising therefore
therefore that
that
indecisive and often lacking In
in coninitment
commitment
policymakers generally seem indecisive
to reduce Inflation.
in-~latior..
Narrowing the range of estimates of output loss
the cost of inflation should be high on the
and developing a measure of the
priorities for macroeconomic research in
in the 1980s.
RULES VERSUS ACTIVISM
two propositions.
The case against activism rests on two
The first
µrivate sector of the economy is inherently
proposition is that the private
stable.
This is a major tenet of monetarism and suggests the absence
of a need for
•or stabilization policy.
Indeed, monetarists generally concon-
tend that the iinstability
nstabi1 i ty observed in the economy results mainly
mainly from
government rather than private sector decisions.
stabilThe inherent stabil-
Ity of the private sector results In
ity
in part from the absence of large and
persistent exogenous shocks and
and in part from the fact that the shocks
that do occur have relatively small and only temporary effects on outout-
put and employment as a consequence of the economys
economy s built-in stability.
1
The second proposition in the case against activism is that even
if the economy were subject to cumulative
<umulative movements In
in output, employemploy-
ment and inflation relative to target levels, discretionary policy
-87-
might only compound the instability rather than dampen it.
The danger
The
that
that policy will turn
turn out to be
be destabilizing follows from the long
inside lag, the long and variable outside lag, and the general unceruncertainty about the
the effect of policy
policy on
on the economy.
economy.
The case for activist
activist policy involves
involves aa rejection of
of the
the two
two proppropositions developed above;
above; the economy needs to and can be
be stabilized by
appropriate manipulation of policy
policy instruments.
The first proposition
in
is that the economy is subject to
in support of policy activism,
activism, then, is
to
substantial
and persistent
sector.
substantial and
persistent disturbances
disturbances arising
arising from
from the
the private
private sector.
In addition,
addition, nonmonetarists contend that policy
policy can
can be implemented with
sufficiently
sufficiently short inside lags and with sufficient precision qiven our
our
understanding
understanding of the
the structure of the
the economy
economy to
to yield an improvement
in economic performance relative to
to aa policy of aa fixed rule.
rule.
Relevant empirical evidence on
on rules versus activism
activism includes:
(1)
the relative size of exogenous
exogenous impulses arisinq
arisinq from
policy
and
nonpolicy
sources
policy and nonpolicy sources
(2)
the degree of persistence in the response to such
such
disturbances
disturbances
(3)
the ability
ability of active policy to improve
improve economic
economic perperformance
in
the
face
of
the
disturbances.
formance in the face of the disturbances.
Stability of the Private
Private Sector
The issue of the stability of
of the private sector
sector has been categocategorized as aa fundamental difference between
monetarists and the convenbetween mon~tarists
conventional Keynesian tenets (See Andersen (1973) and Mayer
Mayer
(1975)).
(1975)).
Nevertheless, it appears to be
be an
an issue on which little, if any, relerelevant
vant empirical evidence
evidence is available.
The evidence that is conventionally
conventionally cited in
in response to the
allegation
the Keynesian
Keynesian position
as
allegation that
that the
position regards
regards the
the private
private sector
sector as
-88-88-
of simulation
simulation experiments with
inherently unstable is the result of
various econometric models.
These experiments suggest
suggest that the models
are stable, usually exhibiting highly damped oscillations back
back to
equilibrium following some
some shock (see Klein (1973)).
Such results
under the postulated experimental conditions
conditions are probably aa necessary
necessary
condition, but not aa sufficient condition to substantiate
substantiate the
the monemonetarist proposition.
We would need
need to look at
at the
the degree of damping
under aa policy of fixed rules relative to the damping under an endogeendoge—
nous policy with feedback from current economic developments.
The case
The
for
rules is
enhanced if
if endogenous
endogenous policy
reduces the
the degree
degree to
for rules
is enhanced
policy reduces
to which
which
disturbances are damped.
Evidence from Model Simulations
Discussions of the effectiveness of
of policies often focus on the
size of po
policy
multipliers.
1icy multi
p1 i ers.
Such
Such measures of the
the leverage
1everage of policy on
goal variables
variables are critical to
to setting
setting policy, but do
do not provide any
evidence on
on the usefulness
usefulness of
of discretionary policy unless they are zero.
Indeed as Cooper and Fischer demonstrate, even
even if
if the
the policy instrument
has aa zero cumulative multiplier it may be useful as aa stabilization
stabilization
has
tool as long
long as it has aa nonzero short-run multiplier.
tool
important
More important
is
is the pj4jç~jjit
predictability of
of the outcome of policy
policy actions which
which is
is more
the goal variables.
closely related to the errors in forecasting the
The
The
case
case for discretion, therefore, has little
little or nothing to
to do
do with
with the
size of policy multipliers,
moving
multipliers, unless
unless there is some concern about
about movinq
the
“penalty
the policy variables too far or too fast such as
as when aa "penalty
function" is added to the "goal
function”
“goal function."
function.”
The time pattern
pattern of the
the
response as well as the predictability
multipliers, on
predictability of the policy
policy multipliers,
—89—
-89-
the other hand, do matter.
discretlon, thereEvidence on rules versus discretion,
fore, generally involve model simulations
simu1ations and these are most useful If
if
allowance is made
made for uncertainty about the
the multipliers.
multip1iers.
Below we review the evidence on the comparison of
of economic
economic perperfonnance
formance under rules and discretion based on simulations with macromacroeconomic models.
First we must define a set of alternative policies;
four alternatives have been Investigated.
investigated.
1)
Actual policy:
instruments
Historical simulations in which policy Instruments
take on their historical values provide the benchmark of actual policy,
discretion as it was implemented as opposed to what would have been
optimal in the context of the model under consideration.
2)
2)
Fixed rifles
r~les or rules without feedback:
FIxed
in which the
Simulations In
policy instrument
instrument is constrained to grow at a constant rate provide
evidence on the effect of fixed rules; for example,
example~ a constant rate of
monetary growth as advocated by Friedman.
In this case
case. the policy
policy inin-
strument is totally independent of current economic developments.
3)
Active rules or rules with feedback:
/J.,n alternative to both disdisAn
cretion and fixed rules is an active rule or a rule which
~thich requires
policy instruments to respond systematically to current economic develdevelinvolving
opments. This approach introduces Phillips type ad hoc rules involving
proportional and derivative controls.
Some experimentation is underunder-
taken to identify “good”
good" rules but short of full optimization.
11
Such
Such
simulations can be viewed as a way of modeling systematic
systematic discretionary
policy without the blatant policy errors that in retrospect always mar
mar
the historical runs.
4) Optimal control:
The benchmark for identifying the best that is
possible under discretionary policy is an optimal control simulation
simulatio~ in
-90—90-
inwhich policymakers are viewed as selecting a time path for their instruments that
that minimizes
minimizes the losses associated with deviations
deviations of their
goal variables from their target levels.
It, therefore,
therefore, requires imim-
posing an
an explicit
exolicit loss
loss function inc1uding
designation of
of relative
including the designation
weights
miniweights on competing objectives and solving the model subject
subject to minimization of
of the losses.
1osses.
The
an ininThe solution allows the selection of an
strument path to
to reflect knowledge of
of the
the structural
structural parameters
parameters of
of the
model and forecasts of future performance based on current and past
values of exogenous
exoge!'lous variables and the dynamic structure of the model.
A superior
eccmomic performance
performance under
superior eccnomic
under such
such circumstances
circumstances hardly
hardly proproA
vides convincing support for discretionary policy,
policy, although
although it
it provides
of tne potential for discretionary policy to
to improve economic
economic
evidence of
performance.
The various policy regimes can be
difbe simulated in aa number of different ways.
In
In aa deterministic
deterministic simulation the error
error terms in the
various
estimated equations are set to zero.
various esti111ated
zero.
This immediately removes
aa potentially important
important source of instability in the private economy
economy
and should be
be expected to bias
bias results in favor
-Favor of fixed rules.
There
are two
two basic types of stochastic
stochastic simulations reflecting the two
sources of random disturbances:
estithe additive error
error terms in the
the esti-
mated equations
equation2 and the estimated coefficients.
Simulations allowing
allowing
for random additive error disturbances are generally labeled stochastic
simulations while those that randomize both
both parameters
parameters and additive
errors are referred to
to as
as fully stochastic simulations.
—91—
-91-
Actual Policy Versus Fixed Monetary Growth Rules
Modigliani reports
reports two simulations
simulations with aa fixed monetary growth
rule
rule over the period beginning in 1959
1959 and ending in
in mid-1971.
mid-1971.
at aa 33 percent
percent annual rate.
case Ml is constrained to grow at
In each
In the
first simulation all shocks are eliminated by
by substituting constant
trends or means for untrended exogenous variables.
variables.
In the second,
second, hishis-
torical values of exogenous
exogenous variables are employed.
In the first case
case
the monetary
monetary rule stabilizes the
the economy,
economy, but, allowing
allowing for historical
shocks the economy "was
“was distinctly less
less stable
stable than actual experience,
by aa factor of 50
50 percent
percent [p. 12].”
12]."
Eckstein investigates the
the implications of
of smooth growth in
in non—
nonborrowed reserves over the period of 1964 through 1975.
1975.
(Nonborrowed
‘64, accelerate
reserves grow at
at aa 44 percent rate in '64,
acce 1erate 1/4
1/ 4 percent
percent point
each year
year until
until they stabilize at aa 66 percent rate
rate during and after
after
1972).
1972).
Eckstein finds that smooth growth in reserves does result in
in
“a more stable growth pattern”
"a
pattern" but does not dramatically
dramatically alter the
overall results for economic performance.
Versus Fixed Rules
Rules
Active Rules Versus
In aa series of papers employing simulations with both the MPS
MPS and
l972b, 1974) compare
St. Louis models, Cooper and Fischer (1972a,
(1972a, 1972b,
Phillips type feedback control rules with
with fixed growth rate rules.
They conclude that
that there are active rules which dominate fixed rules
for both models, under deterministic, stochastic and fully stochastic
stochastic
simulations.
The dominant active rules generally
generally involving
involving strong
derivative controls and some proportional control.
criterion was
The criterion
the average standard deviation
deviation in the unemployment
unemployment and inflation rates.
rates.
—92-92-
For the St.
St. Louis model, for example, the average standard deviations
for each variable were reduced
reduced by about 20 percent
percent in
in the deterministic
deterministic
simulations (over the period 56/1-68/IV),
56/I-68/IV), between
between 50 - 70 percent
percent in
in
-
the stochastic simulations (over the same
same period)
period) and
and by about 50
50 perpercent in the fully stochastic
stochastic simulations (over
(over the period 55/I
55/1 - 7l/IV).
71/JV).
—
The improvement v-ws
was more modest, however,
model , where
however, in the MPS model,
where the
standard deviation
deviation of unemployment fell by
by 44 - 24 percent and that of
-
32 percent in stochastic simulations
simulations over the period
inflation by 77 - 32
-
1956/I - 68/IV.
1956/l
68/l V.
—
Optimal Control Simulations
There have been
been numerous attempts to compare fixed rules with
optimal control simulations including Chow
Chow (1972), Garbade
Garbade (1975),
Cooper and Fischer (1975), Crane, Ravenner
Havenner and Tinsley (1976), and
Crane, Havenner and Berry (1978).
studies find that
that
The first four studies
fixed rules are uniformly inferior to optimal control
control (and generally
inferior to historical
historical policies).
These
simulaThese studies use
use stochastic simula-
tions but actual
actual values of exogenous variables
variables and, with the exception
exception
of Cooper and Fischer, constant parameter values.
Garbade
Garbade for example
example
finds that “discretion,”
discretion, in the form of optimal control, reduces
reduces the
11
11
50 percent compared to aa fixed rule, aa result in close
expected loss
loss by 50
agreement with Chow.
Chow.
Garbade views
views his
his results as
as adding to
to the
“accumulating evidence”
!!accumulating
evidence" of
of the gains
gains associated
associated with discretion “when
when aa
11
valid representation of
of the economy
economy is available.”
available."
But that, after all,
all,
is the major
major element in
in the controversy.
Cooper and Fischer find
find that their active rules perform
perform quite
well in relation to
to optimal control solutions using the St.
St. Louis
Louis model.
-93—93—
Costs are reduced by about 45 percent relative to fixed rules, but
fixed rules outperform historical policy In
in this case due
due in part to
greater instability in instrument movements
movements in the latter case. The
Cooper-Fischer paper produces a possibly valuable insight about the
Cooper—Fischer
relative performance of rules and discretion.
Stochastic simulation
requires multiple simulations for alternative realizations of the
the
stochastic disturbances. They found that the poor overall performance
of fixed rules resulted from
from their “spectacularly
"spectacularly bad”
bad performance
performance in
11
replications where
where losses turned out to be above average for all
all
policies.
Where average performance Is
is good, on the
tf!e other hand,
hand, fixed
control.
rules perform about as well as optimal control.
imoly that
This may imply
optimal policy is nonlinear—restrained
nonlinear-restrained to
t8 fixed rules within
within a band
around target values of goal variables and active only outside those
bands.
Thus, “fine
tuning is rejected, but activism In
in the face of aa
' fine tuning”
1
11
major disturbance
disturbance has a substantial payoff.
This conclusion is reinforced by the Crane, Havenner and Tinsley
1911/1-1974111 period using a condensed version of
study of the 1971/1-1974/11
of the MPS
model, MINNIE.
Optimal policy is not especially volatile after an
initial aggressive expansionary policy
pol-icy in the first
first two quarters to
to
offset the recession implicit in the initial conditions. The optimal
policy again dominates fixed rules, in this case by about 40 percent;
and fixed rules would have increased expected losses by about
abnut 45 perper-
cent relative to historical policies.
Rational Expectations and the
the Limits of Activist Policy
The traditional arguments against activist policy focused on the
Implications
implications of long inside lags, long and variable outside lags, and
-94-
multiplier uncertainty; there was a general emphasis on the limitations
of policy
po1icy in an environment characterized by insufficient knowledge
knowledge of
the economys
economy 1 s structure.
The Lucas—Sargent—Wallace
Lucas-Sargent-Wallace rational expectaexpecta-
tions models suggest aa dramatically different
different basis for fixed rules.
These models suggest that policy is doomed
doomed to ineffectiveness
ineffectiveness in an
environment in which economic
economic agents have knowledge both about the
structure
structure of the economy and the way in which policy authorities rerespond to economic developments.
In this
this case
case too much knowledge rather
than
than too little
little knowledge
knowledge underlies the ineffectiveness of policy.
Real variables according to
to these models respond only to unanticipated
price or
or inflation shocks.
produce
produce surprises.
Systematic policy, by definition, cannot
Therefore, although there exists aa trade-off
trade-off bebe-
be systemsystemtween unanticipated inflation and unemployment, it cannot be
exploited by policy authorities; this
atically exploited
this is
is generally
generally referred to
as the neutrality proposition.
The theoretical structure of these
models and the implications of
of aa number of qualifications,
qualifications, particularly
particularly
of nominal contracts,
contracts, have been
been thoroughly developed
developed in
the existence of
the paper by
by Taylor.
The role, operational
operational specification, and implicaimplica-
tions of rational
rational expectations in macroeconomic models is the
the central
issue
issue in
in macroeconomic theory
theory today and
and empirical
empirical investigations
investigations of
of
these models is certain to
SOs. There
to be
be the growth industry of the
the '80s.
are, however, only
only aa handful of
of empirical studies to date
date that attempt
attemot
to
to test the
the neutrality proposition.
Mccallum (1979)
(1979) in aa recent survey of this literature notes that
McCallum
the formal evidence is not inconsistent
while "the
inconsistent with the neutrality
proposition.
proposition . . . the power of
of existing tests
tests is not high and, in any
any
.
.
event, the
the evidence
evidence is not entirely
entirely clearcut.
clearcut."
-95—
-95-
The two most
most important
important
empirical studies are the Barro
Barro papers (1977, 1978) on the effect
effect of
of
empirical
unanticipated monetary growth on
output and Sargent’s
Sargent's
on unemployment and output
paper (1976) applying Sims and Granger tests for causality
causality to movements
movements
in the unemployment rate, the money supply, government
government expenditures and
other
other macro variables.
Barro estimates
estimates aa reaction
to isolate
Barro
reaction function
function to
isolate unanticipated
unanticipated
monetary
examines the role
role of
of unanticipated
unanticipated and
and anticianticimonetary growth
growth and then examines
pated monetary
monetary change on unemployment and output.
His results
results are
are rere-
markably one sided, supporting the hypothesis
hypothesis that only
only unanticipated
markably
policy actions affect real variables.
But
But his
his empirical methodology
has been
convincingly critiqued
critiqued by
been convincingly
by Small,
Small, Fischer
Fischer (1978)
(1978) and
and Gordon
Gordon
(1979).
{1979).
Sargent is somewhat more cautious in interpreting
interpreting his findings
as indicating
indicating that “the
"the causal structure imposed on the data by the
the
at variance
classical model.
model . . . is not obscenely at
variance with
with the data
data [p. 233].”
233]."
.
.
We think this
this means the results are mixed, which indeed they
they are.
There
movements in
There is some
some evidence,
evidence, for example, that
that movements
in the money
money supply
“cause”
movements in the unemployment rate (using the Granger test) and
"cause" movements
some evidence that it does not (using the
the Sims test).
Summary
The evidence accumulated over the ‘70s
'?Os has has at best only
only aa
consensus over the gains associated with
modest role in increasing the consensus
activist policy.
of the ‘70s
'?Os has clearly eroded
eroded the
the
The experience of
optimism about the potential activist policy that characterized the
apparent success of
of the 1964 tax
tax cut and the long expansion of
of the ‘60s.
'60s.
There is wider recognition today compared
compared to the mid-l96Os
mid-l960s among proproponents
ponents of active policy
policy of
of the limitations of
of active policy
policy and
and the
the
-96-96-
difficulty of “fine
"fine tuning”
tuning" the
the economy by responding
responding to even
even small
difficulty
departures of output and employment from target levels.
Active
Active policy,
policy,
however, continues to have wide support in situations where
where aa sizable
displacement has occurred, as in
in the
the 1973—75
1973-75 recession.
On the other
other
of rules, such
such as Friedman (1968), also allow for
hand, many proponents of
the
14].”
the use of discretionary policy to
to offset “major
"major disturbances [p. 14]."
Therefore, the
the gulf
gulf between proponents of rules and activism is not
nearly so great as
as it might at
at first appear.
The optimal control
aggressive
studies have helped to emphasize the potential usefulness of aggressive
policy action when initial conditions are far away from targets and the
limited potential usefulness of activist policy in response to smaller
displacements.
This lesson is
is perhaps one on
on which proponents of rules
rules
and activism can agree.
CONCLUSION
1
As
As the ‘7Os
70s began, the monetarist—income
monetarist-income expenditure controversy
controversy
in macroeconomics.
was aa dominant theme in
Particularly after the MPS and
Particularly
other large scale
scale models began churning out large values for monetary
the controversy
controversy focused in on the size of fiscal
policy multipliers, the
multipliers, particularly the fiscal multipliers on nominal GNP.
The
econometric evidence of the ‘7Os
'7Os has
has not fully resolved this
this issue,
i.e., there are those who continue
continue to be persuaded
persuaded by the St. Louis
equation results.
And while
while this evidence
evidence questioning the
the reliability
reliability
of the fiscal multipliers in the St.
St. Louis equation undoubtedly
undoubtedly has rereinforced the views of the skeptics,
skeptics, it
it has not necessarily
necessarily shaken the
confidence of
of the equation’s
equation's supporters.
—97—
-97-
'70s began, the orthodoxy
orthodoxy of aa Phillips curve embodying aa
As the ‘lOs
stable trade—off
trade-off was
was under
under an attack it
it did not survive.
After aa trantran-
sitional period, evidence
evidence mounted
mounted in
in support of aa vertical long-run
Phillips curve.
Thereafter, the issues contested
contested have
have been the nature
and sources of
short—run trade-off
of any short-run
trade-off and the implications for the
output
output loss
loss of
of eradicating
eradicating inflation.
inflation.
The
The econometric
econometric evidence from
from aa
wide
range of
and models
models suggests
monetary deceleration
deceleration
wide range
of sources
sources and
suggests that
that monetary
can eradicate inflation, but not quickly
quickly and not
not without large costs in
in
terms of cumulative output loss.
loss.
The major
major unresolved issue is the
significance
significance of
of the credibility effect
effect and the
the degree
degree of
of overestimation
overestimation
in
the cumulative
cumulative output
loss due
to the
the failure
to take
in the
output loss
due to
failure to
take into
into account
account
the
effect of
recent policy
actions and
and expected
expected policy
of recent
policy actions
policy actions
actions on
on ininthe effect
flation expectations.
While fine-tuning may have few advocates,
advocates, the evidence from model
simulations
are likely
likely to
simulations suggests
suggests there
there are
to be
be considerable
considerable gains
gains to
to
activism
activism when the economy is far away from targets and in response to
very large shocks,
shocks.
Rules or
alor activism remains
remains an important issue al-
though the case against activism has been
been broadened by the development
of rational
rational expectations
expectations market clearing models.
-98-
REFERENCES
Allen, S. and T. Seaks, (forthcoming), “Nixon
''Nixon and Ford
Ford as Ex—Post
Ex-Post
Keynesians,”
Journall of Macroeconomics.
Keynes i ans, ~ourna
Macroe~s,_n_om~.
ll
L. (1973), “The
"The State of the Monetarist Debate,”
Debate," Federal
Andersen, L.
Reserve Bank of St.
St. Louis Review
f<e_view (September).
Andersen, L. and K. Carlson (1970),
(1970), “A
''A Monetarist Model for Economic
Stabilization,'' Federal Reserve
Reserve Bank
Bank of
of St.
Stabilization,”
St. Louis Review (April),
7-25.
7—25.
“Monetary and Fiscal
Andersen, L.
L. and J. L. Jordan (1968), "Monetary
Fiscal Actions:
AA
Test of
of their Relative Importance
Importance in
in Economic Stabilization,”
Stabilization,!!
Federal Reserve Bank of
11-24.
of St. Louis Review (November),
(November), 11-24.
“Rational
Nonrational
Andersen, P. ((1979),
1979) , "Ra
ti ona 1 Expectations Forecasts
Forecasts from Non
rat i ona 1
Models,”
Models," Journal of
of Monetary Economics (January), 67-80.
Barro, R. (1977), "Unanticipated
Barro,
“Unanticipated Monetary Growth and Unemployment in
States,” American Economic Review (March),
the United States,"
(March), 101-115.
Barro, R. (1978), "Unanticipated
Level in
in
Barro,
“Unanticipated Money, Output, and the Price Level
States," Journal of Political Economy
Economy (Auqust),
(Auoust),
the United States,”
549-580.
549-580,
“The Reduction of Inflation and
Cagan, p.
P. (1977), ''The
and the Magnitude
Magnitude of
Unemp 1 oyment" in Wi
11 i am Fellner
Fe 11 ner ed. ,, Contemporary_Economic
Contemporary Economic
Unemployment”
William
Problems 1977,
Washinqton, D.C.,
1977, American
American Enterprise Institute,
Institute, vJashinqton,
41-SO.
41-50.
by Slack
Slack Demand”
Demand" in
Cagan, P. (1978), "The
“The Reduction of Inflation by
ed. , Contemporary Economic Problems 1978,
William Fellner ed.,
1978,
American Enterprise
Enterprise Institute, Washington
WashTn-gton D.C.,
D. C., 13-45.
Carlson,
K. (1978), “Does
"Does the
the St.
St. Louis Equation Now
Now Believe in Fiscal
Fiscal
Carison, K.
Policy?”
Policy?" Federal Reserve
Reserve Bank
Bank of St.
St. Louis
Louis Reyiew
Review (February),
(February),
13-19.
1319.
Chow, 0.
“How Much Could Be Gained by Optimal Stochastic
G. (1972), "How
Stochastic Control
Policies?'' Annals of Economic and Social Measurement
Measurement (October),
Policies?”
391-406.
Chow, G. (1977),
(1977), “Usefulness
"Usefulness of Tmperfect
Imperfect Models for the Formulation of
Stabilization Policies,"
Policies,” Annals of Economic
Economic and Social
Social
Measurement (Spring).
~
Christ, C. (1975),
( 1975), “Judging
"Judging the Performance of
of Econometric Models of
International Economic Review (February).
the U.S. Economy,”
Economy,u Jnternational
(February).
Congressional
mdc tandjn Fiscal
Congressional Budget Office (1978),
(1978), Understanding
Fiscal Policy.
—99—
-99-
Cooper, J.
J. P. and S. Fischer (1975), “A
''A Method of Stochastic
Stochastic Control
of Nonlinear Econometric Models
Models and an Application,”
Application/ 1 Econometrica
Econometrica
(January),
147-162.
(January), 147-162.
Cooper, J. P. and S. Fischer (1974),
(1974), “Monetary
"Monetary and Fiscal
Fiscal Policy in aa
Fully Stochastic St.
St. Louis Econometric Model
Model,"
Journal of
of Money,
,“
Journal
Credit and Banking
Banking (February), 1—22.
1-22.
Cooper, J. P. and S. Fischer (1972),
(1972), “Simulations
"Simulations of
of Monetary Rules on
the
the FRB-MIT—Penn
FRB-MIT-Penn Model,”
Model," Journal of Money, Credit and Banking
(May), 384-396.
384-396.
(May),
Cooper, J.
J. P. and S.
S. Fischer (1972), ''Stochastic
Mone“Stochastic Simulation of Monetary Rules in Two Macroeconomic Models,”
Models," Journal of American
American
Statistical
Association (December),
(December), 750-60.
Statistical_Association
Corrigan, E. C. (1970),
(1970), “Measurement
"Measurement and
and Importance
Importance of Fiscal Policy
Changes,” Federal Reserve Bank
MonthjL3~yIeyL(June)
Changes,"
Bank of New York Monthly
Review (June),
133-45.
Crane, R., A. Havenner,
Havenner, and J.
J. Berry (1978),
(1978), “Fixed
"Fixed Rules versus
Activism in the Conduct of Monetary
Monetary Policy,”
Policy," American_Economic
American Economic
Review (December), 769-783.
Crane, R.
R.,, A. Havenner, and P. Tinsley (1976), “Optimal
"Optimal Macroeconomic
Control
Policies,” Annals of
Control Policies,u
of Economic and Social Measurement
191—204.
(Spring), 191-204.
“The Channels of
Policy,”
deLeeuw, F. and E. Gramlich (1969),
(1969), "The
of Monetary
Monetary Policy,"
Federal
Federal Reserve Bulletin,
Bulletin, 472-91.
EconodeLeeuw, F. and E. Gramlich (1968),
(1968), “The
"The Federal Reserve—MIT
Reserve-MIT Econometric Model,”
Model," Federal_Reserve
Federal Reserve Bulletin (January), 11-46.
11-46.
(1969), "Monetary
“Monetary and Fiscal Actions:
deLeeuw, F. and J. Kalchbrenner (1969),
A Test of
A
of Their Relative
Relative Importance
Importance in Economic Stabilization
Stabilization -—
-Conmient,”
Comment," Federal
Federal Reserve
Reserve Bank of St.
St. Louis Review (April),
(April), 6-11
6-11.
Dewald, W. and M. Marchon
“A Modified
Marchan (1978), "A
Modified Federal
Federal Reserve of St.
Equation for Canada, France, Germany, Italy, the
the
Louis Spending Equation
United
add the United States,”
United Kingdom, and
States," Kreditund_Kapital.
Kredit und Kapital.
Eckstein, D.
0. (1978), The Great_Recession,
Great Recession, North Holland, Amsterdam.
L. R.
R. Klein (1968), The Wharton Econometric
Econometric ForeEvans, M. K. and L.
Fore-
casting Model,
Model, Economics Research UnTf,
Unit, University
University of Pennsyl—
Pennsylvania,
vania, 2nd ed.
Feldstein, M. (1979), "The
of Permanent
Permanent Inflation
Inflation and
Feldstein,
“The Welfare Cost of
Optimal Short-Run Economic Policy,”
Policy," Journal of Political
Political Economy
(August), 749-767.
(August),
-100-‘Do-
"The Credibility Effect and Rational Expectations Fellner, W. (1979), “The
—
Implications of the Gramlich Study,”
Study," Brookings_Papers
Brookings Papers on
on Economic
Activity,
Activity, 167-178.
·
·
(1978), "On
“On Activist
Fischer, 5.
S. (1978),
Activist Policy With Rational Expectations,”
Expectations,"
National Bureau of Economic Research
Research Working Paper.
Fischer, S. and F. Modigliani (1978), “Towards
"Towards an
an Understanding of the
Real Effects and Costs of Inflation,”
Inflation," Weltwirtschaftliches
Archiv,
Archiv, 810-833.
Friedman, 8.
"Even the St.
St. Louis Model Now
Now Believes in Fiscal
Fiscal
B. (1977), “Even
Policy,”
Policy," Journal of Money, Credit
Credit and Banking
Banking (May),
(May), 365—67.
365-67.
Fromm,
Fromm, G. and L. Klein (1973),
(1973), " AA Comparison of
of Eleven
Eleven Econometric
Models of the United
United States,”
States," American
American Economic
Economic Reyj~P~~pg4
Review Paper and
Proceedings (May).
“
Fromm, G. and L.
L. R. Klein (1976),
(1976), “The
"The NBER/NSF
NBER/NSF Model Comparison
Comparison
Seminar: An
An Analysis
Analysis of
of Results,”
Results," Annals of Economic
Economic and Social
Social
Measurement, 1-28.
7-28.
“Discretion in
of Macroeconomic
Garbade, K. (1975), "Discretion
in the
the Choice of
Macroeconomic PoliPolicies,” Annals of Economic
cies,
Economic and Social
Socia 1 Measurement
Measurement (Spring),
21
5-238.
215-238.
!I
Goldfeld, S. and A. Blinder
Blinder (1972), "Some
Endogenous
“Some Implications of Endoqenous
Stabilization
Stabilization Policy,”
Policy," Brookings Papers on Economic Activity,
Activity,
585-640.
Gordon, R. J. (1976), "Comments
Ando" in J.
J. Stein,
“Comments on Modigliani and Ando”
Monetarism, North Holland, Amsterdam, 52—66.
52-66.
“New Evidence that Fully Anticipated Monetary
Gordon, R. J. (1979), "New
Changes Influence Real Output After All,"
All ,“ National Bureau of
Economic Research Working Paper
Paper #361.
#361.
Gordon, R. J.
"What Can Stabilization Policy Achieve,”
Achieve," American
American
J. (1978), “What
335-347.
Economic Review (May), 335-341.
(1979), “Macro
"Macro Policy Response to Price Shocks,"
Brookings
Gramlich, E. (1979),
Shocks,” pmYijj9~
Activit, 125—166.
Papers ~gn
on Espnomic
Economic Activity,
125-166.
Gramlich, E. ((1971),
1971 ) , “The
"The Usefulness of Monetary and
and Fiscal
Fi seal PolicieS
Policies as
Discretionary
Discretionary Stabilization Tools,”
Tools," ~
Journal of Money, Credit and
Banking (May),
(May), 506-532.
Hymans, S. H. and H. T. Shapiro (1970), "The
The DHL—III
DHL-III Quarterly
Quarterly Model
“
Economy,” Research Seminar
of the U.S. Economy,
Seminar in Quantitative
Quantitative Economics,
University of Michigan.
11
“Comentary on
on the Current State of the Monetarist
Klein, L.
L. R.
R. (1973), "Commentary
Debate," Federal
Federal Reserve Bank of St.
St. Louis Review (September).
Debate,”
-101—~0~—
R. E. (1975), "Econometric
Policy Evaluations: AA Critique”
Critique" in
in IC,
K.
Lucas, R,
“Econonetric Policy
and A. Meltzer, The Phillips Curve and Public Policy,
Brunner and
North Holland, Amsterdam.
Mayer, T. (1975), "The
“The Structure of Monetarism I,”
I," Kredit_and_Kapital.
Kredit and Kapital.
McCallum,
Policy
Mc Ca 11 um, B. (1979),
( 1979), “The
"The Current State of the Po
1 icy Ineffectiveness
nebate," American
American Economic
Economic Review
Review (May),
(May), 240-245.
240-245.
Debate,”
(1971), “Monetary
"Monetary Policy and Consumption”
Consumption" in Consumer
Modigliani, F. (1971),
~
Spending and Monetary Policy: the Linkages, Federal Reserve
Bank of Boston,
Boston, Boston.
Modigliani, F. (1977),
(1977), “The
"The Monetarist Controversy or Should We
We Forsake
Policies,” American Economic Review (March), 1-19.
Stabilization Policies,"
1-19.
(1976), “Impacts
''Impacts of
of Fiscal
Fiscal Actions on
Modigliani, F. and A. Ando (1976),
Aggregate Income and the
the Monetarist
Monetarist Controversy”
Controversy" in
in J. Stein ed.
ed.,
Monetarism, North Holland, Amsterdam.
Amsterdam.
Disinflationary Policies,”
Policies," American
American Economic
Okun, A.,
A.,”!IEfficient
Efficient Disinflationary
(May), 348-52.
Review (May),
Okun, A. (1972),
(1972), “Fiscal-Monetary
"Fiscal-Monetary Activism: Some Analytical Issues,”
Issues,"
Brookings
on Economic Activity, 123-163.
123-163.
~rgpj
gs Papers
Pap
op_Economic_Activity,
disserPapademos, L. (1977),
(1977), Optimal
Optimal Aggregate
Aggregate Employment
Employment Policy, Ph.D. dissertation, Massachusetts Institute
Institute of Technology.
Perry, 0.
G. (1978), “SlOwly
"Slowly the Wage
Wage Price Spiral: AA Macroeconomic
View,”
Brookings_Papers
View," Brookings
Papers on Economic Activity,
Activit):', pp. z59—29l.
t59-291.
A Classical Macroeconomic Model of the United
Sargent, T. (1976),
(1976), “"A
States,”
Political Economy
States," Journal of
of Political
Economy_ 207—237.
207-237.
“Rational Expectations and
Sargent, T. and N. Wallace (1976), "Rational
and the
Policy,” Journal of_Monetary_Economics.
Theory of Economy Policy,"
of Monetary Economics.
“The Almon
Schmidt, P.
P. and R. Waud ((1973),
1973), "The
Al man Lag
Lag Technique
Technique and the MoneMonetary versus
versus Fiscal
Fi seal Policy Debate,”
Debate," Journal of
of the American
American
Statistical Societ):'
Society (March), 77-19.
11—19.
Silber, W. (1971),
(1977), “The
"The St. Louis Equation:
'Democratic' and
‘Democratic’
1
‘Republican’
Republican 1 Versions and Other
Other Experiments,”
Experiments,!! Review_of
Review of Economics
(November), 362—367,
362-367.
··
and Statistics (November),
$mall, 0.
Barro’ss Unanticipated
Small,
D. (forthcoming), “A
"A Comment
Comment on Robert Barro'
Money Growth and Unemployment in the United States,”
States," American
American
Economic Review.
Journal_of
Stein, J.
J. (1978),
(7978), “Inflation,
"Inflation, Employment and Stagflation,”
Stagflation," Journal
of
Monetary_Economics.
Monetary Economics.
-102—
-102-
DISCUSSION OF
OF THE MEYER-RASCHE AND TAYLOR
TAYLOR PAPERS
Neil Wa
11 ace
Wallace
For us at
l970s constitute ten years of
at this conference, the
the 197Os
of
additional data
data and some theoretical
theoretical developments
developments that
that suggest new
new ways
ways
of interpreting those and earlier data.
The two papers
papers presented this
morning -- in part, because
because of
of the assignments given the authors ---
——
1970s.
contain very
very different
different views about the lessons of
of the 197Os.
II will
come to still
still aa third view and, as it happens, one that does not reprerepresent aa compromise between them.
As II understand it,
it, Meyer-Rasche accepted the task
task of summarizing
l97Os, while
while Taylor accepted the task of
of
lessons from the data of the 197Os,
surmiarizing
l970s.
summarizing lessons from the
the theoretical developments of the 197Os.
That division of labor did not turn out well; it
it encouraged Meyer-Rasche
to proceed as if one could learn lessons from data
data without invoking
theory.
basis of the preliminary draft of the Meyer—Rasche
Meyer-Rasche paper
On the basis
made available to me and on the basis
basis of their
their oral
oral remarks this
this
morning, II am left somewhat in the
the dark about the point
point of view of
of the
Meyer-Rasche paper.
paper.
II know what
Pm not sure
what they did, but I'm
sure what their
message is.
is.
Based on what
what they did, one might
might infer that for Meyer-Rasche,
the
l970s represent no more
the 197Os
more than
than ten years
years of additional data.
They
Neil
is Professor
Neil Wallace is
Professor of Economics at the University of Minnesota
and
and Advisor,
Advisor, Research
Research Department, Federal Reserve
Reserve Bank
Bank of
of Minneapolis.
-103-103-
use those data and earlier data in the same way that most economists
economists
ten
ten years ago used
used the data available to
to them.
In particular, both
their
so—called structural
reduced—form models
their so-called
structural models
models and
and their
their reduced-form
models consist
consist
of regression equations that in form are the same as those most econoeconomists used in the 1960s.
l960s.
Moreover, Meyer—Rasche
Meyer-Rasche extrapolate
extrapolate from those
for the
effects of
different policies
regression equations
equations for
the effects
of different
policies in
in the
the same
same
way that
that many
economists in
the l960s
their estimates.
estimates.
way
many economists
in the
1960s extrapolated
extrapolated from
from their
That
that for Meyer—Rasche,
That is why I say that
Meyer-Rasche, the 197Os
1970s seem to
to represent
represent no
no
more than ten years of additional data.
data.
Even at the level of pure
pure empiricism, aa different
different lesson
lesson can
can be
drawn.
Meyer—Rasche extrapolation procedure applied in the late
The Meyer-Rasche
l970s.
1960s did badly predicting the 1970s.
Why, then, believe
believe that those
same procedures applied now will do well predicting the l980s?
1980s?
Happily, though, we do
do not have to decide on the
the basis of pure
empiricism.
l97Os -- many of
The theoretical
theoretical developments of the 1970s
of which
which
--
are described
described in
Taylor’s paper
convincing arguments
are
in Taylor's
paper -- provide
provide convincing
arguments why
why
-—
we should
should not take seriously
seriously as
as “multipliers”
"multipliers" the correlation coefficoeffi-
cients or the
the functions of them presented in the Meyer-Rasche paper.
paper.
Meyer—Rasche
criticism of the
the multiplier
multiplier interinterMeyer-Rasche are
are aware of the criticism
pretation of
of their estimates.
In effect, they acknowledge the criticriti-
cism and say that they are unwilling to
to defend such an
an interpretation.
That, though, is what leaves me confused about their message.
Nor does
it
it help
help to
to suggest, as Meyer seemed to
to in
in his
his oral remarks, that their
estimates of Phillips curve trade—offs
trade-offs provide
provide upper bounds on
on the
unfavorableness of this trade-off.
quires aa supporting argument.
argument.
ing, or not interesting.
such aa claim
claim also reLogically, such
re-
Moreover, upper
upper bounds
bounds can be
be interestinterest-
All of
of GNP is an
an upper
upper bound on
on the
the output
-104-
loss
loss that accompanies a one percent cut in the inflation rate, but it
is not an interesting upper
upper bound.
Meyer-Rasche
Meyer-Rasche must convince us
us that
their
their estimates are interesting
interesting upper bounds
bounds if,
if, in
in fact, they are
upper
upper bounds at all.
Such
Such convincing
convincing must take the form of aa theoretitheoreti-
cal
cal argument that says why
why it is legitimate to
to extrapolate in
in particpartic-
ular ways from particular
particular correlations.
In the
the 1960s,
1960s, many economists thought that their
their policy extrapoextrapolations from the kinds of models used
used by Meyer-Rasche
Meyer-Rasche were
were legitimized
leqitimized
by existing theory.
The
The theoretical
theoretical developments of
of the 1970s
1970s have
have
convinced many of
of us that that is not
not so.
Although Taylor's
Taylor’s paper
paper
some of those
those developments, his paper stops short of dedescribes some
describing in full generality why we were
were led astray badly by the kind of
of
theorizing that was used.
used.
Since that kind of theorizing still
still perper-
sists, it
it is worthwhile summarizing in
in a general way
way what is wrong with
it.
Whether we are talking
talking about most textbooks in macroeconomics
macroeconomics or
most macroeconometric
macroeconometric models,
models, the
the models from
from which
which policy
oolicy implications
implications
are drawn
drawn consist
consist of aa set of relationships
relationships -- a consumption function,
—-
money demand function, and so
so on.
an investment function, aa money
Let us
label these
these M
M1,, M
M ,, MM , ... ,MN
MN (M
(M for
macroecolabel
for model).
model). The
The style
style of
of macroeco1 22 33
nomics
implinomics textbooks is to
to present the complete model and its policy implications
and also
also to
separate chapters
consumption,
cations and
to present
present separate
chapters -- one
one on
on consumption,
—-
one on
on investment, one on
on money demand, and so on -- that are meant to
—-
justify one by one the relationships of the complete model, the M~.
Mi.
When builders
builders of macroeconometric models try to justify their models,
they also proceed in
in this
this way.
In order to get at
at what is wrong with
this
this kind of theorizing,
theorizing, we must describe the logical relationship
-105—105—
between these justifying chapters and the macroeconomic or
or macroecomacroecoMN.
nometric model consisting of MM
11,, MM
22 , ... ,MN.
Each justifying chapter consists of aa set of assumptions.
Let
Let us
label
label these sets of assumptions ~l’~2
s ,s , ... , SN (S for story), where
where for
i Si is said to justify
M~
each i,
justify Mi.
.
1 2
The most extravagant
extravagant claim made
Si and
and f\
M~is
about the
the relationship between
between Si
following: For each
each i,
is the follnwing:
i'
s. implies Mi. In
particular,
In particular,
-1 impiiesM~.
~i
is also true.
true. In other
other words,
it
it
in
is never
never claimed
that the
the converse
converse
is
claimed that
general, Si and
and Mi
M~ are
are not
equivalent
not equivalent
M~. This nonequivalence
and more is implied by
by Si than
than just
just Mi.
nonequivalence has two
consequences.
First, it
it implies
implies that consistency among the M~
Mi does
does not imply
consistency
among the
the Si.
Si.
consistency among
If
inconsistent, then
If the
the Si
Si are
are mutually
mutually inconsistent,
then it
it
cannot be claimed
M~. Note,
claimed that there is an underlying theory of the I\.
Mote,
in this regard, that consistency
consistency among the Si is never checked
checked and, as
as
II illustrate below, that
that inconsistency is
is easy
easy to
to demonstrate
demonstrate for most
macroeconomic
macroeconomic models.
Second, if
if the
the Si are mutually
mutually consistent, nonequivalence
nonequivalence between
s7
and
Mi implies
we are
are missing
many of
the implications
of the
the
and M~
implies that
that we
missing many
of the
implications of
underlying theory by limiting attention to the M~.
Mi. Thus, for example,
the Si often
often contain at least hints of aa welfare analysis of inflation.
As is
is well known, the typical M~provide
Mi provide no such
such analysis.
II will now briefly
briefly defend the
the nonequivalence
nonequivalence claim
claim and, at the
same
same time, argue that
that inconsistencies are
are present
present in standard
standard macro
macro
models.
And, since this is St.
St. Louis, II will begin by focusing on
money demand.
The usual
usual way to defend the money
money demand functions of most
most macromacroeconomic
models is
is to
to appeal
appeal to
the Baumol
economic models
to aa transaction
transaction cost
cost model
model of
of the
Baumol
—106-106-
(1952), Tobin
Tobin (1956), or
or Miller-Orr (1966) variety,
variety.
Those models exex-
plain money demand in the presence of default-free,
default-free, higher-yielding
securities -- Treasury
Treasury bills, say -- by transactions
transactions costs, for example,
——
trips to
to the bank.
function.
——
But
But the
the models imply
imply more
more than aa money demand
They
public’s means of
They imply that if the ratio of the public's
of paypay-
its holdings of interest—bearing
interest-bearing assets changes as aa result,
ments to
to its
of open-market operations,
operations, then
then there
there is aa change in the amount of
say, of
resources
resources used up
up in transactions.
transactions.
But
But such
such aa change contradicts
contradicts the
the
usual resource-supply
resource—supply assumptions of most macro models.
models.
Those make no
of resources being used
transallowance for an altered amount of
used up
up in transactions.
For this and other reasons, the implications for open-market
open-market
operations of the
the theory of interest in the inventory
inventory models are very
different from those of
of most macro models,
models, particularly monetarist
monetarist
models (see Bryant and Wallace 1979).
1979).
It is also standard to assume that the money
money demand function that
one
one derives for aa closed economy
economy holds
holds with only
only minor modifications
for an open economy in aa world in which each of several countries
issues its
its own money.
It is this
this view
view that lies
lies behind the
the attachment
of) laissez—faire
laissez-faire floating exchange
exchange rates.
rates.
to (the viability of)
But such
aa claim
claim is supported neither by
by an
an acceptable theory (see Wallace 1979),
1979),
nor
by recent
recent experience.
experience.
nor by
That
that the
That experience
experience suggests
suggests that
the demand
demand for
for
aa particular money in a world of many
many monies may be very different
different from
the demand for aa single money in aa closed
closed economy.
In the 1970s,
inconsistencies regarding expectation
1970s, of
of course, inconsistencies
expectation
formation have received the most attention.
Expectation
Expectation formation
formation is
is
important because macroeconomics is concerned primarily
primarily with
with aspects of
behavior that depend upon
upon views about the
the future -- asset acquisition
acquisition
--
-107-
versus current consumption, the composition of assets, or nominal wage
wage
determination
determination in
in those contracts that Taylor discusses at length
length in his
paper.
paper.
macroecoIt has been
been argued convincingly that the M~of
Mi of most
most macroeco-
nomic models contain, either
either implicitly or
or explicitly, forecasting
schemes that are good schemes in some environments
environments and not in others.
(See, for example, Lucas 1976J
1976:)
of the
Moreover, careful examination of
Si reveals
reveals that
that the
the particular
forecasting schemes
imbedded in
in the
particular forecasting
schemes imbedded
the Mi
were chosen because they were good schemes
schemes in particular environments.
environments.
The inconsistency arises because the environment implied
implied by all the
Mi -- including various
various specifications for policy
policy -- may not correspond
--
——
at all to that assumed in the various Si.
This kind
kind of inconsistency
is avoided by using aa perfect
perfect foresight (rational expectations) equiequilibrium concept.
concept.
By
By using that concept, the economist avoids imposing
on
the individuals whose behavior is being modeled any fixed way of
of
on the
extrapolating
attrib—
extrapolating from the past, and
and ensures that
that he
he or she is not attributin~to
uting to them views
views about
about the future that make no sense for the envienvironment they are in.
Now having said that perfect foresight is an equilibrium concept,
its merits or its
it should be evident that it is misleading to discuss its
implications
policy
implications in terms of a particular
particular policy conclusion
conclusion like
like "policy
(whatever
matter.”
(whatever that means) does not
not matter."
The perfect foresight
equilibforesight equilib-
rium concept has been
been around
around for aa long
long time.
It
It would be surprising,
indeed, if
if that
that concept alone implied aa result like “policy
''policy doesn’t
doesn't
matter.’
matter.
11
In genera
general,
1, of course, by themselves equilibrium concepts
imply very little.
The importance
importance of the
the perfect foresight equilibrium
equilibrium
has nothing to
to do
do with the validity
validity of some vague conclusion
concept has
-108-
like “policy
"policy does not matter.”
matter." Why,
Why, then,
then, all the attention
attention to “policy
"policy
1
doesn’t
doesn
t matter”
matter 11 in this
this morning’s
morning's papers?
In 1975, there appeared a paper by
by Tom Sargent
Sargent and me in which
which aa
result of
result
of that
that sort
sort was
was obtained.
obtained.
We
M1, MM
We took
took aa particular
particular Ml'
MN'
22 , ... , MN.
one that we argued resembled in many
many respects standard macro models,
and
and replaced aa fixed forecasting scheme, one of
of the M~
Mi, by perfect
,
foresight.
We argued that the replacement made aa great difference
difference for
the implications of the
the model.
In particular, under
under perfect
perfect foresight
and certain other assumptions, all policies in
in a certain
certain class gave
rise to the same equilibrium
equilibrium values for real variables.
did not follow under
under the fixed forecasting scheme.
This
This result
Our message was,
was,
of forecasting scheme imposed matters
therefore, that the kind of
matters greatly.
greatly.
Such aa message, though, is very different from one
one that says
says that
that the
perfect
seriously as
perfect foresight version should be taken seriously
as aa model
model of
of this
or any other
other economy.
From
From the discussion above -- and from remarks in
in
-—
our 1975 paper -- it should be
be evident that the imposition
imoosition of aa perfect
perfect
——
foresight equilibrium concept
concept does not by itself turn aa hodgepodge of
of
indefensible relationships
relationships into
into aa coherent model.
The Sargent-Wall
Sargent—Wallace
conace “policy—doesn’t-matter”
"policy-doesn't-matter" result is
is to
to be con-
trasted with aa neutrality
neutrality result obtained
obtained by
by Lucas (1972).
(1972).
The Lucas
result was obtained from aa model that is coherent in the sense that its
conclusions are derived from aa mutually consistent (and defensible) set
of
of assumptions, aa single S.
S.
The Lucas neutrality result, however,
applies only to
to alternative deficits_consi~~j~~~of
deficits consisting of money transfers that
individuals know they will receive in proportion to
hoj4jjiyipf
to their holdings
of
money.
This is neither monetary policy in the sense of open market
operations -- there is,
is, in fact, only one asset in the Lucas
——
-109-
model -- nor is it the kind of fiscal policy that any country ever
mddel
—-
The Lucas model is important because it is the first coherent
follows.
model that implies anything like Phillips curve correlations.
The
The
model implies
implies that it is not legitimate to
to extrapolate from these corcorrelations for the effects of
of different
different policies.
What is new
new about the 1970s and what
what offers bright prospects for
the 1980s
1980s is not so much
much the view II have
have set out about
about the illogical
illogical
structure of standard macroeconomics.
That view can, II think, be
be found
in
in Leontief
Leontief (1947) and
and Koopmans (1947) and, II might
might add,
add, in
in the
the attiattitude
of many
nonmacroeconomists toward
tude of
many nonmacroeconomists
toward macroeconomics.
macroeconomics.
What is
is new
and
What
new and
exciting about the 1970s
1970s is
is the progress we have made in devising dedefensible
can explain
explain aa wide
fensible assumptions
assumptions that
that can
wide range
range of
of macroeconomic
macroeconomic
phenomena.
phenomena.
Lucas (1972)
( 1972) is an outstanding example.
In the
the work on
search
search and matching
matching models (see, in particular, Mortensen
Mortensen 1979), we
we see
see
the
the beginnings of aa theory of unemployed resources.
And, perhaps,
perhaps, in
in
new work on money
money (see,
(see, for example, Kareken and Wallace 1979), there
are
are ideas about how to confront long-standing
long-standing problems in monetary
theory.
theory.
Although II think we are making rapid progress, the profession
is very far from
from having reached
reached aa consensus.
First,
First, not
not everyone, by any means, agrees that
that we
we must
must completely
abandon
abandon the style of
of macroeconomic theorizing and modeling
modeling that II have
described
described above.
above.
For
to do
do that
that is
abandon macroeconomics.
For many,
many, to
is to
to abandon
macroeconomics.
This
This is
is right if macroeconomics is defined by aa style of modeling.
But
But
macroeconomics is defined by
if, instead, macroeconomics
by the phenomena it
it seeks
seeks to exexplain and by the policies it seeks to analyze, then this is
is not aa call
for abandoning macroeconomics.
It is a call
call for abandoning aa fallacious
fallacious
style of reasoning that has evidently
evidently gotten us nowhere.
-110-
Second, even
imong those who agree that we must, as it were, start over in macroeco1mong
macroecojomics and
1omics
and monetary
monetary theory, there is little agreement
agreement about how
how to pro—
pro-
:eed.
For example,
example, in my very brief listing of promising developments,
For
[ did not include disequilibrium theory.
In my view, diso~uilibrium
disequilibrium
:heory is not
not very promising,
promising, but many
many economists disagree.
Given the lack
lack of consensus on theory, it would be
be surprising if
:here were consensus on policy.
And there is not.
Academics, of
;ourse, thrive on controversy, which very naturally
naturally accompanies the
levelopment of
of substantially new theories in aa field.
:ontrast, seek consensus.
Policymakers, in
Since the economics profession is far from
iaving reached
consensus on
envy the
1aving
reached consensus
on macroeconomic
macroeconomic policy,
policy, II do
do not
not envy
the
;ask of policymakers in the 1980s.
l980s.
:ask
The absence of professional concon-
ensus leaves policymakers in the position of having to make up their
;ensus
1wn minds.
~n
-111—111—
REFERENCES
REFERENCES
“The Transactions Demand for
Baumol, W. J. (1952) "The
for Cash: An
An Inventory
Inventory
Theoretic Approach,”
Approach," Quarterly Journal of Economics 66,
66, 545-556.
Bryant,
N. Wallace (1979) "The
of Interest-Bearing
Interest-Bearing
Bryant. J. and N.
“The Inefficiency of
National Debt,”
Debt," Journal of ~
Political Economy 87,
87, 365-381.
Kareken, 3.
J. and
and N.
N. Wallace (1979) editors,
editors, Models of Monetary Economies,
Federal
Federal Reserve Bank of Minneapolis.
Koopnans,
Theory,” Review of Economic
Koopmans, T. (1947)
( 1947) “Measurement
"Measurement Without Theory,"
Statistics 29, reprinted in R. A. Gordon and L. R. Klein, eds.,
~
Reac!Tiiijs,n Business Cycles (Homewood:
(Homewood: Irwin 1965).
1965).
Leontief, W. (1947) "Postulates:
“Postulates: Keynes’
Keynes' General Theory and the
Classicists," in The New
New Economics, ed. by
bys.
Classicists,”
S. Harris (New York:
Knopf).
R. E.
E. Jr. (1972) "Expectations
Neutrality of Money,"
Lucas, R.
“Expectations and the Neutrality
Money,”
103—124.
Journal of
of Economic Theory 4, 103-124.
(1976) “Econometric
"Econometric Policy Evaluation:
Evaluation: AA Critique,”
Critique," in
in K.
---,Brunner
Brunner and A. Meltzer, eds.,
eds. , The Phillips Curve and Labor
Markets, Volume 1l of the
Carnegie—Rochester Conference
the Carnegie-Rochester
Conference on Public
Public
Policy, aa supplementary series to the Journal of
of Monetary
Economics (Amsterdam:
19—46.
(Amsterdam: North Holland),
Ho 11 and), 19-46.
Miller,
H. and
Miller, M. H.
and D. Orr (1966) “A
"A Model of the Demand for Money
Money by
Firms,”
Firms," The Quarterly
Quarterly Journal
Journal of
of Economics 80, 413-435.
Mortensen,
Mortensen, D. (1979) “The
"The Matching Process as a Non—Cooperative/BarNon-Cooperative/BarGame.” The Center for Mathematical Studies in Economics
gaining Game."
and Management Science, Northwestern University Discussion Paper
No. 384.
384.
Sargent, T. and N. Wallace (1975) “‘Rational
'"Rational Expectations’
Expectations' the Optimal
Monetary Instrument and the Optimal Money
Money Supply Rule,”
Rule," Journal
of Political
83. 241—254.
Political Economy 83,
241-254.
Tobin, 3.
J. (1956)
(1956) “The
"The Interest Elasticity of
of Transactions
Transactions Demand for
Cash," Review of Economics and Statistics 38,
38, 241—247.
241-247.
Cash,”
Wallace,
(1979) "Why
in Foreign Exchange are Different
Different From
Wallace. N. (1979)
“Why Markets in
Other Markets,"
Reserve Bank
Bank of Minneapolis Quarterly
Other
Markets,” Federal Reserve
Fall.
Review, Fall.
-112—112—
DISCUSSION OF
OF THE TAYLOR
TAYLOR PAPER
DISCUSSION
Hyman Minsky
For
For this conference John
John B.
B. Taylor has prepared
prepared aa survey paper
paper
titled "Recent
“Recent Developments in the Theory
Theory of
of Stabilization Policy.”
Policy."
Such aa survey is useful
useful as it
it develops the critical issues in
in the
field, indicates
indicates what progress has
has been made, defines the questions on
on
the research frontier and serves as guide through an important literaliterature.
Its usefulness depends upon the competence,
competence, taste and vision
vision of
of
the author.
John B.
B. Taylor
Taylor holds aa position and has the credentials that
that bebespeak of
speak
of competence.
competence.
The paper
us is
academic exercise
The
paper before
before us
is an
an academic
exercise that
that
author’s command over a literature which is sometimes
illustrates the author's
sometimes
technically demanding.
The paper also shows that he is able to ignore
the developments in economic theory and the economy which
especially
which are especially
relevant for stabilization policy and the theory thereof.
Hence in
reading Taylor’s
Taylor's paper II was led to question the taste and vision that
guides him and the literature
literature he surveys.
The theory of stabilization policy is important only as
as it
it serves
as aa guide
to action
action in
in an
an unstable
unstable world.
world.
as
guide to
The
The topics
topics and
and the
the literalitera-
ture
ture that Taylor has
has chosen
chosen to
to cover are
are not
not useful to
to anyone
anyone seriously
seriously
involved in stabilization policy; one cannot derive any guide for
action with respect to
to the serious issues of
of stabilization policy from
Dr.
Or. Minsky is Professor of Economics at Washington University in
St. Louis.
-113-
this survey or from the underlying papers.
papers.
Therefore the paper serves
1
showcase for Taylor’s
Taylor s talents.
talents.
no useful purpose aside from being aa showcase
In
aa similar vein, the underlying literature may be
be best interpreted as
the products
products of aa game
game played for academic advancement.
In selecting what to discuss Taylor ignores the literature which
quite clearly demonstrates that nee-classical
neo-classical aggregate
aggregate economics,
which focuses on price or wage rigidities and which introduces
introduces money as
an exogneous variable, will not
not do.
The
The literature he
he focuses on
on looks
to refining and making more precise the very neo-classical formulations
whose
whose logical consistency and empirical relevance
relevance has been
been demolished
by developments in
in theory
theory in recent years.
However one rule of the
the
game
and the authors of the reviewed literature play is
game Taylor
Taylor and
is that rere11
search is to be
be carried on
on “within”
within 11 the neo-classical
nee-classical model; thus taste
taste
search
and vision conspire to rule out the relevant and the serious
serious because
because it
it
is unorthodox.
The most important developments
deve 1opments for the theory of sstabilization
tabi 1i zation
l970s were not
policy during the late
late 1960s
1960s and 1970s
not in the literature
literature but
in the "world."
“world.”
The
The observations that theory has to explain
explain and the
stabilization policy has to
to contend
developments in the economy that stabilization
with
with changed
changed radically in the mid 196Os.
1960s.
In particular, stabilization
stabilization
policy now has to
finanto deal with threats and
and partial realizations
realizations of
of financial instability as well as with stepwise
stepwise increasing unemployment and aa
stepwise acceleration of inflation.
For
For aall
11 who take “our
"our economy”
economy" rather than the literature
1i terature about
economics derived from the neo-classical (monetarist
(monetarist and pseudo—
pseudo-
Keynesian) research program as
as the subject matter of their
their research,
the world underwent aa marked change of
of state around 1965.
-114-
The
The
Instability
instability that policymakers have to contend with after 1965 is of aa
different order of magnitude
magnitude and the
the potential
potential consequences of mismanmisman-
agement of
of stabilization policy are much more
more serious
serious than
than earlier in
the post-war period.
The
The financial system and practices evolved from
1945 to 1965
1945
1965 so
so that the system, which had been
been virtually impervious to
financial instability, became highly susceptible.
Between 1945
1945 and
1965 there were no threats of aa financial crisis
crisis of
of the scale which
which
could usher in aa deep depression;
depression; in the years
years that have followed there
have been
been at least three such
such threats within the United
United States,
States, as well
as aa number of threats to the stability of
of the international financial
system.
Whenever
Whenever financial instability threatens to trigger
trigger aa debt-
deflation process, policy interventions by both the government
government and the
make aa difference in the path of the economy
economy
central bank can really make
through calendar time.
time.
Nothing in
an apin the paper before us exhibits an
ap-
preciation of the change in the
the character
character of the
the “stabilization
''stabilization probproblem”
lem11 over the years
years surveyed.
Once the potential consequences of the
the mismanagement
mismanagement of policy
becomes so
so much more serious,
serious, the importance of
of economic
economic theorizing
about stabilization policy increases.
In
In particular, economic theory
needs to be relevant in the sense that the critical
critical situations
--
- in
this case financial instability and the way
way in which
which financial varivariables affect aggregate
aggregate demand -- are well defined
defined within the theory.
--
If theory is based upon
upon misspecifications
misspecifications of
of the
the economic
economic process and
the problems faced by policymakers, then theory cannot be
be relevant:
garbage in -- garbage out applies to
to theory construction
construction as well as
as to
garbage
——
computer modeling.
—115—
-115-
The problems of the economy have been exacerbated because policymakers have been guided by insights and conclusions drawn from neoneo-
classical theory.
Neo—classical
for
Neo-classical theory is an
an inappropriate tool for
dealing with instability,
instability, for financial or any other instability is
foreign to this theory.
In neo-classical theory any deviation from
equilibrium must be due to exogenous developments and any sustaining of
a disequilibrium must be
be due to “barriers.”
barriers.
11
11
Neo-classical theory is
able to explain instability only by postulating the existence of
of one
one or
more devils, be they trade
trade unions,
unions, OPEC, monopoly, the
the central bank,
government or democracy.
Because economic policy
policy advising over the
past decades has been largely monopolized by practitioners of neoneoclassical theorizing our current economic malaise is in good measure
iatrogenic.
The physicians, including our hosts, have served to
to make
the disease worse.
worse.
AA theory of stabilization policy is needed if and only if
if the
economy is unstable. There is no sense whatsoever to
to the concept
‘stabilization
if the
is stable.
stable. When
"stabilization policy’
policy" if
the beast
beast is
When Wallace,
Wallace, Sergent,
Sergent,
al play their games
et al.
games by positing aa system
system whose behavior is deterdeter.
mined by
by elements that are independent of the variables that, in
in their
specification, stabilization policy directly affects, then the propoproposition that
policy does
does not
sition
that policy
not matter
matter is
is true
true not
not by
by demonstration
demonstration but
but by
by
assumption.
As the instability that is so evident in the world cannot
occur within their models, the games they play only serve to show that
their models and the empirical tests that they perform
perform are irrelevant
for our economy.
In my view the strong proposition that emerges from
one literature
literature surveyed by Taylor, is that this large body
body of work is
irrelevant for
for the world in which we live.
live.
-116-
If economics is to be
anything more than an academic nit—pick,
nit-pick, theory and theorizing has to
go in other directions than those represented in the literature
literature Taylor
surveys.
If economic theory is to be relevant for stabilization policy in
“why and in
be addressed are "why
in what
our economy, the questions that must be
way is our economy unstable?”
unstable?
11
Note the phrase “our
our economy.”
economy."
11
subThe sub-
ject matter of any theory that aims to
to be relevant is not an abstract
economy devoid of institutional detail but rather an economy
economy that is
rich in specific institutional detail and which exists at aa particular
time and has aa special history.
The problem of
of economic theory is to
select the essential details of the institutional framework
framework to
to model:
the aim of the theorizing is to show causal connections that lead to the
observed instability.
instability.
The hope is that by showing how instability is
generated the theory will indicate policy interventions which can atatif not eliminate
tenuate if
eliminate instability.
Although the lines of argument examined by
by Taylor
Taylor are largely
irrelevant to the topic of this conference, “Stabilization
"Stabilization Policy:
Policy:
Lessons
l970s and Implications from the 1980s,"
1980s,” there were
Lessons from the 1970s
developments in theory over the past decade that are relevant to stabistabilization policy: Taylor either is ignorant of these developments or
chose to ignore then.
them.
The developments in economic theory in recent
years that are relevant to the theory of stabilization policy
policy are:
1)
l)
Progress in general equilibrium theory
2)
The two-Cambridge
two—Cambridge debate
3)
3)
The
The recovery of the “lost”
"lost" financial elements
elements in
in Keynes.
Keynes.
Because II am writing a comment rather than a survey article II will just
just
devote one paragraph to each of these developments.
-117—117—
years progress
progress in general equilibrium theory
theory made
During recent years
be satisfied for the key propositions of
the conditions that need to be
this theory to be
be valid precise.
One conclusion of these developments
equiis that the coherence and coherence—seeking
coherence-seeking theorems of general equilibrium theory are not unconditionally valid for aa decentralized set of
of
institutions
markets with capital assets, money, banking and financial institutions
such as we have.
have.
An implication of
of this conclusion
conclusion is that the introintro-
duction of
of money as an "exogenously
“exogenously determined"
determined” instrument designed to
facilitate trade into
into a general equilibrium model in which relative
prices determine consumption and production decisions throws no
no light
whatsoever on the behavior of aa capitalist economy with
with aa “money”
"money" that
that
is created in aa banking process.
There is no established nicroeconomics
microeconomics
that can serve as a basis for aa macroeconomic or monetary theory that
is relevant to
micro—
to stabilization policy as long as
as the results
results in microeconomics depend upon highly artificial constructions to explain the
1
existence of and changes in money.’
money.
1Of the general equilibrium theorists, perhaps F. H. Hahn has
1of the general equilibrium theorists, perhaps F. H. Hahn has
l imi tat i ans of theory. See F. H. Hahn: “On
"On
been most open about the limitations
Some Problems of Proving the Existence of an Equilibrium in
in aa Monetary
Economy," in R. Clower (ed.), Monetary Theory
Theory (Penguin,
(Penguin, 1969).
7969).
Economy,”
Professor Friedman
Money," Economica,
EconOmica, February 1971, 38
“Professor
Friedman’ss Views on Money,”
61—80.
(149), pp. 67-80.
On the Notions of Equilibrium in Economics (Cambridge:
(Cambridge: Cambridge
Cambridge
University
Press7i~7~T~~
University Press,
7973).
11
1
Also see
K. Arrow
Arrow and F. H. Hahn, General Competitive Analysis, (San Francisco
Holder Day, 7977),
1971), especially Chapter 14,
74, The
The Keynesian Model, pp. 347note that in their earlier
earlier
369. In introducing their discussion they note
proof that aa temporary equilibrium always exists they
they"“...supposed
... supposed that
that
at the moment
moment an
an equilibrium was shown
shovm to exist, economic agents
agents had no
capital assets as we know capital assets. It is interestinq to
to note
note
that Arrow and Hahn head
head Chapter
Chapter 14
74 with aa quotation
quotation from H.
W. B. Yeats,
Yeats,
The Second Coning,
hold.”
Coming, “Things
"Things fall
fall apart, the centre does
does not hold."
-118—118—
The two-Cambridge debate, ostensibly about capital theory, was
really about the validity of the integration of Keynesian theory with
the earlier neo-classical theory.
The critical issue that the debate
clarified centered around the pricing of
of capital assets.
economy is characterized
characterized by two
tv,o price systems.
systems.
capitalist
AA capitalist
One is
is the price system
of current output,
output, the second is the price system
system of capital assets.
The price system of current output largely depends upon wages and markmarkupon current
ups, whereas the price system of capital assets depends upon
estimates of future expected profits, current estimates of
of the
the unceruncer-
tainties involved over various horizons, and current capitalization
rates of
of profit
profit streams.
streams.
In
an economy
monetary, banking
and
In an
economy with
with the
the monetary,
banking and
financial systems that characterizes capitalist
capitalist economies the capitalcapitalnmonetary" phenomena and the two price systems can
ization rate is a “monetary”
and do
do vary relative to each other.
Inasmuch as the ratio of the
capitalized values of expected future profits
profits to the supply price of
investment
investment output is aa determinant of
of investment demand,
demand, aggregate
aggregate dedemand is sensitive to
to the ratio
ratio of
of these two
two sets
sets of prices.
prices.
The
The two—
two-
Cambridge debate is of vital importance for the
the theory of stabilization
stabilization
conclusion that if the ongoing processes
policy because it leads to the conclusion
of an economy affect this ratio it
it will lead
lead to
to endogenous
endogenous change in
the performance
performance of
of the economy:
i.e., variations in
in the ratio of
employed to available resources will result.
The two—Canbridoe
two-Cambridae debate
made it
it clear that the “proofs’
"proofs" in the literature that aa growth
grm;th equilibriequilibrium of an investing capitalist economy exists depend upon the
the assumpassumption that the present
present value of future profits
profits always equals the perperpetual inventory valuation of
of capital assets.
assets.
But the
the equality
equality of
of the
two valuations
valuations of capital
capital assets in
in an
an attribute of
of equilibrium.
-119—119—
The
"proofs" of the coherence of an investing capitalist economy does not
“proofs”
coherhold; the proofs depend upon first assuming that aa condition of coher. t s. 2
ence ex1s
exists.
The third theme in economic theory in the 1970s
l970s that is relevant
to
to stabilization policy is the recovery of the financial and monetary
monetary
aspects of Keynes'
aspects
Keynes’ revolution in economic theory.
There is
is something
There
very queer about the standard interpretation of Keynes as embodied in
in
the various IS-LM models.
This essentially non-monetary
non-monetary view of
of the
economy is paraded as aa representation of the theory of the major ecoecolife’s work was almost entirely on
nomic theorist whose
whose life's
on money and
finance.
In the recovery of what lost in
Hicks-Hansen—Kleinin the Hicks-Hansen-Klein-
Modi gl i ani-Pati nkin tradition it became clear
cl ear that underlying Keynes’
Keynes'
Modigliani-Patimkin
theory was the premise that to understand capitalism it is necessary to
model capitalism.
This means that
that it is necessary to model the way
positions in capital assets and investment are financed, the dependence
banking and financial system, and the effects
of this financing upon the banking
of financing relations first upon
upon investment
investment and then on
on income, employemploy-
ment and prices.
In this analysis, in aa capitalist
capitalist economy unemployIn
unemploy-
ment exists when the long run expectation of profits
profits by
by business men
together with capitalization
capitalization rates that reflect portfolio preferences
in an uncertain world lead to
to demand prices for capital assets that
that are
"too
low" relative to the supply prices of investment
investment output.
output.
“too low”
The dede-
of investment
mand price for capital assets as well as the supply price of
2This is the outcome of the two-Cambridge debate on Capital
2
This is the outcome of the two-Cambridge debate on Capital
Theory, although the standard discussion and summary of the debate,
11
6. C. Harcourt, “Some
Cambridge Controversies in the Theory of Capital:’
G.
Some Cambridqe
Capital."
(Cambridge,
(Cambridge, England:
this clear.
The Cambridge University Press) does not make
-120-
output depend upon financing terms.
te;y;s.
terms. which cannot
Financing terms,
fu1ly
be captured by
by aa single interest
or not
fully be
interest rate, reflect whether or
term expected behavior of the
recent and near terrn
the economy lead
lead to
to suffisufficient realized and expected profits
profits that almost all
a 11 of
of the payment
commitments on outstanding obligations are expected to be
be fulfilled.
By integrating money, finance, expected profitability and the supply
price of
of current output into
into aa theory of effective demand, Keynes dedeveloped the basis for aa theory of the economic processes of aa capitalcapital-
ist economy that explained why such
“so given to
such an economy is
is "so
to fluctufluctuations.”
ations."
Instability is an inherent characteristic of
of aa capitalist
Keynes’ theory.
economy in Keynes'
Furthermore, Keynes’
Keynes 1 theory is rich, for
even though it does not lead
lead to a set of policies which eliminate
instability, it does lead
lead to policy moves ((fiscal
fi seal policy) which offset
3
the effects of instability upon employment
employment and aggregate income. 3
l970s natured,
As the 1970s
matured, history advanced the argument from the
simple question of “why
"why is our economy unstable?”
unstable?"
The question that
economic theory had to address if
if it was to be
be relevant
relevant to
to stabilistabili11
zation policy
po 1 icy became “why
why is it that our economy is so much
much more unun-
l9SOs?”
stable in the ll970s
970s than in
in the 1950s
?"
The issues that theory had to
3Miong
works” in the reintegration of money are:
Among the “key
key works
Joan Robinson, Economic Heresies, (London: MacMillan,
MacMillan, 1971).
1971). P.
Davidson, Money and the Real
Real World, (New York: John Wiley && Sons,
1972).
J. A. Kregal
Kregal, The Reconstruction of Political Economy, (London:
l97~)7
MacMillan, 1973
.
S. Weintraub, ~j~y~nesian
~nesian Theory of Employment, Growth and Income
Distribution, (Phil
adelphia, Chilt~5~FW66JT
(Philadelphia,
Chilton 1966).
Victoria Chick, The Theory
Theory of Monetary Policy, (London: Gray-Mills
Publishing Ltd., 1973).
i973).
H. P. Minsky, John~y~nard~ynes
John _Maynard Keynes (New
(New York: Columbia
Columbia University
Press, 1975).
11
11
,
—121-121-
address can be made even more precise by
by dividing the
the question
question Into
into two
parts:
“Why is it
"Why
it that
that our financial system seemingly is more
more unstable,
more vulnerable to
to threats and partial
partial realizations of
of financial crises
(both
“Why is
(both domestic and international)
international) since the middle
middle l960s?”
1960s?" and "Why
it
1g705 progressed?”
it that
that inflation became more serious as the 1970s
proqressed?'
1
Once economic
economic theory moves from the study
study of an
an economy to
to the
the
of the instability of
economy and once the
the various faces of
of
study of
of our economy
our
our economy are taken as the problems theory must address
address then the
the need
to
to model money, banking and capital-asset pricing moves to
to the
the foreforeground.
ground.
In Taylor’s
Taylor's survey, which presumably
presumably deals
deals with
with stabilization
policy, banks and banking are nowhere discussed.
discussed.
We
l✓ e all know
know that in
our economy money
money is created
created by
by the actions
actions of profit seeking banks and
other financial institutions,
institutions, that the assets acquired
acquired and
and liabilities
liabilities
issued by
banks evolve
evolve in
to profit
and that
issued
by banks
in response
response to
profit opportunities,
opportunities, and
that
the ~ix
nix and activities of
of financial institutions also
also evolve.
This
This
implies that
am economic theory applicable
app 1i cable to our
our economy will
wi 11 intei ntet:1at an
grate banking and financing markets into the determination
capital
determination of
of caoital
and the determination of
asset prices, investment decisions und
of the domain
of
of stability
stability of the
the economy.
economy.
ignoring it.
You
You cannot
cannot understand
understand something
something by
by
The literature discussed by Taylor’s
Taylor's paper
paper iqnores
ignores
his selection
selection of
of the
the literature to discuss, apapbanking and Taylor, by his
parently believes you can understand and
and give guidance for stabilization
policy for our economy by ignoring banks and banking.44
policy for our economy by ignoring banks and banking.
4
It would
today’s economists were acquainted
would be useful if today's
acquainted with
H.
H. Simon’s
Simon's “Rules
nRules vs. Authorities in Monetary Policy,”
Policy,u Journal of
Political
l~37.
Political Economy.
Economy, 1937.
—122—
-122-
It has
has long been argued that the instability
instability of the economy
economy is
related
related to the
the structure
structure of liabilities by which
which positions
positions in
in capital
assets are financed.55 Experience during the l97Os lends substance to
assets are financed.
this argument.
Experience during the 1970s lends substance to
The
The relation between
between the debt financing of
of capital
assets positions and the need to
cormiitments on maturing debt
to fulfill commitments
by rolling over debts - by
by issuing new debts - is aa critical deterdeter—
-
minant of the stability
stability of
of an
an economy with sophisticated finance.
As
As aa
the maturing of the
the flow of
of funds data (poorly designed as
result of the
the
the set of
of accounts may be) it is possible to relate the evident
evident instainsta-
bility of our economy to the growth of the debt structure
structure relative to
to
income
income and the increased complexity of financial relations.
In
In order
order
to answer questions about why our economy is unstable it
necessary
it is necessary
to fully integrate the monetary
monetary mechanism
mechanism with system behavior.
The
literature Taylor
“silent” on
Taylor surveys is “vague”
"vague 11 or "silent"
on the processes by
which
capita] assets are financed.
which positions in capital
One striking characteristic of our economy that became evident
evident in
in
the 1970s
l97Os is the
the link between financial instability and accelerating
accelerating
iinflation.
nfl a ti on.
Since
l96Os whenever
Si nee the
the mid 1960s
v,henever the Federal
Federa 1 Reserve follows
fo 11 ows
the rules for monetary
monetary policy to constrain inflation that were
were develdeveloped on
on the basis of
of the experience of the l94Os
1940s and ‘SOs,
'50s, aa financial
crisis develops; when the Federal Reserve and the government succeed in
containing the crisis so
so no deep and long recession follows, the
the finanfinancial base
base is laid down for inflation at
at aa higher rate.
Since the
Since
middle 1960s 1t.'e
have had three “cycles”
llcyclesu of inflation,
inflation, constraint,
middle
we have
5 1. Fisher,
~I.
“The Debt-Oeflectiun
Great Depressions,”
Fisher, "The
Debt-Deflection Theory
Theory of
of Great
Depressions,"
Econometri
cas 1933.
Econometrica;;_,
,
—123—
-123-
Incipient
incipient financial crises, lender of last resort
resort intervention, federal
renewed expansion, financial innovation and accelerated infladeficits, renewed
infla-
tion.
"sequence" took
took four to
to five years
years to work
In each cycle this “sequence”
its way
way through the system.
Any theory that is useful
useful for stabilization policy
policy will need to
explain
explain why the economy reacted to variations in
in the rate of
of growth of
the reserve base, or
or in one manner in the years
years prior to 1965
in
1965 and in
another manner in the years since 7965.
1965.
For an
an economic
economic theory
theory to do
this it need
need contain aa sub—theory
sub-theory of
of “financial
"financial stability and instai nstability.”
bi
l ity."
Nowhere in
Taylor’s survey or
in Taylor's
or in
in the literature he
he surveys is
this aspect of the stabilization problem addressed.
Any theory of the capitalist process needs to focus in
in the decideci-
sions to own
own capital assets, the techniques
techniques used
used to
to finance control
over capital-assets and
and the investment and investment financing propro-
cesses.
Obviously
Obviously aa theory,
theory, if
if it
it is
is not merely mechanistic, which
decisions today that are based upon future revenues and costs
explains decisions
will include aa theory of
of expectation formation.
The fundamental
problem in the making of
of decisions today that involve revenues and
time horizon is that
costs over
over a significant time
that the future is uncertain;
the future cannot be represented by
by aa set of nice stable probability
functions over well-defined outcomes.
The need to make decisions in an uncertain world
world leads to
to one
how does one behave rationally in an irrational
irrational world?”
world? 11
question, 11“how
An
"irrational
in which what happens is not explained
explained with
“irrational world"
world” is one in
precision by
by the accepted theory.
theory.
the requisite precision
As
As long as theory does
not explain a phenomena
phenomena with
with the exactness required
required for decision, then
irrational.
the world of that phenomena is irrational.
-124-124—
If, for aa capitalist
economy, the world conforms to expectations derived
derived from standard
theory aa large part of the
(instathe time, even as it behaves in aa manner (instability) inconsistent with this theory aa part
part of
of the
the time, then decision
making formulas that use the accepted theory will not determine
determine the
rational man.
behavior of aa rational
behavior
In aa world where diverse types of behavior
behavior
excan occur,
occur, theory is effective
effective as
as aa guide to
to decision and policy
policy ex-
actly as it yields information as to
to which of
of the diverse
diverse types of
behavior of the economy is
is likely to rule.
If
If economic theory is to be
an
an ingredient in the formation of expectations
expectations by
by aa rational man, it
it
needs to relate the expected
expected behavior of the economy to
to history
history and the
evolving
evolving institutional arrangements.
The Franklin National bankruptcy of 1974
1974 and what
what followed is aa
concrete
situation in which
which policy actions truly affected
concrete example of a situation
the
the behavior of the economy.
In May of 1974 the Federal
Federal Reserve, under
under
Arthur
National so
so
Arthur Burns, opened the discount window wide to Franklin National
that all of Franklin National’s
National 's overseas and money
money market liabilities
were validated.
The Federal Reserve by
by this
this action
action aborted aa wave of
of
withdrawals from the international banks
“world of
banks and assured the "world
international finance''
finance” that the offshore liabilities of
of large,
large, if not
respectable, American
American banking
banking institutions
institutions were implicit
implicit contingent
liabilities of the Federal Reserve.
Reserve.
This and related
related interventions by
the
cooperating institutions
institutions in
in 1974-75
1974-75 together
the Federal
Federal Reserve
Reserve and
and cooperating
together
with massive government deficits made it virtually
virtually certain that
that the
recession of 1974-75 would be contained
contained and that the subsequent rerecovery would lead
lead to serious balance of payments difficulties and ininflation at an accelerated
accelerated rate.
rate.
Policy
Policy may not always matter, but
but
-125—125—
there are junctures
junctures in
in the history of
of an economy when policy really
matters:
1974-75 was one such
such juncture.
juncture.
It is the duty of economists
economists who
who parade as
as knowing something
issues.
about stabilization policy to be aware of such issues.
the
Neither the
literature Taylor discusses nor Taylor in
in his
his paper
paper seem to
to be aware of
problems.
these problems.
policy needs
Theory that is useful for stabilization policy
to offer
offer guidance to central bankers and other
other policynakers
policymakers when
when they
to
situation such as ruled in
in 1974—75.
1974-75.
are faced with the need to act in aa situation
Taylor’s paper nor the literature he
By this criteria, neither
neither Taylor's
he chose to
report on are useful.
-126-126—
THE CASE FOR GRADUALISM IN POLICIES TO REDUCE
REDUCE INFLATION
INFLATION
Allan H. Meltzer
Inflation is usually defined as aa sustained rate of
of increase
increase in aa
broadly
of prices.
prices.
broadly based index
index of
Whatever meaning
meaning one gives
gives to
to the
the imim-
precise term !!sustained,!!
sustained, the past fifteen years seem to
to meet
meet the
standard.
Both the all—item
all-item consumer price index and the implicit GNP
deflator have increased in every
every quarter since late 1965, and neither
seems likely to reach
reach aa zero rate
rate of change in the near future.
Sustained
Sustained inflation
inflation at the rates of recent years is rare, even
even if
not unique, in the histories
histories of developed economies.
economies.
It seems
seems useful,
at aa conference summarizing the lessons of the seventies and
and drawing
to look back on the path we
we have travtravimplications for the eighties to
elled
elled and to
to explore
explore the path we might take to restore
restore price stability.
stability.
II shall use the
the opportunity
opportunity to
to discuss
discuss some of what
what has been
been learned
learned
about
about monetary policy.
The list is
is aa long one, particularly
particularly if we inin-
clude
known but later
clude propositions that once were "known"
later forgotten or rerejected
jected in the years of Keynesian
Keynesian orthodoxy, so II shall not attempt
attempt to
be complete.
Any
gain from ending inflation depends on
Any long-term gain
on aa negative
negative
relation between inflation and real output.
The
The most common reason for
suspecting that aa gain will occur is the
the observed association between
inflation and
and changes in relative prices.
See Cukierman (1979).
(1979).
The
Or. Meltzer is Professor of Economics and Social Sciences
Dr.
Sciences at
at CarnegieCarnegieMellon University.
University. The author is
is grateful
grateful to
to Alex
Alex Cukierman
Cukierman and
and
Jerry L.
Jordan for
for helpful
comments on
draft.
Jerry
L. Jordan
helpful comments
on an
an earlier
earlier draft.
—127—
-127-
principal problem for monetary policy at present is to achieve this
gain
gain by ending inflation at minimum
minimum transitional loss of output.
Every six months, II join
join with my colleagues on the
the Shadow Open Market
Committee
Cammi ttee in recommending
recommending aa policy
po 1icy of pre-announced, gradual,
gradua 1 , sustained
sustained
reductions in the growth of money as aa means of restoring price stabilstability.
ity.
A clear
A
clear statement of
of the
the reasons
reasons for
for aa policy
policy of this
this kind
kind
often called gradualism
gradual ism
--
has
has not been provided.
provided.
—-
II will
will try to parpar-
tially fill that gap and to
to relate the case for gradualism to
to some of
the lessons we have learned from recent experience with
with sustained
sustained ininflation.
The history of recent inflation is
is surrounded
surrounded by myths that
that obobits persistence.
scure the origins
origins of the inflation and the reasons for its
persistII begin with an account of the origin and an explanation of persist-
ence.
Much of
of the case for gradualism depends
depends on
on the way in which
which inin-
dividuals
dividuals form anticipations
anticipations of the future.
I present one view of
rational
rational expectations,
expectations, in
in the
the sense of
of Muth (1961), and
and use
use this
this model
model
of
of expectations to show how Federal Reserve policy
policy procedures can conconvert real shocks into permanent
permanent changes in the rate of
of price change.
change.
Then II present the
the case for gradualism in aa world
world in which persistent
and transitory
transitory changes in monetary
monetary policy cannot
cannot be identified quickly.
THE
THE ORIGIN AND PERSISTENCE OF CURRENT
CURRENT INFLATION
The most
most enduring myth about the origins of the current
current inflainflation is that the inflation started during the
the Vietnam war.
viar.
According
According
to
of history, President
President Johnson rejected the
the recomrecomto aa standard version
version of
11
mendations of his advisers by refusing to
to choose
choose between
between “guns
guns and
butter.”
butter."
The
delayed
asking Congress
The President
President de
1ayed asking
Congress for
for increased
increased taxes
taxes
-128-
(or for smaller expenditures for redistribution) and allowed the budget
1967.
deficit to overstimulate the
the economy in 1967.
has been intractable.
has
Since 1967, inflation
some estimates,
or more years
years
According to some
estimates, ten or
of recession would be required to eliminate inflation
inflation by monetary
monetary and
fiscal policies.11
The facts do not correspond to
to this capsule history.
The rate of
increase of
of consumer
consumer prices
prices reached the
the 33 to
to 4% range
range at least aa year
before the Vietnam deficits.
deficits.
Spending by the federal government in
in doldol-
5% below the 1962
1962 level
lars of constant purchasing
purchasing power
power remained
remained 33 to 5%
level
during most
most of 1965.
Budget deficits and government spending
spending did not
start the inflation or
or encourage the Federal Reserve to expand in
in 1965
1965
1966.
or 1966.
1966.
The budget had
had aa small surplus in
in 1965,
1965, and aa small
small deficit if
in
The Federal
Federal Reserve slowed
slowed the growth rate of the monetary
monetary base
late in
in 1966
1966 in aa sudden burst
burst of
of concern
concern about
about rising
rising inflation.
inflation.
The
The
1967 deficit of more than $13 billion comes after these first steps
steps to
slow inflation and much too
too late
late to explain the start
start of the inflation.
inflation.
A surtax was
was added to
SO the Vietnam defA
to the income tax in
in 1968,
1968, so
deficit proved
proved to
to be temporary.
By late 1968, the budget again was in
1969.
surplus, and
and the surplus persisted
persisted in 1969.
The 1969
1969 surplus
surplus of
of $8.5
billion is one of the largest of the past
past thirty years in
in real
real as well
as in nominal terms.
terms.
To
To sustain
sustain the thesis that the Vietnam deficits started the curcurrent inflation,
inflation, one must not only ignore
ignore the
the problem of the timing of
of
1See Perry
more complete statement
statement of this
this view and
see
(1978) for aa more
for an extreme form of the
the argument that
that inflation
inflation is intractable.
Perry’s
Perry's Phillips curve implies
implies that it costs $200
$200 billion
billion dollars of
of
real output for each percentage point reduction in
in the rate of
of inflainflation.
tion.
—129—
-129-
corrmented earlier, but must accept
the start of inflation, on which I conniented
generdeficit generthe improbable
improbable proposition that six quarters of wartime deficit
irreversible.
ated anticipations that were irreversible.
Credulity
Credulity is
is strained
strained
comfurther when the 1967
1967 deficit is expressed in
in constant
constant dollars to compare
pare with
with the deficits in
in earlier
earlier and later
later years.
years.
The
The 1967 deficit is
is
almost identical to
to the 1958 deficit when both
both are expressed in dollars
of the same purchasing power.
power.
of sustained
inflation.
sustained inflation.
The 1958 deficit did not initiate years
On the contrary, inflation fell from the 33 to
to
4% range
of 1956-57
range in
in 1958—59
to less
1%
4%
range of
1956-57 to
to the
the 11 to
to 2%
2% range
1958-59 and
and to
less than
than 1%
by 1961.
The 1975
1975 nominal budget deficit of $70 billion is four times
larger than the deficits of 1958
1958 and
and 1967 when
when the
the three are expressed
expressed
in dollars of comparable
comparable purchasing
purchasing power.
power.
The
The 1975 deficit is not
not
balanced budget
budget or aa surplus but
but by sustained deficits.
followed by a balanced
Yet, most broad
broad measures of the rate of price change declined in 1976.
1976.
deflator rose by less
The GNP deflater
less than 4.5%,
4.5%, on average, for the first
three quarters
the year.
less
three
quarters of
of the
year, and
and the
the consumer
consumer price
price index
index rose
rose by
by less
than 5% for the year as aa whole.22
The proximate
proximate cause of the start of the current
current inflation
inflation is the
monetary
of the
196Os.
monetary policy
policy of
the early
early 1960s.
Inflation persists
persists because
Inflation
because policy
policy
of future inflation by producing
to sustain
sustain anticipations of
continues to
persistent inflation.
Bursts of anti-inflation policy, and announceannounce-
ments of firm congnitments
inflation, are not followed by
by
corrmitments to reduce inflation,
policies that reduce money growth.
2The
The decline in
in the rate of inflation affected more than just
food prices
prices as
as is sometimes claimed. The wholesale price
price indexes of
of
consumer finished goods rose by less than 2.5% for the year.
—130—
-130-
>
0
CD
0
tO 0
(D~
—‘C
<0
r
—~
tOED
—
~1
CD°
o
tAG)
I
8 .0
CD
d,
o
I
Rate of Growth of the Monetary Base (3yr. moving avg.)
-
CD
0
Rote of Growth of the Monetary Bose
( 3 Year Moving Average)
ojo
9.0
w
CHART 1
c,,
C:
>
5
I
—4
~ 7.0
..:
>.
r<)
o
I
6.0
Cl>
f/)
0
Q)
-
5.0
5
...>Cl>
C:
0
b
I
0
4.0
o
3.0
‘
(iJ
o
-...
-
-Is
~
..c:
3
b
I
ro
0
e>
2.0
0
Cl>
1.0
b
I
CD
Year
0
-a
1980
(0
CD
-.4
c_fl
-a
o
1975
(0
—4
(0
1970
-a
1965
o
—
1960
(0
1955
o
L....l---'---'-...,_J........1---'---'-...,_J........1----'---'-...,_L....l__,__.__,_L....l__,__.__,__.J....J
(0
0
01
01
0::
0
0
Chart 11 uses aa twelve quarter moving average of the growth of the
adjusted
of the
effect of
adjusted monetary
monetary base
base as
as aa measure
measure of
the long—term
long-term effect
of monetary
monetary
policy.
Using this measure as an index of the sustained thrust
thrust of
policy. we can divide the monetary history
monetary policy,
history of the past twenty—
twentyfive years into
into five episodes.
The
The first, from 1955 to 1960, has
has aa low
average rate of monetary
1.1%.
monetary growth, 1.1%.
transition.
The second is
three—year
is aa three-year
The
The twelve quarter
quarter moving average rises steadily
steadily toward
the 5.5% range.
In the third period, 1964-71, the growth of the base
remains in the neighborhood of
of 5.5%.
5.5%.
The fourth period is a one—year
one-year
transition,
transition, 1972, during which the maintained growth of the base moves
from
about 5.5%
5.5% to
from about
to 8.5%.
8.5%.
Since 1973,
of the
base
Since
1973, the
the moving
moving average
average of
the base
has grown at aa maintained rate
rate of about 8.5%.
8.5%.
AA number
number of
of studies,
studies, including
including my
my own
own Meltzer
Meltzer (1977),
(1977), suggest
suggest
that inflation follows money growth
growth with an average two-year lag.
The
mean of the three—year
three-year moving average ending in
in year t, shown in Chart
1, is
is an
an unweighted average
average centered
centered in
in year
year t—l
t-1.
-
If
If we
we impose aa twotwo-
year lag, inflation
inflation in
in year t+1
t+l is influenced by
by the twelve quarter
rate of
of growth
growth of the monetary base ending in
in year t.
t.
To
To measure perper-
sistence, I have
have computed
computed the standard deviation of the percentage
rates of
change of
the consumer
price index
the percentage
rates
of change
of the
consumer price
index and
and the
percentage rate
rate of
of
change of
1956-61, 1965-72
1965—72 and 1974-78
1974—78 that
of money wages for the years 1956-61,
correspond to
to the two-year lag
lag of
of prices behind the maintained
maintained growth
of the monetary base.
base.33 The data are shown in
in Table
Table 1.
1.
3
The rates of price and wage change are one-year
one-year averages
averages of
of the
all—item consumer price index for six-month
six—month spans and average
all-item
average hourly
earnings over six-month
BCD. Wage data
data are not available
available
six-month spans
spans from BCO.
before 1965.
1965.
—132-132-
TABLE 11
TABLE
Mean
Mean (p)
(w) and Standard Deviations (cv)
(o)
Rate of
of Price
Change t+l
Growth of Adjusted
Adjusted
Monetary Base in
intt
Years ((t)
t)
1955-60
1964—71
1964-71
1973-78
Omitting 1974
1974
pµ
cva
pw
1.1
1.1
5.7
5.7
8.4
. 18
.18
.44
..31
31
1. 9
1.9
4.0
7.5
1.00
1. 26
1.26
2.42
6.4
.85
of ,iage
Rate of
Wage
Change t+l
CT
cv
p
ll
N.A.
5.9
8.0
7.7
CT
cv
1. 13
1.13
.79
.79
.36
The
The data show aa tendency for the standard deviation of the
the rates
rates
change of money and wages to
to fall
fall in recent years.
of change
Removing the
standeffects of the oil shock, by
by omitting 1974, further reduces the stand-
ard deviations.
deviations of
of the rates of change of
of
The standard deviations
wages and prices
deviaprices are not startlingly different
different from the standard deviations of the maintained growth of
of the adjusted base.
The
The persistence
persistence
of rates of price change from year to
to year appears to
to be
be related to
to the
persistence of maintained rates of money
money growth.
To
To examine
examine further the relation between
between the persistence of money
growth and the persistence of
of inflation, Table 22 compares the two
quarter average rates of growth of base
base money to the quarterly averages
averages
of the rates of
1.
of change of
of prices and wages
wages used in
in Table
Table 1.
As
As before,
before,
II imposed
imposed aa two-year
two-year lag of rates of
of price change behind rates of money
growth.
The data
the variability
variab1lity of base money
money growth
data now suggest that the
is of approximately the same magnitude
magnitude as
as the variability
variability of the
the rate
rate
is
of wage change.44 The standard deviations
deviations of
of the rate of price change,
4The time periods for the base
base differ from those in Table
Table 11 bebecause Table 11 has
has aa three-year moving average. II have
have kept the periods
for rates of price and wage change the same as in Table
Table I.
1.
-133—13
3—
however, are not closely related to the standard deviations of rates
of base money growth.
Short-term variability of the rate of
of price
change reflects more than the variability of
of monetary growth.
growth.
TABLE 22
Mean (v)
(ci) and Standard Deviations (ci)
(0)
Two quarter
moving average
average
of
of growth of
monetary base
Period
t
1954-59
1954—59
1963-70
1.1
1. 1
5,7
5.7
1972—76
1972-76
8.2
p
ci
a
0.87
1.10
1.
l0
0.91
Period
Period
t+2
1956—6]
1956-61
1965-72
196572
1974—78
1974-78
Standard Deviations
Deviations (ci)
( 0)
Standard
quarterly average
rate of change
over six—month
six-month spans
Consumer
Money
prices
wages
wages
p
µ
ci
a
1.55 1.61
1.
4.00 1.
1.34
4.00
34
8.2
2.61
2.61
p
µ
ci
a
5.9
5.9
1.19
1.19
8.0
8.0 0.90
0.90
1953-70 and 1972-76
1972—76 include several periods
The data for 1963-70
oeriods in which
inflation was given "highest
‘highest priority
priority" as
as aa goal of public policy.
Careful
Careful inspection of the data
data shows that periods of slower growth of
of
the base coincide
coincide with these announcements in
in 1966,
1966, 1969-70
1969-70 and 1974—75,
1974-75,
periods of
of slower growth is long
long enough to have any
but none of these periods
marked
effect on
on the
~tandarddeviation
deviation of
of the
the growth
rate of
base.
marked effect
the standard
growth rate
of the
the base.
Table 22 shows that the standard deviation
deviation of
of the two quarter moving
moving
growth rates is
is independent of the rate of growth of the base and not
very different
different in
in the three sample periods.
The data suggest two reasons for the persistence
persistence of inflation
inflation and
the slow
slow response of inflation to changes
changes in the growth rate
rate of
of money.
First, short—term
short-term rates
rates of price change are relatively variable, so
growth from
people have difficulty separating the effects of money growth
other influences on short-term
short-term price changes.
—134—
-134-
This
This is
is particularly the
case for recent
recent years,
years, when announced changes in oil prices have had
on measured rates
considerable influence
influence on
rates of
of price change and their
variability.
last.
not last.
Second, the commitment
commitment to anti-inflation
anti-inflation oolicies
policies does
unwilling to buy
buy long-term
long-term contracts based on the
People are unwilling
assumption that the slower rate of money growth will persist
persist long
enough to reduce the trend
trend rate of inflation.
inflation.
In the next section, I
relation
offer an
an explanation of the rel
a ti on between
between the variability of money
growth and the persistence of inflation.
5
THE BASIC INFERENCE PROBLEM
PROBLEMS
Each week
week the Federal Reserve reports
reports the
the growth
grm,th rates of various
monetary aggregates.
Market participants try to infer the future
course of
of money
money growth, interest rates, prices and exchange
exchange rates from
the announcement.
Their problem, and ours as economists, is to sepasepaTheir
rate transitory changes in money growth (or other variables) from perpersistent changes.
II call this problem
problem of separating permanent
permanent or perper-
sistent changes from ephemeral or transitory changes
changes the basic inferinference problem because it
it arises for most economic variables
variables and is aa
major
major problem for people making decisions.
illustrate
in aa given week
To il
1us tra te the problem,
prob 1em, suppose that ·in
week the
the anannounced change in money
money is large relative to past changes.
Few obob-
servers will use the observation for a single
single week to
to predict
predict the
growth path, and fewer still will predict an
an equiproportionate
equiproportionate change
in the rate of inflation.
Let the increased rate of money growth perper-
sist, for aa month or two, and the balance of
of opinion will start to
55This
This section
section owes
debt to
owes aa large
large debt
to Brunner,
Brunner, Cukiernan
Cukierman and
and Meltzer
Meltzer
(1979).
-135—135—
change. More observers will infer that there has been aa persistent
change in
in the growth rate
rate of money.
The effect of the
the first week’s
week's observation on market prices,
prices,
interest rates and exchange rates differs from the effects of
of aa change
that is
is perceived
perceived to be
be permanent.
Although the change in money is rere-
ported, and therefore is known, the correct inference to be
be drawn from
from
the information is uncertain because
because the content of the
the information
information is
is
uncertain.
uncertain.
AA rational investor
investor who
\\lho uses all available
available information,
must
must first decide what he knows; that
that is to say,
say, he must decide how
he has
has observed can be expected to
to persist.
much of the changes he
This view of the world in which
which monetary and other policies
operate differs in an important way
way from the usual model of rational
expectations developed
developed by Lucas (1975) and others.
There, people are
uncertain about whether the changes they observe
observe are the result of
shocks
shocks that change relative prices
prices or
or shocks that
that change the
the absolute
absolute
price level; once information becomes available, there is no
no doubt
about its meaning.
Given the speed with which information becomes available,
available, the
be the prinprinconfusion between aggregative and relative changes cannot be
cipal source of confusion.
The
The main aggregates in our
our models -- money,
-—
debt and deficits or
or GNP, prices and output -- are observed
observed within aa
-—
month
month or aa quarter.
abOnce they are observed, the confusion between
between ab-
solute and relative changes disappears.
The permanent-transitory confusion does not disappear when data
are published.
published.
The principal
principal uncertainty that individuals face arises,
in this
this model, from an
an inability to
to properly
properly interpret
interpret information, not
from lack of
of information.
information.
People observing
observing the price index must decide
decide
—136-136-
whether a reported increase or decrease in an aggregate is aa one-time
change that will soon be
be reversed or the
the start of
of aa higher or
or lower
of change.
maintained rate of
change.
Expectations remain rational, but the use
of all available
available information does
does not solve the inference problem and
and
does not eliminate
eliminate error.
A simple model
A
model brings out the source of the permanent-transitory
permanent-transitory
confusion.
It is,
is, of
of course, only one of many ways in which the
problem can
can be formulated, but it is the way
way that has been
been used in
in an
application to the problem of stagflation where it
it produces changes in
prices and employment that resemble the aftermath of the oil shock.66
prices and employment that resemble the aftermath of the oil shock.
An
An observable variable
variable Xt can be
be divided into
into two components, aa
permanent
component. X~,
X~, and aa transitory component
component X~.
Xi. X~and
Xi and AX~are
"X~ are
permanent component,
normally distributed random variables with mean zero and known, concon2 andd J~q•
2
.
stant
~
People cannot observe X~ or
s t an t variances,
axp
an axq·
Xi
but must
X~.
infer the permanent value by observing current
current and past values of Xt.
xt
=
+
The expectation of X~,conditional
Xt, conditional on all information available
available in
. d t, 1s
. X~.
xPt·
period
is
per10
The inability to
to separate
separate permanent
permanent and transitory components
makes the optimal forecast
forecast of XX aa distributed
distributed lag
lag of
of past
past observations.
observations.
Contrary to much of the rational expectations literature,77 we find that
6Brunner, Cukierman and Meltzer (1979).
(1979). This
application conThis application
considers the effects of real shocks.
shocks. The role of the permanent-transitory
confusion in the transmission of monetary shocks to real variables
introduces additional problems.
7Benjamin Friedman (1979) is an
an exception.
exception.
—137—
-137-
distributed lag
1ag of past observations is an optimal method of
using a distributed
forecasting.
The reason is that repetitive observation
observation of an
an aggregate
aggregate
are required to
to learn whether
whether aa permanent change has occurred.
If
perIf per-
manent changes are frequent, and transitory changes
changes are infrequent, a
l·ikely to be treated as
as permanent soon
soon after
after it
it
change in XX is more likely
occurs.
At
At the opposite extreme, transitory changes are frequent and
permanent changes
changes are rare, so
so it is optimal
optimal to observe aa relatively
long series
series of observations before concluding that aa permanent change
2
axp
has occurred. In more technical
technical terms,
terms, the larger the ratio -2- the
axq
xq
faster people correctly infer that aa permanent change
change has occurred; the
smaller the ratio, the larger is the number of
of observations
observations required to
sustain
sustain the inference that aa permanent change has occurred.
II
11
We
We can put more content into the terms
tenns “frequent”
frequent 11 or “infrequent”
i nfrequent 11
by using the computed standard deviations for the two quarter and three-
year moving averages in Tables 1l and 22 to estimate the relative varivariof permanent
permanent and transitory components and to find the implied
ance of
length of the lag in reaching rational
rational judgments
judgments about permanent
permanent shocks.
The permanent
permanentS variance
van ance of the growth rate of the monetary base is set
equal to the variance
variance of the three-year growth rates.
The two quarter
moving average growth rates include both
both permanent
permanent and transitory comcomponents.
ponents.
We assume that permanent and transitory variances are indeinde-
pendent and compute
compute the
the transitory variance by subtracting the variance
of
of the twelve quarter average from the variance
variance of the two quarter
average.
Muth (1960, pp.
pp. 302-4) shows that the best (minimum variance)
Muth
linear
linear estimator of the
the permanent value of aa variable can be computed
from past actual values using the
the variances of the permanent
permanent and
-138-
transitory components.
For the problem at hand, the calculations for
the three periods of relatively
relatively constant growth of the monetary base
show that the relative variances of the growth rates of the base are:
are:
1955—60
1955-60
1964-71
71
1964-
1973—78
1973-78
2
°Ji__
.04
2
. 19
.19
. 14
.14
oq
These ratios imply very different lags in the adjustment of the
expected growth of the base.
base.
In 1955—60,
1955-60, only
only 55% of the adjustment of
expectations occurs within three years.
The
The reason is that the very
very low
three—year average growth
variance around the three-year
growth of base money obscures
the change in the maintained rate of growth, when it occurs.
Rational
individuals interpret most of the permanent change
change as transitory and
fail to adjust fully for several years.
In
In the two remaining
remaining samples,
the variance of the
the permanent
permanent component is higher
higher relative to
to the
variance of the transitory component.
variance
Expectations adjust more quickly;
more than 95% of the full adjustment occurs in the first three years.88
are related to the growth of
of money
money that
Expectations of inflation are
individuals expect to
to be
be maintained.
maintained.
The expected growth of
of base money
money
can be reduced permanently only
only if the actual growth of base
base money is
reduced.
The speed of
of adjustment of expected
expected to actual
actual growth
growth can be
growth rate
rate of
of the base is rerereduced, also, if the variability of the growth
duced.
For example,
examp 1e, if
if the Federa
va ri anee of the
For
Federal1 Reserve reduces the variance
to equal
equal the variance
variance of the twelve quarter
two quarter growth rate to
8BtTransitory1
variances are
moving
uTransitory variances
are computed
computed from
from two
two quarter
qua,~ter moving
averages, so
so two
two quarters are used as
as one period when computing
computing the
lags.
11
—139—
-139-
about the permanent
growth rate, 85% of the adjustment of expectations about
growth occurs in the first year.
growth
of inflation
inflation respond
Expectations of
more rapidly to monetary policy; the length of
of the lag of inflation
inflation bebehind
hind money growth declines.
It is, no doubt,
doubt, aa mistake
mistake to use
use these numbers as
as precise
precise estiestimates
mates of the expected length of the lag.
lag.
Fortunately, the principal
implications do not depend on the
the precision with
with which we measure the
speed of adjustment of expectations.
expectations.
short—tern policies are
If short-term
are less
less
variable, the speed of
of adjustment
adjustment increases.
Faster adjustment of
exof ex-
pectations
pectations lowers
lowers the length of time between changes
changes in the growth rate
rate
of the
the monetary base and changes in
in the expected growth of the base
and,
and, therefore, in the expected rate of inflation.
The
The shorter
shorter the
lag, the smaller,
jbus, is the persistence of inflation.
smaller, ceteris pparibus,
A related,
A
related, but distinct, implication
implication explains why
why short—term
short-term
changes in
in the
the growth
growth rate
rate of
of the
the base
have little
effect on
on maintained
maintained
changes
base have
little effect
inflation.
inflation.
The larger the transitory variance
variance of the growth rate
rate of
of the
The
base, given the long-term or
or permanent variance, the longer
longer is the lag.
Short-term
rate of the base have
have little
little effect
Short—tern reductions in the growth rate
on
on long-term expectations
expectations if the short—tern
short-term growth of the base is
highly variable.
The
The real
real costs of reducing inflation are higher,
higher,
under these circumstances.
circumstances.
rising
rising unemployment.
unemployment.
The costs take the form of
of recession and
Recession
encourages the
Recession encourages
the Federal
Federal Reserve
Reserve to
to shift
shift
to aa policy of
of monetary expansion
expansion thereby reinforcing expectations
expectations that
that
rate of the
the base will
will not
not be reduced.
the maintained average growth rate
Chart 1,
anti—inflation policy have,
l, above, shows that past periods of
of anti-inflation
in fact, had little effect on
on the maintained
maintained growth rate of the base.
base.
-140-
The calculations in Tables 1l and 22 imply that the lag in
in the
fifties.
is shorter now than in the fifties.
formation of
of expectations
expectations is
The data
that the
the reason for the shorter lag is the increase
suggest, however, that
in the measured variance
variance of the
the permanent
permanent component, not
not aa reduction
reduction in
the measured variance
variance of the transitory component.
THE POLICY PROBLEM
the short-term variance of
of the
The Federal Reserve can reduce the
growth of the monetary base by adopting targets expressed in
in terms of
base.
the base.
currency, the uses
uses of
of the base, are approxiReserves and currency,
approxi-
sum of reserve bank
bank credit and international reremately equal to the sun
serves.
serves.
With floating (or adjustable) exchange
exchange rates, the
the Federal
Reserve can control the two quarter
quarter growth rate of the
the base by controlcontrol-
ling the stock of Reserve bank credit.
credit.
ling
To control the base the Federal
Reserve need not solve an impossible or even a difficult problem.
All
they must do is control the asset side of
of their balance sheet.
As is well-known, the Federal Reserve cannot control both
both interinterest rates and the growth rate of the
the base.
By specifyinq
specifying short—term
short-term
targets in terms of values (or ranges) of the Federal funds rate, the
Comittee surrenders control of short-term
short—tern changes
Federal Open Market Committee
in the base.
base.
The problem of separating permanent
permanent and transitory
changes helps to explain
short—term control of the base
explain how loss of
of short-term
contributes to persistent movements
movements of the base even if
if the
the dominant
dominant
are real, not nominal shocks.
shocks in the economy are
To
illustrate the
problem, II use
use the
the three
three equation,
equation, equilibrium
equilibrium
To illustrate
the problem.
model based on
on Brunner, Cukierman and Meltzer (1979).
-141-
All variables
are natural
natural logarithms.
neoProduction or output, Yt, is given by aa neo~
classical production function
(1)
~
=
ut
+
with ~
lt, the number of
of man hours of labor
labor and ut aa productivity
productivity shock;
6o is the elasticity
elasticity of output with respect to
to labor.
Real aggregate
aggregate
spending is always equal
equal to
to output, ~
Yt, and depends on expected
expected or per—
pernanent
y~,on
agmanent incone,
income, y~,
on the real
real rate of interest and on shocks to agdiffergregate demand,
demand, c~.
et. The anticipated
anticipated rate
rate of inflation is the differ-
ence between
between the iogarithns
logarithms of
of the price level
level anticipated for next
period ~t~t+i~
(tPt+ll and
and today’s
today's prices ~
(pt).
The market
market rate of interest
is
is i~.
it.
(2)
~
=
a+by~ + c[it
-
~
+
b>0;ccO
b > 0 ; c < 0
Equation (3) equates
+ 1t, to the
equates the current stock for base
base money, BB +
denand
demand for base money, where
where ft is the shock to
to the level of
of nominal
9
ct. affects
money balances.
balances. 9 Some
Some part
part of the
the shock to
to spending,
spending, et,
affects the
demand for money; the rest affects the demand
demand for bonds and the supply
supply
of labor.
Increases
Increases in spending are financed by
by reducing
reducing the demand
for money
money so ae is
is positive
positive and
and increases in c2 reduce the demand for
money.
(3)
B
B ++ y
t
=
+
+
~
+
y~ +
y
~
-
Bet
8<0
B
< 0
1>
> y
y,, 8
a>
> 0
0
l
99The
analysis can
terms of
The analysis
can be
be cast
cast in
in terms
of growth
growth rates
rates of
of money
money by
by
naking
adjustnents.
making minor adjustments.
-142-142—
IS-LM model.
The three equations form an augmented IS—LM
The principal
novelties are the distinction between permanent and current income and
the introduction of
of permanent and transitory shocks.
soocks,
The three shocks,
~ ~
ut,
ct and 't' have permanent and transitory components, but people are
not able to distinguish the permanent and transitory components when
observing the shocks.
For example, ut
u~== u~
u~ with known variances
ci~q~
Eu~and
EAu~
o~P and o~
and Eou~
, normal distributions and expected values Eui
9
equal to zero.
ci~
Substituting eq. (l)
(1) into (2) and (3) and
i~reduces
and solving
solving for it
reduces
the system to
to two equilibrium relations.
The
The money market equilibrium
equilibrium
eq. (4) and the IS curve,
curve, eq. (5) relate i~to
it to the three
three
or LM, in eq.
shocks, to
to the price
price level
level and to other variables.
variables.
For
For the
the current
100
treat y~and
i~as
shocks.~
analysis, I treaty~
and lt
as given and
and independent
independent of the shocks.
(4)
si~
Bit = 8B ++ 't - Pt
-
yu~++ Bet
SEt - ~
yut
y•lt - (l(l-y)y~
1)y~ - a
(5)
cit
+
ut
-
c(tpt+1_pt)
=
-
-
Ct +
-
-
ai~ b y~ a
-
-
of its existence, the Federal Reserve used the market
During most of
interest rate (or some surrogate like the level of free reserves) as
as
the operating
operating target.
terest
rate at
terest rate
at ii
0
Suppose the Federal Reserve sets the target inin-
and
supplies or
to keep
and supplies
or absorbs
absorbs base
base money
money to
keep i~
it== i.
i •
0
10A full solution is given in Brunner, Cukierman
lOA
Cukierman and Meltzer
(1979) by specifying the labor market equations. The
The additional
additional detail
detail
would not alter the conclusions of
of this discussion. The
The principal difdifexferences that have been
been neglected
neglected are the dependence of y~ on the expected values of
of the real shocks
shocks and the dependence of 1lt on the actual
actual
values of the real shocks.
shocks. The reader who
who is disturbed
disturbed ~y
oy the partial
solutions can substitute permanent
permanent and actual values of
of shocks -- real
y~and
shocks -- for y~
and ~
lt· For the analysis that follows what matters
matters is
that the responses
responses of IS
IS and LM to the shocks cause
cause i~to
it to differ
from ii .
0
4
-—
--
0
-143-
The stock of base money B
B ++ rt changes only as required to maintain the
~
interest
that the
the stock
interest rate
rate at
at ii,, which
which is
is to
to say
say that
stock of
of money
money now
now dedeo
pends on the real
real shocks.
~t
(6)
~(et~
ut)
Equations
Equations (4) and (5) are shown as solid lines in Figure 1.
slope of LM
LM from eq.
eq. (4) is positive in the i, pp plane.
plane.
IS is -1.
The
The price level is p.
0
The
The
The slope
slope of
of
The policy
policy of fixing interest rates,
at i,
temporarily at
i o , makes the interest rate pre—determined
pre-determined at i.
i0 .
MoneMone-
tary policy keeps the interest rate constant by changing money.
WhenWhen-
ever there are real shocks to productivity or to spending and the
demand for money, the Federal Reserve changes the stock of money enough
to hold interest rates fixed until it decides that the shock is perpermanent.
Consider the effect of aa negative
negative productivity shock,
shock, dut
<
0.
From (4) and (5) we compute the elasticities
0 and
>
tILM
~t
dut
=
1
c
IS
AA negative shock shifts both the LM
LM curve and the IS
JS curve to the right
right
in Figure 1.
small,
If y~ is sma
11 , the
the demand for money changes very 1little,
i tt 1e,
and interest rates rise.
rise.
The
The Federal Reserve offsets the rise in inin-
terest rates by
by increasing the money stock.
dit
1
1
=
- <
B
<
0
0
-144-
ii
' \'
\
'\.
12
i2
‘1
i1. l
io0
•••••• ,.
/I2~12
'
/
I
'
L--~er::--~:--'-.
\
\
I
I
is
2
Is I
'rs
1
p
p0p1 p2 P
3
FIGURE 11
FIGURE
-145-
If the negative productivity shock is
is transitory, Federal Reserve
If
policy eliminates any effect on interest rates
rates but increases
increases the price
level by more than
than the increase resulting from the transitory decline
in productivity.
The dotted lines 1s
151 and LM11 in Figure 1l show the
1
effect of the transitory change in u~.
ut. Prices and
and interest rates
rates rise;
p1 is
of the
price level
level at
at the
the intersection
intersection of
IS 1 and
and LM
LM 1,,
is the
the log
log of
the price
of IS
1
1
and i11 is the interest rate.
rate. Federal Reserve policy shifts the LM
curve further to
to the right, shown by LM22,, restoring
restoring the
the interest
interest rate
i 0 and increasing the price level
level to pp22; p22 - pp11 is the relative
relative rate
-
of change in the price
price level
level resulting from Federal
Federal Reserve policy, and
pl - pp0 is
is the
the rate
price increase
increase caused
by the
in producrate of
of price
caused by
the decline
decline in
produc—
t iv i ty.
tivity.
—
The mean values of
of the transitory shocks
shocks are zero so the effect
of Federal Reserve’s
Reserve's response to transitory shocks is
is on
on the variance
variance
of rates
rates of
of price change and not on their average
average over time.
AA policy
of pegging interest rates increases
increases the variability
variability of the measured
rates of price change resulting from transitory
transitory shocks.
Our earlier
earlier
finding that
that the variance of the rate
rate of
of price
price change rose
rose during the
shocks is consistent with this impliperiod in which there were oil shocks
implication.11
cation
11
Suppose, however, that the negative productivity shock is
is perpermanent, or
or persistent, not
not transitory.
In this case,
price level
case, the price
fluctuates around pp2 following
the increase in
in money to LM22..
2 following the
Because
Because
permanent
be observed separately, or
permanent and transitory shocks cannot be
11
~There
There are, of course, other causes of variability including
including the
shocks
for money (Ct)
(ct) and the Federal
shocks to spending and the demand for
Reserve’s
Reserve's response
response to these shocks.
shocks.
-145-146-
separated reliably,
reliably, people
people must
must decide
decide whether
whether the
the observed
observed rate
rate of
of
separated
price increase, p22 - pp,, the change in money, s’~,
~t' and other
other changes
0
—
have caused aa one-time price change or aa persistent change in
in the rate
of price change.
If the inferences
inferences drawn from available
available information
to believe that some part of the change in the measured
lead people
people to
measured
rates of price change and money are persistent
persistent changes in the
the rates of
change,
change, instead
instead of one-time changes
changes in level, the IS curve shifts
shifts
to the right.
further to
The size of the shift depends on the degree to
The
12
12
which the anticipated rate of inflation, tPt+l - pt, rises.
rises.
~
-
~
The Federal Reserve policy
policy of fixing the interest rate at
at ii 0
sustains the inference that the observed changes in prices and money
reflect
reflect aa persistent increase in rates of
of change,
change, not aa one-time
one-time change
change
in levels.
levels.
The reason is
is that, when
when IS
IS shifts to the right the policy
policy
of fixing interest rates
rates requires the Federal Reserve to again increase
increase
the money stock, shifting
shifting LM further to
to the right.
right.
The additional changes
changes in
in money and prices reinforce
reinforce beliefs
about the persistence of the changes in
in money
money and prices.
As
As the perper-
ceived and measured
measured rates of
of inflation rise, anticipated inflation
rises, and there is aa further
further rightward shift in IS.
inAdditional in-
creases
creases in money are now required to
to hold
hold the market interest rate at i 0
Each
Each increase in the stock of money
money reinforces the belief that
has been aa persistent change in
in the rate of money
money growth.
there has
Each
increase in
in the
the equilibrium price level
level reinforces
reinforces the belief that the
run of transitory, negative shocks to productivity produces a
12 A run
of transitory, negative shocks to productivity produces a
similar result. tPt+l is today’s
period’s price.
today's expectation of
of next period's
The rational expectation takes the form of
indiof aa distributed lag, as indicated
earlier,
so
expectations
adjust
gradually.
cated earlier, so expectations adjust gradually.
-147—147—
rate of price change has increased.
The Federal
Federal Reserve’s
Reserve's policy
policy of
The
maintaining
one—tine change
maintaining the level
level of interest rates converts aa one-time
change in
in
the price level into
into aa series
series of
of price changes that strengthen percepperceptions that there has been
been aa change in the rate of change.
Rational investors
anticipainvestors “know”
nknow the model, so they know that anticipa11
tions about the price level
level adjust slowly because they and others
others are
unable to
to separate persistent and transitory changes.
holding the interest rates at ii
0
as long
long as
as the money stock grows.
The policy of
of
implies that
that the price level
level will
will rise
‘~
That is,
is, as long
long as
as ~
tPt+l- Pt is
positive,
positive, the policy of
of fixing interest rates will
will require
require the Federal
Reserve to let the money
money stock rise.
rise.
The Federal Reserve can eliminate
eliminate the bulge in the money stock
and in the measured
measured rate of price change
change by
by raising
raising the target rate of
II have drawn aa dotted line at the intersection of
of IS
1s22 and
LM22 in Figure l1 to show the rise in interest rates required
required to
to keep
keep the
price level from exceeding pp33.. The dotted line shows that the required
interest.
interest rate is i
i 22; ii 22—i
-i11 is the additional increase in interest rates
resulting from Federal Reserve policy. The increase ii2—i
-i 11 is
is temporary,
not permanent.
Once people recognize that the
the money stock is constant,
anticipations of rising prices decay; IS shifts to
to the
the left;
left; the market
rate of interest falls to i
i 11; and the price level
level falls between pp22 and
p33.. (The precise level of prices is at the value of i11 on LM22.)
.)
The combination i1,, pp1 is the interest rate and price
price level
level comcom1 1
bination to which the econony
economy moved following the permanent
permanent loss of
productivity.
It is not an
an accident
accident that the economy eventually
settles at the rate of interest
interest i11 following the “anti—inflationary”
"anti-inflationary"
increase in interest
interest rates
rates to i
i2;
it is an
an implication of the neutrality
2 it
-148-
of money.
money.
of
allowed the money stock to rise,
Monetary policy, at first, allowed
then held the money stock constant,
constant, eliminated the anticipation of
rising prices and allowed the interest rate
rate to decline.
The
The lasting
effect of the interest
interest rate policy is
is aa higher price level.
The amount
of increase depends, of course, on the speed with which the Federal
i~= i.
interest rate target it~
i •
Reserve abandons the interest
0
This discussion of policy has
has neglected
neglected many
many complicating
complicating feafeatures.
The adjustment of
of prices and interest
interest rates has
has been analyzed
as if these changes occur
occur without real effects.
effects.
The gradual adjustadjust-
perment of
of employment when rational individuals cannot distinguish
distinguish per·
manent and transitory productivity changes has not been emphasized.
The case for fixing the level of
of interest rates is not strengthened
strengthened by
by
these omitted effects.
A
principal result of the policy of fixing market
market interest
interest rates
A principal
is that additional changes in prices
prices (and output) are induced by
by monemone-
tary policy.
policy.
People
People are forced
forced to
to decide how
how much of the
the observed
observed
and how much
much is
is transitory.
change in money is persistent amd
The
The deterdeter-
mination
mew permanent price
mination of the new
price level
level is
is made
made more
more difficult.
The permanent decline in
in productivity
productivity produces aa temporary
temporary inincrease in unemployment
unemployment and aa permanent loss
loss of
of real income.
UnemployUnemploy-
ment rises because people do
do not recognize instantly
instantly that
that the shock is
permanent.
Hence, they do not instantly
instant 1y adjust
adjust their real
rea 1 incomes ((and
and
real
rea 1 wages) to the level
1eve 1 they eventually reach.
Monetary
Monetary policy
po 1 icy can
reduce this
this cost of adjustment only if the monetary authority cam
can sucsucceed in reducing real wages to their new, permanent
permanent level without set~
set-
ting off
off anticipations of
of rising prices.
prices.
The monetary
monetary authority must
have superior information
information on the speed
speed with which people recognize the
th,
0
-149-
permanent loss
loss of
of real
real income
income and
and the
the speed
speed with
with which
which anticipations
anticipations of
of
permanent
price changes
changes form and
and decay.
moneThere is no reason to believe that mone-
tary authorities have information of this kind or are able to
to set marmarket
interest rates in
in a way that minimizes
minimizes the cost of adjusting
adjustin9 to
ket interest
real shocks.
shocks.
On the contrary,
contrary, monetary
monetary policy produced persistently
persistently
higher rates of price change following the productivity
productivity shocks
shocks of this
decade.
THE CASE FOR GRADUALISM
Reliance
Reliance on market interest rates as
as the operating target
target of
monetary policy produced high rates of
of growth of the
the monetary base and
sustained inflation.
sustained
The
The low variance of
of the long-term
long-term average growth
of the base suggests that
that the 8.5% growth rate of the
the base is perceived
perceived
“permanent” rate of change.
as aa "permanent"
To end inflation the rate of
of growth
of the base must be reduced.
If
If expectations form and decay quickly in the presence of
of new
information,
information, the problem of ending inflation is made easier.
AA credicredi-
ble
expectations.
ble policy
policy to stop inflation
inflation causes prompt revision of expectations.
Revised expectations,
expectations, and slower
slower growth of
of base money
money bring inflation
to
to an
an end.
Rational
Rational individuals recognize that
that sunk costs
costs or
or contracts
must be forgotten, so as contracts are revised, they enter into
into agreeagree-
ments or
or commitments that reflect their revised expectations.
Even in
this case, there are benefits to gradualism if costs of adjustment can
to learn about the new environment.
be reduced by permitting people to
The analysis in the preceding
preceding section
section suggests
suggests some of the diffidiffiThe
culties people face when
when forming judgments about
about the
the persistent
persistent rate of
of
change of money.
Some of these difficulties cam
can be
be reduced if policy
—‘SD—
-150-
V
makers announce
announce the intended rate of
of money growth.
makers
not sufficient to
to change anticipations permanently.
Announcements are
Announcements
AA principal reason
is that policymakers statements are not entirely
entirely credible.
credible.
Past
promises to slow money growth and
and reduce
reduce inflation
inflation have been
been followed
within aa few quarters by renewed expansion.
Consequently, rational
rational
exindividuals treat any initial reduction in money growth (or budget expenditures)
penditures} as temporary,
temporary, not permanent,
permanent, changes.
changes.
An announced reducreduc-
tion in the growth of money, initially,
initially, will not be interpreted as
as aa
reduction in the maintained rate of money growth.
Gradual reduction in money growth can reduce the cost of lowering
the
three ways.
the rate
rate of
of inflation
inflation in
in three
ways.
First,
of
First, maintaining
maintaining the
the growth
growth of
the base at aa steady rate lowers
lowers the variance of the transitory comcomponent and reduces the lag in
in the formation of expectations.
expectations.
Second,
the maintained average
average rate of money growth falls gradually, so people
have time to adjust future commitments to reflect revised expectations.
Third, if
if costs of adjusting to aa lower rate of
of inflation
inflation are
are not
proportional to the total adjustment but increase
increase with
with the rate per
period,
period, costs of adjustment are reduced by lowering
lowering the rate per
per
period.
If the rate of
of adjustment of money growth is very low,
low, the varivariance
of the
permanent component
component is
is low,
ance of
the permanent
low, so
so the
the lag
lag in
in adjustment
adjustment of
of
expectations
expectations increases.
If the rate of adjustment of money growth
growth is
component increases,
rapid, the variance
variance of
of the transitory component
increases, so
so costs of
adjustment
adjustment rise.
rise.
The
The optimum rate
rate of
of adjustment
adjustment is achieved by inin-
creasing
component and reducing the varicreasing the variance
variance of the permanent
permanent component
variance
ance of the transitory component of
of money growth.
This
This is
is equivalent
equivalent
to
to finding the minimum lag
lag in the formation of
of anticipations.
-151—
-151-
pol icy of gradual, pre-announced
pre-announced reductions In
in money growth
The policy
advocated by the Shadow Open Market Committee
Committee did not emerge as aa solsolution to the problem of finding an optimal lag.
The choice of am
an
optimal policy depends on
on information that is not
not yet available.
Our
like most policies, depends more
more on empirical judgments about
proposal, like
the length of lags
lags and costs of adjustment than
than on
on hard evidence.
I
have no doubt that future research will find aa better path.
path.
SOME FINAL SPECULATIONS
SOME
The
The chief difficulty in
in the policy of gradualism is
is the length
length of
time required to
to reach the rate
rate of
of growth consistent
consistent with mom-inflanon-inflationary growth
growth in the economy.
economy.
If we use the long—run
long-run growth of
of real
real
output as aa guide, the rate
rate of base money growth must
must fall from the
current rate
rate of
of 8% to
to no more than 3%.
3%.
If payments
payments technology concon-
tinues to improve, base
base velocity
velocity will
will rise in the future as it has
has for
at least the past quarter century.
The non—inflationary
non-inflationary rate of base
money growth is then no more than 11 or
or 2%.
in money growth
Is a seven year program of sustained reductions in
the best that can be done?
I expect not.
There is reason to believe
believe
that policymakers cam
can increase their credibility by
by meeting
meeting pre—
preannounced targets.
targets.
Increased credibility
credi bi 1ity permits
permits policymakers
po 1 i cymakers to
to lower
1ower
the maintained growth rate
rate while lowering the relative variance
variance of the
transitory component of
of money growth.
Credible
announcenents mean
Credible announcements
mean that
individuals distinguish permanent
permanent changes closer
closer to
to the time they occur
occur
by using announcements of proposed changes as
as a reliable indicator of
of
future money growth.
-152—152—
No one can be very certain about these issues. The evidence on
from experience in Germany, Switzerland, the United
which we rely comes from
Kingdom and our own experience in the middle seventies.
sevent1es. Each of these
anticexperiences suggests that within two to three years at most, the anticipated rate of inflation declines.
dee lines. The rate of price and wage change
falls; long-term interest rates decline, and real output rises or
accelerates.
accelerates.
Those who desire “incomes
incomes policies”
policies!! to reduce the lag for adjustadjust11
ment might
might find pre—announced
pre-announced monetary policies more attractive
attractive than
pnent
either the failed incomes policies of the past or present, or complicated, inefficient programs to tax wage and price changes.
Instead of
announcing the rate of price and wage changes that the government
favors, the government can announce the rates of monetary and fiscal
expan.sion that the government intends to maintain. These announcements,
expansion
if they are credible, help individuals to form expectations about
future rates of inflation.
Analysis of the length of the lag
lay in the adjustment
adjustment of anticianticipermanent values
pations relates these adjustments to the adjustment of pennanent
maintuined rates of change.
change,
or maintained
v-Je have is
is neither
neither inconinconThe evidence we
expectati ans that I have sketched nor more
sistent with the theory of expectations
a;1y other
o~her explanaion I have seen. This is not a strong
consistent with any
inclaim, but it is considerably better founded than the belief that In-
flation is intractable.
-153-153—
REFERENCES
REFERENCES
(1g79) "Stagflation,
Stagflatiom,
Brunner, Karl, Alex Cukierman and
and Allan H. Meltzer (1979)
Persistent Unemployment and
and the Permanence of Economic Shocks,”
Shocks,"
multilithed, cC- ‘negie—Mellon
·negie-Mellon University.
Relative Price Variability, Inflation
Cukierman, Alex (1979) "Relative
Inflation and the
Allocative Efficiency of the Price System,
System," multilithed,
multilithed, CarneqieCarneoieMellon University.
Equilibriun Model of the Business Cycle,"
Cycle,’
Lucas, Robert
Robert E.
E. (1975)
(1975) “An
"An Equilibrium
Journal of Political Economy, 83 (Deceriber)
(December) pp. 1113-44.
Meltzer, Allan
Allan H.
Anticipated Inflation and Unanticipated Price
H. (1977)
(1977) ''Anticipated
Change’ Journal of Money, Credit, and Banking, g9 pt.
Change"
pt. 22 (February),
pp. 182—205.
182-205.
Muth, John F.
‘Optimal Properties
F. (1960) "Optimal
Properties of Exponentially Weighted
\leiqhted
Forecasts,’
Statistical Association, 55
Forecasts/ Journal of American Statistical
55
(June) pp. 299-306.
299—306.
1
Muth,
F. (1961)
Expectations and
Theory of
Muth, John
John F.
(1961) “Rational
"Rational Expectations
and the
the Theory
of Price
Movements,”
Movements,'' Econometrica, 23
23 (July) pp.
pp. 315—35.
315-35.
‘Slowing the Wage-Price
Wage—Price Spiral: The MacroPerry, George L. (1978) "Slowing
Macroeconomic View,"
View,” Chapter
Chapter 22 in Okun
Okun and Perry (eds.)
(eds.) Curing
economic
Curi nq Chronic
Inflation, Washington: Brookings Institution.
—154-154-
FEDERAL BUDGET POLICTES OF THE l97Os
FEDERAL BUDGET POLICIES OF THE 197Os:
l9SOs
SOME LESSONS
LESSONS FOR THE 198Os
Michael
E. Levy
Levy
Michael E.
At the close
1970s, the public and the politicians alike
close of the
the 197Os,
alike
perceive inflation as
as the foremost economic challenge of
of the day.
Other important econonic
economic and
and social
social issues will
will carry over
over into the
the
l98Os; forsaken claims will be revived and new demands are bound to
198Os;
surface.
the
But our effectiveness in
in coping with all
all these -- in fact the
—-
very survival of
poof this
this countrys
country 1 s traditional economic, social, and political structures -- nay
may well depend on
on our ability
ability to
to contain and
—-
control inflation in the coming
coning decade.
There
inflation control
control nay
may require
There is a growing belief that inflation
fiscal restraint, aa slowing of government spending, aa reduction in the
size of the realized budget deficit.
Yet, as
as we approach the threshold
of
of the l9BOs,
198Os, II can think of at
at least five major policy
policy issues
issues in
in
solutions, each
each of which would place new claims on our fiscal
search of solutions,
resources~
resources.
oo
Half aa decade after the initial “energy
llenergy crisis,
crisis/' we are still
in
in search
search of
of an energy policy
policy that generates widespread
widespread public
public and
and
political support
support for economically viable solutions.
oo
to channel the hardcore unemployed into the mainmainOur efforts to
stream of
of our economy have yet to succeed.
succeed.
Michael E. Levy is Director,
Director, Economic Policy Research at the Conference
Board.
Board.
—155—
-155-
a.rms race -- even If
if attained
oo Success in slowing the nuclear antis
—-
through SALT
SALT II
II -- may have to be bought at the cost of accelerating
-—
defense spending for years to come.
oo
Welfare reform has been the subject of several aborted propospropos-
it is bound to resurface as
l980s.
als of the 1970s;
7970s; it
as a major
major issue in the 1980s.
unfulfilled social promoo National health insurance -- aa major unfulfilled
prom--
ise of the 1970s
1970s -- is high on the public agenda of the coming decade.
—-
It
It is all
all too easy
easy to add to this list
list of enlarged public
—claims -even at aa time when inflation
inflation control
control is our top priority
priority
and budget restraint
restraint is promulgated.
(Note that I have omitted any
11
mention of 11“safety’
safety" or “environmental
envi ronmenta 1 issues.”)
issues.")
Such are the complexcomp 1 ex-
ities
iti es and contradictions
contradictions of
of budgetary policy which would
would seem
seem to place
inflation
inflation control practically
practically beyond our reach.
Yet my monetarist friends
friends are able to collapse
collapse the social and
political
conplexities of inflation
political complexities
inflation control
control into the simple issue of
“monetary
"monetary integrity.’
integrity."
To them, the deep—seated
deep-seated inflation
inflation of the last
last
decade-and-a-half is strictly
strictly a monetary phenomenon.
Its "cause"
“cause” (like
(like
that of every inflation)
inflation) was excessive monetary growth
grov-,th reinforced,
11
perhaps, by aa few nasty “shocks,”
oil price escalations
escalations of
shocks/ 1 such
such as the oil
1973 and 1979.
1979.
Its “cure”
inflation) is secured
"cure" (like
(like that of every inflation)
through aa persistent
persistent slowdown in money growth.
On aa purely technical
level,
all the answers.
level, the monetarists have, of course, all
In fact, some
of my own econometric exercises have tended to reconfirm their
their valuable,
if somewhat simplistic,
simplistic, generalizations.11
if
1Michael E. Levy, assisted by Steven Malin, International InfluInfl,!Michael
ences on U.S.
U.S. Inf1atiop~l97L-1976,
Inflation 1971-1976, a study prepared
prepared for the U.S. DeDepartment
September 7, 1977
1977 (unpublished, available
partment of Commerce,
Commerce, September
available from
the author).
-156—156-
moneHowever, even if they were formally correct, these simple monetary propositions would tell
the changes in social
tell us nothing about the
attitudes and national priorities which generated
generated the political prespressures that bent the economic
economic structure and drove the monetary printing
press.
They
They provide no
no clues as to how and why the economic and social
social
structure was changed and whether this
this process
process is
is reversible
reversible or
.
2
2
cumu
l at1ve.
cumulative.
11
budgetary policy,”
policy/' such as it
it is,
is,
By contrast, analysis of “budgetary
promises to shed some light
light on
on these unanswered questions, because the
government budget is aa fulcrum of social and political change.
UnforUnfor-
tunately, it is difficult, at best, to chart aa course of fiscal and
budgetary policy over years and decades.
In fact, one may even
question the existence of aa meaningful "course"
than the drift
drift
“course” other than
created by the complex and contradictory forces and events that shape
the federal budget from year to year.
if this "drift"
“drift” were governed by a powerful current
Obviously, if
current
11
if “bends”
bendsn in this current could be discerned, we should expect
and if
far—reaching economic implications,
far-reaching
implications, because the
the federal budget powerpower-
fully touches all
all social groups, all segments of our
our economy.
I have
interpreted my assignment as the search for such
such bends in the current.
22For
For more formal analyses that question the independent contricontribution
bution of
of money growth in 11“explaining”
explaining 11 the inflationary process, see,
see,
example, Franco Modigliani and Lucas Papadenos,
Papademos, "Targets
Monefor example,
“Targets for MoneYear," Brookings Papers on Economic Activity,
Activity,
tary Policy in the Coming Year,”
‘Slowing the Wage-Price Spiral:
l1:1975,
:1975, pp. 141-63; George L. Perry, "Slowing
esp. pp. 45-46,
Inflation,
The Macroeconomic View,”
View,' esp.
45-46, in Curing Chronic Inflation,
Arthur N.
M. Okun and George L. Perry, eds.
eds., The Brookings Institution,
Institution,
1978; also Martin Neil Baily,
Baily, ibid.,
ibid., p.
p. 58.
Washington, D.C. 1978;
1
,
—157—
-157-
VIETNAM:
THE ORIGINS OF U.S. INFLATION
There is widespread
widespread agreement
agreement that the
the persistent
persistent U.S. inflation
of the last decade-and-a-half got under way
‘Keynesian’ exexway in
in 1965 as "Keynesian"
3
cess
expendirapidly escalating defense expendicess demand inflation. 3 In 1965, rapidly
tures for the Vietnam War were superimposed on
on aa full-employment
economy that was
was on the verge of aa private investment boom.
Not only
to enact timely tax increases (until the belated
belated ten—
tendid we fail to
but our exuberant
exuberant "guns
butter"
percent surcharge of 1968-1969),
1968—1969), but
guns and butter’
guns and Great Society’)
(or "guns
Society") policy
policy added new
new and rapidly escalating
civilian programs (Medicare,
(Medicare, Medicaid,
Medicaid, Food Stamps,
Stamps, Job Corps, Model
Cities).
Cities).
Vietnam War costs rose rapidly from about $100 million
million in
in fiscal
4 Total del969.~
de1965 to almost $29 billion at their peak, in fiscal 1969.
67 percent, during
fense expenditures
expenditures rose by nearly $32 billion, or 67
this period; and the
the share of GNP devoted to national defense advanced
from 7.2 percent in
in fiscal
fiscal 1965
1965 to 9.5
9.5 percent in fiscal 1968 -- its
its
——
high for the decades of the 1960s
l960s and 1970s.
Yet it
it would be a mistake to attribute the persistence
persistence of U.S.
inflation first and foremost to
to the
the Vietnam
Vietnam War -- even if one’s
one's time
—-
preceding the oil
oil crisis of late 1973.
horizon is limited to the period preceding
3E.g., see Perry, bc.
lac. cit., p. 23. Note, however, that some
monetarists have pointed out that the onset of this inflation was prepreceded by about two years of what was considered at
at that time rapid
monetary
monetary growth.
4These are "full-cost"
full_cost estimates.
estimates. For further details and for
“incremental—cost”
incremental-cost estimates, see Michael E.
E. Levy with
with Juan de Torres,
Torres,
11
11
Delos R. Smith and Vincent Massaro, The Federal Budget:
Budget: Its Impact
Imoact on
the Economy, fiscal 1973
1973 edition, The Conference Board,
Board, iT~iYbFk,
New York, 1972,
1972,
~p.-27.
esp. pp. 26-27.
-158-
From fiscal 1969 through fiscal 1973 annual expenditures for Vietnam
dropped by about $18 billion
billion in current dollars -- the decline in r~I
real
--
terms
tei:-_ms was, of course, much greater -- while total defense expenditures
expenditures
--
declined by
by nearly $5
$5 billion.
billion.
The share of GNP devoted to national
national
defense dropped from its 1968 peak of
of 9.5 percent to 66 percent in
fiscal 1973 and continued to decline to 55 percent by fiscal 1979.
1979.
Yet
the large Vietnam “peace
"peace dividend”
dividend" of
of the early 1970s
1970s brought no end to
U.S. inflation.
inflation.
When the 1970 recession barely
barely reduced the inflation
inflation
rate,
rate, a ninety-day
ninety-day wage and price
price freeze was introduced on
on August 15,
1971.
1971.
It followed by four phases of wage and price controls that
lasted through the third quarter of 1973.
April 1974.)
ended in April
(The final decontrol phase
Yet these controls
controls brought, at best, a modest
and inadequate respite,
respite, before the quadrupling of OPEC oil
oil prices
pushed the economy into double-digit inflation
inflation in 1974.
1974.
“SHOCKS” AND THE INFLATION OF THE l97Os
"SHOCKS"
1970s
A significant
if not a major
A
significant part,
part, if
major one, of the inflation
inflation surge of
1973—1974 that resulted
1973-1974
resulted in double-digit
double-digit inflation
inflation has been
been attributed
attributed
to special factors -- "shocks"
“shocks” of aa largely
largely international
international nature.
nature.
——
Three
Three distinct
distinct inflationary
inflationary influences
influences deserve to be
be distinguished:
oo
The depreciation of the external value of the dollar.
dollar.
(It
(It got
under way around mid-1970 and accelerated after
after the closing of
of the
“gold window"
window” on August 15, 1971, hitting
"gold
hitting bottom in July 1973.)
oo
The escalation of agricultural conriodity
commodity prices,
prices, particularly
particularly
grains,
1973.
grains, from late
late 1972
1972 through 1973.
(It
(It was caused largely by the
prior depletion of U.S.
disappearU.S. agricultural stocks,
stocks, the temporary
temporary disappearance of the Peruvian anchovies,
anchovies, bad
bad weather and poor crops
crops in
in many
—159—
-159-
parts of
of the world in 1972, the “Russian
"Russian wheat deal”
deal" of 1973, and the
worldwide boom that raised consumption of high—protein
high-protein foods.)
oo
The sharp rise in the prices of fuels and some industrial
industrial
commodities, but mainly the quadrupling of OPEC oil
commodities,
oil prices during the
last quarter of 1973.
Elsewhere II have described these special
special events and reviewed the
5
best available evidence as to
to their impact on
on U.S. inflation.
inflation. 5 This
significant before
combined inflationary
inflationary impact seems not
not to have been significant
mid- or late 1972.
It increased rapidly thereafter,
thereafter, appears to have
peaked during the second half
half of 1974,
1974, and faded durinq the second half
of 1975.6
1975. 6 On
On the basis of econometric estimates, I concluded that “the
"the
joint
major identifiable
identifiable 'international
acjoint impact of these major
‘international shocks'
shocks’ accounted for about 5.5 percentage points -- or roughly 60 percent -- of
—-
——
the dramatic increase in the inflation
inflation rate of the implicit
implicit GNP deflator
from about 3.5 percent (annual rate) in the second half of 1971 to
around 12.5 percent in the second half
half of 1974.
1974.
The elimination of
this shock—induced
shock-induced inflation
inflation during 1975 accounted for over 70 percent
of the decline
dee 1i ne in the inflation
i nfl ati on rate
rate of the GNP deflator
defl a tor to
to an average
7
l976.”~
of about 55 percent by
by the second half of 1976."
Research evidence developed more recently
recently leads me to believe
that these estimates of international
international influences
influences on U.S. inflation
inflation may
“shock effects.”
well represent upper limits
limits of these "shock
effects."
In any
any case, the
5
5Michael E. Levy, assisted by Steven Malin, International InMichael E. Levy, assisted by Steven Malin, International Influences on
1971—1976, op. cit.,
on U.S. Inflation,
Inflation, 1971-1976,
cit., esp. chap. 1.
1.
6lbid., chap. 4,
1bid.,
esp. Table 10.
7lbid., p. 8.
Ibid.,
-160-
evidence suggests that U.S. inflation would have remained substantial
throughout the first half of the l970s
197Os -- though well below the double—-
digit
digit level
level -- even in absence of
of these special
special price-escalating
price-escalating interinter——
national developments.
In fact, aa convincing case could be made
made that
the “basic”
"basic" inflation rate embedded in the U.S. economy was trending
trendfog
higher, irregularly but persistently, during the last decade-and—a—half
decade-and-a-half
and that
that this
this uptrend was masked mainly
mainly by temporary deviation
deviation caused
by the controls of the
the early b970s
797Os on the one hand, and by special
other.88 Not even the 1974-1975
international shocks
shocks on the other.
1974-1975 recession
recession
by far
far the
-- by
the most
most severe
severe of
of all
all postwar
postwar declines
declines
——
——
was
was abbe
able to
to brake
brake
this bong—term
long-term (1965-1979) uptrend of U.S. inflation
inflation rates.
“INFLATIONARY
EXPECTATIONS” AND “INFLATION
"INFLATIONARY EXPECTATIONS"
"INFLATION INERTIA’
INERTIA"
Most econometric models designed to explain this persistence
persistence of
11
U.S. inflation
inflation have assigned aa major robe
role to
to ‘inflationary
inflationary expectations”
expectations11
that infbuence
influence future wage agreements
agreements and pricing patterns, and to inincreased “inflation
"inflation inertia”
inertia" (a
(a concept which
which implies simply that the
inflation persists,
persists, the more persistent
persistent it
it becomes).
longer inflation
In the
words of one
beading expert “the
one leading
"the significance of ongoing inflation has
has
risen together with the rising rate of inflation.99
risen together with the rising rate of i nfl at ion."
To the layman, this
this may seem aa bit like
bike aa dog chasing its own
tail, but for the econometrician, the loop
loop has
has been
been closed:
8
econometric
econometric
This uptrend is clearly illustrated by Perry, bc. cit., esp.
8This
uptrend is clearly illustrated by Perry, loc. cit., esp.
1, when the two periods
periods babebbed
labelled “Controls
"Controls (1972—73)”
(1972-73)" and
p. 24, Table l,
“Food-fuel
explosion (1974-75)"
(1974—75)” are excluded.
"Food-fuel explosion
excluded. The
The batest
latest international
international
shocks came from the rapid slides
slides in the value of the dollar in 1978
(until November) and in 1979
1979 (May
(May through October),
October), and from the 1979
round of OPEC oil
oil price increases.
9Perry, loc.
cit., P.
p. 37~
37.
boc. cit.,
-161-161-
requirements for aa technical “explanation”
explanation have been satisfied. The
11
11
end result of these elaborate econometric exercises is a widely acacPerry ca
calls
it aa nmai
“mainline
cepted model
mode 1 ——
-- Perry
11 s it
n1 i ne model”
modeP
—-
that explains 15
15
inflation on
years of accelerating U.S. inflation
on the basis of aa few initial
initial
price escalations
years of excess demand, aa few years of orice
escalations caused by
‘inflationary expectations”
special ‘shocks,”
"shocks," and a lot
lot of ;'inflationary
expectations" and “infla"infla-
tion inertia”
inertia" designed to link and extend these inflationary spurts and
to bridge all
all the intervening years when inflation
inflation should have subsided
10
but did not.10
--
II would
like to propose aa somewhat different
different approach:
would like
a search
for fundamental changes in our economic and social system that appear
mid-1960s and persisted -- if
if not gained
to have originated in the mid-196Os
—-
momentum -- during the past decade—and—a-half.
decade-and-a-half.
-—
If
If such structural
structural
inflationary
changes could be identified,
identified, and if
if they carried strong inflationary
implications,
persistimplications, they would go
go a long way toward explaining the persistinflationary expectations
ence of inflationary
expectations and the increase in inflation
inflation
inertia.
inertia.
Analysis
of U.S.
of the
last two
two decades
Analysis of
U.S. budgetary
budgetary policies
policies of
the last
decades
be extremely
extremely useful in this search.
proves to be
10Leading
lOLeading supporters of the “mainline
umainline model”
modelu are well
well aware
aware of
“From 1975 through 1977, all
this difficulty.
difficulty. Thus, Perry notes: "From
all
available measures of tightness in either labor markets or product
product marmarkets registered ample slack.
slack. And no
no large
large upward movements
movements have ococthe price level since the Organicurred in particular components of the
Organioil prices in 1974.
zation of Petroleum Exporting Countries increased
increased oil
Yet despite all
all these disinflationary
disinflationary developments, the rate of inflainflation, by
by any broad measure, has continued at aa historically high rate
and now shows signs of creeping still
still further upward."
upward.”
-162—
-162-
U.S. BUDGETARY
BUDGETARY POLICY:
POLICY:
U.S.
LOOKING FOR TRENDS
Analyses of budgetary
budgetary policy often tend to be
be too global in
in apapproach, focusing mainly on what
what is perceived to be
be the overall expanexpan-
sionary (or
(or restrictive) impact of the
the budget on the
the economy.
economy.
Because
Because
of
neo—
of our narrow
narrow preoccupation with
with “fiscal
!!fiscal policy”
policy 11 as a major neoKeynesian tool for economic stimulation (or restraint), we
we have tended
to
to lose
lose sight
sight of the more complex ways in which the
the size, composition,
and rate of
of growth of
of the federal budget may affect the economic
system.
tendency to
to focus on short periods -- usually aa
Moreover, the tendency
——
single fiscal
fiscal year or two -- and excessive reliance
reliance on simple, rather
rather
——
inadequate,
“fiscal impact”
inadequate, measures of "fiscal
impact" (such as the “full—employment
"full-employment
budget surplus")
has compounded the
the myopia of traditional fiscal
surplus”) has
analysis.
Since II have
have chosen U.S. inflation
inflation as the focus for the present
present
review of federal budgetary policies,
policies, II am
am concerned
concerned mainly
mainly with
longer—term
longer-term trends and their
their implications,
implications, rather than with
with short—term
short-term
fiscal impact.
Such an
an analysis
analysis should pay
pay special attention to those
budget components that tend to create special inflationary
inflationary pressures.
pressures.
It seems to me that national defense spending and transfer payments to
individuals
individuals deserve special attention
attention in this context.
context.
Defense expenditures
expenditures have an
an inherent inflationary
inflationary tendency.
They create employment and income, but do not produce
produce any “market
"market
goods,”
goods," nor do
do they yield the
the kind of
of “public
"public benefits”
benefits" that are perper-
average consumer
as an
an immediate
enhancement of
of wellceived by the “average
consumer” as
imediate enhancement
well11
11
care, education,
education, or
or police
being (as, say, public spending for health care,
and fire protection).
This inflationary
inflationary tendency of defense spending
-163—163-
pronounced in the case of war expendiexpendibecomes, of course, particularly pronounced
tures.
Among civilian
civilian programs, transfer payments to individuals give
rise
rise to special inflationary pressures.
Designed to
into redistribute in-
sector (often in favor of
of the poor and the
come within the private sector
needy), transfer payments tend to increase short-tern
short-term inflationary
inflationary
pressures if
if the income gainers tend to
to spend aa higher proportion
proportion of
of
their
“contributors” (as is usually
their marginal income than the "contributors"
usually the case).
genMore important
important for
for the
the present
present analysis,
analysis, these transfers
transfers tend
tend to
to gen-
erate longer-term inflationary pressures in at
at least
least two distinct
distinct ways:
ways:
o
They impair incentives to
to work and to
to invest among the “conllcon-
11 Reductions
tributors,”
gainers.~
tri bu tors, n if
if not also
al so among the income gainers.
Reductions in
productivity
and in
growth of
GNP are
the more
inproductivity gains
gains and
in growth
of real
real GNP
are the
more obvious
obvious inflationary consequences.
o
If the ucontributors
“contributors” consider themselves reluctant losers
11
(rather than “voluntary
donors”) -- as may often be the case -- they
"voluntary donors")
——
--
will strive to recapture what they consider their
their “rightful”
"rightful" (e.g.,
(e.g.,
traditional
traditional or expected) share
share of real income, or real growth.
If
If the
“losers”
losersl! are concentrated in the productive
productive sector of the private
private
11
economy,
nonproducers, this
economy, while the income gainers are mainly nonproducers,
this attempt
at ''recapturing
“recapturing rightful
itself in wage and price
rightful shares”
shares'' will manifest itself
escalations.
11
~The
The list
list of
of theoretical studies
studies and empirical research on
on disdisincentive
; ncenti ve effects on
on “income
u income gainers”
gainers from
from unemployment
unemployment insurance and
11
too extensive for review here. Lately,
welfare payments is too
Lately, additional
additional
evidence on this subject has become available from analyses of
of various
“negative
negative income—tax
income-tax experiments.”
experiments.
11
11
-164-
With these analytical considerations in mind.
mind, II have reviewed
trends in total federal
well as national
federal budget outlays as we11
national defense
defense exex-
12
Denditures
Jenditures and
and transfers to individuals.12
in Chart 1l and Table 1.
The results are summarized
Unemployment compensation has been
been excluded
excluded
from transfers to
to individuals as shown there (but not from my own dedetailed analyses) because its large cyclical fluctuations tend to mask
the trends that
that concern us here.
FOUR PHASES OF NATIONAL DEFENSE
DEFENSE SPENDING
National
l960s and 1970s may
National defense
defense expenditures of the 1960s
may be
divided into
into four distinct phases:
(1) the “cold
war” phase
"cold war"
phase preceding
Vietnam;
Vietnam; (2) the
the escalation phase of the Vietnam War (fiscal 1966
1966
through 1968);
1968); (3) the de—escalation
de-escalation phase until
until the completion of the
1973; and (4) the
the recent post-Vietnam
troop withdrawal in February, 1973;
Only during the escalation phase did defense spending grow
grow much
much
phase.
phase.
faster than
l960s, it
than GNP;
GNP; during the
the pre-Vietnam
pre-Vietnam phase of the early 1960s,
it
barely advanced, and during the deescalation phase it
it declined rapidly
(see
(see Chart 11 and Table
Table 1).
1).
More recently, the
the growth rate of defense
spending has accelerated, but it
it has remained below
below the growth rate of
GNP.
latest uptrend continues
continues (as is
is suggested
suggested by the current
If this latest
political
political climate and initial
initial congressional debates of the SALT
SALT II
II
12 For the analysis of
of transfers to
to individuals,
individuals, unpublished tabtabulations from the Office of
of Management and
and Budget on direct and inindirect “payments
payments for individuals”
individuals were
were used, rather than
than federal
transfer payments to individuals
individuals as
as tabulated for the national-income—
national-incomeaccounts (NIA) budget.
budget. The former data are more appropriate for the
the
analysis at hand, since they include,
include, for example, both
both Medicare and
Medicaid,
Medicaid, while the NIA data
data treat Medicaid as a purchase of
of health
services
services by state and local governments.
11
11
-165—165-
cnart I.1
Chart
FEDERAL BUDGET OUTLAYS BY MAJOR COMPONENTS,
1961-1979
COMPONENTS, FISCAL 1961-1979
Annual Growth Rates
Outlays as aa Percent
Percent of GNP
GNP
w,
Total Budget Outlays
total
30 , - - - - - - - - - - - ~ - - - - - - ,
Outlays
Total Budget Outlays
25
, - - - - - - - - - - - - - - - - - - - , 25
25
20
20
20
,0
5
Of---->---------------,
Payments for Individuals *
35 , - - - - - - , - - - - - - - - , - - - - - , - - - - ,
Payments
Payments for
for Individuals*
IC
30
25
20~
20
15
10
"New
Ec.oriom,cs··
Social Actv,an,
Eoonom,cs
I
11110,
"'Social
Activism·'
Soc,al Act,v,sm
I
C
0
National Defense
20
15
10
0
-5
National Defense
Defense
7Un
“cod
War”
t,on Ceescalauion
“Vietnam War”
H
‘
Post Vietnam
-1 0 ' - - ' - ' - ' - ' - ' - ' - ' - - ' - - ' - ' - ' - - - - ' - - ' - ' - - '
1977
1965
1973
1969
1977
1961
1969
c
“Cold
War”
,951
‘961
lion Oe°scaIation
Deescalation ,
“V,etr,am War”
"Vietnam
War"
1965
965
1969
1969
Post’V,etnsm
?ost·Vietnam
1973
c~de,ei1a.q,~,,~,d,act
h,~*aseac’,~a~
ha,* as !M
ii,. ,a,o~
** ,nckides
a,I (l~ecl at><! •nd~ect ‘,,,,ie’
trnt1sier oay,,el,’s,.caaiL~,.mØoyrne,,~c,nen,ai,.,.
payments ••ceot urn,mp<oymen, comoensatt<Jn, ·,,tnc~
was e.cSuded Mere
maier cyc;’c*
cyc,«:aa compOf'Ml
5o,,,casI
,dS,oeqet,
Oo,ini,,Oa Board
Sou,casc Onion,,
Office of Ma,ia~ernei’t
Management o
ar'\d
S~et ma
The come,ence
—166-166-
IC
1977
1977
Table I1
Selected Data for
Analysis of Federal Budget Policy, Fiscal 196l~1979
1961-1979
forAnalysis
7961-65
1961-.65
1966-79
19$6.79
7967-65
196).1;!5
Averaoe
~
Average
Average
Average
Tome)
Total Budget Outlays
Payments
lndiv,duaia’
Payments for Individuals"
National
National Defense
5,2
5.2
6,6
6.6
11
1.1
10.8
10,8
15.3
5.7
6.7
5.2
6.6
1.1
Productivity
Productivity
Real ONP
GNP
Inflation
(Implicit GNP
GNP Della
tor)
Intlelion (implicit
Deflator)
3,2
3.2
4,2
4.2
1.5
1.5
1.6
2.9
2.9
5.9
5,9
3,2
3.2
4,2
4.2
1.5
Total Budget Outiays
Tote)
Outlays
Payments
Payments for individusls’
lndividua1s·
National Defense
Deticit
Budget Deficit
19.1
19.1
4.3
8.5
5,5
0.8
21.0
7.1
6,8
6.8
1.7
19.1
Fiscal Thrust
Thrust
Expenditure
Component
Expenditure component
comoonent
Revenue Comoonent
1.4
1.4
i1.0
.0
0.4
2,0
2.0
2.1
0.0
0.0
'1.4
1.4
1965’Se
,-,
t97o’73
1970-73
Average
Avernge
1974’79
1974-79
Aiterste
Averaoe
11.9
11,9
16.1
14.1
14.1
7.6
7-6
15.9
-1.6
—1.6
12.3
12,3
14-5
14.5
7.4
2.4
4.6
3.6
3.6
2.0
3.2
3.2
4.9
4.9
0,7
0.7
1.6
1.6
8.0
20.3
20.3
5,0
5.0
8,7
8.7
1.1
1.1
20.5
20,5
6.7
67
7.~
7.1
1.5
21.7
21.7
8.8
5.4
5,4
2,3
2.3
;_5
1.5
2.2
‘—0.7
-0.7
18
1.8
1.7
0,2
0.2
2,5
2.5
2.3
2,3
0.1
01
Avervge
Average
Annual
Annual Growth
Growth Rates
Percent o!
GNP
Percent
ol GNP
4.3
8.5
8.5
0.8
0.8
10
1.0
0.4
0.4
‘tnclucex
elI dorect
paymetttt, except unemployment compensauon,
compentatlon, whicn
wticfl Wee
Ina,or cyclical
com•includes all
direct and
and ,nditect
,ndirect Ireotter
transter payments,
was excluded
e~cluded here sa
as the
the ma1or
cycl!cal component.
conference soatd.
Sources: Oftice
Office of
of Manegement
Management Soc
anc euoget:
8uoget The
The Conference
Soard
-167—167—
be reached when the share off GNP devoted
agreement), aa point may soon be
0
to national
national defense will be rising again.
with the exception
exception of the early Vietnam
Vietnam War escalation -- its
But with
--
contribution
l960s was discontribution to
to the
the inflation of the second half of the 1960s
discussed earlier
earlier -- defense spending as aa percent of
of GNP has
has been declindeclin--
ing.
The decline in the share of GNP devoted to national defense could
have been expected to moderate
moderate (rather than stimulate) inflationary
l970s.
pressures during the 1970s.
TRANSFER PAYMENTS:
PAYMENTS:
THE BEND
BEND IN
IN THE TREND
Transfers to individuals
individuals present
present aa drastically different
different picture.
Fiscal 1965
1965 marks aa clear dividing line between the moderate growth of
these transfers during the first half
half of the
the decade and the much higher
growth rates that began with fiscal 1966
grov1th
1966 and lasted at least through
1).
fiscal 1977 (see Chart l).
During fiscal years 1978
1978 and 1979, the
growth
grov1th of
of transfers to individuals slowed significantly.
The
The share
share of
GNP redistributed through federal transfer programs
programs rose rapidly and
4.2 percent
persistently from 4.2
percent in fiscal 1965
1965 to 9.1 percent in
in fiscal
1976
1977; it
it declined
slightly during
during fiscal
fiscal years
l978 and
1979.
1976 and
and 1977;
declined slightly
years 1978
and 1979.
Clearly,
it is much too early to
Clearly, it
to tell whether fiscal 1977
1977 marked the
end of the rapid-growth
rapid-grov1th phase of these
these transfers
transfers and the beginning of
of aa
,ihether it
it represents simply
simply aa
new phase of relative containment, or whether
brief "pause."
“pause.”
Whether pause
pause or change, this is the first noticeable
13
downward deflection
1966.13
deflection in aa trend that started in
in fiscal 1966.
13 Note that payments for individuals grew
gre,i at
at an average annual
rate
1966-1979, compare
rate of 15.3 percent during fiscal 1966-1979,
comoare with
v1ith 6.1 percent
during fiscal 1961-1965. As
As aa percent of GNP, these payments averaged
4.3 percent in fiscal
1961—1965, 5
5 percent
fiscal 1961-1965,
percent in fiscal 1966-1969, 6.7
6.7
percent in
1970—1973, and 8.8 percent in 1974-1979
1974-1979 (see
(see Table 1).
in fiscal 1970-1973,
-168—
-168-
Clearly, fiscal 1965 marked a watershed for transfer programs: it was
“New Economics”
mew “Social
the end of the uNew
Economics!! and the beginning of aa new
!!Social
Activism."
Activism.”
The relatively
relatively moderate growth
growth of transfers to individuals during
first half
half of the 1960s
1g5os reflected the basic policy
policy approach to the
the first
Kennedy Administrations
“New Economics.
Economics.”
Administration's "New
11
The acceleration of real
growth
the reduction in the unemployment rate
growth amd
and the
rate were to be
be achieved
through stimulation of the private sector, rather
rather than through public
public
programs and an expansion of the government sector.
sector.
The
The major policy
tools were the liberalized
liberalized depreciation of 1962, the investment tax
tools
credit
credit of 1963, and the corporation and personal
personal income tax cuts
cuts of
1964 and 1965.
1965.
The New Economics proved remarkably successful.
During
During
1961—1965, the unemployment rate declined gradually toward the
fiscal 1961-1965,
4 percent full-employment target (as defined in the l96Os),
1960s), real GNP
grew at
at an
an average annual rate of 4.2 percent and annual productivity
productivity
gains averaged 3.2 percent.
All
All these were far better performances
1970s, yet price
price stability
stability was preserved
than those obtained during the 1970s,
right up to the onset of the Vietnam War.
The assassination of President Kennedy in 1963 and, in its
its wake,
Johmson, the passage of the Civil
the assumption of power by Lyndon B. Johnson,
Rights Act in 1964, and the burning
burning of
of the inner
inner cities
cities during the
long, hot summer of 1965, ushered in aa new era of “Social
nsocial Activism.”
Activism."
President Johnson -- one of the great parliamentarians of
of this century
--
and aa great admirer of President Roosevelt’s
Roosevelt's New Deal -- secured the
--
passage of far—reaching
far-reaching new social and economic legislation; this inincluded the Economic Opportunity Act of 1964, the Permanent Food
Food Stamp
Act of 1964, the Social Security Amendment of 1965 which created
—169-169-
Medicare and “Medicaid,”
Medicaid, and the Demonstration Cities and Metropolitan
“Medicare”
11
11
11
11
Cities”
Development Act of 1966 which established the new “Model
"Model Cities"
program.
program.
Many of
of the
the new
new federal
took the
transfers to
Many
federal programs
programs took
the form
form of
of transfers
to
1966—1968
individuals and
and expanded
expanded at aa very rapid pace even during the 1966-1968
expansion
of the Vietnam War.
expansion phase of
In fiscal 1965, federal expendiexpendi-
tures for Food Stamps,
Stamps, Medicare and Medicaid were negligible;
negligible; by fiscal
1968, they amounted to $0.2 billion,
billion,
billion, $5.3 billion,
billion, and S2.0
$2.0 billion,
1978. the latest year for which
respectively; and by fiscal 1978,
which actual
actual data
(rather than estimates)
estimates) are available,
available, they had risen to $5.5 billion,
$25.2
billion -- for aa combined total equal to 2.0
$25.2 billion,
billion, and $10.7 billion
——
percent of GNP.
This rapid expansion
expansion of social programs with
with heavy reliance on
on
transfer payments extended from the second half of the 1960s
1960s through
the
the 1970s.
After
beneAfter repeated large adjustments in Social Security bene-
fits far in excess of inflation, the entire Social Security
Security program
program was
put under the umbrella of
of aa cost—of-living
cost-of-living escalator clause in
in 1975,
while real after-tax take-home
take-home pay of many workers
workers and real returns
returns on
on
investment were lacking such protection and declined during aa major
major
part of
of the 1970s.
1970s.
the producing
producing to the nonnonRapidly growing transfers, mainly from the
retired, the disabled, the nonworking
producing sectors (such as the retired,
were financed in what would appear to
to be highly inflationary
inflationary
poor), were
ways:
oo
By frequent large increases in Social Security taxes which
are, in the view
view of
of many economists,
economists, among
among the most
most inflationary
inflationary taxes.
—170—
-170-
oo
contributed to excessive money
By large budget deficits that contributed
14
growth.14
oo
By inflation itself which fattened the
the federal
federal government’s
government's
income—tax take,
eroding real
of
take, while
while eroding
real after—tax
after-tax purchasing
purchasing power
power of
income-tax
workers and real after-tax
after-tax return on investment.
The limited statistics available on
on the subject
subject tend to
to confirm
this erosion
erosion of real
real purchasing
purchasing power
power of
of the
the producing sector.
sector.
For
For
after—tax weekly earnings of nonfarm production workers
example, real after-tax
workers
-- the
the best measure
measure available
available from
from the
the Bureau
Bureau of Labor
Labor Statistics ---
——
grew at
rate of
of 22 percent
grew
at an
an average
average annual
annual rate
percent during
during 1948—1965,
1948-1965, as
as comcompared
with 0.1
1966-1978 (see
(see Chart
Chart 2).
pared with
0.1 percent
percent during
during 1966-1978
2).
Even
Even after
after
allowing for all the limitations of these data, the sharp erosion since
since
14
‘4While
While there
there is no simple, positive, short-term
short-term relationship
between budget
(e.g., deficits
deficits may
may be
between
budget deficits
deficits and
and inflation
inflation (e.g.,
be induced
induced or
or
enlarged by
by aa recession which also tends to
to curtail inflation), perpersistent high budget deficits during relatively prosperous periods
periods exert
exert
strong upward
growth. This
This linkage
was illuminated
strong
upward pressure
pressure on
on money
money growth.
linkage was
illuminated
during
1979 testimony
testimony of
of Paul
Volcker, Chairman
of the
during the
the September
September 5,
5, 1979
Paul Volcker,
Chairman of
the
Federal Reserve Board, before the Rouse
House Budget Committee.
Representative
"There are those who say
say there
Representative Simon: “There
is no relationship between money
money supply
and the money supply
2
policies
the Fed
deficits .... How
depolicies of
of the
Fed and
and our
our deficits?
How do
do you
you describe it and what
what kind
kind of
of relationship is there between
that increase in the money supply and the deficits?"
deficits?”
Mr.
Mr. Volcker:
Volcker: “The
"The degree
degree to
to which the budgetary
budgetary defideficit puts
on the
Federal Reserve,
cit
puts pressure
pressure on
the Federal
Reserve, puts
puts pressure
pressure on
on
the credit markets and
and through the credit markets pressure
pressure
on the
the Federal Reserve to increase the money supply, depends
deoends
great deal
on what
else is
is going
deal on
what else
going on.
on. And
And the
the relationship
relationship
aa great
becomes
much more
difficult in
boom period
than in
rebecomes much
more difficult
in aa boom
period than
in aa recession period.
period. But all things
things equal,
equal, over aa period of
time, the
the deficit means at the
the very least that credit marmarkets will be
be tighter than
than they otherwise would have been
with a constant
constant Federal Reserve
Reserve money—supply
money-supply tarqet and that
that
the money-supply
money-supply target will have to be
be increased,
increased, which in
turn has inflationary
inflationary repercussions.”
repercussions. I!
—171—
-171-
chart
Chart 2.
ANNUAL PERCENT
ANNUAL
PERCENT CHANGES IN REAL AFTER-TAX WEEKLY EARNINGS
Private Nonfarm Production Workers
5
7
-
:.....
•.
'
-- i
3
no Dependents
Dependents
Workers with
with no
-5
c__J
5
--,
nfl
rfl
-5
—5
'50
‘50
'55
‘55
'60
‘60
'65
‘65
Sources:
Surnau of
Laoor Slat1st1cs:
The Conference
Conlerence Board.
Soard.
Sources: Bureau
0’ Lacor
Stat,st,ca: Tb
5
—172—
- 172-
70
‘70
~
~
Workers with Three
Three Dependents
Dependents
1948
~
75
‘75
1978
1965 is obvious.
In Its
its 1979 Annual Report, the Council
Council of Economic
Advisers discussed
discussed the erosion of investment incentives and stressed
15
the need for stimulating investment. 15 After
After reviewing four alternate
measures of
of profitability, the CEA concluded:
“Of
"Of the
the four measures of
profitability, only one,
one, the rate
rate of
of return on stockholders’
stockholders' equity,
has
has regained the 1955—70
1955-70 average.
The other three are well below
below the
1955—70 average and still further below the average for 1962—66,
1955-70
1962-66, when
when
16
16
investment outlays rose very strongly.”
strongly."
Not only were investment
investment incentives eroded in the 1970s, but a
investment had to be devoted to “nonlarge and increasing
increasing amount of
of investment
''nonproductive
enviromental regulaproductive uses”
uses in order to
to meet new safety and envirornental
regula11
tions.
In
In this setting
setting of poor real after-tax gains for workers and
low investment incentives, productivity
productivity and real growth could be
be exexpected to
to suffer.
In fact, average
average productivity
productivity gains have
have been dede-
clining steadily
steadily since the first half of the 1960s
of
l960s and real growth of
1970s averaged well below that of the
the previous
previous decade.
GNP during the 1970s
(For details, see Table 1.)
Thus, not only did the federal government redistribute
redistribute aa steadily
rising share of real
real income -- mainly from the producers
producers to nonpro—
nonpro--
ducers -- but this
this redistribution appears to have contributed to, and
-—
was in turn affected by, aa slowdown in real growth.
Thus, workers
to sizable real—income
real-income
conditioned during the l950s
1950s and early 1960s to
150p._cit.,
op. cit., pp.
pp. 124-34. The CEA concluded: “If
"If the investment
1983 is to be realized, policy
needed to
to reach our economic goals in 1983
actions are required
required that will strengthen investment
investment incentives and
and rererisks” (p.130). It went on to
duce investment costs and risks"
to recommend
recommend
11
“tax
tax reductions
reductions designed
designed to
to strengthen
strengthen investment
investment incentives.”
incentives. 11
16
p. 129.
Ibid., p. 129.
—173—
-173-
gains were doubly disappointed as they received a smaller part of a
more slowly growing pie.
In such
such an environment, attempts to restore
real gains of
of workers through higher wage demands,
demands, and to
to shore up
profitability through price increases, could be
be expected
expected to
to recur frefre-
quently, since
since they were
were bound to fail against the power of
of the federal
government to enforce its
its own priorities.
11
11
In the struggle to
to recapture aa ‘fair
fair share”
share of real
real income growth
the patterns of an
an earlier
earlier and happier period),
(probably based on the
strongly positioned groups could be expected to do better than
than those
those in
in
relatively weaker
weaker bargaining positions.
Thus, highly paid skilled
workers
workers and strong unions
unions would experience less erosion of real gains
unskilled or unorganized labor.
than unskilled
evidence presented by
Some recent evidence
Perry indicates that this is precisely what happened in the
the l970s.
1970s.
He
concludes that "for
“for the eight years
years as
as aa whole
whole (1970-77), union wages
have risen an average of
of 11 percent aa year faster [than average wages].
But
But while
while they
they have
have outpaced average
average wages
wages over this period,
period, the
the 1.7
percent
percent average annual increase in real wages
wages in the union sector
during the 1970s just
just maintained the average rate
rate of real wage increase
77
of the previous decade.’
decade. ,,17
During the 1970s, the federal government -- unwilling
unwilling to
to adad——
just its own inflationary policies and priorities -- applied wage and
—-
price
price freezes and controls intermittently.
These "incomes
“incomes policies”
policies"
were
were intended to suppress
suppress inflationary pressures from the
the private
private proproductive
sector that had
had been created, or at least intensified, by
by the
the
ductive sector
government’s
government's own policies.
In order to
to minimize
minimize the political
17Loc._cit., pp. 31-32.
Loc. cit., pp.
-174-
pressures that arise from large and frequent tax increases (and that
ultimately led to
to the “taxpayers’
"taxpayers' revolt”
revolt" of the late
late 1970s),
1970s), the
federal government
Social Security taxes
mainly on increases in Social
government relied mainly
(which
(which are less “visible”
"visible" and create less popular resistance than
than
income
income—tax
income taxes), on the inflationary feedback that swells income-tax
receipts
it erodes
erodes real
receipts as
as it
real after-tax
after-tax buying
buying power,
power, and
and on
on deficit
deficit
financing.
financing.
During fiscal
1961—1965, annual
annual federal
During
fiscal 1961-1965,
federal budget
budget deficits
deficits as
as
percent of
percent; this
of GNP
GNP averaged
averaged 0.8
0.8 percent;
this percentage
percentage rose
rose steadily
steadily to
to
aa percent
19701966-1969; 1.5 percent during fiscal 19701.1 percent during fiscal 1966-1969;
1974—1979 (see
1973; and 2.3 percent during fiscal 1974-1979
(see Table
Table 1).
1).
FISCAL
FISCAL POLICY:
POLICY:
THE
“FISCAL THRUST"
THRUST” OF
OF THE
1970s
THE EXPANSIONARY
EXPANSIONARY "FISCAL
THE 1970s
sketched some
some of
the processes
processes through
II have
have sketched
of the
through which
which the
the diversion
diversion
of an increasing share of GNP to
to transfers
transfers (mainly from the producing
producing
to the nonproducing sector) added inflationary
inflationary pressures
pressures after
after 1965.
1965.
Implicit
Implicit in
in this
this analysis were the
the following two propositions:
propositions:
oo Direct
Direct and
and indirect transfers to individuals, jointly with
of fiscal growth
growth over
national defense speniding,
spending, dominated the patterns of
the last decade-and-a-half.
decade—and—a—half.
(But
(But except
except for the Vietnam
Vietnam escalation
phase, transfers were by far the most prominent component
component shaping fisfiscal
ca 1 growth.)
growth.)
oo
The budgetary
The
budgetary policies
policies and
and processes
processes described
described here
here resulted
resulted
in
expansionary budgets
budgets in
in the
l970s than
had been
in far
far more
more expansionary
the 1970s
than had
been the
the case
case in
in
the
the previous decade.
Moreover,
expansionary thrust
Moreover, this increased expansionary
originated from rapidly growing spending programs (mainly transfers),
rather than from tax reductions.
—175-175-
The extent
extent to which
which the first proposition is true may be gleaned
from Chart l.
1.
To my knowledge, the second proposition
proposition is
is new and has,
so
so far, been unproven.
Therefore, it calls
calls for empirical investigation
and evidence.
Until recently, I had suspected but had
had been unable to document
document
satisfactorily
satisfactorily that, on the average, fiscal policy of the l970s
1970s had
been
been more expansionary.
With the
the cooperation of the
the Bureau
Bureau of Economic
Analysis of
of the Department of Commerce, II have been able to develop
reasonably consistent (preliminary) quarterly
quarterly and annual
annual estimates of
‘fiscal thrust”
"fiscal
thrust'' back to fiscal 1959
1959 -- just
just in time for this meeting
meeting
(see Table 2).18
“expenditure component"
component”
2) . 18 This measure
measure consists of
of an "expenditure
19
which measures
measures change in
in autonomous
autonomous government
government expenditures, 19 and aa
--
‘revenue componentn
conponent” which measures the initial
!lrevenue
initial revenue loss (expansion(expansionary (+)) or revenue gain (restrictive
ctiirajchan9es
(restrictive(-)) from a
structural
changes in
in
(+))
(-))
tax
tax provisions (rates or base).
base).
Each component, as well
well as
as total
“fiscal
"fiscal thrust”
thrust" (their sum)
sum) is best measured as
as a percent of
of GNP, in
18
.
“.
18 I coined the term fiscal thrust
1974 when II published
1
the
"fiscal thrust" in 1974
published my
first annual estimates
estimates in
in ~
The Federal Budget: Its Impact on the Economy,
The Conference Board, New York,
1975 edition, p. 12.
York, 1974, fiscal 1975
12. My
quarterly estimates were published in 1976
1976 (op. cit., fiscal 1977
first quarterly
1977
edition, p. 11).
11). The measure
course7d~FT~ed
rom
measure itself is, of
of course,
derived ffrom
Keynesian macroeconomic analysis.
analysis. Previous uses
uses of similar
similar measures
may be found in vlilliam
William H. Oakland, "Budgetary
‘Budgetary Measures of
of Fiscal PerPerEconomic Journa 1 (April 1969),
348-58; E.
formance,"
formance,” Southern Economic_Journal
1969), pp. 348—58;
Gerald
Gera
1d Corrigan,
Corrigan, “The
"The Measure
Measure and Importance of
of Fiscal
Fi sea 1 Policy
Po 1icy Change,”
Change,"
Federal Reserve Bank
Bank of New York ~p~jjjyReview
Monthly Review (June 1970), pp.
pp. 135—45;
135-45;
“Federal Budget Discipline and National Priorities
Paul W. McCracken, ''Federal
of the
the 1970s,”
in Michael
E. Levy,
çonpmic Issues
of
1970s," in
Mi chae 1 E.
Levy, editor,
editor, ~Ma ior Economic
Issues of
of the
the
l970s, The Conference Board,
1970s,
Board, New
New York, 1973,
1973, esp.
esp. p.
p. 9.
19National-imcome-accounts (NIA)
National-income-accounts
budget data were used; induced
expenditures (mainly regular unemployment compensation) are excluded;
and long—lead
long-lead defense expenditures are
are adjusted from their “delivery
"delivery
basis”
basis" to aa timing that reflects more
more closely
closely actual production.
-176-
Table
Table 2
·auarterly
"Fiscal Thrust”
Thrust" and Its Major
Quarterly and Annual Estimates (Preliminary) of “Fiscal
1959-1980 1
Components, Fiscal 1959~198O’
(NIA budgetdata;
budget data;$
at seasonally adjusted annual rates)
(NIA
$ billion at
Expenditure
Contribution’
Conmbution'
ny 1959
1959
FY
3.5
111tI
Iv
IV
—0.9
-0.9
5.3
1,6
1.6
—2.3
-2.3
—1.1
-1.1
I
II
1.5
F? 1960
FY
1960
F? 1961
FY
10.2
10.2
F? 1962
1962
FY
7.1
-1-2
1.2
III
Ill
Iv
IV
I
II
F?
FY 1963
1963
4.3
4.3
III
/II
I
It
II
F?
FY 1965
1965
IIII
Ill
III
tV
IV
I
ItII
F?
FY 1967
21.2
212
(
IIII
20,4
20.4
FY 1968
1968
III
Ill
Iv
IV
I
IIII
F?
FY 1969
1969
F?
FY 1970
1970
—0.3
-0.3
2,7
2.7
1,5
1.5
3.4
16.2
Ill
Iv
IV
I
IIII
2.74
15.2
15.2
2.46
2.46
0.25
0.68
0.39
0,39
1.10
0.81
-7.5
—7.5
—6,6
-6.6
1.7
-5.6
—5.6
3.0
22.0
22.0
3.6
—0.4
-0.4
3.1
-0.5
—0.5
0.94
0.62
0.82
0.89
0.30
2.0
2.0
5.8
-22
—2.2
9.6
—6,3
-6.3
—1.0
-1.0
-7.1
—7.1
‘—0.4
-0.4
5.8
0.1
4.3
4.3
0.9
10.9
10.9
0.83
0.83
0.48
0.98
0.47
6.6
4.6
5.0
5.0
2,7
2.7
0.0
0.0
0.2
—5,5
-5.5
0.1
0,1
-14.8
—14.8
7.3
2.75
2.75
18.9
18.9
—5.2
-5.2
0.00
—0.11
-0.11
0.03
0.03
0.38
0,38
8.7
3.6
2.2
3.3
—0.5
-0.5
-0.2
—0.2
-1.9
—1.9
0.3
0.3
2.0
2.0
5.6
3.3
9.5
III
Ill
Iv
IV
I
IIII
0.32
17.8
—2.3
-2.3
-0.17
—017
0.87
0.10
0.06
-0.1
—0.1
—0.7
-0.7
1.9
44
4.4
2.9
0.2
-5.0
—5.0
-02
—0.2
7.1
4.8
4.8
6.9
6.9
2,4
24
111
IV
0.86
5,5
5.5
—2.1
-2.1
0.38
0.38
0.33
0.33
-0.03
—0.03
0.07
0.07
—1.0
-1.0
54
5.4
5.2
5,2
5.8
-0.1
—0.1
0.0
1.7
1.8
5.8
5,8
3.4
7,2
7.2
3.5
0,23
0.23
0.18
018
0.98
0.98
-0.09
—0.09
0.75
0,75
15,4
15.4
3.4
19.9
FY 1966
1966
1.30
2.2
2.1
-2.7
—2.7
0.3
0,3
0.0
0.1
0.1
,.6
4,6
5.4
0.0
—0.7
-0.7
0.2
02
2,6
2.6
0.38
0.36
0.40
0.53
0.71
0.71
1,2
1.2
1.0
5,7
5.7
-0."!
—0.1
1.9
10.1
10.1
2.1
(
7.8
0.0
0,2
0.2
—2.5
-2.5
-0.1
—0.1
—1.0
-1.0
5.3
5.3
0.6
0.6
0.4
OA
lIt
Ill
IV
200
2,00
1.7
17
2.0
2.4
3.7
3,7
0.0
0.0
0.3
0.4
04
—2,4
-2.4
5.3
III
11:
Iv
IV
I
IIII
9.8
g.e
0.7
0.25
0.25
0.04
0.04
-0.10
—0.10
0.12
0.12
-2.5
—2.5
0.6
0.0
—0.3
-0.3
0.0
0,0
2.2
2.2
1.9
—0.2
-0.2
0.4
0.4
IV
0.30
1.2
-0.4
—0.4
—0.03
-0.03
0.30
0.16
016
0.37
1.69
3.7
3.9
4.0
10.4
10.4
-177—177—
Asa •Jo of GNP
Asa%oiOWP
tat’cnan
ge
T,u·cnange
Contribution
Contnbut,on
{S)
151
-0.18
—0.18
1,17
1.17
0.34
0.34
-0.48
—0.48
-0.22
—0.22
i_s
—0.1
-0.1
1.0
5,4
54
—0.5
-0.5
,,,
0.74
-1.1
—1.1
—0.4
-0.4
Exoendirure
Expenditure
Contribution
Conrnbuuon
(4J
5.6
5.5
1.6
-3.7
—3.7
-1.1
—1,1
0.0
-0.6
—0.6
-2.0
—2.0
0.0
1.8
18
2.0
2,7
2.7
3,7
3.7
/II
IV
I
((
II
,.,
2.8
-2.6
—2.6
0.2
0.2
-0.5
—0.5
0.6
0.6
I
IIIt
.",seal
Fiscal
Thrust
~
(3)=/li+/2!
(3)stt)+12)
0.3
0.2
-1.4
—1.4
0.0
1.2
12
II}
IV
F?
FY 1964
1964
'"
Tau’change
Tax·change
Contribution’
Conmbut1on'
(2}
(2)
0.01
0.45
0.09
1.12
1.12
0,07
0.07
0.05
-0.29
—0.29
0.00
0.00
-0.52
—0.52
0.00
0.00
-0.12
—0,12
-0.39
—0,39
0.00
0.00
-0.07
—0.07
—0,02
-0.02
0.00
—0.06
-0.06
0.00
0.13
0.00
0.00
0.00
0.00
0.05
0.05
0.08
0.08
-0.42
—0,42
0.00
0.00
0.04
0.04
-0.44
—0.44
-0.02
—0.02
1.64
0.00
0.01
0,01
0.74
0,74
0.86
0.52
-0.01
—0.01
0.00
0,26
0.26
0.27
-0.29
—0.29
0.42
0.02
0.02
-0.68
—0.68
-0.G3
—0.03
—0.30
-0.30
-0.07
—0.07
-0.02
—0.02
-0.25
—0.25
0.04
-0.63
—0.63
0,00
0.00
0,03
0.03
-0.65
—0.85
0.01
0,01
-1.64
—1.64
—0.72
-0.72
-0.11
—0.11
-0-77
—0,77
—0.06
-0.05
0.60
0.80
0.38
-0.04
—0.04
0.32
-0.05
—0.05
Fiscal
Theusi
--1!!!.E!_
(4/e,,(5)+!6;
0.55
0.55
1,24
1.24
0.39
0.39
-0.77
—0.77
-0.22
—0,22
-0..22
—0.22
0.25
0.25
-0.08
—0.08
-0.49
‘—0.49
0.12
U3
1.93
0.34
0.34
0,40
0.40
0.47
0.47
0.71
0.71
1.43
0.23
0.18
018
1.03
-0.01
—0,01
0.33
0.38
0.37
-0.47
—0.47
0.05
2.50
2.50
-0.17
—017
0.88
0.84
0.92
0.84
0.64
-0.01
—0.01
—0.11
-0.11
0.29
0.29
0.65
2.46
2,46
1.25
1.25
0.50
a.so
0.30
0.30
0.44
0.44
2.44
0.87
0.60
0.64
0.34
1.83
1.83
0.25
0.71
0,71
-0.26
—0,26
1.11
-0.83
—0.83
—0.75
-0.75
0.19
0.19
-0.61
—0.61
0.32
0.32
2.29
2.29
0.39
0.41
0.4,
0.41
1.07
1.07
Table 22 (continued)
Table
Thrust'' and Its Major
Quarterly and Annual Estimates (Preliminary) of "Fiscal
“Fiscal Thrust”
Components, Fiscal 1959·1980'
1959~198O’
(NIA budget
budget data~
data;$
seasonally adjusted
adjusted annual
annual rates)
rates)
jNIA
$ billion
billion at
at seasonally
Expenditure
Expenditure
Contribution'
contribution’
/1)
F?
FY 1971
1971
14.9
7.9
—1,7
-1.7
III
Ill
lv
IV
I
IIII
23.6
III
Ill
IV
I
ItII
18.3
111
III
IV
I
IIII
31,0
31.0
III
111
IV
I
II
2.3
2.3
7.0
7,5
7.5
14.2
14,2
FY 1975
1975
F?
60.0
III
Ill
IV
9.0
13,0
13.0
4.4
—0,4
-0.4
5.9
46.9
46,9
F? 1977
1977
FY
IV
I
IIII
III
111
IV
iv
I
IIII
III
111
eat.
est.
sat.
est.
eat.
est.
eat.
est.
39,8
39.8
48.8
—17.6
-17.5
3,8
3.8
3.9
3.9
1,2
5.2
0,17
0,17
0.52
0.52
0.55
1.01
0.98
0.98
0,23
0.23
1.64
1.64
1.25
1.25
1,60
1.60
22.7
1.09
1.09
0,07
Q,07
0.51
0.5~
0.88
0.88
1,55
i.55
8,8
88
4.1
4,1
2.5
7.3
55.7
0.65
0.65
0.22
0.13
0.56
0.56
1.99
15.4
15.4
—2,4
-2.4
15.8
15.8
—1.7
-1.7
0,80
0.80
0,26
0.26
0,34
0.34
0.59
0.59
21,8
21.8
6,3
6.3
12,2
12.2
—1.9
-1.9
33.6
—6.8
-8.8
9,7
9.7
—0.6
-0.6
—0,4
-0.4
0.58
0.58
0.81
0.81
0,27
0.27
—0.02
-0.02
0.34
2.54
15.9
'.5.9
5.2
5.2
12,1
12.1
20,7
20.?
—4.1
-4.1
—0,3
-0.3
—02
-0.2
—4.7
-4.7
—0.1
-0.1
10,8
10.8
7.1
7,1
4,8
48
11.0
1~,0
38.9
53.9
9.8
2.28
4.12
4,12
—3,2
-3.2
4.0
2.5
2.5
3,7
3.7
17.8
17,8
6.0
8.0
14.1
14.1
33.7
2.5
7.5
3.1
14.4
14,4
-0.4~
—0.41
1.75
1.75
0,12
0.12
0.02
Q.Q2
142
14.2
0.6
22,1
22.1
21.7
—0.6
-0.6
4,3
4.3
—0.7
-0.7
—9,3
-9.3
45.9
IV
I
IIII
Ill pre!.
prel.
27.5
—30.6
-30.6
12,9
12.9
•.4
4,4
2.7
12.0
12.0
F? 1979
FY
1.48
58.6
Sa.6
7.0
32-0
32.0
IV
I
IIII
III
Ill
F? 1980
FY
1980
12.9
12.9
0.13
0.13
0,34
0.34
1.04
1.04
0.59
0.59
-3.0
—3.0
21.8
-5.9
—5,9
0.6
0,2
0.2
—2.8
-2.8
—1.8
-i.8
3.0
3.0
19.1
1.2
9.6
9.6
17,0
17.0
F?
1978
FY 1978
0.2
0.2
0,5
0.5
—4.4
-4.<:
0,2
0.2
—1.4
-1.4
26.0
III
Ill
IV
I
IIII
Ill
Ill TO.
T.Q.
2.12
2.12
13.5
1,36
1.36
2,0
2.0
16,8
16.8
4.2
4.2
10.6
10.6
As
3% ofCNP
Asa¼ofGNP
rxu-cnange
Tax•cnange
Contribution
Conlribulion
(Si
/Si
0.77
0.77
—0.17
-0.17
0.36
0.65
0,60
0.60
4,0
4.0
1.4
1.4
3,7
3.7
9.4
1.9
1,9
0.4
04
—7.5
-7.5
0.4
04
14.0
14.0
3.4
3.4
23.9
18,7
18.7
II
18.5
—8.0
-8.0
2,6
2.6
—3.5
-3.5
(4)
!')
1.46
1.46
2.6
2.6
—4.8
-4.8
Expenditure
bpena,ture
Contribution
Contribution
5,0
5.0
3,3
3.3
8.5
6,0
6.0
—2-3
-2.3
-4.9
—4.9
21.4
1.6
0,2
0.2
F? 1974
1974
FY
22.8
22.8
—5.1
-5.1
1,4
1.4
3.7
3.7
11.7
11.7
6,8
6.8
F?
FY 1973
FisCa)
Flsc.ii
Thrust
~
)3)s)t)+f2)
i3)=f1i.,.i2)
6,7
6.7
—0,3
-0.3
1.8
1.8
—0,3
-0.3
3.6
3,6
6,7
6.7
6-3
6.3
F?
FY 1972
1972
F? 1976
FY
1976
Tax-change
Tax•change
Conmtn//ion'
Contnbulion’
(2)
(2;
0.45
0,29
0.29
0,19
0.19
0.43
Fiscal
Fiscal
Thrust
~
(4)e(EJ+181
!4i,,,(S)+/6!
2.23
0.67
0.50
0.33
0.33
0.82
0.82
0.57
0.57
—0.03
-0.03
0.17
—0.03
-0.03
—0.46
-0.46
0,24
024
—0.21
-0.21
—0.71
-0.71
0,22
0.22
—0.39
-0.39
0.16
0.16
0,04
0.04
—0.59
-0.59
0,03
0.03
—0.28
-0.26
0.02
0.02
0,03
0.03
—0.32
-0.32
0.02
o.oz
-0.10
—0,10
0.01
0.0.
—0.19
-0.19
—0.12
-0.12
0.20
0.20
0,79
0.79
2.54
2.54
—1.91
-1.9"
—0.04
-0.04
0.25
—0,04
-0.04
0.38
—0.18
-0.18
0.22
0.13
0.13
0,19
0.19
—0.45
-0.45
—0.20
-0.20
—0.02
-0.02
—0.01
-0.01
—022
-0.22
0,42
0.42
—0,11
-0.11
0.69
—0,07
-0.07
—0.08
-0.08
0-00
0.00
—0,37
-0.37
0,40
0.40
—0.02
-0.02
—0,02
-0.02
1.66
1.66
0.37
0.37
0.13
0.13
0.33
0.33
0.81
0.81
1.08
-0.25
—0,25
1,79
1.79
—0.47
-0.47
0.05
2.02
202
019
0.19
0.55
0.55
0,23
0.23
1.03
1.03
4.02
0.99
0.04
1.52
1.52
1.45
1.45
2.39
2.39
3.12
3.12
—1,10
-1.10
0.23
0,23
0.23
0,30
0.30
2.92
2.92
0,91
0.91
0.29
0,64
0.64
1.07
1.07
1,10
1.10
0,45
0.45
0.20
0,12
0.12
0.34
2.41
2.41
0.69
0,95
0.95
0.27
0.27
0.51
0.51
1,36
1.36
0.08
0.69
0,17
O.H
0,41
0.41
‘Author’s
derived Itory
pubiianec and
‘01
'Author·s preuntinary
prel1m,nary estimates
esl!mates derived
trom the
the best available
avsHable publ,sned
and unpubt~uhedsources.
unput,i she-c sovces. Dale
Data revisions
rev,s1ons and
and retinement,
refinements hare
have not
tmen cornpleteo.
yet Deen
complelec.
‘Increases I( —- icr
i—I ln
ents mdv0=
subltaclions ci
cvangea i,i,i ‘‘regular”
'Increases
1or reductions
reductions 1-)
!n “adlusled”
"adJusted" NIA
NIA budget expenditures.
expenditucss. Adiuslm
Ad1ustments
1nciuoe subtract,ons
of changes
"regular" unempicyunemploysent benell!s
benetitsand
“delense timely
ment
and oI
of toe
tne 14IA
N!A "defense
timely adiuslment
adjustment"
'Initial
(-) or reductions I( - is
l ,n tax
tax revenues resu!t,,.,g
from sttuclural
stre.Jctural Cflaii~e5
cnan~es it1n tao
tax bases
bases or
on best
best ouDIlshea
ouol!shea and
un•
‘Initial increases I—Icr
resultrrgirorn
cv rates. baseo
bssec Ott
anus,gublisneo estimates
tom the Treasury
Timing ci
ncreases in
inthe
;:,uo!isheo
estimates from
Treasury Department and
and she
the Bureau ot
of Economic
Ecoriomic Anaiysis.
Analysis. T,ming
oi the ettect
eftect nt
,:,f the mcreases
the tan
tax base
base
on
tile employee's
emoioyee’e pan
contrioutions to
social secunty
secunty has
beet, chat
ged by
by author
ccncenl,ale this
increase mainly
on the
part ol
of contnoutions
to soc,al
has been
changed
author to
to concent,ate
this ,ricrease
ma,rHy it1n the
the lastiwo
last two calet’
ca!en•
darquartars
darquanem.
TO —Transitional
T.Q.Trans1t1onal quartet
quarter
Sources; Bureau oI
Economic Analysis; The Conference
Conlereece Board.
Sources:
of Econom,c
-178-
to permit
permit historical comparisons
comparisons and minimize inflation—induced
inflation-induced
order to
20
distortions of these measures
distortions
measures2°
In short, fiscal thrust and its
its components are designed
designed to
measure the initial
initial expansionary impact originating
originating from the federal
budget to which the traditional
traditional Keynesian multipliers
multipliers could
could be
be applied
aoplied
(or which could trigger
trigger fiscal
fiscal simulations
simulations in
in econometric models.)
models.)
What concerns us for the present
not so much the
present analysis
analysis are not
quarterly, or even the annual, levels or
or changes in fiscal thrust, but
but
rather the average degree of stimulation of
of the budget over the broad
broad
longer time periods distinguished here.
The
suirmiarized in
The results,
results, summarized
Table 1, confirm the proposition that, on balance, the budgets of
of the
1970s were more expansionary
expansionary than
than those of the 1960s, largely as the
result of much
much faster spending growth.
Fiscal
1961Fiscal thrust averaged 1.4
1.4 percent
percent of
of GNP during fiscal 1961-
1965, compared with a 2.0 percent average
average for fiscal 1966-1979.
Within
the latter
latter period, average
average fiscal thrust rose from 1.5
l .5 percent of GNP
during fiscal 1966-1969
1966-1969 to 1.8
1.8 percent during fiscal 1970—1973
1970-1973 and 2.5
percent
1974-1979.
percent during fiscal 1974-1979.
nant
nant throughout.
The expenditure
expenditure component was domidomi-
But tax cuts
cuts provided significant stimulation during
the period of
of the “New
New Economics;
Economics
tax increases provided belated and
11
11
;
restraint during the escalation phase of the Vietnam
7limited
imited .r_estr_ii_i.'1.!.
Vietnam War (par(parguns and Great
Society” spending);
tially off—setting
off-setting the "guns
Great Society"
spending); and tax
changes were nearly neutral over the course of
of the l970s.
1970s.
20
be constructed
(analogous to
ZOAA “weighted
''weighted fiscal
fiscal thrust
thrust'' could
could be
constructed (analogous
to
the "weighted
“weighted full-employment
full-employ:;1ent budget surplus’),
surplus"), but
but the complications
created by such
such aa refinement are hardly
hardly warranted in the light of
of the
use of any simple overall
overall measures
measures of fiscal impact
impact and the crudeness
crudeness
o~the
O' the basic estimates.
—179—
-179-
THE NEW "SOCIAL
REGULATION"
“SOCIAL REGULATION’
Changes
Changes in the composition and growth of
of the federal budget and
its components were not the only inflationary manifestations of what
has been
been termed here aa new “social
social activism.”
activism.
11
11
The same emphasis on
social welfare and on the
the consumer, rather than
than on
on real
real growth and
and the
producer, gave
regulation” in the mid—
gave rise to
to aa new wave
•.,;ave of “social
social regulation
mid21 The
1960s and the early 1970s.
l970s.21
1960s
The impetus came from consumer
consumer groups,
groups,
11
11
environmentalists, labor unions, civil
civil rights advocates and diverse
diverse
public
public interest groups, who felt that the traditional
traditional regulatory
agencies were not achieving “social
social goals,”
goals, such as
as product safety,
safety,
11
11
clean air
afr and water, equal
equal employment opportunities, safer and healthhealthier working conditions.
In response to these public pressures, twenty
“social regulatwenty new "social
regulation’
tion" agencies have been created since 1970.
Among these,
these, the most
important
important ones are the Consumer
Consumer Product Safety Commission, the EnvironEnviron-
mental Protection Agency,
Agency, the Equal Employment Opportunity Commission,
and the Occupational Safety and Health Administration.
These
These new
agencies charged with social regulation
regulation were among the most prominent
prominent
“growth
l970s; their full—time
industries" of the 1970s;
full-time staff
staff increased from
"growth industries”
17,324 in fiscal 1970 to
17,324
to 69,258
69,258 in fiscal 1979 (86 percent of the
regulatOry staff).
federal governments
government's total
total regulatory
The
The administrative
administrative and
21
21 For
“social regulaFor further discussion of the evolution of
of new "social
regulation”
tion11 and some cost estimates, see Michael E.
E. Levy,
Levy, assisted by Delos
R. Smith and Steven Malin, The Federal Budget:
Budget: Its Impact on the
1980 No.
No. 2, pp.
pp. 12-14. For
For an
an encompassing
encompassing critical
critical
Economy, fiscal 1980
of the impact of government regulation, see Murray L. Weidenbaum,
review of
Business,
Prentice-Hall, Inc., Englewood
Englewood
Business. Government, and the Public,
Public. Prentice-Hall,
1977; also Murray
Cliffs, N.J., 1977;
Murray L.
L. Weidenbaum, The Impacts of GovernGovernment
ment Regulation, Working Paper
Paper No. 32,
32, Center
Center for the Study of
of American
Business,
Business, Washington University, St. Louis, July 1978.
-180-180-
reporting costs imposed on businesses grew
grew accordingly. More Important,
important,
business had to divert large and increasing
increasing amounts of cash
cash flow and
capital into investments designed
designed mainly to achieve
achieve compliance with
with new
social regulation.
A
major part of
of these investments -- regardless
A major
——
11
of whatever their social
social benefits -- was “unproductive”
unproductive in terms of
of our
our
11
-—
traditional measures of real output and productivity.
In fact, accordaccord-
estimates, productivity of the nonresidential
ing to
to the best
best available estimates,
nonresidential
it would
would
business sector was 1.4 percentage points lower in 1975 than it
have been under the regulatory conditions
conditions of 1967.22
1967. 22
The tendency of the new “social
"social activism”
activism" to pursue socially dedeThe
sirable goals without any proper regard for economic
economic implications,
implications,
without due consideration
consideration of benefit-cost
has been
been
benefit—cost relationships, also has
felt in
in the regulatory area.
Excessively short deadlines for meeting
regulatory standards, detailed
detailed prescriptions of specific
specific technological
solutions, absolute prohibition of the use of certain
certain substances or
processes have often raised marginal compliance
compliance costs well
well in excess of
23
marginal benefits. 23 Consequently, the new social
social regulation
——
regardless of whatever its
its social
social merits -- has
has been highly inflationinflation——
ary.
In its 1979
1979 Annual Report, the Council of
of Economic Advisers
22
.
,,
22 Edward Denison, "Effects
InstituEffects of Selected Changes in the Institutional and Human Environment upon Output per Unit of Input,”
Input," Survey
of Current Business.
Business, January 1978, pp. 21—44.
21-44.
23 For
For aa discussion of these problems,
problems, see the section on "Regu“Regulatory Reform"
Reform” in
in the 1978 Annual Report of
of the Council of Economic
Economic
206—216); also the section
Advisers (pp.
(pp. 206-216);
section on “Regulatory
"Regulatory Policy”
Policy" in the
~1979 Annual Report of the CEA (pp. 85-91).
-181-
described the dynamics of the inflationary process
process induced by the new
social regulation in
in the following way:
actions enter
Once incurred, the costs of regulatory actions
into
wage— and price—setting
into the
the wageprice-setting mechanisms of the economy.
economy.
Most
costs of
of regulatory
action show
Most of
of the
the costs
regulatory action
show up
up not
not as
as
governmental budget expenditures, but as
as increased
increased costs to
industry.
relative to
and
industry. Acceptance
Acceptance of
of higher
higher prices
prices relative
to wages
wages and
other money incomes is the way in which
which society pays
pays for the
benefits
social regulation.
however, our
ecobenefits of
of social
regulation. In
In fact,
fact, however,
our economic institutions and
and measures of
of prices do not distinguish
between
between these
these sources of
of price increases
increases and others.
others. IndiIndividuals and groups try to
to escape
escape paying the costs of
of reguregulation by
by increasing wages and other forms of income to
match
match the
the higher
higher prices.
prices. The
The result is an additional round
of
increases. But
regulation cannot
of price
price increases.
But the
the costs
costs of
of regulation
cannot be
be
avoided, and
widespread
attempts
to
do
so
add to
asd
widespread
attempts
to
do
so
simply
24
inflation.24
inflation.
SOME
SOME LESSONS FOR THE 1980s
My journey along the inflation road
road of the last
last decade—and-a—half
decade-and-a-half
has ended with aa thesis, rather than with solid
solid conclusions.
The
search for an explanation of
of the largely unexplained aspects of
of our ininflation (or of the "excessive"
you will) -- its
its duradura“excessive” money growth, if you
--
tion, persistence
persistence and steady escalation -- uncovered basic changes in
in
-—
social
social and political orientation and in our public policy.
These
These
changes -- II referred to them as aa new social
social activism -- originated
originated in
—-
——
the mid-1960s
mid-1960s and gained momentum
momentum in the 1970s.
1970s.
This social activism
activism
manifested itself in
in increased
increased reliance
reliance on
on the federal
federal government
government to
achieve socially desirable goals through new,
new, or enlarged, budgetary
budgetary
and regulatory
regulatory programs.
The consumer and
and “social
nsocial benefits”
benefits 11 were
stressed,
stressed, often at the expense of
of higher costs, slower real growth and
lower productivity
gains.
lower
productivity gains.
24
24
Among
consumers -- many
Among consumers
many of
of whom
whom are,
are, after
after
——
Op. cit., p. 87.
op.
-182-182-
all, producers as well -- these new social benefits were often to be
—-
focused on the nonproducers (who tend
tend to be perceived
perceived as
as nmore
more needy'
needy’
1
11
s oci al benefits).
benefi tstl).
and, hence, more deserving of “social
The ‘costs”
costs of this new
new social activism included increased
increased disdis11
11
incentives to work and to
to invest, slower growth of real GNP,
GNP, amd
and lower
lower
productivity gains.
AA main result was
was aa persistent
persistent increase
increase in’infla—
in-infla-
tionary pressures of our entire economic system.
system.
If this thesis has any merit, if it
it contributes in any signifisignifi-
cant way to the explanation of the ongoing U.S. inflation,
inflation, the impli—
implications
catioms are clear:
inflation control depends
depends on removal of
of
Successful inflation
the fundamental causes of U.S.
U.S. inflation.
inflation.
Fiscal and monetary
monetary policy
policy
Fiscal
restraint,
restraint, while
while necessary, will not be sufficient.
sufficient.
New policies to
encourage
encourage greater productive
productive efforts
efforts and faster real growth will
vii 11 be
essential, if price stability
stability is to be
be restored in
in the 1980s.
-183-
DISCUSSION OF THE LEVY AND MELTZER
MELTZER PAPERS
Poo·1e
William Poole
Michael Levy has
has assumed the task
task of
of explaining
explaining the persistence
persistence
of inflation.
of
confess, though, that II got off to
to aa bad start at the
II confess,
very beginning of
of his paper.
His second sentence
sentence reads:
reads:
“Monetarist
"Monetarist
explanations
deep—seated inflation provide no insights as
explanations of this deep-seated
as to
to
its economic,
economic, social,
social , and political
pol iti cal causes’
causes"
(emphasis added).
sentence towards the end of
of his summary reads:
And
And aa
“Fiscal and monetary
"Fiscal
monetary
policy restraint will be necessary, but may not be
be sufficient [to
control
contra 1 inflation].’
i nfl a ti on J."
1lieve
i eve these claims.
Fortunately, however, Levy
Levy does not really bebeOn page two of his paper
paper he says
says that, “on
"on aa
anpurely technical level, the monetarists have, of course, all the answers.
In fact, some of my own econometric
econometric exercises have
have tended to
reconfirm
reconfirm their valuable, if
if somewhat simplistic, generalizations.”
generalizations."
If
If we strip away the loaded words
words such
such as
as “simplistic,
"simplistic," then it
clear that Levy accepts the basic argument that inflation
inflation cannot
cannot
is clear
occur in the absence of
of excessive money
money growth.
Accepting this propopropo-
sition,
inflasition, Levy surely does not believe that successful
successful control of
of inflation would be possible
possible without slowing money growth.
Indeed, II cannot
cannot
believe
believe that Levy would
would claim
claim that slowing money growth would
would fail to
reduce inflation.
He simply does not in fact
fact believe that monetarist
explanations provide no insight into the economics
economics of
of inflation.
Dr. Poole is Professor
Professor of Economics at Brown
Brown University.
-184-
Levy s paper is not about monetarist
monetarist propositions linking money
Levy’s
1
growth to inflation,
excessive money growth.
inflation, but about the causes of excessive
This
obviously important.
This issue is obviously
But the reasons
reasons monetarists have
not paid much attention
attention to
to this issue to date ~re,
are, first that
that it
it was
important to
is indeed
to gain
gain agreement that inflation is
indeed aa monetary phenomphenom-
enon -- aa proposition not widely accepted thirty years ago -- and, secsec-—
——
ond, that the methods of economic analysis may not provide great ininsight into the causes of
of excessive
excessive money growth.
growth.
Levy feels that the
important issues
issues concern
and social
important
concern changes
changes in
in the
the economic
economic and
social structure
structure
that have produced
produced an
an inflationary environment.
let me
me introduce
introduce aa qualification to the
Before commenting further let
simple nionetarist
monetarist view.
Clearly, insofar as changes in the economic
and social structure, in the
the average tax rate, and
and in the regulatory
burden affect incentives and
and productivity,
productivity, the rate of productivity
productivity
growth nay
may slow
slow down.
Reduction in the growth of real output,
output, given
the rate
of money
growth, will
raise the
the rate
inflation.
the
rate of
money growth,
will raise
rate of
of inflation.
As
As aa first
first
approximation, what matters
matters is the money
money stock per unit of real GNP.
But the slow-down in productivity
productivity growth and therefore in
in output growth
can directly account
account for only aa very small
small part
part of our inflation.
We
might be able to explain one to two percentage points of
of the inflation
in recent years by
by the
the slow-down in output growth.
what all the
the shouting is about.
But that
that is not
If the current
current rate
rate of
of inflation were
only one or two percentage points above the rate of the early
early sixties,
sixties,
then the subject
subject of
of this
this conference would not be
be inflation
inflation but rather
productivity or
or some other
other issue.
If II understand Levy correctly, he
he feels that sociological
sociological and
non-monetary factors have accelerated the rate of inflation
inflation and
and that
-185-
the monetary authorities have been dragged along -- forced to accoswnoaccommo-—
infladate with money growth the more fundamental factors producing inflation.
Even on
on this view, however,
however, Levy should be much
much more interested
Even
than he is in what he calls the simplistic monetarist explanation.
If the price level were very closely linked to
to the money
money stock,
stock,
with
with a very small margin
margin of
of error, then it
it would
would be
be absolutely
absolutely clear
clear
that non-monetary
non-monetary factors could work to
to increase
increase inflation only insofar
as they operated
operated quite directly
directly on the Federal Reserve.
The greater
greater
the amount of slack or imprecision in the money/price relation, the
more credible
credible Levy’s
Levy's argument
argument becomes.
becomes.
If
If the relation
relation is imprecise,
in the
the short run, there is much
much room for non—monetary
non-monetary facfacespecially in
tors to produce an acceleration in the rate
rate of inflation directly,
directly, and
and
for the Federal
Federal Reserve to be drawn into
into monetary expansion
expansion later by
pressures to
to sustain the ongoing inflation process.
process.
The very word "ac“ac-
commodation"
to an inflainflacommodation” has the flavor of the central bank responding to
tion that
that has already occurred
occurred in order to prevent longer—run
longer-run forces
from reversing the inflation through aa process
process involving unemployment.
showing changes in defense spending, governgovernLevy presents data showing
ment transfers,
transfers, and so
so forth.
But
But he presents
presents no evidence whatsoever
such as these
that even bears on the validity of his claim that factors such
are responsible for the inflation.
Surely time series
series evidence
evidence on
United States inflation relative to government spending
spending would be relerele-
vant.
Also, cross
cross section
section
evidence
evidence relating the rate
rate of inflation to
the size
size of the government budget or its
its rate of growth in different
countries would be relevant.
Does Levy dispute the common finding that
inflation follows rather than leads money growth, aa finding that seems
inconsistent with the accommodation argument?
-186-
I'm forced to make a few
Since Levy has presented no evidence, I’m
comments based on casual
casual empiricism
empiricism and aa priori plausibility.
plausibility.
cormients
One of
Levy’s
Levy's claims is that the erosion in
in the growth of real incomes has
has led
led
to seek higher wages and prices in an attempt to reworkers and firms to
relost income growth.
coup their lost
If this
this argument is true, why
why did wages
and prices fall sharply
sharply as people became poorer between
between 1929 and
and 1933?
Is
short—run Phillips curve -- which shows that wage inflation
Is the short-run
inflation
——
slows as people become poorer through unemployment -- consistent with
-—
Levy's proposition?
Levy’s
If growth in taxes has been aa major
major factor in rere-
ducing growth in disposable income, then why have we not seen more
more acac-
spending and
and taxes rather than the activity
tivity to reduce government spending
claimed
claimed by
by Levy to raise nominal
nominal wages and
and prices?
II may be
be wrong, but
but
Proposition
at the federal level
seem awfully weak to
to me
me at
level,
Proposition 13 pressures seem
and in
in any event
event seem to be a lagged
lagged result of
of the inflation
inflation process
rather than part of aa process that can explain
explain the inflation.
What
What other
other evidence beside nominal wage and price increases can
we look at?
What about strike activity, or union membership, or concon-
centration
centration ratios in industry?
industry?
All
All of
of these would
would seem to
to have some
some
possible
possible connection to inflation, or at least as symptoms of
of the procprocess
ess Levy is talking about.
My impression is that these factors all cut
in the wrong direction
direction in the United States.
Most fundamentally, how
can real factors, other
other than through productivity effects and effects
on Federal Reserve behavior, have anything to do with nominal magnimagnitudes?
Levy seems to recognize
recognize the importance of
of explaining Federal
Federal
seventeen deals at some length
Reserve behavior; his footnote on page seventeen
-187-
with the question of the role of budget deficits in explaining Federal
Reserve money creation.
I believe that aa number of factors, some
some of which
which are closely
connected to the ones
ones Levy has emphasized,
emphasized, should be examined in terms
connected
of their effect
effect on Federal
Federal Reserve behavior.
My list of important
important
items is this:
l.
1.
Since the mid—sixties
mid-sixties there have
have been consistent underestimates
underestimates of
of
the natural rate of unemployment by
by the Federal Reserve and by
by the
economics
economics profession.
These
These underestimates
underestimates have led
led to money
money
growth that on
on average has been
been too
too high, even accepting
accepting the
the view
that monetary policy should aim for an unemployment rate close to
the natural
natural rate.
2.
There
There has been aa great over-emphasis on nominal interest
interest rates and
doesn’t really matter.
aa view that short-run money
money growth doesn't
Although
Although
the Federal Reserve has
has long recognized the importance of long-run
long—run
money growth,
operating in
growth, it
it always seems to be operating
in aa series of short
short
runs that never
never add up
up to aa long run.
3.
The Federal
Federal Reserve has
has from time to time made political miscalcumiscalcuThe
lations based on a combination of overly optimistic forecasts of
of
the effectiveness of fiscal actions and overly optimistic
optimistic forecasts
of when fiscal actions would occur.
Probably the
the best example of
this point is the Fed’s
Fed's delay in
in tightening money in
in 1967
1967 while
while
this
pass aa tax increase.
waiting for Congress to pass
4.
The Federal Reserve’s
short.
Reserve's policy
policy horizon
horizon has been
been too short.
OrdinariOrdinari-
ly,
ly, the Fed
Fed looks
looks ahead long
long enough to see significant
significant impacts of
monetary policy on employment and
and output
output but
but not
not long enough to see
any important impact on prices.
-188-
5. The Federal Reserve has used aa poor control
control mechanism based on the
federal funds rate.
has produced
produced aa procyclical
procycl ical monmonThis mechanism has
etary policy because it
it makes persistent procyclical mistakes so
easy.
6.
6.
The Federal Reserve is obviously
obviously responsive
responsive to political
po 1it i ca 1 pressures,
especially from the administration.
administration.
These political
political considerations
considerations
nay
may have reflected concern, from time to time, over
over reelection
reelection of
of aa
President
and over
over reappointment
President and
reappointment of
of aa Federal
Federal Reserve
Reserve Board
Board ChairChairman.
JI continue
continue to
to believe
believe that Federal Reserve behavior
behavior is
is not
not at
all aa simple function of broad societal trends.
Accidents of hishisAccidents
tory
such as
as assassinations
happen and
tory such
assassinations do
do happen
and are
are important.
important.
While
While II
certainly
rule out.
out~the importance of research on general
certainly would not rule
principles of political behavior, II still
still feel
feel that neglecting the
interplay of personalities
personalities and events
events is aa mistake.
In an endeavor
of this type, traditional
traditional historical analysis can provide very
very subsubstantial
stantial insights.
In summary, II believe that Levy
interpretaLevy provides aa misleading interpretation
tion of what monetarism
monetarism is all about.
Monetarism involves the economeconom-
ics
ics of
of the relations between money, output, prices, and interest rates,
and
and the economic processes
processes responsible for these relations.
It
It does
not
not pretend
pretend to
to offer
offer an economic
economic explanation
explanation of
of money
money growth
growth and should
should
not, therefore, be criticized
criticized for not doing so.
Now that monetarist
monetarist propositions
propositions -- at
at least in
in their
their long—run
long-run
-—
form -- are so
so widely accepted, it
it clearly makes sense to move
move on to
——
issues concerning why
why the
the monetary
monetary authorities
authorities behave
behave the way
way they
they do.
do.
Levy has offered aa number of interesting
interesting hypotheses on
on this
this question,
but has not provided any evidence.
To
To my
my taste,
taste, his approach
approach is
is less
-189-189-
productive than it might be because he pays so little
little attention to
to the
monetary authority
authority itself.
itself.
Surely the Federal Reserve
Reserve should be
be the
focal point of
of the political and sociological analysis.
caretaker function.
functinn.
far more than a caretaker
The Fed
Fed has
has
If
If the factors Levy discusses are
important,
important, we need to
to know how they impinge
impinge on
on the Federal Reserve in
order to have much
much confidence
confidence in
in the
the argument.
Now let
to the
Allan Meltzer.
Meltzer.
Now
let ne
me turn
turn to
the paper
paper by
by Allan
II will
will start
start with
with
an outline of his
his argument
argument as II understand it.
it.
First,
affect output.
Fi
rs t, Meltzer
Meltzer believes
believes that
that expectational
expectational errors
errors affect
output.
The
The expectational errors
err"'ors that he stressea
stressei are
are those between the normal
normal,
or
permanent plus
plus the transitory components.
components.
or permanent
He mentions in passing
He
that this
of Lucas.
this view is different
different from that of
While it is true that
Lucas
Lucas uses a spatial rather
.-a_ther than aa temporal model, II think that it
it
really comes
cones to much the sane
same thing.
Additional output can be
be obtained
obtained
in the Lucas model
rnode1 only if labor is willing to substitute hours inter—
inter-
temporarily.
In any event, the Meltzer
Meltzer view is that when prices are viewed as
temporarily high
high the level of output is expanded,
expanded, and when
when prices are
temporarily low the level of output is contracted.
Actually, it
it is
probably
probably better
better for me to state Meltzer’s
Meltzer's proposition aa bit differentdifferently: the permanent level of prices depends
depends on the permanent level of the
money
stock and
and it is
is deviations
the actual
from the
money stock
deviations of
of the
actual money
money stock
stock from
the
level that are most clearly
clearly related to deviations of
of output
permanent level
from normal full employment output.
Since deviations of output from potential output are related to
errors, it is important to investigate
investigate the formation of
of
expectational errors,
these expectations.
To
To illustrate the basic idea, Meltzer uses aa
-190-
simple model
model from the statistics literature in which
which aa time series has
has
known properties consisting of permanent
permanent and transitory variations.
inference problem is to use the past data to make the best guess as
The inference
component in the next period.
period.
to the
the permanent
permanent component
to
to the
The solution to
problem requires knowledge
knowledge of the permanent
permanent and transitory variances.
Given that information,
information, the next—period
next-period forecast depends
depends on
on aa distribdistributed lag of the past observations of
of the series,
series, with
with the distributed
distributed
lag weights depending on the pernanent
permanent and transitory variances.
variances.
lag
This
be generalized
generalized easily -- although the technical problems
basic idea can be
-—
may not be solved easily
easily -- by considering more complicated time series
series
--
models including multivariate frameworks.
frameworks.
However,
However, the basic
basic idea
idea
comes through quite clearly in the univariate model analyzed
analyzed by
Meltzer.
His tables 11 and
and 22 provide the flavor of how the means
His
and permanent and
and transitory variances might
might be
be extracted
extracted from the data
for different periods.
periods.
Now let me make
make an
an important distinction that does not seem
seem very
clear
clear in
in Meltzer’s
Meltzer's paper.
When we
we examine aa policy of gradualism there
are two analytically
considerations.
analytically distinct
distinct considerations.
One concerns the time
series of agents’
agents' forecasts of permanent
permanent values and the magnitudes of
expectational
expectational errors under the
the assumption that agents’
agents' estimates of the
permanent
permanent and transitory variances remain fixed and given an assumed
money growth path.
Here,
Here, it
it is clear that if money growth slows sharpsharp-
ly,
ly, then
then the market will interpret the initial
initial slow-down
slow-down as being
largely transitory; if the slow-down is
is in
in fact permanent, then there
will be aa large and persistent
persistent expectational error.
sumptions, the case for gradualism is compelling.
-191-
Under these asasOnly with
with aa gradual
gradual
decline in money growth would It
it be possible to avoid large expectaexpectational errors and the accompanying losses in
in output.
An entirely
entirely separate issue -- and one that II think is
is at
at the
--
heart of the problem -- concerns the way in
in which
which agents
agents change
change their
their
-—
estimates
estimates of the permanent and
and transitory variances
variances over time.
Meltzer’s
Meltzer's discussion is much less helpful on
on this
this issue.
issue.
If the
Federal
Federal Reserve could convince
convince agents that the money
money growth process had
changed and could convince agents that it would slow money growth
sharply, then
then forecasts of the permanent
permanent money stock would not be dedeon past observations.
observations.
termined by the old distributed lag on
Under these
these
slow money growth abruptly without producing
assumptions, the Fed
Fed could slow
producing
expectational errors and there
there would be no
no case
case for gradualism.
gradualism.
Meltzer has not offered
offered any formal analysis of how
how agents learn
learn
permanent and transitransifrom experience to change their estimates of the permanent
tory variances.
Nor has Meltzer offered an analysis of
of how agents
might
might be led to change their estimates of these variances by the Fed
introducing
introducing aa new policy,
policy, aa process which would
would not require any
learning from past money stock observations at all.
My
My comment on
on this
this
to reflect
reflect aa criticism
criticism of Meltzer’s
Meltzer's paper;
paper; II do
do not
point is not meant to
know
know of any interesting models of learning and II do not have the fogfoggiest idea of how to
to go about modeling this process.
process.
My point is simsim-
ply that it
it is important
important to separate the issue of
of calculating permanent
permanent
values qj~y~p
of how agents
given estimates of the variances from the issue of
form new estimates
estimates of these variances
variances over
over time.
The only constructive thought II can offer is that prescriptions
the best path for the money stock in the future might
might be
be based in
as to the
part
part on
on an analysis of
of the effects
effects of
of reducing transitory variance.
variance.
-192—
-192-
Money growth has been high
high in
in the recent past; if the actual rate of
Money
money growth is brought down only slowly from this
this high initial
starting point and if
is compressed by making
if the transitory variance is
this
slow—down smooth and in
this slow-down
in accordance with announced intentions,
then it is possible that the initial effects would actually
actually be to
to raise
rctise
agents'estimates
of the permanent rate of money growth for the next few
agents’estimates of
periods.
This result would occur
This
occur if aa significant
significant part of the
the recent
high money growth had been
been regarded by
by agents
agents as transitory and
and theretherefore had not been
been built into their estimates
estimates of the permanent part of
money growth.
The likelihood of the perverse result could be inVestiinvesti-
on money
gated by
by examining
examining the effect of aa reduced
reduced transitory
transitory variance
variance on
growth
growth expectations
expectations for next year in a time series model applied
applied to
to
money growth over the past
past few years.
actual money
actual
II have two final comments.
First,
First, as
as John Taylor has emphasized,
there is considerable uncertainty about the relative
relative validity of
of purely
expectational
expectational theories of the business cycle and
and theories
theories that
that stress
lagged adjustment due to contracts and similar
similar types of institutions.
institutions.
As Meltzer has noted but not emphasized, the case for gradual reduction
As
of
of money growth is considerably strengthened by this uncertainty
uncertainty bebe-
cause insofar as the
the contract
contract view has validity, aa sharp reduction in
money
money growth -- even if
if fully anticipated
anticipated -- would
would produce
produce a sharp dede—-
--
cline in output.
Secondly, although
although we have concentrated on economic
economic factors, II
think it is worth mentioning political processes.
It
It is
is not obvious
obvious to
me that maintainance
maintainance over aa long period of time of aa gradual reduction
of the money
money stock is politically
politically feasible.
It is certainly
It
certainly conceivconceiv-
able
able that aa quick
quick and
and dirty reduction of
of money growth, accepting
accepting the
-193-193—
output effects that would occur, is the only solution that is
severe output
politically
politically viable.
II am not
not sure whether
whether or
or not II believe that a
quick purging of inflation
inflation would be better politically,
politically, and even if I
did know what II believe II would not have
have any idea of why II believed
believed it.
it.
Nevertheless, this issue is surely important for aa full policy analysis
of winding
winding down inflation.
inflation.
An economic analysis of the minimum cost
method of reducing inflation is obviously
obviously important, but unfortunately
cost—benefit calculation that
it is not at
at all clear that the cost-benefit
that governs
governs
the political process is
is very closely connected
connected to the economic
economic costs
and benefits, however firmly we
we may
may establish them.
-194-
DISCUSSION OF
OF THE LEVY AND
AND MELTZER
MELTZER PAPERS
Albert E. Burger
What did the experience of
of the last
last half of
of the
the l96Os
1960s and
and the
the
l970s teach us about the effects of monetary
decade of the 1970s
monetary policy
policy
actions?
It did not teach us anything "new."
new.’
It
only gave
It only
gave us
us another
another
long-standing proposition
proposition
set of empirical observations to support the long—standing
11 generate
an acceleraexcessive growth of money wi
that aa maintained excessive
will
generate an
acceleration
tion in inflation and
and will
will raise
raise inflationary expectations.
The policy
price stability
stability in
actions that engineered the move from
from price
in the
the first
first half
half
of the 1960s
to aa 66 percent rate
rate of maintained inflation by 1973 were
were
of
l96Os to
an accelerated rate of purchase of government securities by the Federal
Federal
Reserve which resulted in
in aa faster growth of monetary base and bank
reserves and, hence, aa rise in the trend growth of money from 1—2
l-2 perpercent
cent to
to 66 percent.
nid-l96Os there already existed
Prior to the mid-l960s
existed aa very
very large amount
of evidence
evidence that this would be the expected
expected result of
of these types of
of
policy actions.
Indeed, one does not have to use highly sophisticated
sophisticated
methods of analysis to come to this
this conclusion.
conclusion.
Simply
Simply aa close look at
the
the data
data should be
be enough to convince most
most people of
of this strong rere-
between the growth of money and inflation.
lationship between
The experience since 1973
1973 has reminded us
us that price theory can
be useful in analyzing macroeconomic
macroeconomic developments.
Severe supply
Albert E. Burger is Assistant Vice-President
Vice—President and Economist,
Economist, Federal
Federal
Reserve Bank of St. Louis.
—195—
-195-
shocks raise the level of prices and, hence,
hence, contribute
contribute to the measured
rate
rate of inflation.
1975—1976 illustrate, these effects
However, as
as 1975-1976
effects do
not result in sustained inflation.
Both Mel
Meltzer
and Levy point
point out that sustained
sustained inflation
inflation is a
Both
tzer and
phenomenon.
monetary phenomenon.
They differ with
with respect to whether monetary
monetary
actions are the “fundamental”
"fundamental" cause of iinflation.
nfl at ion.
Meltzer puts the
blame
of the
blame for inflation and its acceleration
acceleration directly
directly at the door of
Federal Reserve.
Reserve.
He rejects the assertion that the Vietnam
War, deVietnam War,
de-
ficits, and government spending of the mid—l96Os
mid-1960s were
were the origin
origin of
of ininflation or were the motivating force causing the Fed to
to expand
expand money.
II agree with Meltzer that the Fed
Fed must accept the blame for starting
and maintaining inflation.
The
The money stock grew at steadily more rapid
rates because the Fed allowed
allowed it to do so by providing
providing the necessary
bank reserves.
If the Fed had not supplied more reserves, money growth
would
would not have accelerated
accelerated and, hence, inflation
inflation would
would not have
have accelaccel-
erated.
The Fed can make excuses
excuses about why
why it followed such a policy,
but the fact remains that it did follow such
such aa policy.
of why policy moved from one
one
Levy raises the interesting question of
inthat underwrote price stability
stability to one that underwrote
underwrote accelerating
accelerating inflation.
His
His conclusion is that, in
in the mid—l96Os,
mid-l960s, there were
were major
political and social changes
changes that led
led to greater social
social activism on
on the
part of the government (such as aa shift toward increased “nonproductive”
unonproductiven
transfer payments
payments and
and regulation) that
that reduced productivity
productivity and
and
set off the inflationary spiral.
I would interpret
interpret his conclusion as
as
meaning the Federal
Federal Reserve was caught up in
in this process and
and essenessentially pulled
pulled along the path
path it followed by forces over which it
it had no
control.
-196-196—
idea that the
There is a growing body of evidence supporting the Idea
factors Levy discusses operated
operated to lower potential real
real output
output growth.
However, if
moneif these factors had not been
been accompanied by aa surge in
in monetary
is considerable
considerable doubt
have had
tary expansion,
expansion, there
there is
doubt we
we would
would have
had the
the
acceleration in inflation that we experienced.
leaves open the question of
of why, despite repeated
This still
This
still leaves
statements
inflation, the Federal Reserve
statements of policy intent to halt inflation,
allowed its policy actions
actions to feed inflation.
If the Fed had actually
actually
planned an acceleration in
in inflation,
inflation, it
it could not have followed aa
that was better grounded in theory and supported by
by empirical
program that
evidence.
I have difficulty
difficulty accepting the explanation that the Federal
Federal
was simply pulled along by the tide of expansionary sentiment.
Reserve was
To some extent, that
that may have been the case.
Especially, one can point
to the repeated failure of certain
certain members of Congress
Congress to
to accept the
interest rate consequences of their deficit spending.
However, the
basic cause of the high and rising interest rates that have charactercharacter-
ized the last
last 15 years has been
been the
the inflation generated
generated by
by Federal
Reserve actions and the resulting rise in inflationary expectations.
I would
would ascribe the failure of monetary policy to achieve its obob-
jective of stable overall prices to aa failure to accept and remain comcommitted
mitted to a few very basic principles.
These are:
(1) the primary job
of aa central bank is to
to prevent
prevent an acceleration in
in the basic rate of
inflation and monetary policy cannot fine tune real output; (2) excessexcess-
ive money growth means an acceleration in inflation; (3) money grows at
sustained, faster rate only when the central bank
moneaa sustained,
bank provides more
more mone-
tary base; (4) if there is aa surge in government demand for credit or
private demands
for credit
surge in
private
demands for
credit or
or aa surge
in measured
measured inflation,
inflation, short-term
short-term
—197—
-197-
interest rates will rise and Federal Reserve attempts to prevent this
rise
only ensure
ensure that
at these
rise will
will only
that interest
interest rates
rates remain
remain at
these higher
higher
levels;
levels; (5) the Federal Reserve can control
control the trend growth
growth of
of money;
and (6) although in theory, money growth can be controlled by operating
on the federal funds rate, in
in practice this is aa very unsatisfactory
procedure.
procedure.
If the Federal
Federal Reserve had remained committed
committed to
to these six
basic
principles, it
policy would
basic principles,
it seems
seems very
very unlikely
unlikely that
that monetary
monetary policy
would
have followed the path that characterized
characterized the last 15 years.
Of the above
above six principles, the last •two
two have been the
the hardest
for the Federal Reserve to accept:
flaws in aa federal funds target.
ability to control
control money
money and the
More
More than anything else, these
these two
items have contributed to the failure to achieve
achieve policy objectives.
objectives.
Too
“can the Fed
Too often the question of "can
Fed control money?”
money?" has
has gotten mixed
up with the question of "should
money?"
up
“should the Fed control money?”
central
If the central
bank can control the growth of the monetary base, it can control
control the
supply of money.
This should be
be aa lesson that
that is learned in an
an introintro-
ductory money and banking
banking course.
During the past
past 15 years the Federal
Reserve has tried to control the federal funds rate, not growth
growth of
monetary
monetary base and bank
bank reserves.
reserves.
Hence, the Federal
Federal Reserve has
has not
“control
led” money.
controlled
11
11
This is
is why the most important
important aspect of
of the policy actions
actions
announced by
the Feds
6, 1979,
announced
by the
Fed's Open
Open Market
Market Committee
Committee on
on October
October 6,
1979, was
was
the part announcing aa change in operating
operating procedures.
procedures.
Primary emphasis
was shifted from the federal funds rate to growth of aa reserve aggreaggregate.
gate.
If the Federal Reserve remains committed
committed to
to this
this change, monemone-
tary actions
actions may
may start
start to
monetary policy.
tary
to match
match the
the intent
intent of
of monetary
policy.
198-198-
-
we got into our current predica—
predicaIt is much easier to analyze how we
ment than it is to
to state how to
to get out of
of it.
it.
Obviously, to lower the
trend rate of
of inflation,
inflation, the growth rates of the monetary base and
money
money must be
be reduced.
However, the objective of monetary policy
policy is
so with
not just to
to slow inflation, but to do so
with aa minimum loss
loss of
of real
output.
As other
other papers at this conference have emphasized, there is
is aa
“slowing”
great deal of
of uncertainty about the effects of alternative "slowing"
policies
policies on real output and employment as well as their short—tern
short-term
effects on
on the financial markets.
Traditional macroeconomic
macroeconomic models
models
usually assign a fairly large and prolonged
prolonged real output effect
effect to anti-
inflationary monetary policy.
However, as
as Taylor points
points out in his
paper, recent developments in
in economic theory raise serious questions
about implications of traditional models.
Despite our
of the effects
our uncertainty about the exact magnitude of
on real output, it is becoming generally
generally accepted that the
the less the dedethe less effect
gree of uncertainty about the path of monetary actions the
these actions will have on real output and the larger and quicker their
effect on
on inflation.
Meltzer discusses this
this issue under the heading of
of
11
the “basic
basic inference problem.”
problem. 11
He shows that,
that, if transitory changes in
the
the growth of
of money
money are frequent, it
it is optimal to observe aa relatively
long
long series of observations before
before concluding that aa permanent change
has
has occurred.
The
The past behavior of the Federal Reserve with respect
respect to
the growth
money has
has made this aa good rule to follow.
growth of money
The Federal
Reserve has announced monetary
monetary targets
targets and then repeatedly
repeatedly failed to
to
hit these targets.
The Federal
Federal Reserve
Reserve has announced
announced major policy
policy
actions
designed to
money growth,
as it
November 1978,
1978, and
actions designed
to slow
slow money
growth, as
it did
did in
in November
and
then actually substantially reduced money growth for five months.
-199—199—
!.~i
transitory change in money growth,
However, this was apparently only aa transitory
as the last six months have completely reversed the pattern of slow
money growth.
lesson that the Federal Reserve
Reserve has learned
learned is
is
Hopefully, one lesson
that it
it must make its policy
policy announcements credible to the public.
Credible means taking actions, and maintaining
maintaining those
those actions that are
consistent with its stated policy intent,
intent.
Also, when the Federal
Federal
major change in
in its method of implementing policy, it
Reserve makes aa major
should
should clearly
clearly explain this
this new procedure.
procedure.
The immediate case
case in point
is the October
October 66 announcement of aa move toward aa reserve
reserve targeting
targeting
procedure.
To minimize disturbances in financial markets and to have a
on inflationary expectations, the Federal
Federal Reserve should
maximum effect on
clearly
clearly explain the new rules of the
the game.
game.
How much
much more short-rum
short-run
flexibility does the Fed
Fed plan
plan to
to allow
allow in
in the federal
federal funds rate?
Exactly which reserve aggregate is going to
to be the new target? What is
the Federal Reserve's
Reserves growth target for this
this reserve aggregate?
How
How is
the Federal Reserve going to project the
the relationship between the
reserve aggregate and money?
An improved monetary policy
policy for the 1980s
An
must include answers to
to these questions.
questions.
-200-
FLEXIBLE EXCHANGE RATES IN THE l970s
1970s
Jacob A. Frenkel
INTRODUCTION
Our
Our recent experience
experience with
with aa system
system of flexible exchange rates
iad led
1ad
led to
to aa renewed interest in the
the operations of foreign exchange
exchange
narkets and
and in
in studying the
the principal
principal determinants of
of exchange rates.
rhe 1970s
l97Os witnessed
evolution of the international
fhe
witnessed the dramatic evolution
international monemone-
tary system from aa regime of pegged exchange
exchange rates which
which prevailed
prevailed for
about aa quarter of a century since
since the Bretton Woods conference into
into aa
regime of flexible (though managed) rates.
The emergence
emergence of the new
legal
legal and economic system confronted
confronted traders, national
national governments
governments and
and
international
international organizations
organizations with new economic problems, choices and
instruments.
instruments.
During the 1970s
1970s exchange rates have fluctuated widely
widely
and inflation rates accelerated.
The international monetary
monetary system had
had
to accommodate extraordinarily
extraordinarily large oil related shocks
shocks which affected
trade flows in goods and assets.
Huge oil payments had to be recycled.
Uncertainties concerning future developments in international
international politics
reached new heights and the prospects
prospects for the world economy got
gloomier.
on
These developments have placed unprecedented pressures on
the markets for foreign exchange as well as on other
other asset markets.
Or. Frenkel is Professor
Professor of Economics
Economics at the University of
of Chicago and
Dr.
Research Associate at the National
National Bureau of Economic Research.
Research. The
author is indebted to
to Lauren Feinstone
Feinstone for efficient research assisassisNational Science Foundation for financial support.
support. In
In
tance and to the National
revising the paper he would like
having benefited from
like to
to acknowledge
acknowledge havinq
useful comments by Sebastian Edwards,
Edwards, Stanley
Stanley Fischer, Craig S.
S. Hakkio,
Paul
Nussa and
Paul Krugman,
Krugman, Nichael
Michael L.
L. Mussa
and Nasser
Nasser Saidi.
Saidi.
-201-201-
They have been associated
associated with aa large slide in
in the value of the U.S.
dollar,
dollar, and have resulted in
in speeding up
up the creation of new instituinstitu-
tions like the European Monetary System which
which provides the formal
framework for the management of
of exchange
exchange rates
rates among
among members.
The inin-
creased interdependence among countries and the
the recognition that exex-
change rate policies by one national government
government exert influence
influence on
other economies have also induced aa legal
legal response from international
organizations.
For example, in
in late April
April 1977,
1977, the Executive Board of
the International Monetary
Monetary Fund
Fund approved the details of
of the second
amendment
amendment to Article IV
IV of the
the amended Articles
Articles of
of Agreement
Agreement dealing
procedures for surveillance
surveillance of member
member countries'
with the principles and procedures
countries’
exchange rate policies.
These developments provide the background for this paper which is
intended to present aa brief survey of key issues and
and lessons from the
197Os.
experience with floating rates during the 1970s.
The main orientation
orientation
of the paper is
is empirical and the analysis is based on the experience
of three exchange
exchange rates:
the
Dollar/DM.
the Dollar/OM.
the Dollar/Pound,
Do 11 ar /Pound, the Dollar/French
Do 11 ar / French Franc and
In the second
second section II analyze the efficiency of
of the
foreign exchange markets by examining the relationship between spot and
context I also examine
examine and
and interpret
forward exchange rates; in that context
the extent of exchange
exchange rate
rate volatility.
volatility.
The
The analysis of
of the foreign
exchange
quesexchange markets
markets is important
important because it sheds light on
on several questions like: l)
1) have exchange rates fluctuated "excessively?"
‘excessively?
2) is
there evidence
evidence that speculation in
in the
the foreign exchange markets is
is dedestabilizing?
3) is there evidence
evidence that there is “insufficient”
insufficient specuspecu11
lation
lation in the foreign exchange markets?
4) is there evidence for a
market
are unexploited profit
market failure in the sense that there are
—202-202-
11
opportunities? These Issues
issues are relevant for assessing the performance
opportunities?
of floating rates as well as for discussing whether there is aa case for
government intervention in the foreign exchange markets.
The analytical
framework that is used for interpreting
interoretinq the
the volatility
volatility of exchange
exchange
rates and the association between spot amd
and forward rates is the modern
modern
theory of
of exchange rate determination.
Within this perspective
perspective exex-
change rates are viewed as the prices of assets that are traded
traded in
organized markets and, like the prices
prices of
of other assets, are
are strongly
influenced by expectations about future events.
interest rates is
The relationship between exchange rates and interest
analyzed
analyzed in
in the third section from the perspective
perspective of
of the monetary
monetary
approach to the exchange
exchange rate.
rate.
This
releThis analysis is of
of particular rele-
vance in view of
of the new policies of the Federal
Federal Reserve Board,
Board, which
which
were announced on
1979, that are
on October 6, 1979,
are intended
intended to curb inflation
and to support the
the dollar.
One of the key issues
issues that is
is raised in
in
this section
section is the distinction between anticipated
anticipated and unanticipated
changes in
in rates
rates of
of interest.
interest.
tion is obvious.
The
The policy implication of this distincdistinc-
As an
an analytical matter this
this distinction
distinction is important
because the modern approach to exchange rate
rate determination implies that
that
exchange rates
rates are strongly influenced by
unpredicted.
11
news
which by definition is
news’11 which
Therefore, unanticipated
unanticipated rather
rather than anticipated
anticipated changes
changes
in interest
on changes in
interest rates should have a strong effect on
in exchange
rates.
This
This prediction is tested empirically.
The fourth section analyzes the
the relationship
relationship between
between exchange
exchange
rates and prices by examining
examining the patterns
patterns of
of deviation from purchasing
power parities.
parities.
This
This examination
examination is relevant
relevant for assessing
assessing whether the
flexible exchange
exchange rate
rate system was successful in insulating
insulating national
-203—
-203-
economies from foreign shocks,
shocks, and whether It
it provided policyinakers
policymakers
an added instrument for the conduct of macroeconomic
macroeconomic policy.
with an
The
purchasing power
evidence on
on deviations
deviations from purchasing
power parities is also relevant
for the discussion of
of whether there is aa case
case for managed float.
The
fifth section concludes
concludes the paper with some concluding remarks.
THE EFFICIENCY
THE FOREIGN
MARKET
THE
EFFICIENCY OF
OF THE
FOREIGN EXCHANGE
EXCHANGE MARKET
AND
THE MOVEMENT
ANO THE
MOVEMENT OF
OF EXCHANGE
EXCHANGE RATES
RATES
In this section I analyze the principal characteristics of the
relationship
spot and
relationship between
between spot
and forward
forward exchange
exchange rates
rates which
which seem
seem to
to
l97Os.
emerge from the
the experience of the 1970s.
Following an analysis of
of the
the
efficiency of the foreign exchange
exchange market
market I discuss the more
more general
general
issues underlying the relationships between spot
spot and forward rates and
their volatility.
cy of the Foreign
Forein Exch
an e Ma
rket
The Efficiency
Exchange
Market
One of the central insights of the monetary
monetary (or the asset
asset market)
exchange rate is the notion that the exchange rate,
rate,
approach to the exchange
being aa relative
relative price of two assets, is
is determined in aa manner similar
to the determination
determination of other
other asset prices and that expectations
expectations conconcurcerning future course of events play aa central role in affecting cur1
rent exchange
exchange rates)
rates.
exchange market is efficient
efficient and if the exchange
exchange
If the foreign exchange
rate is
is determined in a fashion similar
similar to
to the determination
determination of other
asset prices, we
we should expect current prices to reflect all currently
currently
1‘For
For collections of articles
articles summarizing
summar1z1ng this
this approach see the
Scandinavian ,Journal of. E.c.orulmii;_,_, no.
no. 2, 1976, and Frenkel and Johnson
Johnson
(1978).
-204-
available information.
Expectations concerning future exchange rates
Expectations
should be
be incorporated
incorporated and reflected in forward exchange
exchange rates.
Thus,
of the market, II first regress the logarithm
to examine the efficiency of
of the current
current spot exchange rate, an
tn St, on the logarithm
logarithm of
of the oneoneof
month forward exchange rate
Fti
rate prevailing at the previous month, an
£n Ft-l,
as in
in equation (l).2
(1). 2
(1)
an
=
a
+
b an Fti
+
If the market
market for foreign exchange is efficient and if the forforward exchange
exchange rate is an unbiased forecast of the future spot
spot exchange
exchange
rate, then
then we expect that:
l) the constant term in equation
equation (1) should
1)
not differ
differ significantly from zero,33 2) the slope coefficient
coefficient should
not differ
differ significantly
significantly from unity and, 3)
3) the residuals
residuals should be
serially
serially uncorrelated.
II examine three
three exchange rates:
Pound, the
the Dollar/Franc
Dollar/Franc and the Dollar/OM.
Pound,
Dollar/DM.
the Dollar!
Dollar/
Equation (1)
(1) was estimated
Equation
estimated
period June 1973
1973 —- July
July 1979.
1979.
using monthly data for the period
The
The beginning
beginning
of the
the period was
was determined by the attempt to concentrate on
exon the experience of
of the current exchange
exchange rate regime (following the initial
post Bretton-Woods
Bretton-Woods transition period).
period).
The resulting ordinary leastleast-
22For an
efFor
application of the same methodology
methodology in analyzing
analyzing the efficiency properties of the foreign exchange market
market during the German
1921—1923 see Frenkel
hyperinflation of
of 1921-1923
Frenkel (1976,
(1976, 1977,
1977, 1979). For an
application to other exchange rates during the
the 1920s,
Frenkel and
l920s, see Fremkel
Clements
l92Os and the 1970s,
l970s, see
Clements (1980), for an application to the 1920s
Krugman (1977); for an interesting analysis using time—series
time-series and crosssection data,
data, see Bilson (1979), for an
an analysis of market
market efficiency
using novel econometric techniques, see Hakkio (1979a), and Hansen and
Hodrick (1980),
1979)..
( 1980), and for surveys, see Levich
Levi ch (1978,
( 1978, 1979)
3More precisely, if
if (assuming risk neutrality) the forward rate
spot rate
measures the expected value of the future soot
rate, then
then the constant
constant
0~5 2
term in the logarithmic equation (l)
(1)should
°~~ see
should be
be --0.5ag;
see Frenkel (1979).
-205-
squares estimates are reported in Table 1.
l. Also reported in Table 1
are additional regressions which will be
be analyzed shortly.
are
As may
may be
be
seen for the Dollar/DM
Dollar/OM exchange rate, the hypotheses that (at the 95
confidence level) the constant term
term does not differ signifisignifipercent confidence
cantly from zero and that the slope coefficient does not differ signifisignificantly from unity cannot be rejected.
These hypotheses are
are rejected
for the Dollar/Franc
Dollar/Franc exchange rate and are rejected
rejected (marginally) for the
Dollar/Pound exchange rate.
Dollar/Pound
rate.
The joint
joint hypotheses, however, that the
reconstant is zero and that the slope
slope coefficient is unity cannot be rejected at
at the 95 percent for the Dollar/Pound and the Dollar/OM
Dollar/DM exexchange rates and at
at the
the 99 percent for the
the Dollar/Franc
Dollar/Franc exchange
exchange rate.
The test
test statistics for testing the joint
joint hypotheses are reported in
in Table 1.
the column headed by
by FFin
l.
an efficient
efficient market, expectations
It was argued above that in an
concerning future exchange
exchange rates are
are reflected
reflected in
in forward rates, and
that
that spot
spot exchange rates reflect
reflect all
all currently available
available information.
If forward exchange rates prevailing at
at period t-l summarize
summarize all relerelevant information available at that period, they should also
also contain
the
the information that is summarized in data corresponding to period t-2.
It thus
thus follows that including additional lagged values of the forward
(1) should
rates in equation (l)
should not greatly affect the coefficients
coefficients of
determination
signifidetermination and should not yield
yield coefficients that
that differ significantly from zero.
The results
results reported in Table l1 are consistent with
an Ft
this hypothesis; in all cases the coefficients of in
Ft-Z
do not differ
2 do
of the additional
significantly from zero
zero and the inclusion of
additional lagged
variables does
does not
not improve
improve the
the fit.
fit.
Furthermore, in
in all
all cases
cases the
Durbin-Watson statistics are consistent with the hypothesis
hypothesis of the
—206—
-206-
Foreign Exchange Markets
Mllrket:s
Efficiency of Foreign
Monthly Data: June
June 1973
July 1979
19)3 —- July
1979
(standard errors
(standard
errors in parentheses)
parentheses)
Depend,cnt
Dependent
·variable:
Variable: tu
in S
St
L
Dollar/Pound
DoJlar/l'ound
Es
r tarn Lion
Estimation
l'k,thod
Method
Constant
constant
in tF,
t-1
th r—l
OLS
.033
(.017)
(.017)
.956
.956
(.
O24)
(.024)
OLS
OhS
.031
(.038)
(.018)
1.047
1.047
((.116)
.116)
tv
lV
.035
(.018)
(.018)
OLS
OLS
in F
F~
9.n
/F2
s.c. .
s.e
D,W.
0.8.
.96
.027
. 027
1.72
1.
72
.96
.027
.027
1.94
1. 94
..953
953
(.024)
(.024)
.96
.027
1.72
1. 72
—.237
-.237
( .078)
(.078)
.843
(. 051)
(.051)
..79
79
.029
2.23
2.23
OLS
OLS
-.225
—.225
(. 082)
(.082)
.706
(.117)
(.117)
..79
79
.029
1.90
IV
rv
—.219
-.219
(.079)
( .079)
.855
(.052)
(.052)
..79
79
.029
2.25
-.023
.023
(.027)
(.027)
.971
( .032)
(.032)
.93
.93
.032
2.12
OhS
OJ.S
-—.019
.019
( .028)
(.028)
.913
( .119)
(.119)
. 93
.91
.032
1.96
IV
rv
— .022
-.022
..972
972
(.033)
( .033)
.93
.93
.032
2.12
t-22
—.088
-.088
(.113)
(.113)
F
m
m
1.86
.90
Dollar/Franc
Dollar/Franc
'
N
0
~
'
.146
(.
117)
(.117)
4.83
2.73
2.
73
Do
tIn r/DM
DolL.ir/DM
OLS
OhS
—
(.028)
(.028)
No
to:
~,Tote:
.063
(.
122)
(.122)
..51
51
.02
2
s.c. is
the c:;taudard
standard error
ont of determination;
s.o.c.
i.s Ll1(c
error of the equation
i:.quat~on and F
R is the coeffici
coeffl~_ient
determination; in
in the case of
i.nslrurnvntal
was computed as l-Var(ut)/Var(£n
The FF statistic.
tests the
ins trustee taT variables
variables estimation
estimation the R.
R was
1—var (ut) /Var( In St)..
statistic rests
joint
restrict ion that the constant equals zero and the slope equals nil
ity.
statistic ts
joint r(.'strictiou
unlt:y.
The rest
test statistlc
i.s distributed
distributed
as F(2,
Critical
F(2,7l)
(95 percent)
percent).
The
instrumental
as
F(2, 71).
Criti.cal values for F(2,
71) are 3.13 (95
percent) and 4.92 (99
(99 percent).
The instrumental
variable
tisrntion method is used in order to allow
possibility of errors
variables arising
arising
variable; (iv)
(lV) es
estimation
allow for
for the possibility
errors in
In variables
from using
‘sing Kn
in F
F~
rate; the instruments
instruments are a constant
constant and
ourhin a
_ as a proxy
proxy for the expected future spot rate;
and Durhin's
1
2
T~en~statis
tic which tests for the
in variables
dist rihuted XX with
with
rank variable.
variablt~._ tThe
rn--statistic
the absence of errors in
variables is distributed
22 2
degrees of
of freedom.
critical value for X~(2)
x(
) is
22 dcgrel'.s
The critical
:is 5.99 (95 percent).
percent).
absence of first—order
first-order autocorrelated residuals and an examination of
higher order correlations (up to 12
12 lags) shows that no
no correlation of
of
any order is significant.
To
To further examine
examine the relationship
relationship between the various exchange
exchange
rates we
we note
note that one of the
the assumptions underlying equation (1) was
the notion that the forward exchange
exchange rate measures the unobservable
value
spot exchange rate.
value of
of the expected
expected future spot
This
This assumption propro-
vided the justification for using equation (1)
(1) instead of the more
fundamental relationship that is
is embodied
embodied in equation (2):
(2)
( 2)
an St
in
=
=
an(S~
a ++ bb rn
(S~ It - 1)
1) ++ st
-
where (S~
(S~ I t—l)
spot exchange
t-1) denotes the expected
expected spot
exchange rate for period t
available at period t - 1.
based on the information available
1.
-
If, however, the
If,
at tt - 11 is a "noisy"
forward exchange
exchange rate at
noisy’ proxy for the expected
-
future value of
of the spot rate, (i.e., it measures
measures it with a random
random
error) then we would obtain that
error)
(3)
an Ft1
=
am(S~
t
-
1)
+
vt,; E(vt)
=
0
and substituting equation (3) into equation
equation (2) yields:
(4)
an St
=
a
+
b an Ft1
+
(ct
-
bv~i).
In this case the
the error
error tern
term in equation (1) would be
be ut
ut
=
=
st - bvt-l'
—
and the assumption that the covariance between
between an
£n Ft,
Ft-l and lit
ut is
is zero
would entail aa specification
specification error,
error, and
and the application of the ordinary
ordinary
least-squares
(OLS) procedure
procedure would yield inconsistent
inconsistent estimates
estimates due to
least—sguares (OLS)
the classical errors in variables bias.
-208-
In order to examine the possibility that the OLS estimates might
be subject to
to the errors in variables bias, one needs to
to test the hyhypothesis that cov(ut.
cov(ut, In
in Ft_ 1 ) = 0. This test follows the specifispecifi4
(l978).~ To perform the test
cation test
test outlined by
by Hausman (1978).
equation (1) was estimated by applying the OLS procedure as well as by
using an instrumental variables (IV) estimation method.
Under
Under the
null—hypothesis of no misspecification the OLS coefficients vector bb0
null-hypothesis
0
is an efficient and an unbiased estimate of
of the true coefficient
vector.
Under the alternative hypothesis of misspecification
misspecification the vecvec-
.
.
tor bb0 is biased and an unbiased coefficient vector b can be obtained
0
1
estimation procedure. The test—
testby applying an instrumental variables estimation
statistic relevant for testing the null—hypothesis
null-hypothesis can be written as
( 5)
(5)
m
=
~l
-
b0) (var b1
-
var bY1(b
-
b0)
where var(b11)) and var(b
var(b00)) denote the variance-covariance
variance-covariance matrices of
.
.
is distributed
b1 and b0,, respectively.
respectively. Under the null-hypothesis mmis
(in large samples) as x22 with two degrees
degrees of freedom. Table
Table 11 reports
the results of estimating equation (1) by applying the instrumental
variables estimation method.
As may be seen for all exchange rates the
two vectors of coefficients bb11 and b are very close
close to each other.
0
For example, for the Dollar/Pound exchange rate the constants are .033
consequently, the resulting
and .035 and the slopes are .956 and .953, consequently,
m statistic is .90 which is well below 5.99 -- the critical value of
m
--
4This test was recently applied by Obstfeld (1978) to the analyanalysis of
l970s and by Frenkel
of the foreign exchange market during the 1970s
(1980a, 1980b) to the analysis of
of the foreign exchange markets
~arkets during
the 1920s.
—209—
-209-
2
(2) at
percent confidence
level.
xx2(2)
at the
the 95
95 percent
confidence level.
Themm statistics
correspondThe
statistics correspond-
ing to the other exchange
exchange rates are also below this
this critical
critical value.
value.
It
It
is concluded, therefore, that the use of
of the forward
forward exchange
exchange rate
rate as
as a
proxy for expectations does
does not introduce aa significant
significant errors in
in
variables bias
bias and thus the use of the OLS estimation procedure
procedure seems
appropriate.
The efficiency of the foreign exchange
exchange market and the
the rationality
rationality
to measure expectations
expectations can also
of using data from the forward market to
be analyzed from aa different angle.
Consider equation
equation (6):
n
(6)
xt
=
a
0
+
+
~ t +i=1
5~xt~+ ~t_~
Cll
I
+
w~
i=l
where
xt denotes the percentage change in the spot exchange
exchange rate
where x~denotes
(an 5~
~t-ldenotes
(in
St - an
zn St_ 1 ), rrt-l
denotes the forward premium on
on foreign exchange
exchange
(an
an St_
~tl~’
(in Ft
Ft-l
t denotes time, nn denotes
denotes the number of lags, and
1 ), ~
1 - in
-
~
-
ww denotes an error tern.
term.
If
~t-1summarizes
If rrt-l
summarizes all available
available information
concerning the future evolution
evolution of
of the exchange
exchange rate, then qiven
given the
the
value of
of the forward premium nt-l, the past history of the percentage
percentage
~
change of
of the exchange rate should not "help"
prediction (i.e., the
change
“help” the prediction
past history should not be viewed as Granger-causing
Granger-causing future changes),
and the joint
joint hypotheses
hypotheses that a 1 and Si are zero should not be rejected.
rejected.
The results of applying these tests to the three exchange rates for
various number of lags are reported in
in Table 2.
Also
Also reported
reported in
Table 22 are the
the results of testing the
the joint hypotheses that aa11 and Si
are zero and
and that y,
y, the coefficient
coefficient of the forward premium, is unity.
unity.
The
The relevant statistic for testing the null-hypothesis is an
an F—
F-
statistic which is reported in Table 2.
2.
As is evident in all cases
cases the
null-hypothesis cannot
cannot be rejected at
95 percent confidence level
null-hypothesis
at the 95
-210-
since the values of the various F—statistics
F-statistics fall well below the
the corcorsince
responding critical
critical values.
It is concluded, therefore, that the
the forward
premium on
on foreign exchange
exchange may be viewed as aa rational expectations
measure of
of the percentage depreciation of the currency in that it
it inincorporates
corporates the available information that
that is contained
contained in the
the series
of past depreciations.
that may
may be
be drawn from the previous
previous
The principal conclusions that
discussion are that the behavior
of the foreign
behavior of
foreiqn exchange market
market during
been broadly consistent
consistent with
with the general implications of
the 1970s
l970s has been
the efficient market
market hypothesis
hypothesis and that the forward exchange
exchange rate
summarizes the relevant available
available information concerning the future
evolution of
of the rate.
Exchang_<e_Rate
Movement: Volatility and Predictability
E xc han ~
In this
this section II analyze the volatility of exchange rates and
the
the extent to which this
this volatility is predictable.
To
To set the stage
quarterly perfor the
the analysis, II present
present in Figure l1 the daily and quarterly
per-
in the three exchange
exchange rates.
rates.
centage changes in
This figure indicates
exchange rates have been
been very volatile and that the
that the various exchange
degree of volatility of day-to-day changes in the exchange
exchange rates have
been extraordinarily
extraordinarily high
high and has been much smaller when averaged over
over
longer periods.
Further, the standard errors of the regressions in
indicate that the forecasts of future spot exchange rates
rates based
Table 1
1 indicate
on the forward rates are imprecise:
the standard errors of the equaequa-
tions are about 33 percent per month.
These characteristics of price changes are typical to auction and
to organized asset markets.
In such markets current prices
prices reflect
reflect
—211—
-211-
Table
Table 22
Test of Rationality
Prediction of currency
Rationality of Forward
Forward Premium
Premium Prediction
Currency Depreciation
Depreciation
1973 —- July
Monthly Data,
Data, June 1973
July 1979
1979
Dependent Variable
Dependent
5
an
S~-— an
t—l
in st
in st-l
Null Hypothesis
Null Hypothesis
Number of Lags
Number of Lags
F—statistic
3
F(4, 64) =• 1.680
1.680
F(4,64)
44
F(5,62)
P6,62)
5
F(6,60) • 1.231
fl6,6o)
1.231
66
F(7, 58)
P(7,58)
33
55
=
= 1.555
1.555
P(6,62)
F(6,62) =• 1.518
1.518
P0,60)
F(7 ,60) • 1.207
1.207
66
P(8,58)
F(8,58)
3
3
P0,64)
F(4,64)
44
P3,62)
F(5,62) =• 1.327
1.327
5
F(6,60) • 1.087
P(6,60)
1.087
P0,58)
1.014
F(7 ,58)
1.014
F-statistic
Dollar /Pound
Dollar/Pound
a = 0, 13 • 0
a 1 = 0~Bj1 — 0
1
44
= o, r = 1
0, ~ 1
0.1"" 0, e1
a = 0, ~
1
Dollar/Franc
Dollar/Franc
"1
o, ~i
0, $..
=
= 00
66
a"1
1
0, 3 . •
i
0, 5.
o.
y = 1
0, y
1
Dollar/DM
Dollar/OM
0, 8.0• 0
1 "' 0, Sia
o,
0, 6.
Si
=
0, y
1
=
1.131
1.131
1.175
1.175
P6,64)
F(5,64)
1.436
1.436
P3,62)
F(6,62)
1.519
1.519
5
66
5(7,60)
F(7,60)
1.146
1.146
P(8,58)
F(8,58)
1,063
1.063
3
P3,64)
44
P6,62)
F(5,62)
S5
66
P3,60)
F(6,60) == 1.321
1.321
5(7,58)
1.342
F(7 ,58) = 1.
342
3
P(5,64)
44
5
P3,62)
F(6,62)
P0,60)
66
5(8,58)
F(8,58)
5
—212—
-212-
P6,64)
F(5,64)
33
3
aa
"1
1.141
1.141
44
3
a
0.1
1.610
1.
610
F(4,64)
=
1.262
1.262
1.183
F(5, 64)
F(7 ,60)
1.123
1.123
1.183
=
=
=
1.287
1.287
1.403
1.403
1.525
1.525
Pigure
Figure 11
,HORT-RUN
VARIABILITY IN
EXCHANGE RATES
HORT-RUN VARIABILITY
IN EXCHANGE
RATES
APRIL
2,
1973
DECEMBER
APRIL 2, 1973- DECEMBER
fOuanerty pe,centage
percentage changeal
changes)
lOuarterly
Daily percentage changeel
(Daily
changes)
t
~
[U~iTED~INGDOM
UNITED KINGDOM
tE-
I~
~1 • 11i!tt /. , rs 11,,
10
10
~
'
30
:UNITED KINGDOM
I
r~ '~1 j j ~
h~j
1
,~,-~
1
0
0
1
'
7975
978
977
7979
1973
lif l~ .... ,. . . ~ ~ l,
,~i*LLLJ ~
SitLb~
4
976
977
1978
~
j
1975
978
1
4
FED.REP OF GERMANY
~11,,t l
978
0
-10
-20
~
973
1974
978
978
1977
978
30
-
OF GERMANY
20
-
A ~
10
~=
-10
~
975
10
978
1977
20
0
'--,-es,--'-,,,,..........,ai,--'=a~---cc--::,,,.---'-L10
1973
1974
197S
1976
1!1n
19711
1974
1975
FRANCE
I
~-10 I
-!
973
1974
30
j'
,
1974
0
‘\j~y1i==~=~i
100
f-FRANCE
FRANCE
1973
I 10
"j
f
974
20
-
.10
1*7~~ IM+
½,,,=,~"',,,~.~~.,"-",.~....c,,,,.~~.-,,,-'-~,.9;,a~......J
-10
973
IN TERMS
TERMS OF
OF U.S.
U.S. DOLLARS,
DOLLARS,
IN
31,
1978
31, 1978
-20
1977
197$
973
1973
Source: Artus and
and Young (1979).
source:
(1979).
—2 13—
-213-
1974
1914
1973
1915
978
1976
1977
19n
1378
""
expectations concerning future course of events, and changes In
in expectexpectations are immediately reflected in corresponding changes in prices.
Periods
Periods which are dominated
dominated by uncertainties, mew
new information, rumors
and announcements
announcements are likely to be
be periods in which
which changes in
in expecexpec-
tations are the
the prime cause of fluctuations in asset prices.
Further,
Further,
information which alters expectations must be new, the
since the information
resulting fluctuations in
in price cannot be predicted
predicted by lagged
lagged forward
. h are based
base d on
. f orma t·ion. 55 Therefore,
Therefore, during
exchange
which
information.
on past
pas t 1n
h1c
exc hange rates
ra t es w
such
exchange rates to
to exhibit
exhibit large fluctufluctusuch periods, one should expect exchange
ations and to be unbiased but imprecise forecasts of future spot rates.
gain further insights
To gain
insights into the implications of this perspecperspective on
on the relationship between predicted and realized changes in exexchange
change rates,
rates, II present
present in Figures 2-4 plots of
of predicted
predicted and realized
changes in exchange rates
rates for the
the three pairs of currencies where the
predicted change is measured by the lagged forward premium.
Also prepre-
sented in these figures are the differemtials
differentials in national
national inflation
inflation
rates which are discussed
discussed in the
the fourth section.
The key fact which
emerges from these figures is that predicted
predicted changes in exchange rates
6
account
account for aa very small fraction of actual changes.6
55The
The analysis of the role of “news
"news" in determining
determining current exexchange rates and in explaining forecast errors from the forward
forward rate
has been made
(l976a, 1976b,
l976b, 1977, 1979a). See
made forcefully by
by Mussa (1976a,
Dormbusch (1978). The large degree of volatility is also
also Dornbusch
also analyzed
by McKinnon (1976) who attributes it
it to
to insufficient
insufficient speculation.
speculation.
6These
and the
the following empirical regularities are analyzed
analyzed in
in
interdetail in
in Mussa (1979a). See also Frenkel and Mussa (1980). An interesting extension
extension would
would examine the relationship
relationship between the variances
of predicted and actual changes in exchange
exchange rates
rates in aa manner analogous
that of
of Shiller
(1979).
to that
to
Shiller (1979).
-214—
-214-
P4
tO
!(
US/UK
C
—
N
—~--~
N
~
~
I
I
I
I 1,1
I I
I1'
lI
II
,,
I
,,
1I
II
I
,,Ii
1,,,
i
I I
',,1
,,,,
,,
1,
i
I
I
-
,
‘
1977
N’~Ca
1976
to
N.
1975
1978
1979
L4,a
a)
(-
a)
“C
a)
Ci
4,)
0--co
P
a)
01
—I
I
H
a,J
0
U
a)
“‘—‘
Na
Ci
00 ci
Ca
•
‘C
CO
0
4’)
N’
Ca
Ca
LI’)
C’)
-215-
a)a)
,C
4,) (4,4
0
Ti
000.
“’a
0 a) H
“a
I
‘-“
0”,0
a)
~
Ci
(00)04
‘~‘), Pa
H
Monthly percentage changes of the U.S./U.K. consumer price
i~d~ces [~(,;',n COLLS/COL;K)], of the $/5. excha~ge rate,
(L.. ,,:.,n S ) , and the montklv forward premium; f"n(F
/s t- )]
t
•
t- 1
1
July 19/3 - July 1979.
a’)
Ci 0Ci ‘0Ci “a
a)
H
a)0’,
0
a)
p
a)
Figure 2:
Li,
1974
—
~
C~’-,
r—
2?
1973
:
i
,,
,,,, ,,,,
I I,
11
':',l;
i,,!
I1
11
11
11
I 1
·,,!\,.
, , ,,
I
\ /
I
I I
— — •_.—___————
— __
~
.r-a’t.~.—
I
I
I
:
==
I,''1
I
I
~=_.n—
COLUK
I
'11
)
I
.—
/:,.ln ( _sl§
i'
~-
“~Thgi
o o
00
COL
_.r..t—r.—
I
I
=
j
I
I
0
\j I
'-.!I 1
— —
II
p
:
I ,,
i'i
I/
1 I
——
'
1·
— ..=•.—
:
—
==
F....
__= =
II
'
..v-
—
11
,Cl
‘
(0
“‘~:::::::“~
—
/
—.
l
.
I
IJ ' ' · I I\ /\ . 1
l; I . : : \
1'1 ~ ; .1 I I i; i t; I
I I I ·· 1 !, I 1 I.. ; , , '
I I i\J ;/11 1I ;f ; ! }
I I I 1, ' I I j,,:!
11lU \,I II:; !i :\ i 11'' I 1 I .I \: \'
I I l ;J
If , ; l f JI 1 ii I I I
I 11
:;
~
.
I I I II
I
I I
'
\
I
I
;,
I
: I I I I II
I I I I \
,v~
I
I
I
I
I
11
I
I I I
11 I I
II ,\ I
11 1 1 I
—
[':1'
I
—
't \. I Ii\ '
fi l..._ .J \ ! I
-‘-
I
I!\
,,. .{ i
CO
_.
a)0~’)
—
.~“
:..._—~.-
ll
I
I
I
2?
I
:
I~
C)
'I
n, '
'.;-1,
: I
—.
. . .
11 ~
:I I \
I
~
;11r\ V\
I
•
I\
I
I I I II I I I
I
I 1I I ii I I I
I 11 1 11 I \ I
1\ : I I
\\A 1'1
I
I I 1,11,1 I
I\ J· \_, I
:f
···I
.. I •.
Y lJ :'·. ·II _.. ·. 'I ·.. .' } I . ' ''
;, .. ! ·. 1( I ·, I ... ,·•.1 _tt,
I
I I
I
——
: : i1
I
I
1
~..
—~
!:
' !I
ii
I
,1 ,1, I
/ I
lI
—
"
IflMVHI 1N1IU1A
r
I
-‘--
I /":
II
·,
I I i i
I I : ll
I..' I ! ~\
1
,C
:\ l
ri
~
II II
/ II
I I
I I
1
I
I I
l
I
Ii 11
II I 1
I1 1
1 I I
I I I
I I II
,,, I
III I
I II I
jl
I
I
1
1I
II I
ii I:
I I 11
I I 11
11 1 1
I I /I
I I I I
1 1 I
1
'
I
I
— —
1'
II
,1
I I
I
!
11
11
11
____
i———
C
b. lnSt
i/
1'
II
I II
II ii
II II
Ii 11
price
Monthl y percen tage change s of the U.S./F rance consum er
rate,
ge
exchan
.
$/F.Fr
the
of
,
indice s, [6(£n COLU5 /COLF]
)]
(6 1n St), and the monthl y forwar d premiu m; [1n(Ft _ 1 /St_ 1
July 19/3 - July 1979.
Ci
Ci
-216-
~
Ct
ID
C,
—-cotlCa
Ct
1979
CO
,.~
CO
“i
~°
1978
“‘-Ca ID
0)11)41
—
Ci
11) 0
Ci
m
Ci
Pt Ci
(-C
I 00
ID
-4
“
1977
0’
~
H. Pt Pt
Ci Ca Ca
coPt.
‘0Ot.
Ci
-~
0)
CO
~
__
1976
n-rn
~
hI’
0
414
o o
~1 “i
ID
H’t
~‘
‘--i
CO
1975
coCCi
‘-‘4~~
Coj~ CD
ID
CO
‘C
0’
,—,
1974
Ca Ci
Ci Ci’
0-ri
-
‘0cfiW’<
-.
an’Dn.fl..r
~_—_—_:t_-=.=fl:••
Cl.
~
‘~
7
COrn
-~
—
—.
..
.
.••
.:
.
‘
0
~_%
—
—
—
— — —
L
—
r•..
.,.•
~..
.
..
~
...t—.
~_7
———
.nr
‘~‘
—c
N
N
“~-~t_.
—a
>1
—~~-rr_
“r,~.
,,,_~
PERCENT CHANGE
:z
4.’
Figure 3:
CO
....z
""
=
....
,_
0
..,
0-’Ci
<
:,::
Ci
1973
I ~
....'-"
N.)
Pt
Ca)
C
0
=
:t~.
3
—CA
0
0
~rJ
CA
US/F ranc e
l
11
11
Ii
11
11
0H
“S
‘I)
p
±J
US /Ge rma ny
—
•
4*’-” _,—
..-“
t.~-~—
—
—•
• ~
—— —
~..
,—..
‘c__F..
•,
a.—~ — INIDHId
—_•4•a..
— S’”
InIJWII9
-c~~T~ r— ~
C!)
<1a-~——~
c...:
7
‘-a.----_—.—
C I
-~
..
yO
C
:
.
Germ an consu mer pric e
Mont hly perc enta ge chan ges of the U.S./
exch ange rate ,
indi ces, [6(£n COLU 5 /COL ~)], of the $/DM
lln(F t- 1 /s t- 1 )]
ium;
prem
ard
forw
hly
(6 Zn s ), and the mont
July 19 3 - July 19 79.
7
0)
‘—
to
at
—
C”
Ct
—
_
I”-
Pa
Ca
.‘—
-‘~
p00a)P04004~”)
a)
II
a)Cl’
a)”-’Cl) 0)
0%
P
—
(0
CiWNa
r-.
000”
III
C) HCl,)
0-’ CO
Ca
a)
a)$a)
2?
toti
1H
2?
,~i”’-’
oo.,o.c
C)
00a)
~4
(1) H
0’
H
~,, (001-P
a)
000
0%
0
a)
Ci
0’ci
a)H
04<01
ciCJ’1$N’
Ci”)
i
Figu re 4:
-217-
1979
1978
1977
1976
1975
1974
1973
This fact
rate changes seem to be
fact suggests that the bulk of exchange rate
11
11
due to “new
information” which, by
new information
by definition, could
could not have been
been
anticipated and reflected in the forward premium or discount
discount which
which preprevailed in the
the previous period.
In order to examine this hypothesis,
hypothesis, II present in
in Figures 5—7
5-7
plots of the spot and the contemporaneous
contemporaneous forward exchange rates for
the three pairs of currencies.
Also
Also presented
presented are the ratios of
national price levels which are discussed in the fourth section.
section.
If
If
the dominant factor underlying changes in rates
rates is new information,
which alters views about current and expected future exchange
exchange rates by
which
approximately
approximately the same amount,
amount, then one should expect aa high
high correlation
correlation
between
rates,
between movements
movements of spot and forward rates.
This fact is clearly
demclearly dem-
exonstrated by Figures
Figures 5—7
5-7 where
where it
it is
is seen
seen that spot
spot and forward exchange
change rates
rates tend
tend to move
move together
together and by
by approximately
approximately the same ampliampli-
tude (the vertical difference between the two rates correspond
correspond to the
forward premium of discount on foreign exchange).
The high correlation
between movements in spot and forward rates is expected
expected since the two
rates respond at the same time to the same
same flow of mew
new information.
This characteristic is
is typical to the foreign exchange
exchange market
market as
as well
as to other markets for stocks and durable assets.
The recent pattern
gold prices provides aa useful example of
of gold
of this
this general principle.
principle.
and the
the future price of
of gold as
as recorded
recorded rereTable 3
3 reports the spot and
on four recent consecutive
cently in the
the New York Commodity Exchange
Exchange on
days.
The
The two key facts which are illustrated by this
this table are
are the exex-
tent of day-to-day
day—to—day volatility in gold prices
prices and the uniformity
uniformity by
which these changes are reflected in the price of gold for immediate
immediate
delivery
delivery as well as
as in the prices for the twelve future delivery
delivery dates.
-218-
/··--.... __________'\
\......... .
I
/I
I
I
I
~ I
\I
•
'
I
a)-,
'
"'
;::
0C
<
a:
tnFt
LI
=
0-
=
C
‘F,
-J
4—
C
=
(0
C
=
LI
>1
US/UK
US/UK
1973
Figure
Figure 5:
1974
1976
1975
!
1976
1976
!
1977
1978
1978
!
1979
1979
Monthly
Monthly observations
observations of
of the
the Dollar/~spot
Dollar/E spot (S-n
(:n S ) and
and Forward
Forward
(Zn F)
F~) Exchange
Ratio of
the U.S./U.K,
Cost
(tn
Exchange Rates
Rates and
and the
the Ratio
of the
D.S~/U.K. Cost
of
Living
Indices
[Zn
(CCL~J
ICOL
)(scaied
to
equal
the
of Liiing Indices [Zn (CO½J 5 !COL1TK)(scaled to equal the spot
spot
initia~monE~)]:June
exchange rate at the initial
month)]:June 1973 —- July
July 1979.
1979.
—219-219-
!-...
. \·-~~
~
....
\
· · · •. I···.....\
i CCL
ln(
~
\ CCLF
·•••·••··.·............··...
—Inst
.•·······..
I
I
/I
I I
I I /
I
I/
I V
I
I
I
"'lnf:t
1973
1973
Figure 6:
1974
US/France
US/France
1976
1976
1975
1975
1977
1977
1978
1979
1979
Monthly
observations of
the Dollar/F.Fr.
St) and
Monthly observations
of the
Dollar/F.Fr. spot
spot (Zn
(£n S)
and ForwarC
Forwar<
(Zn Ft)
(in
Ft) Exchange
Exchange Rates
Rates and
and the
the Ratio
Ratio of
of the
the U-S./French
U.S./Fr~nch Cost
Cost
of
of Living
Living Indices
Indices [Zn(COLTTS/COLF)(scaled
[tn(COLUS/COLF)(scaled to
to equal
equal the
the spot
spot
exchange rate at the initial month)]:June 1973 -— July
July 1979.
1979.
-220-
I
LnFt
In
US/Germany
US/Germany
1973
Figure 7:
1974
1975
‘1976
1976
-
!
1977
-
-
1978
1979
Monthly observations of the Dollar/PM
Dollar/DM spot (Zn
(tn SS)) and Forward
(Zn
F~) Exchange
Exchange Rates
Ratio of
Cost
(2..n F)
Rates and
and the
the Ratio
of the
the U.SJGerman
u.sJGerman Cost
of Liiing
Living Indices
[Zn(C0L~ /COL~)
(scaled to
equal
the spot
Indices [2n(COLUS/COLG)(scaled
to
equal
the
spot
5
exchange rate at the initial month)]:June 1973 —- July 1979.
1979.
—221—
-221-
Table 33
Table
Futures Price of Gold
Gold on Consecutive Days
Futures
Daily Data;
Data: October
October 1, 1979 —- October 4, 1979
---~--------··-Price (per
(per ounce)
change from
day
Price
ounce) and
and change
from previous
previous day
___
Delivery Date
__
Oct. 1, 79
Change
Oct. 2, 79
Change
Oct. 3,
79
3, 79
Change
Oct. 4, 79
Change
,,
416.0
1;16.0
21
21
411.0
411.0
—5.0
-5.0
393.5
393. 5
17.5
-17.
5
369.7
369.7
23.8
-23.8
November
419.5
20
416.0
—3.55
-3.
397.7
397. 7
—18.3
-18.3
377.7
377. 7
—20.0
-20.0
December
424.5
20
20
421.00
421.
—3.5
-3.5
402.5
—18.5
-18.5
382.5
—20.0
-20.0
1980 February
432.8
432
.8
20
429.6
—3.2
-3.2
410.7
18.9
-18.9
390.7
N
'
20.0 0J
~j
-20.0
April
440.9
20
438.0
438
.0
—2.9
-2.9
418.8
418.8
—19.22
-19.
398.8
—20.0
-20.0
June
June
448.6
20
41,6.0
446.0
-2.6
—2.6
1,26.6
426.6
-19.4
—19.4
406.6
-20.0
—20.0
August
456.3
20
20
454.0
—2.3
-2.3
434.4
—19.6
-19.6
Lfll, .4
414.4
—20.0
-20.0
October
464.0
20
462.00
1,62.
—2.0
-2.0
1,42.
442.22
—19.8
-19.8
422.2
—20.0
-20.0
December
471.5
20
469.88
1,69,
—1.7
-1.7
449.8
—20.0
-20.0
429.8
—20.0
-20.0
1981 February
20
20
20
477.5
477
.5
485.0
—1.4
-1.4
—1.1
-1.1
457.5
465.0
1,65
.0
-20.0
—20.0
—20.0
-20.0
437.5
437. 5
April
478.9
486.1
1,86.1
445.0
445.0
—20.0
-20.0
—20.0
-20.0
June
493.3
493.3
20
492.5
—0.8
-0.8
472.5
—20.0
-20.0
452.5
L152
.5
—20.0
-20.0
August
August
500.5
20
500.0
—0.5
-0.5
480.0
—20.0
-20.0
460.0
—20.0
-20.0
October
1979 October
Note: These prices are settlement prices at the Commodity Exchange, New York as reported in the Wall Street
Journal October 2-5,
2—5, 1979
c’.J
N
'
Another feature
feature which
which is
is revealed
revealed by
by Figures
Figures 5—7
5-7 Is
is that
that the
the
Another
approximately equal
equal
ontemporaneous spot and forward exchange rates are approximately
hus indicating that the markets
market's best
best forecast of the future spot
spot rate
ss (approximately) the current spot rate.
This phenomenon reflects
reflects the
act that, as an empirical matter, exchange rates have followed
approximately) aa random walk process.
For such aa process, current
rices are indeed
indeed the best forecasts
forecasts of
of future prices.
To
To the extent
extent
hat the exchange rate had some drift,
drift, the above statement
statement should be
drift.
nterpreted in reference to that drift.
empirical phenomenon
phenomenon
This empirical
exchange rates even
even though
eems to correspond to the actual paths of exchange
t does not reflect aa theoretical
theoretical necessity.
characteristic of the foreign exchange market is deThe final characteristic
de-
cribed
of currencies
cri
bed by Figures 8-10,
8—10, which plot for the three pairs of
he spot exchange rate and the forward premium on
on forward exchange.
ince
funda—
ince the units of the spot rate and the forward premium
premium are funda-
different, the two series were normalized by subtracting from
1entally
entally different,
,ach series its mean and by dividing by the corresponding standard
‘ach
‘rror.
,rror.
The fact which emerges from these figures is that generally
though not always) there is aa positive
ex—
positive correlation between
between the
the ex-
,ected
currency (as
(as measured
measured by
by the forward premium
ected depreciation of the currency
,nn foreign exchange) and the spot exchange
exchange rate.
rate.
This positive
positive cor—
cor-
·elation may
may be
be rationalized by
by noting that currencies
currencies which are ex—
ex-elation
Iected
•ected to depreciate are traded at aa discount in the forward market
nd, on average, these currencies
.nd,
currencies also
also command aa lower foreign exchange
exchange
alue in the spot market.
•alue
This correlation is interpreted further in
he next section.
:he
-223-223—
I
\‘
\
\
\
~·
I
I
'-
I’
/
‘
I
\,J
I
I
I
l
\ ~
\ I
0-
I
I
C
=
C
I
I
I
I
I
I
C
U-
~I
I
\
'r,
\
Q
C
:z
C
<C
.....
....
U-a
\
\
I
I
I,.
I
'-I
I~
I
_,,I
C
<C
....=
I
I
/
(\
I
=
I
I
IV
_,,......,✓
=
U-a
\ I
I I
:z
"'
C
<C
\
=
::r:
LI
'-'
‘C
X
....
U-a
US $/UK£
$/UK £
‘1973
1973
Figure 8:
Figure
‘1974
1974
‘
-
1975
‘
1976
1976
-
1977
1977
‘
‘
1978
‘
-
1979
1979
Monthly ob~ervations
normalized Dollar/b
Dollar/L spot
Monthly
obftervations of the
the normalized
spot excha~
exch~~
rate
S~) and the normalized
rate (Zn
(2.n S~)
normalized forward
forward premium [£n(Ft/S ) l]
Both series
series are normalized
normalized by
by subtracting
series ij
subtracting from each seri~s
mean
error:
mean and
and by
by dividing
dividing by
by the
the corresponding
corresponding standard
standard error:
June 1973 -— July
July 1979.
-224-
I
,,~
,,,,
~
/ I
Ln(.t)
/ I
/
/
I\ I
I
I
I
I
I
I
I
/
I
I
I
'v
I
I
II !
11
I I \
11 ,) I
I
I
I
I
I
I
\
I
I
I
,,
I
I
I
I
I
I
\
\ I
I
~,
'-1
I
I
I
~,
I,—’
/
i'..1
,.....- NM
InSt
I
I
I
I
'
I
L-,
I I
I
,
,
—
..........
/
"""
lnS 1
1973
Figure 9:
9:
Figure
1974
1974
1975
1975
-
1976
1976
-
1977
1977
US $/FFr
-
1978
1978
-
1979
1979
Monthly
of
Monthly observationj
observatio~
of the
the normalized
normalized Dollar/F.Fr.
Dollar/F.Fr. spot
spot
1
exchange
(Zn
51)
and
the
normalized
Forward
exchange rite
r~te
(tn
S
)
and
the
normalized
Forward Premium
Premium
1
[Zn(F~/S
normalized by
[£n(F /S ?].
) ] . Bothtseries
Bothtseries are normalized
by subtracting
subtracting
from ~acfi
each series
series its
its mean and by dividing
dividing by the
corresponding
corresponding standard
standard error:
error: June
June 1973
1973 —- July
July 1979.
—225—
-225-
In(.1)
r
U-a
=
0-
C
(
C
I
I
=
=
C
\
\.,
\
I
U..
C,
C
z
C
<
La-a
C
=
U-a
C0
C
=
LI
DC
U-
US $/DM
1973
1973
Figure
Figure 10:
1974
-
-
1975
1975
- 1976
1976
‘
-
‘
1977
1977
-
-
1978
1978
‘
-
1979
1979
-
I
("~n SN
Monthly
observations of the normalized
spot (Zn
Monthly observations
normalized Dollar~DM
DollarNDM
soot
and the normalized
normalized Forward Premium [.Ln(F
[Zn(F~/S~Y
]. Both
/St)~].
Both seria
serit
are normalized
from each series
normalized by subtracting from
ieries its mean and
dividing by the corresponding
standard error:
1973 —by dividing
corresponding standard
error: June 19i3
July 1979.
19 79.
July
-226-
EXCHANGE RATES, INTEREST RATES AND INNOVATIONS
In this
this section II analyze the relationship
relationship between exchange rates
and interest
interest rates from the analytical perspective
perspective of the
the monetary apapproach to the exchange rate.
To
To set the
the stage
stage for the analytical
analytical dede-
velopment
useful to recall the
the typical
typical analysis which
which generally
generally
velopment it is useful
predicts aa negative association between the rate of
of interest and the
exchange rate.
According to that analysis
analysis, aa higher
higher rate
rate of
of interest
,
attracts foreign capital which induces aa surplus in the capita
capital1 account
of the balance of
of payments and thereby induces an appreciation
appreciation of
of the
exchange rate).
rate).
domestic currency (i.e., aa lower spot exchange
Another variant
variant
of the popular approach states that the higher rate of
of interest
interest lowers
spending and thus induces a surplus
surplus in the current account
account of
of the balbal-
ance of payments
payments which results
results in aa lower spot exchange
exchange rate.
AA third
variant of this approach
approach claims
claims that
that the higher
higher rate of interest
implies (via the interest
interest parity
parity theory) aa higher forward premium on
on
foreign exchange and to
to the
the extent that
that at aa given point in time the
forward exchange rate is predetermined
predetermined by past history,
history, (am
(an assumption
that is
is clearly rejected by the evidence on the
the comovements of
of spot and
and
that
forward rates), the reguf
red rise in the forward premium
required
premium will
will be
brought about by aa lower spot
spot rate
rate (i.e.,
(i.e., by an appreciation
appreciation of
of the
domestic currency).
Whatever
Whatever the route, this approach predicts aa
p~g~~jve
exthe rate
rate of interest and the spot
spot exnegative relationship between the
change rate (or alternatively, aa positive
positive relationship between the rate
rate
of interest and the foreign exchange value of the
the domestic currency).
currency).
These predictions, however, do not seem to
to be in
in accord with
with the
the
broad facts.
Over the recent period the rise in the rate of
of interest
in
in the U.S. (relative to
to the foreign rate of interest) has been
been
—227—
-227-
Figure
Figure 11
Foreign Exchange Value of the
U.S. Dollar and Interest Rate Differentials
Ratio Stale
Scale
March
1973:100
MarCb 1973
=100
110
105
_,
~
Weiqhted
Foreign
Currency
thee Dollar
w·htdA
e1g e Ayeraoe
veraae F
ore1an C
urrency Value
VI
a ue of
ofthDII
o ar
-
Ratio Scale
Ratio
MauI
1913:100
M
arcb 1973
=100
110
los
105
~
100
''
95
90
100
A
95
\
v- -
8S
80
Percent
5
90
90
,/
as
85
80
at
Percent
5s
U.S.-Foreign Interest Rate Differentials
f\.
4
J
Short-Ter: 3/
3
2
-
44
\'
33
22
r'
J---, ----r-,-. ...
/ ' , ...'-
0
., -~
-2
long Termll
..... ,- I
••
~
.... , _ .... 4
\.
.3
...-
.J.-'-'I
—1
-i
!
-2
·2
:~
.3
'-'1'
.4
0
1976
i
1977
1978
1q79
1979
.4
-4
Sources:
Federel Reserve Statistical Release H.13;
Federal Reserve
Reserve Bulletin;
Monetary
Sources; Federal
4.13; Federal
Bulletin International Monetary
Fund, International
!nternationol Financial Statistics.
-Statistics,
ii
in the
the
Ll Secondary
Secondary market
market rates
rates for
for 90-day
90-day large certificates
certificates of deposit
deposit in
the united
United States
States less
less the
weighted
wl!lighted average
overage of foreign
foreign three-month money market
market rates.
:1_ U.S.
2.
us.
long.term
gov~rnment bond
bond yields
less the
overage of
Ieng.term
Iong.tern~govornment
yielas less
the weighted
weighted average
of foreign
foreign long-term
government bond yields.
yields,
Latest data
plotted~May
latest
data plotted:
May
Source:
Source:
D. R.
1. Mudd (1979).
(1979)’.
-228-
depreciassociated with aa rise
rise in
in the spot
spot exchange rate (I.e.,
(i.e., with aa depreci-
ation of the dollar).
ll illustrates the point by plotting the
Figure 11
against the interest
interest rate
foreign exchange
exchange value of the U.S. dollar against
differential.
As
As is evident, in contrast with the popular prediction,
the higher (relative)
(relative) rate of interest in the U.S. has
has been associated
with
(i.e., with aa lower foreign exchange
exchange value
with a higher exchange
exchange rate (i.e.,
of the
the dollar).
dollar).
This
exThis contradiction,
contradiction, however,
however, does
does not
not arise
arise when
when the
the ex-
monetary (or an asset market) perspecPerspecchange rate is analyzed from aa monetary
tive to which we mow
now turn.
monetary approach
approach are hypotheses
hypotheses
The major building blocks of the monetary
concerning the properties of the demand
demand for money and money market
equilibrium and hypotheses concerning the link between
between domestic and
foreign prices. 77 Consider
Consider first the equilibrium
equilibrium in the money
money markets.
markets.
N/P and M*/P*
The supplies of domestic and foreign real balances are M/P
where NMand
and PP denote
denote the
supply and
level,
where
the nominal
nominal money
money supply
and the
the price
price level,
respectively, and where variables pertaining to the foreign country are
indicated by an asterisk.
Denoting the demands for real balances by
LLand
and L* (both of which are functions which are specified
specified below),
equilibrium
equilibrium in the
the money markets is attained when
(7)
(7)
L
L = M/P
N/P and
(8)
L* = M*/P*.
7For theoretical developments and applications of the approach
see, for
for example, Dornbusch (1g76a,
(1976a, 1976b), Kouri (1976),
(1976), Mussa (1976a),
(1976a),
Frenkel (1976), Frenkel and Johnson (lq78),
(l~78), Frenkel and Clements (1980),
Bilson (1978), Hodrick (1978), and Frankel (1979).
—229—
-229-
From equations (7)-(8), equilibrium in
in the money markets implies that
the ratio of the two price levels is:
(9)
(9)
PP
~
P*
MM
1*
M*
L*
IL ·
The second building block links domestic and foreign prices.
8
Assuming
purchasing power
Assuming the simple version of purchasing
power parity implies that:8
(10)
pP == SP*
Using equation (10) in (9) yields
((11)
11}
5S
-
MM
M* LL*
M* I
which
which expresses
expresses the exchange
exchange rate
rate in
in terms
terms of domestic
domestic and
and foreign supsup-
plies and demands for money.
To gain further insight into the deterdeter-
minants of
of the
the exchange
exchange rate and to set the stage for the
the empirical
estimation, assume
assume that the demand for money
money depends on real
real income (y)
(y}
and the rate of interest
interest (i)
(i} according to:
(12)
L = ay~e~
8For
For aa discussion of the choice of
of the relevant
relevant price
price index to be
be
used in equation (10},
(10), see Frenkel (1978). This simple version of the
purchasing power
power parity theory is used here to simplify
simplify the exposition.
To the extent that there are systematic
systematic deviations from
from purchasing
power parity they can be incorporated into the final
final exchange rate equaequation. Similarly,
Similarly, to
to the
the extent
extent that purchasing power
power parities holds in
in
the long
long run but not in the short run, the final exchange
exchange rate
rate equation
equation
will reflect these dynamic characteristics.
characteristics. To the extent
extent that purpurchasing power
power parity pertains to traded goods only, the exchange
exchange rate
equation would
would also contain terms which relate
relate to
to the relative prices
orices
non—traded goods; for aa formulation along these lines, see
of traded to non-traded
Dornbusch (1976b) and
and for an
an empirical application, see Clements and
Frenkel
(1980). AA more
Frenkel (1930).
more refined
refined specification
specification would allow
allow for the
effects of tariffs on
on the relationship between domestic
domestic and foreign
prices as well as
as for short-run effects of
of unanticipated
unanticipated money on outoutput
on prices
and the
the exchange
exchange rate.
put rather
rather than
than only
only on
prices and
rate.
-230-
(13)
L*
=
Using equations (12)-(13)
in (11)
(11) and assuming
assuming for simplicity
simplicity of exex(12)—(13) in
momey
position that foreign and domestic parameters of the
the demand for money
~
we
are the
the same,
same, i.e., that
that c~
a= c~,
a*, and that ~n = n*,
we obtain:
((14)
14)
zn S
~n
S
=
v*
zn M ++ nn tn
i *)
CC ++ ~n
~n ~+a(~(
i - i*)
M*
-
y
where CC s in(b*/a).
in(b*/a).
Equation (14) relates the exchange
exchange rate
rate to
to the ratios of domestic to
foreign money supplies and incomes and to
to the interest rate
rate differendifferential.
tial.99 Most
Most pertinent to
to the present purpose and in agreement
agreement with the
facts summarized by Figure 11,
relation11, equation (14)
(14) yields aa positive relation-
rate of interest and the exchange
exchange rate.
ship between the rate
The economic
economic
The
interpretation of this
this association in
in the context of the U.S. dollar
dollar
and the inflationary environment
environment is
is as
as follows: aa rise in the
the domestic
domestic
(relative)
(relative) rate
rate of interest is primarily
primarily dominated by
by aa rise in the exex-
pected (relative) rate of
of inflation which induces aa decline
decline in the dedemand for real cash balances; for a given path of
of the nominal money
supply, asset market equilibrium requires a price level which is higher
than the price which would
would have
have prevailed
prevailed otherwise.
Since the
domestic price level
level is linked
linked to
to the foreign price
price through some form
of purchasing
purchasing power
power parity, and since the path of the foreign price is
91t should be
rt
noted that aa similar
similar set of
of variables would also
appear in the reduced form of aa variety of alternative
alternative models. The dedependence of the demand for domestic money
money on
on the domestic rate of ininterest and the dependence
dependence of
of the
the demand for foreign money on foreign
A more
rate of interest is assumed only for simplicity of
of exposition.
exposition. A
general formulation would recognize that the demands for domestic
domestic and
foreign monies
monies depend on all margins of substitution.
substitution. See
See Frenkel and
Clememts
Clemen
ts ((1980).
1980) .
—231-231-
assumed to be given, the higher domestic price can only
only be achieved
achieved
through aa rise in
in the spot exchange rate (i.e., through aa depreciation
of the currency).
of
of the
positive association
association between
between interest
interest
This
explanation of
This explanation
the positive
rates and exchange
exchange rates has
has an intuitive appeal in
in that it implies
implies
an inflationary environment, aa relatively rapid rise in
in prices
that, in an
is associated with high nominal
nominal rates of interest as
as well as with aa
depreciation of the currency
currency in
in terms of foreign exchange.
The traditradiThe
tional prediction of
of aa negative relationship
relationship between interest
interest rates and
the exchange
exchange rate
rate may, however, be reconciled with the monetary
monetary approach
under the assumption that it
short—run liquidity
it concentrates on the short-run
effects of monetary
monetary changes.
Accordingly, in the short-run, aa higher
rate of interest
interest may arise from tight money which induces
induces an appreciaappreciathan aa depreciation of the
the currency.lo
should be
be ememtion rather than
currency.1° It should
phasized, however, that during an
an inflationary environment
environment (like
(like the one
prevailing
prevailing in
in the U.S. in recent years) the variations in the rate of
of
interest are most
most likely
likely to be
be dominated
dominated by variations in inflationary
inflationary
expectations rather
rather than by
by liquidity effects associated with
with changes
in the ratio of money to bonds.
environment the rate
In such an environment
rate of
interest is expected to be positively correlated with the exchange
exchange rate.
The discussion provides an illustration of
of the difficulties
rate of interest as the relevant 111onetary
associated with using the rate
monetary
indicator.
Traditionally, the
the height of the rate of
of interest
interest was the
!OThe short—run liquidity effects is emphasized in Dornbusch
lOThe
short-run liquidity effects is emphasized in Dornbusch
(1976b). The role of inflationary
inflationary expectations in
in dominating exchange
exchange
rate developments
developments is emphasized
emphasized in Frenkel
Frenkel (1976).
(1976). Frenkel (1979) and
and
Edwards (1979) attempt to integrate these two factors.
-232—232—
aiterion
monetary policy
iterion for
for assessing
assessing whether
whether monetary
policy has
has been
been easy
easy or
or tight:
tight:
0
high interest rate was interpreted as indicating a tight monetary
licy while aa low interest rate was interpreted
,licy
interpreted as indicating an
an easy
easy
netary policy.
,netary
By
By now it is well recognized that during inflationary
?riods it is vital to draw aa distinction between
sriods
between nominal and real
inflationary periods
periods the
ates of interest and, as aa result, during inflationary
ate of
of interest may provide aa very misleading interpretation
interpretation of the
tance of
of monetary policy.
The same
same logic
logic applies
applies with respect to the
nalysis of the
the relationship
relationship between exchange
exchange rates and interest rates.
The foregoing analysis also provides
provides the
the explanation
explanation for the obThe
ob—
ervation (which was noted previously) that generally there is
is aa
ositive
ositive correlation between the forward premium on
on foreign exchange
exchange
nd the level
level of
of the spot rate.
rate.
Since
Since the spot rate is
is expected
expected to
to be
ositively
ositively correlated with interest rate differential and since,
ccording to
to the interest parity theory, that differential must equal
on foreign exchange, it follows that the forward
he forward premium on
remium is also expected to be
be positively correlated
correlated with the level
level of
he spot rate.~
rate. 11 That positive
positive correlation may
may also be
be rationalized by
currencies which are expected to
oting that currencies
to depreciate are traded at
at a
11
~For
For evidence on the robustness
robustness of the interest
interest parity relation—
relationhip,
Frenkel and Levich (1977). The positive association between
hip, see Frenkel
he spot exchange rate and the
the forward premium
premium has
has been interpreted
interpreted in
in
erms of an explicit
posiems
explicit monetary model. It is noteworthy that this posiive association would be
ex.ive
be predicted
predicted by any model in which
which current
current exdepreciahange rate
rate reflects immediately the expectations of
of future depreciaion. See, for example, Mussa (1976a) and Frenkel
.ion.
Frenkel and Mussa (1980).
(1980).
ince the rate
,ince
rate of interest and
and the exchange
exchange rate are dimensionally
ncommensurate, their
their association raise questions
questions that
that are familiar
familiar
rom the discussions of the Gibson
·rom
Gibson Paradox. In aa separate paper,
paper, II ininend to examine the relationship between exchange rates and the
.end
the forward
remium
interest differenti
differential)
the various
,remi
um ((or
or the
the interest
a 1) in
in 1light
i ght of
of the
various exolanaexo 1anaiions
ons of
of the
the Gibson
Gibson Paradox.
Paradox.
—233-233-
discount
in the
the forward
forward market
market and,
and, on
on average,
average, these
these currencies
currencies also
also
discount in
command aa lower foreign exchange
exchange value in the spot market.
Prior
Prior to proceeding with the empirical evidence on the relationrelationship between exchange rates and interest rates it might be
be useful
useful to
highlight
highlight some of the main
main features of the monetary approach which are
are
reflected
reflected in equations (11) and (14).
First,
First, these equations
equations demondemon-
strate the symmetric roles that are being played by the supplies of
of
domestic and foreign monies and the
the demands for these monies.
Since
Since
variables like
like real incomes as
as
the demands for monies depend on real variables
well as
as on
on other
other real
variables which
underlie expectations
rates
real variables
which underlie
expectations and
and rates
well
of
of interest, it is
is clear that the monetary approach
approach does
does not
not imply that
the exchange rate depends only on the relative supplies of money; nor
does
exthat real variables do not affect the
the equilibrium exdoes it imply that
rate.
change rate.
Second, from the policy perspective
perspective the monetary approach
approach
of the homogeneity
homogeneity postulate:
brings to the forefront the implications of
ceteris paribus
paribus aa rise in the quantity of
of money
money results in
in an
an equiproequiproportionate rise in
in the
the exchange
exchange rate.
rate.
This illustrates the intimate
intimate
connection between
connection
between monetary policy and exchange
exchange rate
rate policy.
Third,
the
the positive
positive relationship between interest
interest rates and exchange
exchange rates and
the
the central
central role played by
by inflationary expectations
expectations imply that policies
of the currency
which attempt to induce an appreciation of
currency should aim at
reducing inflationary expectations.
expectations.
The reduction in inflationar9
inflationary exex-
pectations would halt the depreciation of
pectations
of the currency
currency in
in terms of
goods
goods and in terms of
of foreign exchange, and
and would result in lower
nominal rates of interest while maintaining (or even raising)
raisinq) real
rates
rates of
of interest.
-234-234-
The discussion in the second section and, in particular, the conconThe
~ibutions
l979a) and Dornbusch (1978) emphasized that
"ibutions by
by Mussa (1977, 1979a)
~e
1e predominant
predominant cause of exchange rate movements is news
news which
which could
)t
it have been anticipated.
It
It was also argued in the second section
iat the forward rate
1at
rate seems to summarize the information that is availavail-
' le
le to
to the market when the forward rate is being set.
set.
We may
may there—
there-
Dre express the spot rate at period t as aa function of factors which
ire
which
ave been known in advance and are summarized by
by the lagged
lagged forward
ate, as well as aa function of the “news.”
:1.te,
news.
11
11
15)
in
St = a+
a + bb tn
in Fti
+ "news"
“news”
zn st=
Ft-l +
The empirical difficulty is in identifying the variable
variable which
easures
easures the “news.’
news.
11
11
Assuming that asset markets clear
clear relatively fast
nd
nd that the “news’
"news" is immediately reflected in
in (unexpected) changes in
in
he rates of interest
interest we may write equation
equation (15) as
16)
in St = a
+
b in Ft1 +~[(i
-
i*)t
-
Et1(i
-
i*)t]
here the bracketed term denotes the innovation in the interest differdiffer—
ntial
Eti (i - i*)t
ntial and where Et-l
i*)t denotes the interest
interest differential which
-
in period tt based on the information available
available
as expected to
to prevail in
t tt - 1.
—
The expected
expected interest rate differential was computed from aa
egression
egression of the interest differential on a constant
constant and on
on two lagged
12 The previous analysis of the relationship
alues of the differential.12
alues of the differential.
The previous analysis of the relationship
12An alternative way
An alternative
to compute the
the expected differential would
se data on
on the term structure
structure of interest rates. Since data on the
ifferential
computaifferential of
of 2—month
2-month rates are not readily
readily available, this computaion would require interpolations.
—235—
-235-
between interest rate differential and the exchange rate implies that
between
the coefficient cia is expected to
to be
be positive.
Table 44 reports the OLS estimates of equation
equation (16)
(76) for the three
exchange rates over the period June
,June 1973-July 1979.
As may be seen, in
all
coefficients of the unexpected interest
all cases
cases the coefficients
interest differential
differential are
positive and in most cases the coefficients
coefficients are statistically
signifipositive
statistically significant.
In order to
to verify the importance of
of using the series of innoinno-
differential, Table 44 also reports
reports estimates
estimates of
vations in the interest differential,
which replace the innovations by the actual series of the
regressions which
interest differential as well as regressions which
which include
include both the
innovation
actual differential.
innovation and the actual
In all cases the coefficients
coefficients
of
of the actual
actual interest differential do not differ significantly from
1~ To allow for a simultaneous determination of interest rates
zero.13
zero.
To allow for a simultaneous determination of interest rates
two—stage—
and exchange
exchange rates,
rates, equation
equation (16) was also
also estimated
estimated using aa two-stageleast—squares
least-squares estimation procedure.
These results
results are reported
reported in
Table 5,
5, and again in all cases the coefficients of the unexpected
interest differential are
are positive.
These
These coefficients
coefficients are highly sigsig-
nificant in
in the Dollar/Pound exchange
exchange rate
rate but insignificant
insignificant in the
nificant
other two rates,
rates.
On the whole, the record shows that during
during the 1970s
l970s
exchange rates and interest rate differential have
have been associated
associated
positively and thus indicating that during that
that inflationary
inflationary period the
the
rise in
same factors which induced a rise
in the interest differential
differential also inin-
exchange rates.
rates.
duced aa rise in the spot exchange
Furthermore, consistent with
13In order to check whether the
In order to
the dollar rescue policies of
November 1978
1978 have had aa systematic
systematic effect on the
the estimates, these reregressions
gressions were also estimated for
for the period up to
to September
September 1978.
1978. The
results did not change materially.
-236-
Table 4
interest Late
Interest
Rate Differentials and Exchange Rates
Monthly Data:
Data: June
1973 -— July
1979
Monthly
June 1973
July 1979
(standard errors
errors in
parentheses)
(standard
in parentheses)
Dependent Variable
,n
S
Constant
in
Ft_i
in F
t-l
(1-i*)t
(i_i*)~
t
2
R.2
D,W.
D.W.
.027
.95
1.73
.388
(,165)
(.165)
.026
.96
1.77
.546
(,199)
(.199)
.026
.026
.96
1.80
1.80
.029
•.79
79
2.11
2.11
..377
377
(.132)
.028
,81
.81
2.22
2.22
.540
(.206)
.028
.81
2.32
2.32
.032
.93
.93
2.02
2,02
.601
.601
(,271)
(.271)
,031
.031
.94
2.08
2.08
.583
(.349)
.031
.94
2.08
2.08
a,e,
8.e.
f(i-i*)t-Et_ 1 (1-i*)t)}
{(i_i*)~_E~i(i_i*)~)l
Dollar
/Pound
Dollar/Pound
.959
.017
(.096)
.032
.032
(.019)
(.019)
(,025)
(.025)
.030
(.
017)
(.017)
( .024)
(.024)
.019
.019
(.019)
.968
(. 024)
(.024)
—.155
-.155
(,
111)
(.111)
—.335
-.
335
(.100)
•.776
776
(.067)
(.067)
.184
(,125)
(.125)
—.301
-. 301
(.079)
(.051)
—.231
-.
231
(.104)
.851
(.070)
(.070)
—.070
-.070
(.
(. 43)
.926
.926
.237
. 237
(,045)
(.045)
(.229)
-.037
—.037
(,027)
(.027)
,955
.955
(.032)
-.040
—.040
(.046)
(.
046)
.952
(.047)
( .047)
.961
.961
Dollar/Franc
'
w
"'
"'
N)
U)
—4
.801
.801
.195
(.188)
(, 188)
Dollar/OH
Dollar/OM
NOTE;
NOTE:
.024
.024
(.290)
Interest
one—month (annualized) Euromarket rates
differential Et_ (1-i*)t
Interest rates are the one-month
rates. The expected interest
interest rate differential
1
computed from
from a regression of the interest differential on a constant and on lagged values of the
was computed
differential.
(I —- i*)t
i*) denotes
denotes actual
actual interest
differential where
where ii denotes
the rate
of interest
interest on
on securities
different.ial. (i
interest rate
rate differential
denotes the
rate of
securiti.es
denominated in U.S.
u.s. dollars and i* denotes the rate of interest
securities denominated in foreign currency.
d(;'1\0minated
interest on securities
currency,
the unexpected
interest rate
differential.
((1-i*)t-Et-l(i-i*)tl denotes
denotes the
unexpected interest
rate differential.
Table 55
table
lntna.t Bat.
Oifferential sad
Interest
Rate Differential
and takings
Exchange Beta
Rates
Instrumental Variables
Monthly Data: June
Jia. 1973 —- July 1979;
1979; Istaseatal
(standard errors in parentheses)
Dependent
Da~n~u~
Variable
J!.n st
R22
ft
D,W.
D.W.
.027
.027
.95
1.69
.024
.024
.97
.97
1.79
1.79
.023
.97
1.78
.036
.69
2.18
2.18
.216
(.165)
.029
.80
.80
2.22
2.22
.165
(,199)
(.199)
,034
.034
..73
73
2.21
.034
.034
.92
2.11
.034
.034
.92
2.07
.034
.92
.92
2.07
2.07
p.i(i_i*)t)
a...
l((i—i*)~—t
(i-i*) t-Et-1
{i-i*) t) I1 s,e,
Sn Ft-1
in
(i..i*)~
(i-i*) t
(.020)
.966
(.026)
-.153
—.153
(.117)
.027
(,
016)
(.016)
.963
.965
.435
.43.5
(,022)
(.022)
(.164)
(,164)
Constant
poller/Pound
Oollar/!i?omi.ll
.020
.020
.017
.017
(.017)
(.017)
Dollar /Franc
Dollar/Franc
Dollar/I}l1
Dollar/$t
—.143
-.145
-·.142.
—.142
( ,098)
(.098)
,')09
.909
-.301
—.301
(.234)
(.125)
(.085)
(.085)
—.279
-.279
(.081)
(.081)
((.053)
.053)
—.184
-·.184
(.126)
.883
((.086)
.. 086)
—.260
"".260
(, 225)
(.225)
-.006
—.006
(.047)
(.047)
.987
.987
-.135
—.135
(.049)
(.049)
(.307)
—.040
-.040
(. 031)
(.031)
.951
((.036)
.036)
—.040
-.040
.931
.951
( .055)
(.055)
(.053)
BOUt
NOTE:
.971
.971
(.022)
.826
.816
,425
.425
(,160)
(.160)
.555
(. 380)
(.380)
—.001
-.001
((.003)
.003)
.555
.555
(.
402)
(.402)
t
are the
the one—month
one-month {annualized)
Euromarket rates. The
1be expected
expected interest
interest rate differential
differential IE
(i-i*))
Interest rates
rate. are
(annualized) luroinrtst
(i—i
interest differential
differential on a constant sad
and or,
on two lagged values of the
§Ilferential.
was computed from
tree aa regression of
of the interest
the 5iher.ntL.
Two-stage least square.
squares astianion
estimation method was used.
instruments for the
the interest
interest dIfferential
differential were a
a content
constant and
Two-stage
nsed. The imanants
two lagged valves
of the
the differential
imstr,aents for the unexpected differential
two
values of
differential and the
the instruments
differential were a
a constant
constant and
and Dufl.in’a
Durbin's
(1 —- iS)
i*)f denotes
denotes actual interest ate
rate differential
differential where
where ii denotes
denotes the rat,
rate of interest on securities
rank variable (i
denominated in U.S. dollars
do lars sad
and j*
i* denotes the
the rate
securities denominated in foreign
foreign currency.
currency,
demoatmatad
rat, of interest on securities
(i..i*) rc-1”~~ ~1denotes
demotes the
iat.rest rate
((i-i*)t-Et-l(i-i*)t]
the —e--nted
unexpecte--4_ interest
rate differential.
differential.
'
co
M
N
'
the hypothesis that current changes in exchange rates are primarily a
response to
response
to new information,
information, the evidence
evider;ce shows
snows the
the importance
importance of
of the
the
innovations
Innovations in the interest differential.
The principle that current exchange rates already reflect expecexpec-
tations concerning the future course
course of events implies that changes in
in
exchange rates are
are primarily due to innovations. In the present secsecapp"lied to the analysis of the relationship
tion this principle was applied
between exchange rates and interest rate differential.
differential.
however, is general.
The principle,
For example, it implies that the relationship
between a deficit in the balance of trade and the exchange rate depends
crucially on whether the deficit was expected or not.
A deficit that
was expected may have no effect on the exchange rate since the latter
already reflected these expectations.
In contrast, an unexpected
deficit in the balance of trade may contain significant new infonnation
information
4
that is likely to induce a strong
strong effect on the exchange rate)
rate. 14
EXCHANGE RATES AND PRICES
One of the striking facts concerning the relationship between
prices and exthange
exchange rates during the 1970s is the extent to which the
evolution of prices and exchange rates have not coincided.
The origiorigi-
nators and proponents
proponents of the purchasing power parity doctrine (Wheatley
and Ricardo during the first part of the 19th century and Cassel during
the 1920s) have viewed the doctrine as an extension of the quantity
14
‘4For aa further elaboration on the relationship between
between exchange
rates,
rates, and
and the current
current account,
account, see Dornbusch
Dornbusch and Fischer (1978) and
Rodriguez
special emphasis
Rodriguez (1978).
(1978). For
For aa special
emphasis on the
the role
role of
of innovations
innovations in
in
the trade balance, see Nussa
Mussa (l979c)
(197gc) and for empirical evidence, see
Hakkio (1979b).
-239-
theory of money to the open economy.
By now
now the consensus seems to be
that
that purchasing power parities can be expected to hold in
in the long-run,
lonq-run,
if most of the shocks to the system are of aa monetary origin which
which do
if
not require changes in relative prices.
prices.
To the extent that most of
of the
shocks
“real” changes (like differential growth rates among
shocks reflect "real"
sectors), the required changes in sectoral relative prices
prices may result
in
in a relatively loose connection
connection between exchange
exchange rates and aggregate
price levels.
The experience during the
the 1970s
1970s illustrates the extent
to
to which real shocks
shocks (oil embargo, supply shocks, commodity booms
booms and
shortages, differential productivity growth)
shortages,
growth) result in
in systematic devideviations from purchasing power parities.
As illustrated
illustrated in
in Figures
Figures 2—4,
2-4,
short-run
short-run changes
changes in exchange rates have not been closely linked to
short-run
short-run differentials
differentials in the corresponding national
national inflation
inflation rates,
rates,
particularly
particularly as
as measured by consumer
consumer price
price indices.
indices.
loose link seems
seems to be
be cumulative.
Furthermore, this
As illustrated in
in Figures
Figures 5-7,
As
5—7,
divergences from purchasing power parities,
parities, measured in terms of
of the
relationship between exchange rates and the ratio of consumer
consumer price
indices, seem to persist.
The
The loose link between prices and exchange rates is
is illustrated
illustrated
in Table
Table 66 which reports the
the results
results of regressions of changes
chanoes in
in the
exchange
exchange rates on changes in (wholesale) prices.
As may be seen, for
the Dollar/Pound and the Dollar/Franc
Dollar/Franc exchange rate, the slope coefficoefficients are very close to
to unity; for the Dollar/DM
Dollar/OM exchange rate the
slope coefficient
coefficient is
less close
is less
close to
to unity.
unity.
slope
Furthermore, in
all cases
Furthermore,
in all
cases
the parameter estimates are extremely imprecise.
imprecise.
The results are even
poorer when the wholesale price
price indices are replaced by the cost
cost of
of
living
living indices.
indices.
It
be noted,
noted, however,
extent this
It should
should be
however, that
that to
to some
some extent
this
-24D-240-
Table
Table 66
Relative Purchasing
Purchasing Power Parity;
Relative
Parity; Instrumental
Instrumental Variables
Variables
July 1979
Monthly Data:
Data: June 1973 —- July
1979
(standard errors
errors in parentheses)
(standard
parentheses)
Dependent
Variable
)ependent Variable
iinSt
I:::.
in st
Dollar/Pound
)ollar /Pound
Constant
Constant
6 tn(P /P*)
/P*)
w
w
.999
w
s.c.
s.e.
D.W.
D.W.
.039
1.71
.003
(.005)
(.
005)
(.653)
(. 653)
Dollar/Franc
)ollar/Franc
—.001
-.001
(. 004)
(.004)
.891
(.682)
(.
682)
.030
2.38
Dollar/3M
lollar/DM
—.001
-.001
(.008)
(.008)
1.313
1.313
(2.057)
.036
.036
1.92
1.92
~ote:
iote:
~I:::. in
/p*) denote,
percentage change
change
in SSt and iI:::. Zn(P
tn(P /P*)
denote, respectively,
respectively, the percentage
Ln the spot
spot exchange
rate an’s in
wholesale price
e~change ratewan~
in the ratios
ratios of the wholesale
price indices.
indices,
;.e.
error of the regression.
stage least-squares
least-squares
;,e. is
is the standard
standard error
~Rgression. Two stage
estimation
constant, time,
~stimation method, is
is used; the
the instruments
instruments are
are a constant,
time, time
time
quared, and lagged values
>quared,
values of the dependent
dependent and independent
independent variables.
variables.
—241—
-241-
phenomenon is specific to the 1970s. During the floating rates period
of the 1920s, the doctrine of purchasing power parities
parities seems to
to have
5
been much
reliable.’15
much more reliable.
The discussion in the second
second section
section emphasized
emphasized that
that in periods
The
which are dominated
dominated by “news,”
news, which alters expectations, exchange
exchange
11
11
rates (and other
other asset prices) are
are expected to be
be highly volatile.
volatile.
Aggregate
Aggregate price indices,
indices, on the other
other hand, are not
not expected
expected to reveal
reveal
such aa degree of volatility since they reflect the prices of goods and
services which are less durable and,
and, therefore,
therefore, are likely to
to be less
sensitive
sensitive to the news
news which alters expectations concerning future
course of
of events.
It follows, therefore, that in
in periods during which
there is amp
ample
fluctuations
1e “news”
news which cause large
1 arge fl
uctuat i ans in exchange rates,
16
there will also be large deviations from purchasing power parities.16
II
II
there will also be large deviations from purchasing power parities.
The different degrees of volatility of prices and exchange rates are
illustrated in
in Table 7, which
which reports
reports the average absolute monthly perpercentage changes in the various exchange
exchange rates and prices.
As is evievi-
dent, the
the mean absolute change in the various spot exchange
exchange rates has
been about 22 percent per month (and even slightly
slightly higher for the changes
in the forward rate).
The magnitudes of
of these changes have been more
more
than double the magnitudes of the changes
changes in most of
of the various price
indices, as well as in
in the ratios of
of national
national price
price levels.
For
For example,
example,
15 For evidence see Frenkel (1976, 1978,
1978, 1980b)
198Db) and Krugman (1978).
16Dn this, see
on this,
Mussa (1979a). It is
is noteworthy that the emphasis
been on
on the words large fluctuations;
fluctuations; this should be
in the text has been
contrasted with periods during which
which there are large secular changes
changes in
in
the exchange rate (like the changes which occurred
occurred duang
during the German
sten
hyperinflation). During
During such
such periods the secular changes do
do not stem
necessarily from news
news and need not be
be associated with deviations from
purchasing power
power parities.
purchasing
parities.
11
11
-242-
Table
Table 77
Absolute Percentage
Changes in
in Prices
Prices and Exchange
Exchange Rates
Mean Absolute
Percentage changes
Monthly
Monthly Data:
Data: June 1973
1973 -— July
July 1979
1979
Variable
Variable
ritry
n.try
COL
coL
WPI
k
Stock
Ma~t
Market
Exchange Rates
Against the Dollar
Dollar
~g~inst
spot
forward
spot
coL/c0L~~
COL/COLDS
‘~
.
.009
.007
.037
—
—
.
.014
.012
.066
.021
.021
.007
.007
rice
nee
.011
.009
.054
.020
.021
.003
many
.004
.004
.030
.024
.024
.004
e:
—
All variables
variables represent
absolute values
percentage
represent the absolute
values of monthly
monthly percentage
changes in
in the data.
data.
denotes the wholesale
wholesale price
index and
WPI denotes
price index
coL
COL denotes
denotes the cost
cost of living
living index.
index.
Data on prices
prices and exchange
rates
IMF tape
tape (May
(May 1979
version)..
rates are from the IKF
1979 version)
The stock
stock market
market
indices
International Perspective,
Perspective, monthly
issues.
indices are from capital
Capital International
monthly issues.
-243-
the mean monthly change
change in the cost of living price index was .4
.4 perpercent in Germany, .7 percent in
in the U.S., .9
.9 percent in
in France and 1.2
percent in the U.K.
These differences are even more
more striking for the
detrended series.
exchange rates have been volatile is clearly
clearly
The notion that exchange
illustrated by Figures 2—4
2-4 and by Table 7.
7.
The comparison of the magmag-
nitudes of the changes in
in the exchange
exchange rates with the magnitudes
magnitudes of
of the
in the
the ratios
ratios of national price
price levels
changes in the price indices and in
may suggest, according to aa narrow
narrow interpretation
interpretation of the purchasing
power parity doctrine, that
that exchange rate
rate fluctuations have been
“excessive.”
excessive.
11
11
The
The previous discussion,
discussion, however, has emphasized that exex-
change rates, being the relative prices of assets, are fundamentally
different
different from the price
price indices of goods and services and, therefore,
therefore,
are expected to
to exhibit aa different degree of volatility in
in particular
during
during periods that
that are
are dominated by
by “news.”
"news."
An
An alternative
alternative yardstick
yardstick
for measuring
measuring the degree of exchange rate fluctuations would
would be aa comcom-
assets.
parison with prices of other assets.
while exchange
exchange rate
Indeed, while
changes
changes have been large relative to
to changes
changes in
in national price levels,
they have been
been considerably
considerably smaller than changes
changes in the prices of other
assets like
like gold, silver, many other
other commodities that
that are traded in
in
organized
organized markets, and common stocks.
For
For example, Table
Table 77 also
also rere-
absolute monthly percentage change in stock market ininports the mean absolute
dices.
As may
nay be seen, the mean monthly
monthly change in these indices
indices ranged
from over 33 percent
percent in
in Germany to over 66 percent
percent in
in the U.K.
By
By these
it is difficult to argue that exchange
exchange rates have been exexstandards it
cessively
cessively volatile.
-244-
the characteristics of
of exexThe fundamental difference between the
also reflected
reflected in
in their
their time
change rates and national price levels is also
series properties.
The monthly changes in exchange rates exhibit
little or no
little
no serial correlation while national price levels do
do exhibit
exhibit
aa degree
degree of serial
serial correlation.
The
The serial correlation
correlation of national
national
price levels has been rationalized in recent macroeconomic theorizing
existence of nominal
conin terms of costs
costs of
of price
price adjustment,
adjustment, the existence
nominal contracts, confusion
absolute prices
tracts,
confusion between
between relative
relative and
and absolute
prices and
and confusion
confusion
changes.
between permanent
permanent and transitory changes.
This
This difference
difference between the
time series
series properties of exchange rates and prices
prices is reflected in the
low correlation between the
the practically random month-to-month
exchange
month—to—month exchange
rate changes
changes and the serially correlated differences between national
rates of inflation.
Given the short-run
short—run deviations
deviations from purchasing power parities,
parities, it
is relevant to explore
explore whether
whether these deviations tend to diminish with
with
time or tend to persist or even grow in size.
In order to
to examine
examine the
the
patterns of the deviations, II have computed the autocorrelation funcfunc-
tions and the partial autocorrelation functions of
of these deviations for
the wholesale and the cost of
of living price indices.
The deviation from
purchasing power parities during month ttis
is denoted by
A
6
and is defined
defined
as:
(17)
At
= An St
-
£n(PIP*)t.
12—14 illustrate the
Figures 12-14
the patterns
patterns of the
the deviations for the three
exchange rates.
As may be seen, the
the general pattern is very similar
similar for
the three exchange
exchange rates and
and for the two price indices.
In all cases
the autocorrelation
autocorrelation function tails off at
at what seems to be
be an
an
—245—
-245-
Figure 12
The
from PPP
PPP with
The Dollar/Pound:
Dollar/Pound: Deviations
Deviations from
with Wholesale
Wholesale Price
Price Indices
Indices
..
01,, ··•~"
'
~;_;•~ 'J:1~.1 •• ,
4:9
.hH
..,.
"·"
,.04
..
cir,co,~«
~wr,u,a
,,.,, ' ',... '
~. 02
~
- - - --
_
---~
E~1, SIO <asoa
~J’hr~__~
___ Jc•
~O•
“S
S. S~O.SS
· ·~·
,.,.•c•""•~•
0.00
0.50
.0.0
------~---
~:
-- --.
-—~——‘
-
‘---
•
—
--~----—:——~~-~.—:—--—---—
—~-—--‘
.
-=111=-·-~:-
~.-;...,·:,no;;;;,_.,,o~ ,~~<,,o~- •-,,. ,~. ""'"'
The
Deviations from
cost of
Indices
The Dollar/Pound:
Dollar/Pound: Deviations
from F??
PPP with
with Cost
of Living
Liv~ng Indices
...
""""'~<<
~°~“~Ek~
5005
0000C0S.CL.,00SS
:~.il:N~49:Nf~,1~,~H9i~J~H
!l~~;_ '~:i ...,l-o,
«o ,a~o•
,ck
•o•
u,. sto ,u~•
O!H(O<M<
-.- ···0-.---
11••>
£50.
«1.
o-• ,
—
'""""•--- -
~fl.StSncoSS(S.flSSS
,_
.
,•
------- -"••-~-- ..:___ "•""''- --.:;-,o,o;.,·~.-;,o~-,;;;;,;;;;;;-o;:--,·~; "~"" ·
-246-246-
--
Figure
Figure 13
13
The
from PPP
Pt’? with
Dollar/Franc: Deviations
Deviations from
with Wholesale
Wholesale Price
Price Index
Index
The Dollar/Franc:
...
...
O"rf~H•C<
uo•i •~:~d-o~
2:0
2:22
t:F.
0.05
0.05
0.0’
2:12
tO
2,.I0.0.22.0.40.0.13.0.35—l.32.0.1e.0.
2:22
1:11—2:21-8:11 —1:81
.o.52.0.021,1l
0.22
0.
50’,
OIHUE•C[
000
O•O, II•&
)
£I—0I’I~s0
~02
0.69
-
2
.0.02—0.02
00c051cs,705,,Fosc110s
10
2
— -.
o,56.50510,
0—0
-
5.2
~
-—
500 50 ~0S
‘05 Sb
0.02
-
0,0
—
-
0.5.C0-0.l2.o.t0.0.t3
0.0
—~ ——
0.0
0.0
-— ——---:
—_
—
The
Deviations from
P1’? with
of Living
Living Indices
The Dollar/Franc:
Dollar/Franc: Deviations
from PPP
with cost
Cost of
Indices
,.,.,~,"".. ~,'"~""
01"("'""
0—0)
•
00
0—~1
~
"·~ •> "O"••
" Q5(•U
I
•••
-·. ·--·o ----.; o·-"""
—1:D—188.1:i1—1:1 .8:44 .1:1: :1:81 :1:14
0.03
Ol'H•••u
.,.~,
.8:81 4:4? .1:84
5005
~•,s
)
0.00
0.02
I .155.151.55
0.22
0.12
2,0Ij.s105100000s.TI0s5
--—-
1.21
5.1’
--
050. 500
—
-
>•12
~--- £0’0c0l,,0LbT0c05l1,c5lls100,E,0~it5
•utoi:ol!,iC,, ,o,. ,u.,;1 ,o,. o, ,~, " " ' " -
-----§: -------··--- ·-·--,· ... ·····.
. , ,0
•O,O
•O••
•Ool
~.O
···.
O,<
O.•
.......
O••
•-••·· ........ ------OoO
) ,0
,.
------lt·-~--------
)g- -·-
______ -~~::r---
"·
".
wit_______~--
------
...........
• ••,r:u;:-.ut.o••·L•110k--,,Y.~i,o, o, ,..,",,.-,,-.- -
.
:
—247-247-
I
—
Figure 14
14
Figure
The Dollar/DM:
W71olesale Price
Price Indices
Dollar/OM: Deviations from PPP with Wholesale
o,,.n,~«
•
LOSS
00
.---.-.
DlNOOU
,, •• ,
<1•8
,000C000,soOlO’-S
1:1; 1:14 1:U .1:1? .5:42
- ‘~.27 —0.~ _5.1S’5.1S~S.
~;.;•~ 'tt~oi-o,
lb—Il
5250
‘‘02
)
5.~0’0.2a.I.l0
2.02
1:51
1:11
1:11 .1:81 —1:41 .1:18
—1:18
:U~
.l0’0.15’I.2O’5.29
‘1.10.1,21
2lO0I0S0lT0cl0l_5l_~5l 5005
~0.22
5.o~ 0.2 20.20
55’. 500
,.,.
"" 00500
,.~,,.
'"" ""·
1:14
tU
1.50
—0.12.0.05
~5.O0
—-
,~i.o•
__
B l , 510
550.
___ f ~a a O ~ - - - 0.02
0.2*
o.ti
-i:··-· -•-·--··--·•-·---·•-···---:·-·-·-·--':"-------·-·!: - --·-·- - - - ·-·-------·-·.----·-·
5055! 005 5525(051005
·t
:
.
_!!:
--·--- ---
):.'
:a~-
-- -11- -~=~ C· ·,=
~-
-
I
-~-~
if:
·:
-
--—-
-
-:
"
------~i _______
~- - - - - - - - - - -
The Dollar/DM:
Dollar/OM: Deviations
Deviations iron
from F??
PPP with cost
Cost of
of Living
Living Indices
.
01u;oo<f
Loss
.,
Hoa~ '&:i,.l-o.
-
. ...
0000Cl0IlL00lL~0
.1i~.9ii~t~
1
05’
_~9 ~
d21s0o1L15:__~_
:~:~:~il
O!r,fO<M<:E
-
n-•1 ll-8
,
——
------;?--•·:.:•···••··=~·-•---••·---••;·•·•···•-·:·-·H••·"•'.•·•·· ·•·•• ·•·
i
-
it·
-
-
—
'•"---••---•
—
—
-----
,,.
~
______
,s:_ _ _ _ ________,__;,;~-;·;~;:-~;;;-;;:;;;~:~•7,o~
,;;;;1,,o~ •-·~..;,-.,.,.,
o
.o••.
•••·
•0,0
•O••
O,<
0,0
O,<
0,,
—~
0,0
0,.
·--··········-····---·········· ·•···••··············•· ··················•-*•··-··········-~ - - - - - -
-248—248-
funcexponential rate and, in all cases, the partial autocorrelation function shows aa spike at the first lag.
pattern seems to indicate
indicate
This pattern
night have been expected on the basis of the time series properties
(as might
of
arid price indices)
purof exchange
exchange rates and
indices) that the
the deviations from pur-
autoregressive process,
process.
chasing power parities follow aa first order autoregressive
It
is noteworthy, however, that
that in all cases the value of the autoregresautoreqres-
sion tern
term is about 0.9, indicating the possibility
possibility that the series may
not satisfy the stationarity requirement
requirement.0
To allow for this
possibilthis possibil-
ity, II have also examined the autocorrelation functions and the partial
autocorrelation functions of
of6t -6t-l', i.e.,
i.e., of the first difference
-
of the deviations from purchasing power parities.
pa ri ti es.
The results indicate
indicate
that these differences are serially uncorrelated,
uncorrelated, thus
thus implying that
that
7 In view of this
17
the
the deviations 6t follow aa random walk
walk process)
process.
of this
possibility, II conclude
conclude that the
the deviations
deviations from purchasing power
power
parities seem
seem to follow aa first order
order autoregressive process but
but that
the data do not provide sufficient evidence
evidence to
to reject the alternative
alternative
hypothesis of aa random walk.
Finally, it
it may be
be noted that the maim
main
difference between accepting the AR(l)
AR(l) rather than the random
random walk
hypothesis relates to the economic interpretation of
of the two alternaalternative processes.
implies that deviations from
The random walk process implies
purchasing power parities do not tend to diminish with the passage of
of
time
time while the stable AR(l) process implies that there are mechanisms
mechanisms
which operate to ensure that in the long—run
long-run purchasing
purchasing power
power parities
17
the
17 If
1f the
not entail (ex
of equilibrium
equilibrium
deviations follow a random walk process, then they do
deviations follow a random walk process, then they do
ante) unexploited profit opportunities
opportunities.
For an analysis
0 For
deviations from purchasing power parities, see Saidi
(1977).
-249-
are satisfied.
For
For the
the purpose of forecasting the near
near future,
future, how-
between using the AR(l) process
ever, there is aa very little difference between
with an autoregressive coefficient of 0.9 and using the random walk
process.
-
CONCLUDING REMARKS
CONCLUDING
In this paper I examined some aspects of the operation of flexiflexiWe
ble exchange rates. The analysis was based on the experience of the
lglOs.
1970s.
The principle conclusions which may be
be drawn from the empirical
work are:
(1) In spite of the extraordinary turbulence in the markets for
exchange1 it seems that to aa large extent the markets
foreign exchange,
have operated efficiently.
that in
It should be noted, however, that
the concept of “efficiency”
"efficiency" is somewhat narrow in
this context the
that it only refers to the notion that the markets do not
not seem
enta i 1 unexploited profit opportunities. AA broader perperto entail
spective should deal with
1<ith the social cost of volatility in
terms of the interference with the efficiency of the price
tenns
system in guiding resource allocation, as well as with the cost
of alternative outlets for
for the disturbances that are currently
reflected
reflected in the volatility of exchange rates.
(2) The high
high volatility of exchange rates (spot and forward) rereflect an intrinsic characteristic of the relative price of
monies and other assets. The price of gold and the price of
stocks as well as exchange rates between national monies depend
critically on expectations concerning future course of events,
and adjust rapidly in response to new infonnation.
information. In this
—250—
-250-
perspective the exchange rate (in contrast with the relative
price of national outputs) is being viewed as
as aa financial
variable.
(3) During
During inflationary periods
periods variations in nominal rates of
interest are dominated
dominated by changes in inflationary
inflationary expectations;
as aa result, high
expected to be
high nominal rates of interest are expected
be
associated with
elith high exchange
exchange rates (a
(a depreciated currency).
This relationship was demonstrated
demonstrated within the analytical frameframework
work of
of the monetary
monetary approach to the exchange
exchange rate, and was
supported
supported by
by the empirical work.
In this
this context the key
finding was the dependence of exchange rate changes
changes on the
interest.
changes in the rates of interest.
in accord
This finding is in
current exchange
exchange rates
with the analytical prediction that current
already reflect current expectations about the future while
primarily reflect
reflect
changes in
in the current exchange rates primarily
changes in these expectations which, by
by definition, arise from
new information.
(4) The experience of the
the 1970s does
does not support the predictions
of
power parity doctrine
doctrine
of the simple version of the purchasing power
which relates the values of current
current prices to current exchange
exchange
rates. The empirical
empirical work showed that deviations from purpurpower parities can
can be characterized
characterized by aa first order
chasing power
autoregressive process.
One of
of the
key analytical
analytical insights
insights that
that is
is provided
provided by
by
One
the key
the
the monetary (or the asset
asset market)
market) approach to the exchange
exchange
rate
circumrate is that exchange rates reflect not
not only
only current circumstances
those circumstances
circumstances which
stances but
but also
also those
which are
are expected
expected to
to
-251—
-251-
exprevail in the future. This anticipatory feature of the ex197gb) does not
change rate (which is emphasized
emphasized by
by Mussa, 1979b)
to such
such aa degree) the prices of national
characterize (at least to
outputs.
outputs.
As a result, during
during periods
periods which
which are dominated
dominated by
frequent changes in expectations about
about the future, one may exexdeviations from purchasing
purchasing power
power parities
pect to find frequent deviations
18
when the
the latter
latter are computed using current prices.18
when
recognized by Gustav Cassel
18This phenomenon
phenomenon was
was recognized
by Gustav Cassel -- the
the most
most
recognized proponent of
of the purchasing power
power parity doctrine. Since
this paper was prepared for presentation
presentation on October
October 20, 1979 -- the
the
date of Cassel
‘s birthday (Cassel was born on
Cassel 's
on October 20,
20, 1866) it
it
seems appropriate to
to conclude
conclude with the
the quote that reflects this key
idea.
—-
—-
The international valuation of the currency will, then qenaenerally show aa tendency to anticipate events,
events, so
so to speak,
speak,
and become more an
an expression of the
the internal value that
that the
currency is expected
expected to possess in aa few months, or perhaps
in aa year’s
year's time (Cassel, 1930, pp. 149-50).
-252—252—
DATA
DATA APPENDIX
APPENDIX
1.
Exchange Rates
exchamge rates are end-of-month
end-of—month rates obtained from the
The spot
spot exchange
IMF
version, updated to July
July 1979 using the
TMF tape (May 1979 version,
November
November 1979 issue of the International
International Financial Statistics)
Statistics)
obtained from the International Monetary Fund. The
The forward exexchange
rates are
one month
are end—of—month
end-of-month rates
rates for
for one
month maturity.
maturity. The
The
change rates
forward rates for the U.K. Pound and the DM for the period June
1973 - June 1978
1978 are bid prices obtained from the International
International
(1MM). For
Money Market
Market (IMM).
For the period
period July 1978 - July 1979 they
are sell prices obtained from the Wall Street
Street Journal. The
The
forward rates
rates for
for the
the French
French Franc
Franc for
for the
the period June
June 1973—
1973July 1974 are bid
July
bid prices calculated from the Weekly
Weekly Review
publication of the
the Harris Bank
Bank which
which reports the spot rate and
the
the forward premium; in each case the closest
closest Friday to
to the end
end
of the month was chosen. For
For the period August
August 1974
1974 - June
June 1978
1978
the
the rates are bid rates
rates obtained from the 1MM
IMM and for the period
period
July 1978 - July 1979
1979 they are sell
sell prices obtained from the Wall
Wall
St refil_J.p ur11<1lSSI?ItlolAlmaI.
—
—
-
—
2.
Prices
The wholesale
wholesale and
and cost
cost of living price indices are period
averages obtained from the
the IMF
IMF tape, lines 63
63 and 64,
respectively.
3.
Rates of Interest
1—month Eurocurrency rates obtained
All interest rates are 1-month
obtained from
the Weekly
the Harris
all cases
figures
the
Weekly Review
Review of
of the
Harris Bank.
Bank. In
In all
cases the
the figures
used correspond to the last
last Friday of each month.
month.
4.
Stock Markets
The stock market
market indices correspond
correspond to
to the last trading day of
of
the month. The sources are Capital International Perspective.
Perspective,
Geneva,
Geneva, Switzerland, monthly issues.
issues.
—253—
-253-
REFERENCES
KEFERENCES
“Fixed and Flexible
Artus, Jacque R.
R. and John H. Young, "Fixed
Fl exi b1 e Exchange
Exchange Rates:
AA Renewal of
of the Debate,”
Debate," National Bureau of Economic
Economic Research,
Working Paper
Paper Series, no. 367,
367, 1979.
Rational Expectations
Rate.” In
Bilson, John
John F. 0.,”
O., "Rational
Expectations and the Exchange
Exchange Rate."
0. Johnson (eds.),
J. A. Frenkel
Frenkel and H. G.
(eds.), ]c~9!i~1ofExciiafl9~
The Economics of Exchange
Rates: Selected Studies. Reading,
1978..
Reading, Mass: Addison—Wesley,
Addison-Wesley, 1978
'Speculative
Efficiency' Hypothesis."
‘Speculative Efficiency’
Hypothesis.”
-________.,“The
- -nanuscript,
- -. , "The University
manuscript,
University of Chicago, 1979.
1979.
Unpublished
Cassel , Gustav. Money
Cassel,
Money and Foreign Exchange After 1919. London:
London:
1930.
Macmillan, 1930.
“Exchange Rates,
Clements, Kenneth W. and
and Jacob A.
A. Frenkel, "Exchange
Rates, Money and
Relative Prices:
l920s.” Journal of
Prices: The Dollar—Pound
Doliar-Pound in
in the 1920s."
of
International Economics, 10,
10, 1980, forthcoming.
“Capital Mobility, Flexible Exchange Rates
Dornbusch, Rudiger (l976a)
(1976a) "Capital
and Macroeconomic Equilibrium.”
Equilibrium." In E.
E. Classen
Classen and P. Salin (eds.
(eds.),
Recent Issues in International Monetary Economics. Amsterdam:
Amsterdam:
North-Holland, 1976, Pp. 29-48.
29-48 .
North—Holland,
of Flexible
Flexible Exchange
Exchange Rate Regimes
Regimes and
Theory of
_________
“The Theory
- ~ ~ -. (1976b) "The
Macroeconomic
Policy.” Scandinavian Journal of
Macroeconomic Policy."
of Economics 78, No.2
(May 1976):
1976): 255—75.
H. 0.
255-75. Reprinted in J.
J. A. Frenkel and H.
G. Johnson
Johnson
(eds.), The Economics
Economics of
of Exchange Rates: Selected
Selected Studies. Reading,
Mass.: Addison-Wesley, 1978.
1978 .
__________
Exchange Rate Flexibility,”
Flexibility," in
. , “Monetary
"Monetary Policy Under Exchange
Managed Exchange—Rate
--~Mra-n~aged
Exchange-Rate Flexibility: The
The Recent Experience, Federal
Federal
90—
Reserve Bank of
of Boston Conference Series, No.
No. 20,
20, 1978,
1978, Pp. 90122.
Dornbusch, Rudiger
Rudiger and Stanley Fischer, “Exchange
"Exchange Rates and the Current
Dornbusch,
Account,”
Account," unpublished
unpublished manuscript, Massachusetts Institute of
Technology, 1978.
1978.
Edwards,
Determi"A Simple Monetary
Monetary Model of
of Exchange Rate DetermiEdwards, Sebastian, “A
nation in the Short Run
Run - Some Preliminary Results for the
Peruvian
1950—1954.” unpublished manuscript, University
Peru
vi an Experience 1950-1954."
of
of Chicago, 1979.
—
Frankel, Jeffrey. “On
"On the Mark: The Theory of Floating Exchange Rates
Rates
Differential.” American Economic Review,
Based on Real Interest Differential."
610—22.
69, No. 44 (September, 1979): 610-22.
-254-
Frenkel, Jacob A. “A
"A Monetary Approach to the Exchange Rate: Doctrinal
Aspects and Empirical Evidence.”
Evidence." Scandinavian Journal
.Journal of EcoEconomics 78,
2O0—2~4. Reprinted
78, No. 2 (May
(May 1976):
1976): 200-24.
Reprinted in J. A.
A. Frenkel
and H. 0.
G. Johnson (eds.),
(eds.), The Economics of Exchange
Exchange Rates:
Selected Studies. Reading, Mass.: Addison—Wesley,
Addison-Wesley, 1978.
1978 .
_________
The Forward
. , "The
Forward Exchange
Exchange Rate, Expectations and the Demand
---,f"o_r_'Money:
Hyperinflation." American Economic Review
for Money: The German Hyperinflation.”
653—70..
64, No. 44 (September
(September 1977):
1977): 653-70
. , "Purchasing
Purchasing Power Parity: Doctrinal Perspective and EviEvi--~d-e-nc-e
of International
dence from the 1920s."
l920s.” Journal of
International Economics 8,
No. 22 (May 1978): 169-91.
169-9 .
_________
Demand for Money
- - ~ - ~ ·.,• “Further
"Further Evidence on Expectations and the 0emand
during the German Hyperinflation.”
Hyperi nfl ati on." Journal of Monetary Economics,
5, No. l1 (February, 1979): 81-96.
81—96.
(l980a) “The
(1980a)
"The Forward Premium on Foreign Exchange and CurCurrency Depreciation During the German Hyperinflation,”
Hyperinflation," American
American
Economic Review 67, 1980, forthcoming.
__________
(l980b) “Exchange
(1980b)
"Exchange Rates, Prices and Money: Lessons
Lessons from the
l920s’ American Economic Review, 20, 2 (May, 1980), forthcoming.
1920s"
forthcoming.
__________
Frenkel
Frenkel,, Jacob A. and Kenneth W. Clements, “Exchange
"Exchange Rates in the
l920s: AA Monetary Approach”
1920s:
Approach" in June M. Flanders
Flanders and Assaf
Assaf Razin
Development in an Inflationary World, New York: Academic
(eds.), 0evelopment
Press, 1980, forthcoming.
Frenkel
0. Johnson (eds.), The Economics of
Frenkel,, Jacob A. and Harry G.
of Exchange
Rates:
Addison—Wesley, 1978.
Rates: Selected Studies. Reading, Mass.: Addison-Wesley,
1978.
Frenkel, Jacob A. and Richard M. Levich,
Levi ch, “Transaction
"Transaction Costs and
Interest Arbitrage:
Arbitrage: Tranquil Versus Turbulent Periods.”
Periods." Journal
(December, 1977): 1209—1226.
of Political Economy 85,
85, No. 6 (0ecember,
1209-1226.
“Efficiency of Foreign Exchange
Frenkel, Jacob A. and Michael L. Mussa, "Efficiency
Markets and Measures of Turbulence,”
Turbulence ,t' American Economic Review,
70, 2 (May, 1980) forthcoming.
Hakkio, Craig S. (1979a) “Expectations
"Expectations and the Foreign Exchange Market.”
Market."
Unpublished Ph. D.
University of Chicago, 1979.
1979.
0. Dissertation, University
__________
(l979b) "Exchange
“Exchange Rate and Innovations in the Current
- ~ - - (1979b)
Account,”
Account," Unpublished manuscript, Northwestern University, 1979.
1979.
‘Predictors of
Hansen, Lars P. and Robert J. Hodrick, "Predictors
of Future Spot
Spot
Rates: An Economic Analysis,”
Analysis," Journal
Jou ma 1 of Political
Po 1 i ti cal Economy 88, 4
1980) forthcoming.
(August, 1980)
Hausman, Jerry A. “Specification
Speci fi cation Tests in Econometrics.”
Econometrics. Econometrica,
11
H
46
46 (Novermber,
( Novermber, 1978): 1251—71.
1251- 71.
—255—
-255-
Hodrick, Robert J.
J. “An
"An Empirical Analysis
Analysis of the Monetary
Monetary Approach to
Rate.” In
the Determination of the Exchange
Exchanqe Rate."
In J.
J. A. Frenkel and H. G.
Exchan e Rates: Selected Studies.
Johnson (eds.),
(eds.), The Economics of Exchange
Reading, Mass.: Addison-Wesley,
ison- es ey, 1918.
‘The Exchange Rate and the Balance of Payments in
Kouri, Pentti, J. S. "The
in
the Short Run and in the Long
Approach.”
Long Run: AA Monetary Approach."
Scandinavian Journal of Economics 78, No. 22 (May 1976): 280-304.
Krugman, Paul.
“The Efficiency of the Forward Exchange Market: Evidence
Paul. "The
from the Twenties and the Seventies."
Seventies.” Unpublished manuscript,
Yale University, 1977.
Purchasing Power
---.=cc·, !!Purchasing
Power Parity and Exchange Rates.”
Rates. Journal
Journal of
_________
11
International Economics 8 (August, 1978): 397-407.
397—407.
“Further Results on the Efficiency
Levich, Richard M. "Further
Efficiency of Markets for
Foreign Exchange"
Exchange” In Mama
ed Exchan
Managed
Exchangee Rate Flexibilit
Flexibility:: The
ç~p~iem~,Federal
Experience, Federal Reserve Ban
Bank of Boston,
Boston, Conference
Recent
Series, No. 20, 1978
1978..
. , "The Efficiency
Efficiency of Markets for Foreign
Foreign Exchange,"
~.,“The
Exchange,” In
--~R'u'd~iger
A. Frenkel (eds.), International
—
Rudiger Dornbusch and Jacob A.
Economic Polic
Policy:: Theor
Theory and Evidence. Ba1timoriTJNi~i~Tlb~ldns
Baltimore: Johns Hopkins
University
79.
University Press, 11979.
McKinnon, Ronald I. “Floating
"Floating Foreign Exchange Rates 1973—74:
1973-74: The
Emperor’s New Clothes."
Clothes.” In Karl Brunner and Allan Meltzer (eds.),
Emperor's
~Institutional Arrangements and the Inflation Problem, Vol.
Vol 3 of
the Carnegie—Rochester
Carnegie-Rochester Conference Series on Public Policy, A
A
Monetary
Economics, 1976,
Supplementary Series to the Journal
~
1 offMo
et~yEconomics,
1
pp. 79-114.
.
Mudd, Douglas, R. “Do
"Do Rising U.S. Interest Rates Imply aa Stronger
Dollar?”
Dollar?" Federal Reserve Bank of St. Louis Review, 61, No. 66
(June, 1979): 9—13.
9-13.
Mussa, Michael, (l976a)
“The Exchange Rate, the Balance of
(1976a) "The
of Payments and
Monetary and Fiscal
Fiscal Policy under a Regime of Controlled Floating.”
Floating."
Scandinavian
Scandinavian Journal of Economics 78, No. 22 (May 1976):
1976): 229-48.
Reprinted in J. A. Frenkel and H. G. Johnson (eds.)
(eds.),, The Economics
~
of Exchange Rates: Selected Studies. Reading, Mass.: Addison—
AddisonWesley, 1978.
Experience with
with Fixed and Flexible
(1976b) "Our
“Our Recent Experience
---~change Rates: A
Comment." In Karl Brunner
Brunner and Allan
Allan Meltzer
A Comment.”
ExEx(eds.),
Institutional
Arrenagements
and
the
Inflation
Problem,
Vol.
3
~
Vol. 3 of
of
the Carnegie—Rochester
Carnegie-Rochester Conference Series on Public Policy, aa
-ial of Mpnetaryjçonomics,
supplementary series to the ~p
Journal
Monetary Economics, 1976,
pp. 123-41.
__________
11
Uncertainty: Causes, Consequences, and
Exchange Rate Uncertainty:
Policy Implications.”
Implications." Unpublished manuscript, University of
1977.
Chicago, 1977.
—256—
-256-
- ______
-~~ • ,
7
Mussa, Michael, (1979a) “Empirical
"Empirical Regularities
Regularities in
in the
the Behavior of
of ExExchange Rates and Theories of the Foreign Exchange Market.”
Market."
Vol. 11
11 of the
the Carnegie-Rochester
Carnegie—Rochester Conference Series on Public
Policy, a supplementary
~
supplementary series to the Journal
of Monetary
9-57 .
Economics, 1979, Pp. 9—57.
. (197gb)
(1979b) "Anticipatory
__________
“Anticipatory Adjustment of aa Floating Exchange
7
--~R'a
t-e." Unpublished manuscript, University of Chicago, 1979.
1979.
Rate.”
- _________
- ~ ~ - · (l979c)
(1979c) “The
"The Role
Role of the Trade Balance in Exchange Rate
Dynamics,”
1979.
Dynamics," Unpublished Manuscript, 1979.
“Expectations and Efficiency in
Obstfeld, Maurice. "Expectations
in the Foreign Exchange
Market,” unpublished paper, Massachusetts
Market,"
Massachusetts Institute of Technology,
Technology,
1978.
1978.
"The Role of Trade
Trade Flows
Flows in Exchange Rate DeterDeterRodriguez, Carlos A. “The
A Rational Expectations Approach,”
mination: A
Approach," Unpublished manumanuscript, Columbia University, 1978.
1978.
Saidi, Nasser,
Nasser, “Rational
"Rational Expectations, Purchasing Power Parity and the
Business Cycle”
Cycle" Unpublished manuscript, University
University of Chicago,
Chicaqo,
1977.
Shiller, Robert J.,
J
Do Stock Prices Move Too Much to be Justified by
"Do
Subsequent
Dividends?” Unpublished manuscript,
Subsequent Changes in Dividends?"
University of Pennsylvania,
Pennsylvania, 1979.
1979.
.,“
-257—257—
INTERNATIONAL STABILIZATION POLICY UNDER FLEXIBLE EXCHANGE RATES
H.
Heller
H. Robert He
11 er
Being the only speaker
speaker at this conference to represent
represent the busibusiness sector, II will focus my remarks on the effects of the flexible exexchange
change rate system -- as
as it
it has operated throughout the seventies
seventies -- on
on
-—
the business
business sector.
the topic:
topic:
—-
In particular, II will
will discuss three aspects of
of
First of all,
all, I will address myself to
to the question of
exchange rate system and its actual operation in
whether the flexible exchange
the
economy, and in particular,
particular, the
the years since 1971 have served the economy,
business sector well.
Second, II will offer
offer my views as to
to what policy
policy
changes
operation of
of the
the present
present system.
system.
would improve
improve the
the operation
changes would
Third,
Third, II
will adopt aa longer perspective
perspective and indicate what intermational
international monemone-
tary reforms might improve
improve the
the operation
operation of the
the system.
THE CONSEQUENCES
CONSEQUENCES OF
FLEXIBLE EXCHANGE
THE
OF FLEXIBLE
EXCHANGE RATES
RATES
The operation of
1971 has
of the flexible exchange rate
rate system since 1971
has
entailed a significant
significant increase in costs to
to the business
business sector.
In
particular, there are adverse effects on international trade,
trade, internainterna-
tional capital movements, and
and foreign
foreign investment.
investment.
II will also argue
that the
the increased costs
costs to the private sector were not offset
offset by aa
greater freedom for the policymakers to
macroto pursue more
more appropriate
appropriate macroeconomic stabilization policies or other direct savings realized by
by the
public sector.
Dr.
is Vice
President for
for International
Bank of
Dr. Heller
Heller is
Vice President
International Economics,
Economics, Bank
of
America, NT
NT++ SA, San
San Francisco, California.
California.
-258-
But II should like to make it clear
clear at the outset
outset that there are
But
at present no
no viable alternatives to
to the
the flexible exchange
exchange rate system.
As long
long as there are large differences in inflation
inflation rates among nations,
a fixed exchange
exchange rate system will not
not be viable.
What
What we perceive as
the cost of flexible exchange rates is therefore truly the cost
cost assoasso-
ciated with high and differential inflation rates.
Nevertheless, the
nonflexible exchange rate system does little to
to make
make countries adopt
adopt non-
inflationary policies.
It is in that sense that the flexible exchange
rate system
also been associated with fluctuating exchange rates
systen has also
and the costs thereof.
International_Trade
International
Trade
The thesis
thesis has
has been
been advanced that flexible exchange
exchange rates
rates disdis-
courage foreign trade.
trade.
There
There are several reasons
reasons for expecting aa dampdamp-
ening effect on foreign trade under
under aa system of flexible exchange rates.
First
First of
of all, there
there is simply
simply the
the increased
increased uncertainty of
of exex-
will have to be
be borne by one or the other
other
change rate fluctuation that will
party to
to trade transactions.
It is important to
to note
note that we are
are not
not
involved
game here.
here.
involved in
in aa zero—sum
zero-sum game
While in
in simple arithmetic
arithmetic terms one
party’s gain must
party’s loss,
party's
must be the
the other party's
loss, the increased uncertainty
will affect both
both parties
parties to
to the transaction.
transaction.
As long as people are
risk-averse, there will be aa net
assomet loss because the welfare losses associated with aa 50/50 chance of losing one million dollars are greater
than the welfare gains
gains of
of aa 50/50 chance of
of winning
winning one million dollars.
That’s the reason why we find
corporate presidents wagering
That's
find few corporate
wagering last
quarter’s corporate earnings on aa double or nothing bet on
quarter's
on the
the outcome
weekend’s football game or at
at the roulette wheels in Las Vegas.
of this weekend's
—259—
-259-
Second, while it is possible to engage
engage in
in forward currency trans—
transactions to eliminate the foreign exchange risk, one
one should keep in
in mind
that there is not only aa brokerage cost associated with these transactransactions, but that there exist
exist no
no organized forward markets
markets for the vast
majority of
of currencies
currencies -- especially
especially those of the developing
developing countries.
——
The system
system therefore has
has an
an inherent bias
bias against trade with the develdeveloping countries -- some of which have
have the widest exchange rate fluctuafluctua-—
tions due to their high inflation rates.
rates.
Third,
Third, there is the natural competitive instinct that
that makes
makes the
“maybe II should
businessman think -- "maybe
should wait one more
more week before II cover
cover
—-
in the forward market
market to obtain aa better rate.
Or worse yet, if I
obtains aa better forward rate next
cover this week, and my competitor obtains
week, then II will lose the contract
contract altogether.”
altogether."
If the rate turns in aa
If
the businessman may
may then
then not even bid on
on the
disadvantageous direction, the
contract.
Fourth, there are costs
costs for the individual firm associated with
the necessity to collect information on exchange rates, to ensure that
proper accounting
accounting and legal
legal procedures
procedures are followed, to maintain
maintain staff
to call up the banks, make
make appropriate calculations, keep records, and
perhaps even hire an economist or consulting firm to
to prepare aa foreign
exchange forecast.
All
All these costs are deadWeight
deadweight losses
losses to the private sector as
as aa
whole, because we are playing a zero
zero sun
sum game where
where one firm’s
firm 1 s foreign
exchange gains will
will be another firm’s
firm's foreign exchange losses.
Gone
old days when one merely
merely had to outwit
outwit the central bank
are the good old
mind as to
to how much longer it should attempt
attempt
that could not make up its mind
had long
long ago become unrealistic.
to maintain an exchange rate that had
-260—260—
Instead, the
the private
private sector
sector has
has to
to maintain
maintain all
all the
the required
required ancillary
ancillary
Instead,
services just in an attempt to stay even and not
not to wind
wind up on
on the
the
losing side of
currency seesaw,
of the c1rrency
seesaw.
United States businessmen are particularly
particularly affected by the introintro-
duction of flexible exchange rate because most international trade used
used
to
to be denominated
denominated in
in U.S. dollars.
Now,
Now, only half of
of world trade is
denominated in dollars and half in
in other currencies.
currencies.
In particular, firms entering
entering new markets often find that they
have
have to adapt to local
local conditions
conditions if they wish to penetrate new
new marmarkets.
kets.
National pride of some of the newly developing countries
countries may
also play an important
important role in their insistence to utilize
utilize their own
currency
currency to
to an
an increasing extent.
While
While some empirical studies
studies failed to find an effect of
of flexible
exchange
volune of
exchange rate on
on the volume
of international trade, II find this evievi-
dence hard to
to believe.
The simplest
simplest of all possible
possible calculations show that foreign trade
increased at an annual rate of 8.8 percent in real terms during 1963-
73, while
while the rate of increase in 1973-75 amounted
amounted to only
only 4.3 percent
per annun.
annum.
That
of increase in the volume of
That is,
is, the rate of
of internationinternation-
al trade was cut in
in half under the flexible exchange rate system.
While it
it is true that this
this does not imply that
that the flexible exchange
rate system caused this
this decline——
decline.-- and II will have to say more on that
that
topic later -- it is certainly
certainly true that the flexible exchange rate
——
system did not prevent
prevent the decline in the trade volume either.
Finally, II need not point out
excharge
out that the myth that flexible exchange
rate would always
always balance
balance our international trade is nothing but that
-- a myth.
-—
People who
who drew the
the opposite
opposite conclusion from textbooks
textbooks in
in
—261—
-261-
international
international economics
economics forgot to read the fine print:
print:
namely that it
was assumed that there were no
no international capital movements.
Only
in such aa world can perfect purchasing power parity hold
imhold and will
will imports
ports automatically be
be balanced by exports of equal value.
movements are very
very much with
with us,
us, and
In the real world, capital movements
they are not so much determined by actual
actual international price differendifferentials, but by the expectation of price level
level changes at
at some
some future
date.
Capital Movements
Movements
Capital
This brings me to the second
second major point to be
be covered: the imimpact of
of flexible exchange
exchange rate on international capital movements.
First of all,
all , it is clear by
by now that international capital movemovements have not served
served as
as the great stabilizer
stabilizer of exchange
exchange rates that
they are supposed to be.
be.
According to theory, well-informed
well-informed private
speculators will act to
to stabilize
stabilize the exchange rate, buying the currency
currency
when it is low and selling it
it when
when it is high.
informed
informed speculators?
But who are these
these well—
well-
The actors
actors with the greatest amount of
of expertise
in the area, the large cormiercial
commercial banks, are highly reluctant to
to take
open foreign exchange
exchange positions.
tion-oriented.
tion-oriented.
They
They are trade—oriented,
trade-oriented, not
not speculaspecula-
Much
Much more money is to be
be made
made by
by actively trading in
earning a small
small spread on each transaction than
than by
the market, and earning
maintaining
maintaining an
an open
open position and hoping for the best.
The situation
situation is
not
of aa grocery
store owner,
owner, who
not unlike
unlike the
the one
one of
grocery store
who makes
makes his
his money
money buying
buying
and selling tomatoes, and not by hoping
hoping to
to make a killing
killing in
in the market
market
when the tomato crop
crop in
in Mexico goes sour and the price skyrockets.
U.S. Treasury data
data show that on
on average
average U.S. commercial
commercial banks were
—262—
-262-
holding less than $100 million in
in open foreign exchange
exchange positions.
This
This aggregate
aggregate amount for all U.S. banks is not much
much larger than some
of the individual transactions foreign exchange traders are called
called upon
to execute.
This leaves private corporations and
and individuals as
as the potential
market stabilizers.
take foreign exchange posiposiWhile corporations do take
they are typically
typically designed to
to offset some
some commercial transactransactions, they
tions rather than as
as aa deliberate
deliberate attempt to take an
an open
open position.
The corporate
corporate treasurer who attempts to make aa career out of
of realizing
foreign exchange profits
profits is aa rare, and
and probably short-lived, breed.
more properly
properly be described
described
Instead, the typical corporate strategy can more
as one
one of foreign exchange
exchange loss-avoidance
loss-avoidance rather than of foreign exexchange speculation.
The final group -- private individuals
individuals -- is certainly
certainly increasincreas——
ingly
ingly active in
in the market.
market.
--
They are probably
probably more active in the oror-
ganized non—bank
(Internanon-bank foreign exchange
exchange markets, such as
as the 1MM
IMM (International
tional Monetary Market, aa division of the Chicago Mercantile Exchange)
Exchange)
than
than in the commercial bank market.
Prior
Prior to the Herstatt
Herstatt calamity,
calamity,
private speculators had access to the bank
bank market
market largely
largely through
small
small banks.
banks.
Since 1974 most major banks have reduced the foreign
exchange
exchange lines made available to
to the smaller
smaller banks,
banks, thereby sharply
sharply
limiting their access
access to the interbank market.
Consequently, most
Consequently,
individuals are active in the 1MM
IMM and the New York exchanges.
As aa
it may
may be
be said
said that these exchanges
exchanges are equal to
rough generalization, it
inthe transactions carried out by one major U.S. bank
bank as
as far as its in-
fluence on the market is concerned.
-263—263—
Considering all
all this, it
it still
still remains true that
that aa speculator is
is
able to make
make profits
profits more consistently by running with the markets
rather than by taking
taking aa position and hoping for a turn-around in
in the
the
market.
To try to pinpoint market turn-arounds is exceedingly
exceedingly diffidiffi-
cult as everyone who
who has tried his luck at it knows.
The upshot is that the herd
herd instinct in
in foreign exchange markets
markets
still very powerful and the
is still
the well-informed speculating loner is the
exception rather than the rule.
Consequently, speculative
speculative activity
activity may
well accentuate rather than
than reduce exchange rate fluctuations.
Investment
The uncertainty surrounding the exchange value of the currencies
has also taken its toll on
on the willingness of investors to
to engage in
in
foreign direct investments and in long—term
long-term construction
construction activity
abroad.
Increasingly, foreign countries
countries insist on
on denominating long—term
long-term
construction contracts in their own currency,
currency, forcing the American
businessman to shoulder
shoulder the foreign exchange
exchange risk.
risk.
Foreign direct inin-
take five or
or even
vestment and long-term construction projects that may take
ten years to complete are particularly affected by
by the
the exchange rate
uncertainty because
because there are no
no organized
in which
organized forward markets in
such long—term
may be hedged.
hedged.
long-term exposures
exposures may
In addition, many of these
projects
are located
countries for
not even
projects are
located in
in countries
for whose
whose currencies
currencies not
even regular
regular
forward markets or capital markets
markets exist, thereby making hedging an imimpossibility.
Under such circumstances the only options open to the
businessman are
are to
to assume the foreign exchange
exchange risk or to forget about
the contract.
-264-
Foreign investment activity is
is also
also greatly complicated by
by
Foreign
changing currency values.
What might be
be aa profitable
profitable foreign operation
What
at
at one
one exchange rate may rapidly become unprofitable
unprofitable as
as the foreign
foreign exex-
change rate changes.
In addition, arbitrary
accounting rules -- such
th-bitrary accounting
—-
have significant
significant impact
impact on aa firm's
and loss
loss
as FASB 88 -- may have
firm’s profit and
——
position regardless
regardless of
of the profitability of the underlying
underlyinq manufacmanufacturing activities.
At best, the effects of
of exchange rate changes on
on
the balance sheet make
profitamake it much
much more difficult to evaluate
evaluate the profita-
bility of the investment.
investment.
At worst,
worst, it leads to erroneous
erroneous investment
decisions
and ultimately
ultimately aa retreat from international activities.
decisions and
The Public Sector
The question arises whether the additional costs imposed
imposed upon the
The
private sector of the economy are counterbalanced by benefits to
to the
public sector of the economy.
While
While this
this is
is aa difficult
difficult question
question to
answer, II believe that it must be
be answered in the negative.
an economic
economic
Benefits may accrue to the economy by the creation of an
about aa greater
greater freedom to pursue approapproenvironment that would bring about
priate economic
economic policies, foster higher growth, or lessen inflationary
inflationary
pressures.
On all these counts the actual
actual experience with flexible
flexible exex-
has been discouraging.
change rates has
Of course,
course, the ultimate proof of
Of
any of these propositions is
is impossible to
to attain.
It
It would require aa
under a fixed exchange
exchange rate regime -- and that is
is
replay of history under
-—
clearly impossible.
Economic inference
inference makes it
it also difficult
difficult to see why aa floating
exchange rate regime should be
be characterized
characterized by high growth and little
inflation.
inflation.
The fundamental
fundamental point is that the flexible exchange rate
—265—
-265-
system does not lessen the balance of payments
payments constraint -- it merely
——
changes its nature.
:hanges
It is difficult
difficult to decide whether aa loss
loss of foreign exchange rereserves
serves or aa fall in
in the foreign exchange rate provides
provides aa more rigid
policy constraint.
constraint.
policy
But while the loss of foreign exchange reserves
under aa fixed exchange rate system provides not only aa self—limiting
self-limiting
constraint in that no country
country has either unlimited
unlimited reserves or unlimitunlimit-
the loss of reserves
reserves eventually
eventually
ed access to international credit, the
more restrictive monetary policy which
which will
forces the adoption of
of aa more
tend to bring the country in
in line with the global inflation rate.
rate.
Flexible exchange
self—limiting properexchange rates do not have such self-limiting
proper-
it has
has instead been suggested that the depreciation of aa curcurties, and it
rency
circles where curcurrency may well lead to the development of vicious circles
rency
of its immediate
rency depreciation brings about more inflation because of
impact on
on the price of imported
imported commodities.
rekindled inflationinflationThe rekindled
may force aa further depreciation and so on.
ary forces in turn may
While the statistical evidence
evidence on the validity of this theory is
is
far from complete and doubtful, it stands to reason that aa fixed exexequalizer of international inflation
inflation
change rate system operates as an equalizer
accentudifferentials, while
while a flexible
flexible exchange rate system tends to
to accentu-
ate inflation differentials.
As far as the international businessman is concerned, it is
is clear
which one constitutes the more attractive
attractive environment:
environment:
given aa choice
between similar -- even if high -- inflation
inflation rates in all countries and
between
——
——
an environment or widely divergent inflation rates,
rates, the
the businessman is
an
likely to
to choose the former one.
-266—266—
which one of the two alternatives is
is
However, it is questionable which
best for all people of the world.
For the central
central banker,
banker, floating rates do not seem to have
brought aa more relaxing lifestyle either.
Gross foreign exchange marmar-
ket intervention
intervention on behalf of central
central banks amounted to aa record of
of $72
billion
1979.
billion dollars in the half year ending
ending July
July 31,
31, 1979.
To
To put this
this
number into
into proper perspective, let us remind ourselves that the total
foreign exchange reserves of all countries in the world totalled the
sane amount in 1971, the last year of
same
of the
the fixed exchange
exchange rate system.
Increasing, rather
rather than
than less, official intervention
intervention has
has been the hallhallmark
mark of the flexible exchange rate system in
in the seventies.
The International
International Monetary System
System
The exchange
exchange value of the dollar against the DM
OM (deutsche mark) or
or
such aa
SFR (Swiss franc) has been
been cut in
in half over the last decade. That
That such
world’s leading
precipitous decline
decline in
in the value of
of the world's
leading reserve currency
cannot be without impact
impact on the role of this
this currency in the world
world and
and on
the international monetary system itself goes without saying.
official foreign exA
A superficial glance at the percentage of official
exof dollars shows that the market share
share
change reserves
reserves held in
in the form of
of the
the dollar has remained virtually
virtually constant
constant at 80 percent.
percent.
these figures are -- in
in my opinion -- highly misleading.
misleading.
-—
—-
However,
While
While high
U.S. Treasury officials
officials have argued that the dollar purchases
purchases on behalf
behalf
of foreign central banks were
were proof of their
their confidence in
in the U.S.
dollar, it
it is probably more appropriate to argue that these official
dollar purchases were largely the result of
of intervention
intervention designed
designed to
stop an
an even further slide of the dollar.
-267-
The
The foreign central banks
were
the
the reluctant
reluctant victims of aa declining dollar and not the exuberant
investors
investors they are made out to be.
In fact, foreign central banks of floating currency countries
have reduced the share of dollars in their foreign exchange reserve
portfolio from over 90
90 percent in 1970
1970 to less than 75 percent in 1976.
So have the central
central banks of countries other than the main
main industrialindustrialized countries, who
who acquired the dollars
dollars as aa result of their interintervention policy.
In other words, those central
central bankers that were free
to consider the dollar
dollar as a portfolio
portfolio investment instead of am
an interintervention currency did in
in fact switch away from the dollar.
The decline of the dollar in official foreign exchange
exchange portfolios
was also masked to some extent by the even faster decline of
of the BritBritish
ish pound in international significance.
Central banks have switched
pounds and purchased
purchased DM
OM over the last decade, so that the posiout of pounds
position
tion of
of the pound is now held by the mark.
mark.
It
It stands to reason that
central banks would
DM anyhow, and had
would have
have wanted
wanted to
to acquire OM
had it not
been for the fact that the pound was even
even weaker than the dollar, the
certainly have been more pronounced.
switch out of dollars would certainly
In addition to the decline in the value of
of the U.S. dollar,
dollar, there
are other
other reasons that make it attractive for central banks to diverdiversify their
their foreign exchange
exchange portfolios
portfolios to an increasing extent.
First
it is clear that aa diversified currency portfolio increases
increases its
its
of all, it
overall stability.
Second, as exchange
exchange rates fluctuate
fluctuate it may be prupru-
one’s trading
currency of one's
trading partners.
dent to hold reserves in the currency
Third, the same argument applies
applies to the denomination of the currency
in which the country’s
country's external debt is denominated.
connecIn that connec-
tion it
it is
is important
important to
to note the very rapid swing away from
-268-268-
dollar-denominated international bond
bond issues in recent years.
dollar—denominated
In 1976
In
the value of
still
of dollar-denominated international bond issues
issues was still
three times as large as the value of OM bonds, but by 1978 the OM
volume was equal to the dollar volume.
Consequently,
Consequently, the need to
make amortization
amortization and
and interest payments in
in marks will
will continue to
increase
desirability of holding
increase in the future and
and with it the desirability
holding marks
as
as liquid reserve assets.
We may therefore
therefore conclude that:
one, the flexible exchange rate
system
system has been
been associated with
with aa significant
significant increase in costs to the
private sector; two, that
that it
it has not brought
brought about aa climate
climate for the
conduct
conduct of more effective
effective stabilization
stabilization policies; three,
three, that
that it has
not decreased the cost
cost of intervention
intervention for central banks; and four,
it has fostered the decline of
that it
of the dollar as
as the world’s
world's leading
currency.
II will now consider several
efseveral measures that might improve the ef-
fectiveness of stabilization policies under the flexible exchange rate
rate
system.
IMPROVING THE OPERATION OF
OF THE FLEXIBLE
FLEXIBLE EXCHANGE
EXCHANGE RATE SYSTEM
At the present
present time, there exists no viable alternative to
to the
flexible exchange
exchange rate system.
simple:
The
The main reason
reason for this conclusion is
is
as long as differential inflation rates among countries
countries prepre-
vail, it is not possible to
to impose or to
to achieve exchange rate stabilstability.
The framers
framers of the new
new Article IV of
of the IMF
IMF (International
(International MoneMone-
iclesof
cement were fully aware of this
tary Fund) ~
Articles
of Agreement
this point:
point:
exex-
change rate stability cannot be achieved without internal
internal stability
stability in
economies.
the relevant economies.
To blame the flexible exchange rate system
system for
—269—
-269-
especially by the private
private
the additional costs that have to be borne -- especially
—-
sector -- would be to blame the messenger
messenger for the bad news.
——
Nevertheless, there are certain improvements in the operation
operation of
exchange rate system that can be
be made in order
order to
to enhance
enhance
the flexible exchange
its effectiveness and to reduce the costs associated with it.
it.
These
are the lessons we can learn from the experience gained during the
the
seventies to
to enhance the operation
operation of the international monetary system
during the eighties.
It will be convenient to group the suggestions into two broad
categories:
those pertaining to improving
improving U.S. monetary
monetary and exchange
exchange
rate policy and those relevant for the international monetary system.
Possible U.S. Policy
Policy Improvements
It should be
be feasible to improve
improve U.S. monetary and exchange rate
policy with aa view
towards enhancing the stability
stability of
of exchange rates.
viewtowards
The first set of suggested steps pertains to
to the conduct of U.S.
U.S.
monetary policy, and it is gratifying that
that the Federal Reserve has alalready announced
announced the adoption of monetary targets and their supremacy
supremacy
over interest rate targets.
The experience
experience of having to chase
chase the marmar-
higher and
summer of 1979
ket interest rates higher
and higher during the summer
1979 while
real interest
interest rates remained
remained negative
negative and the money
money supply grew out of
of
control was an important
deciimportant factor in influencing the
the October
October 1979
1979 deci-
sion to use bank reserves instead of the Federal Funds rate
rate as an imimmediate operating target.
Of
Of course, both the Federal Reserve and
and the other market participartici-
pants will have to gain experience and confidence in
in the operation of
of
the new system to ensure its proper
proper functioning.
-270-
In
In that connection it
it
is
somewhat
syssomewhat disconcerting to note that the introduction of
of the
the new sys-
tem was
was not
handled in
implementation as
tem
not handled
in aa fashion
fashion designed
designed to
to make
make its
its implementation
as
smooth as possible, but was conducted in an abrupt and disruptive fashfash-
ion that resulted
resulted in the introduction of uncertainty, confusion over
Federal Reserve, and thereby greater market
market ininthe intentions of the Federal
stability
symptoms that
stability -- the very
very symptoms
that the Federal Reserve action should
——
have helped to
to alleviate
alleviate rather than
than to foster.
Nevertheless, the overall thrust of the new policy is good, and
once the dust has settled the
the targeting on the monetary aggregates
should prove to be a significantly
significantly better system
system than the interest—
interesttarget approach used in the past.
The operation
operation of the system could be further enhanced
enhanced by the anannouncement of
of intermediate range monetary targets as guideposts for the
Federal Reserve.
Such three
three to
to five—year
five-year targets could be very helpful
helpful
in signalling to the private sector the
the clear
clear intention
intention of the Federal
levels and
Reserve to reduce monetary growth rates to non-inflationary levels
to provide
provide aa framework for orderly and sustained
sustained economic
economic growth.
Of
Of
course,
course, such targets must be strictly adhered to, so that confidence
confidence
in the policy
of the authorities will
will be
be enhanced.
policy statements
statements of
To use
use
the
the announcement
announcement of official targets to influence expectations without
appropriate follow—through
follow-through and implementation merely creates aa climate
climate
in which all policy pronouncements
pronouncements will be doubted and will therefore
therefore
become less and less
less effective.
it is also important to have a realistic
realistic monemoneIn that connection it
tary growth target supported by
by a coordinated fiscal strategy.
To anan-
nounce aa reduced
monetary growth target while the public sector borrowreduced monetary
borrow-
expected to increase
increase drastically might not
ing requirements are expected
-271—
-271-
policy package in that context.
constitute aa credible policy
policy
Monetary policy
cannot work in isolation
isolation and must be seen as one ingredient
ingredient in
in aa cocoordinated
ordinated poliãy
policy package aimed at achieving
achieving economic stability.
erThe central bank can also play an important role in reducing er-
ratic exchange rate fluctuations as
as the November 1978 policy actions
showed.
There is aa significant difference between intervention to
maintain an
an exchange
exchange rate
rate that has
has become unrealistic, and interveninterven-
tion to turn around aa market trend that has become disequilibrating.
Central banks have now learned the lesson that there
there is
is little
little to be
Central
gained by trying to maintain
maintain an unrealistic exchange
exchange rate.
Not only
only
Not
are the foreign exchange losses incurred staggering, but the domestic
such ill—advised
ill-advised intervention are
are also disadvantageous.
disadvantageous.
consequences of such
A
exchange markets to
A central bank that sells its currency in foreign exchange
keep it from appreciating increases the monetary base by providing
providing more
more
keep
of its own currency.
This in turn
turn increases inflationary
inflationary pressures
pressures at
aa later date, thereby leading to
to domestic instability.
Similarly, aa central bank that depends on
on unrealistically
unrealistically high
losses
exchange rates will soon find that the foreign exchange reserve losses
drastic exchange
are staggering and will be forced to permit a more drastic
exchange controls with
rate realignment at a later date
date or
or to impose exchange
all undesirable consequences
consequences attached to
to such
such measures.
In contrast, central
central bank intervention to turn the foreign exexIn
change market
market around and to end a trend that has
has clearly
clearly become destadesta1978 U.S. policy
bilizing can be
be highly successful as the November 1978
measures showed.
The
The essential ingredient to
to the success of
of such an
intervention policy is the simultaneous adoption of domestic monetary
monetary
root cause of
of the
the exchange rate
rate
policy measures
measures that attack the root
—272—
-272-
movement.
It will be recalled that from November 1978 until April 1979
there was
was virtually no
no monetary
monetary growth in the U.S.
This was taken by
the markets as aa signal that the Federal Reserve was prepared to pursue
aa tight, anti-inflationary
anti-inflationary monetary
monetary policy and the dollar remained
stable during that period.
In April
April 1979 the money
money supply again began
to
excessive rate, driving up interest rates, increasing
to grow
grow at an excessive
increasing ininflationary pressures,
pressures, and
and bringing
bringing the dollar under renewed pressure,
pressure,
thereby necessitating the November
1979 policy actions.
November 1979
Possible Improvements in the International_Monetary_System
International Monetary System
It
It is my belief that the world
world economy could
could function quite
quite well
global unit
under aa dollar standard, where the dollar is the
the dominant
dominant global
of account, transaction currency, and store of
of value.
An indispensable
indispensable
precondition for the functioning of such aa system is the unquestioned
stability of the dollar in terms of real
real purchasing power.
power.
Oomestic
Domestic
inflation and the accompanying erosion of
forof the
the currency’s
currency's value in foruneign exchange and international commodity
commodity markets will have the un-
avoidable consequences of reducing the dollar’s
dollar's international role.
The British inflation and decline in the value of the pound resulted in
the elimination of
of that currency from the reserve currency status that
it once enjoyed.
Continued double-digit inflation in
in the United States
will undoubtedly bring about the demise
demise of the dollar as
as aa reserve curcurrency as well.
It is
is up to the United States to get its own house in
in order
order if
if
It
she wishes to preserve
preserve the international
international stature
stature of the dollar.
The
The
benefits flowing to the international community
community as aa result of such
such acaction would undoubtedly
undoubtedly be great.
great.
-273—273—
The most likely alternative to aa dollar standard is at present
multiple-currency reserve standard, where several
the development of aa multiple-currency
currencies, in addition
addition to the dollar,
dollar, will serve an international
role.
However, it
it should be realized that such aa multiple-currency
standard is inherently unstable and is likely to lead
lead ultimately
ultimately to
to
severe financial and economic
economic disturbances.
For the same reason
reason that
bimetallism proved to be
be unstable,
unstable, it
it will be
be found that relatively
small differences in national
national inflation rates among the different
different key
currencies will lead
lead to relatively large shifts in
in capital flows among
these countries.
countries.
Such capital flows will exacerbate balance of
of paypay-
difficulties, as capital is likely to flow into aa country
country that
ments difficulties,
already enjoys aa current
current account surplus.
exchange rate
rate
Consequently, exchange
movements will be accentuated, official intervention will
will have
have to bebecome even larger, or capital controls will have to be
be introduced.
UlUl-
timately, it is likely that capital controls
controls cannot
cannot be avoided,
avoided, and the
very benefits
benefits of aa liberal international
international financial order
order will be
be
destroyed.
multiple—currency
The only feasible realistic
realistic alternative to a multiple-currency
SDR (Special Drawing
system is at present aa system based on the SOR
Rights).
The recent decline of the dollar
dollar has
has consequently
consequently led to
to aa
renewed interest in the SDR
SOR as an
an international reserve asset.
This
This
turn of events
1968
events is not without irony, because the
the SDR
SOR was born
born in 1968
out of fear that there might be
be aa shortage
shortage of dollars when
when the U.S.
balance of payments would return to surplus.
Instead, the SOR
SDR is now
likely to assume aa larger role on
on the international economic
economic scene bebeperceived surplus of dollars.
cause there is aa perceived
-274-
The renewed interest in
a dollar/SDR
dollar/SOR substitution account is
is the natural outgrowth of these
developments.
To move
SDR firmly to the center of the international monemove the SOR
monetary system would
would require at least three steps:
to base
exbase the IMF ex-
SDR, to make the SOR
SDR usable -- that
clusively on the SOR,
that is transferable
-—
SOR inflation—proof.
inflation-proof.
among private entities, and to make the SDR
Let me
elaborate on
briefly elaborate
on each one of these
these points.
Recently, the Economic Counsellor
Counsellor of the IMF, Mr. J.
J. J.
J. Polak,
set forth a plan
plan that would make the SOR
SOR the centerpiece
centerpiece of all IMF
operations.
would signifiThis innovative and farsighted suggestion would
signifi-
cantly
cantly enhance the importance of the SOR
SOR and make it a more central
asset in the international monetary system.
In addition,
addition, such aa move
would
would also have the advantage of
of unifying
unifying many of the Fund
Fund operations
that are now proliferated among an ever larger
larger and more
more complex variety
11
of 11“accounts”
accounts'* and “facilities.”
facilities. 11
Second, the SOR should be made
made transferable among private as well
as public holders.
When the SOR
SDR is freely traded in
in international fifi-
nancial
nancial markets its
its usefulness
usefulness and
and liquidity
liquidity will
will be
be greatly
greatly enhanced.
The SOR
is not
not likely
to assume
The
SOR is
likely to
assume aa significant
significant role
role in
in world
world financial
financial
markets until
transactions that
until it is also used widely for comercial
commercial transactions
create aa need
need to
to effect payments in
in SORs.
But if SORs are
are not freely
transferable between private and official holders, it is unlikely that
they will assume
assume an
an important role
role as an international means
means of paypay-
ment.
Transferability
Transferability of the SOR among private parties is therefore
therefore
essential
international monetary system is
essential if the international
is to be
be based firmly on
the SOR.
SOR.
—275—
-275-
SOR should be turned into a true global standard of
Third, the SDR
inflation—proof.
value by rendering it inflation-proof.
Traditionally, gold has
has fulful-
acfilled the role of an
an international standard of value but official actions and the recent speculative fever have
have deprived
deprived gold of its
its status
as aa stable
stable measuring rod.
Instead, it has become a highly speculative
speculative
commodity.
As presently constituted, the SDR
SOR offers some protection against
the risk inherent in
in differential inflation
inflation rates by providing
providing the
basket.
holder with a diversified currency basket.
But it
it should be
be noted
noted that
the value
value of this currency
currency basket in
in terms of real purchasing power dedeteriorates along with the weighted average of the inflation
inflation rates exexSDR basket.
perienced by
by the sixteen
sixteen countries represented
represented in
in the SOR
AA
superior inflation
inflation hedge is always available
available to
to the investor
investor -- be
be it aa
——
monetary authority or
or aa private entity
countries.
of high inflation countries.
——
holding the currencies
by not holding
The SDR,
SOR, as
as presently
presently constituted,
constituted, forces
inflainvestor to
to accept the depreciating currencies of
of the high
high inflathe investor
tion countries that do not enter the SOR interest rate calculations
based on the five most important currencies only.
Hence, the SDR
SOR as
presently constituted is not aa particularly
particularly attractive asset.
inflation—proofing of the SOR would
The inflation-proofing
would make it
it a truly superior
international
international asset that could
could play an increasingly important role
role on
on
the world financial scene by
by providing
providing a universal unit of account, a
monetary
of value.
monetary transaction medium, and aa stable store of
Such an inin-
flation—proofing of
flation-proofing
of the SDR
SOR could be accomplished by
by linking it to a
price index of the sixteen countries
countries making up the SOP
SOR basket.
Of
course, there are many operational problems to be considered, but
but these
-276—276—
more complex than those that had to
to be
be resolved when
are not inherently more
the SOR as currently constituted
constituted was created.
Of course, there remains a very disturbing thought:
all the
the
If all
individual
necessary
individual countries are unwilling or unable to take the necessary
steps to bring inflation under control, why
why should we
we assume that all
these nations acting in concert through an institution would
would be any
more willing
willing or
or able to act in
in aa manner that would expose
expose their
their own
own
shortcomings7
shortcomings?
Nevertheless, it may be
be possible
possible to achieve an
an interinter-
national consensus on the creation of such an
an asset
asset because the alteralternative of continued
continued international monetary disruption is
is associated
with high
high costs for all.
The only other
other feasible alternative for the eighties is a rapid
inflation rate, such
such that
that the international role
reduction in the U.S. inflation
of the dollar will be preserved
preserved in
in the decades to come.
Without aa
Without
stable dollar that can serve as the anchor of the international monemone-
tary system
system there is
is not likely to
to be
be exchange
exchange rate stability.
The
The
elimination of inflation in the U.S. and in other countries will theretherefore
fore be aa precondition for the
the improvement of the operation
operation of the
international monetary system.
Stability cannot be
be imposed by the
international monetary system or found by manipulating
manipulating the system.
International monetary
International
monetary and exchange rate stability
stability can be
be achieved
achieved only
by first attaining domestic stability.
by
-277—277—
FLEXIBLE EXCHANGE RATES AND MONETARY POLICY:
AA DISCUSSION OF THE
THE FRENKEL AND HELLER
HELLER PAPERS
PAPERS
David Laidler
Laidler
If aa conference such as this one,
one, dealing with United States’
States'
macro—stabilization
macro-stabilization policy, had been organized ten years ago it is unun-
likely that anyone would have
have suggested devoting an entire session to
to
the operation
operation of the international
international monetary
monetary system.
If
If the
the suggestion
suggestion
had been
been made, it would certainly
certainly have been
been greeted
greeted with aa loud “why?”
"why?"
The very fact that this session is included
included in
in this conference epitoepitomizes the most important lesson of
doof all that we have learned about domestic stabilization
stabilization policy in the last
last decade-—namely
decade--namely that
that it
it cannot
sensibly be discussed without explicit reference
reference to the international
environment within which it
it is being implemented.
By this II do not simply mean that United States
States domestic
domestic policies
theworld
have implications for the rest of the
world that policymakers should be
interested in, or that there are interesting debates about the organiorgani-
zation of
of the international
international monetary
monetary system,
system, the outcome of
of which
which will
influence the ease with which American business
business can operate
operate in
in interinternational markets and which ought therefore
therefore to
to concern American policypolicymakers,
makers, though both
both of these observations are surely true.
Rather II
mean that the way
way in which monetary policy impinges
impinges upon
upon traditional
Or. Laidler is
Ontario. This
grant from the
Canada.
Canada.
Professor of Economics at
at the University of Western
Professor
on work completed under aa research
paper draws heavily on
Social
Social Sciences and Humanities Research Council of
-278-
domestic targets,
targets, employment,
employment, prices
prices and
and the
the like,
like, is
is intimately
intimately linked
linked
domestic
to
to the operation of the international monetary
monetary system.
of the papers
papers that II am
am discussing has much to say
Since neither of
explicitly
explicitly about these domestic
domestic matters,
matters, and that is not to criticize
criticize
either of them, because one can on~ysay
on'.y say so
so much in one paper,
paper, II bebe-
lieve that it will be useful for me to use
0s2 these discussant’s
discussant's comments
comments
to explore this
this area in the
the light of the arguments presented by Frenkel
to
and Heller, rather than to provide a detailed critique
critique of those arguarguments.
Both of
of the papers
papers before us deal with the operati
operation
Both
on of aa system
exchange rates.
of flexible exchange
That is only right and proper, given that
is the system (more or less) under which the world is currently
currently
this is
operating.
However, I believe that it would be
be wrong for anyone to
to
conclude
international factors for United
conclude that
that the new importance
importance of international
States
States domestic policy stems from the
the adoption
adoption of aa system
system of flexible
exchange rates per se.
se.
exchange
In the 1950s
l95Os and 1960s,
l96Os, United States policymakers were able to
operate "as
“as if”
beif" the economy they were
were dealing
dea 1i ng with
with was closed, not because
cause the Bretton Woods system was aa fixed rate system, but because
because it
was a dollar standard system.
As we now
now know, with the
the benefit of
of
hindsight, and as some--notably,
some——notably, for example, Robert Triffin (1961)-(1961)——
argued at
at the time, this did not mean that
that the United States could
indefinitely operate
operate its domestic policies while completely ignoring
countries used to be called “the
"the balance of payments
payments conconwhat in
in other countries
straint.”
straint."
“constraint” operated sufficsufficHowever it did mean that the "constraint"
which domestic
domestic
iently slowly that, relative to the time horizon for which
-279—279—
stabilization policies
policies are conceived,
conceived, it seemed
seemed unimportant.
It would
It
only be
be if the world were to
to return to aa dollar standard that this
happy, for United States policymakers, state of
of affairs would be
be rere-
stored.
However though II understand Robert Heller’s
Heller's nostalgia
nostalgia for such
a system, II am much
much less
less sanguine than is
is he
he about the possibility
possibility of
of
the restoration of aa dollar standard.
standard.
the
The breakdown of the Bretton Woods system
system has forcefully reminded
us that the amount of
can extract from his
of seignorage which aa banker can
clients depends upon their willingness to
to pay up.
If he tries to exexIf
tract too much, they will, not without difficulty to
to be sure, take
their business elsewhere.
At
At the risk
risk of oversimplifying,
oversimplifying, under
Woods, the banker, namely the United States, tried to
to extract
extract
Bretton Woods,
too much seignorage.
The
The current chaotic
chaotic international monetary syssys-
tem is the result of the customers trying, as best they can, to find
somewhere
somewhere else to do their banking business.
AA dollar standard
standard is not
be restored unless it is clear
clear to
to the rest of the world that
going to be
the United
United States has mended its ways, and is not going
going to repeat its
previous policies-—either
policies--either willfully
willfully or inadvertently.
The pjfi~evi_()Jl_l,)' evi-
dence
dence that
that the past decade has produced to support this view is
is the
recent
recent announcement of monetary
monetary policy changes by
by Mr. Volcker.
Volcker.
If
If that
that
announcement
announcement is followed up
up with action, and
and past evidence
evidence suggests
that this cannot be taken for granted, and if the new policies
policies are adadthat
hered to long enough to erase
erase the memories of
inof fifteen years of in-
stability, then the possibility
possibility of restoring
restoring aa dollar standard
standard night
might
arise.
However, II believe that we would be ill advised to
to hold
bold our
breath in anticipation
anticipation of
of the event.
-280-
will Inevitably
inevitably remain on
Now this is not to say that the world will
the present flexible exchange rate system into the indefinite
indefinite future.
The problems of operating
operating under
under such a regime as Heller describes
describes are
real ones, although how much they are
exare the result of the flexible ex-
much of
of the underlying monetary ininchange rate regime per se, and how much
stability
stability that
that forced the adoption
adoption of that regime in
in the first place,
place,
is aa point that one might want to argue about.
There is
is undoubtedly
undoubtedly aa
demand for aa stable
stable monetary unit to
to serve as aa means of
of exchange,
exchange, unit
of account, and store of
marof value in international transactions, and mar-
kets have aa way
way of
of evolving in order to
to meet
meet such demands in a manner
that verges on
on the inherently
inherently unpredictable.
unpredictable.
After
After all, the
the Bretton
Woods system was
was not designed to put the world
world on
on the dollar standard,
nor did or indeed could the
outthe United States in
in any way
way force this
this out-
come; it
it arose as aa result
result of the voluntary choices
choices uf aa host of
of instiinstitutions and individuals and
and the evolution
evolution in
in question only appears
appears ininevitable with the benefit of hindsight.
In the current state
state of knowledge, economic
economic theory enables
enables us
us to
say that, so long as domestic monetary
monetary policies remain uncoordinated
uncoordinated
international monetary system
and unstable,
unstable, then
then the
the international
system will
will also be
be ununstable,
stable, whatever its formal
formal institutional framework, and
and that
that as such
policies become stable
stable and harmonized, then
then so will the international
monetary system become stable and perhaps
perhaps adopt aa new reserve currency,
or indeed currencies.
It does not enable us to say anything positive
positive
about the form that such an evolution is likely to
to take.
Nevertheless,
given the array of
of inflation rates, monetary expansion rates and such
at present
present ruling in
in various parts of the world, one is tempted to conconclude
the first step
step towards
towards reestablishing some sort of
clude that even the
-281-
unified world
world monetary
monetary system has yet
yet to be taken.
unified
The European
European MoneMoneThe
estabtary System is regarded by some as being the first stage in estab-
lishing an
an important
important regional base from which such aa system might
evolve; whether it is or not
not depends upon whether its
its members succeed
in developing the means
means to coordinate their domestic policies so as
as to
make them compatible with the maintenance of
of the System,
System, and they show
show
no signs
signs of doing this.
Be that as it may, as
as aa practical
practical matter
matter any discussion of United
States’ stabilization
States'
stabilization policies that is
is to be
be of
of current relevance
relevance
should take aa flexible exchange
exchange rate
rate system as its
its background.
background.
Thus
Frenkel's paper, though it
it
the theoretical and empirical material in Frenkel’s
will
will look rather unfamiliar
unfamiliar to
to many specialists
specialists in the analysis of
domestic monetary policy,
policy, is of considerable
considerable relevance to
to their conconcerns.
cerns.
II will now turn to some of the issues
issues involved.
It should
should go without saying
saying that if one is going to discuss the
macro—stabilization policies are likely to work against
way in which
which macro-stabilization
the background of aa flexible exchange rate regime,
regime, one ought
ought to know
something
something about the way
way in which
which the foreign exchange
exchange market itself
operates.
Frenkel deals with this
this matter
matter from the point of
of view of
what
what may
may be
be referred
referred to
to as the “Asset
"Asset Market Approach
Approach" to exchange rate
beyond doubt
doubt provides a simple and powerful
theory, an approach which beyond
method of analyzing the problem area.
Nevertheless, anyone reading
Heller’s paper immediately after Frenkels
Heller's
Frenkel's must wonder
wonder where
where many of
of
the concerns he raises, particularly about the large
large amount of dollar—
dollardenominated assets held abroad, fit into
into Frenkel ‘5
's analysis.
II believe
believe
that the answer here is that, although the theoretical framework
framework which
-282-
underlies Frenkel’s
Frenkel's work can deal with these issues, the particular
underlies
“monetary” version of the asset market approach which he
"monetary"
he explicitly
explicitly
sets out does
does so only implicitly,
implicitly, and in a way
way that his
his empirical
empirical evievidence suggests
suggests is inadequate.
itThe basic monetary model of
of the exchange rate
rate is simplicity itself.
self.
With national price levels tied to each
each other
other by purchasing
power parity and aa stable deiiand
denand for real balances function in each
each
country, domestic price levels, inflation, nominal interest rates and
the
the exchange rate are simultaneously determined by the
the behavior
behavior of the
“real”
"real" arguments in
in the demand functions in
in question, and by that of
the
the supplies of
of nominal money
money in
in the two countries.
countries.
What does this
this
analysis tell us about the role
role in
in influencing the exchange rate of
U.S. dollar—denominated
dollar-denominated assets left over
over from
from the period when
when the
dollar was the reserve currency, and
and currently held abroad?
This is a
be of key
problem which many commentators, including Heller, believe to be
importance.
The monetary model
model tells us, II believe, that these assets
have no special significance.
They
They are interest-bearing
interest-bearing assets, and,
according
to the monetary version of the more
more general asset market
market apapaccording to
of return
return on them adjusts to
to compensate
compensate their holders
proach, the rate of
for any anticipated change in their purchasing power over goods and
over assets denominated in
in other
other currencies.
Variations in
in that rate
of return are taken account of in
in the model because the nominal interinterest rate
rate they bear is
is an
an argument in the U.S. demand for money funcfunction.
The
The above reasoning hinges upon purchasing power
power parity always
holding, but Frenkel’s
Frenkel's empirical evidence shows that at the very best
it does
does so only on average over rather long
long time periods, and in
in aa
-283-
rough and ready fashion at that.
This means that variations in the
dollar—denominated assets cannot stimultaneously
rate of interest on dollar-denominated
compensate
compensate for variations in
in their purchasing power
power over goods
goods priced
priced
in U.S. dollars and goods priced in foreign currencies.
currencies.
This in turn
means that, although some U.S. dollar-denominated
dollar—denominated assets may be
be perfect
perfect
substitutes for those denominated in foreign currencies, others are
are
not.
That being the case, the currency
currency in which
which they are denominated
must be
be aa relevant
relevant property of
of at least some classes
classes of securitieS,
securities, and
fluctuations in the supply and demand for such securities are likely to
impinge upon
upon the behavior of
of the exchange rate.
rate.
The behavior
behavior of the
dollar—deutsche mark
mark exchange
exchange rate gives Frenkel more trouble
trouble than any
any
dollar-deutsche
other, and surely that is
is not an accident, given that mark-denominated
mark—denominated
assets
assets have so often been
been the destination of
of funds realized by selling
dollar-denominated securities.
securities.
dollar—denominated
There is another characteristic
characteristic of
of the U.S.
U.S. dollar’s
dollar's place in the
the
There
international monetary system
system worth
worth noting:
it
it is
is the unit of account
for many international transactions, not the least of which are those
involving oil.
That means that many international prices are going to
to
be particularly sensitive
sensitive to the conduct
conduct of U.S.
U.S. domestic monetary
policy, and that that policy still has aa considerable power, for good
or ill depending
depending upon how it
it is used, over
over the international economy, aa
power which it
pf~ices in
it would
would not have were prices
in that economy to be
be set
set in
in
other currencies.
The frequent references in
in U.S. debates to
to oil price
increases as
domestic policy
increases
as being
being exogenous
exogenous to
to domestic
policy shows
shows that
that it
it is
is not
not
yet appreciated that oil
oil prices in the world economy
economy respond to U.S.
domestic policy and that attempts to cushion their
their effect by
by domestic
domestic
monetary
monetary expansion are not just useless but
but actively
actively harmful
harmful.
-284-
To put
put all
all this
this in
in another
another way,
way, if
if goods
goods markets
markets cleared
cleared as
as fast
fast
To
as
long—run equilibrium where the
as asset
asset markets, if we
we were always
always in long-run
prices of national moneys and of
of national outoutconcepts of the relative prices
puts were
were interchangeable,
interchangeable, the above problems would not arise.
HowHow-
asset markets
markets do clear faster, and in
in the short
short run do dominate
dominate
ever, asset
of the exchange rate, so that the distinctions upon which
which
the behavior of
the asset
asset market approach
approach focuses are important.
implication of
of the evidence that Frenkel presents.
That
That surely is one
This
This very fact howhow-
ever
ever seems to
to me to imply that
that the asset market approach to
to exchange
exchange
rate determination must be carried beyond aa simple monetary formulaformula-
tion, as it is, for example, by Boyer (1978), to incorporate explicitly
explicitly
other aspects
aspects of
of portfolio
portfolio behavior, and to
to incorporate other
other aspects
it can
of using aa particular
particular currency as a unit of account, before it
claim to provide us
us with aa complete toolkit for dealing with foreign
exchange rate problems, not least those which Heller raises.
NevertheNeverthe-
less, if our
it is still
our toolkit is incomplete,
incomplete, it
still the best
best one that we
have.
As Frenkel’s
Frenkel's paper shows, the
the asset market approach to
to anaana-
lyzing
lyzing exchange rates
rates is extremely useful,
useful, and its use does enable us to
to aa clearer understanding
understanding of how
how to conduct domestic
domestic policy
come to
against a background of exchange
exchange rate flexibility.
One of the best established pieces of conventional wisdom in
One
international monetary economics is that high interest
interest rates are assoassoweak one,
one,
ciated with aa strong currency and low interest rates with aa weak
but one of the best established
established facts of the last few years is that the
high interest rates in
in fact are
are associated
associated with weak currencies, and
vice versa.
As Frenkel
Frenkel shows, the latter prediction
prediction is what
what follows
-285-
from the asset market approach, and, as he also shows, that theory’s
theory's
predictions in this respect are confirmed
confirmed by
by evidence, generated
generated moremoreover by
by an experiment
experiment whose validity
validity does not, as
as far as I can
can see, in
any way hinge upon assuming that purchasing power parity holds.
holds.
Though
Though
II can find nothing to disagree with in anything that
that Frenkel explicitly
explicitly
says about this
this matter, there are aa few things that he didn’t
didn't say that
do seem to me to be of particular relevance to
to the
the theme of this conconference.
The conventional wisdom about the relationship between interest
otherwise of aa currency has
has its historical
rates and the strength or otherwise
roots in the operation
operation of the gold standard, and in particular in the
role played by
by the central
central bank rediscount rate in
in the conduct
conduct of monemonetary policy under such
well—known,
such aa system, a role summarized in that
that well-known,
but now sadly outdated, aphorism
“Seven per cent will draw gold from
aphorism "Seven
the moon"
moon” (which I have been
been unable to
to track
track down to its original
source).
source).
Under such
such aa system the long-run time paths of money
money and
prices in the international economy were given by the rate of
of change
change of
the stock of gold.
Though this rate of change was not always smooth
and
and steady,
steady, because important new gold discoveries
discoveries were
were from time
time to
time made,
made, on
on average it was.
Given that, and given an unquestioned
unquestioned
commitment of central banks to maintain the convertibility of
of domestic
commitment
money into gold, the anticipated inflation rate was, by
by comparison
comparison with
recent experience, not far short of being an
an exogenous
exogenous constant.
MoreMore-
over the
aim of.monetary
not to
and
over
the principal
principal aim
of monetary policy
policy was
was not
to control
control income
income and
employment but
but simply to maintain
maintain convertibility.
convertibility.
employment
In such
such a world,
world, any
any
In
increase in aa central bank’s
bank s discount
discount rate represented an increase in
1
the real cost of borrowing from the banking system, and hence
hence led to
to aa
-286-
contraction (or
(or at
at least
least aa slowdown
slowdown in
in the
the rate
rate of
of expansion)
expansion) of
of domes—
domescontraction
tic credit.
credit.
tic
The monetary consequences of
of that
that in
in turn
turn led to
to aa balance
balance
The
of payments
payments surplus and hence aa “strong”
"strong" currency.
currency.
of
different than that
The world of
of the last
last ten years
years has been very different
that
whieh
which I have just
just described.
standWith nothing to replace the gold stand-
ard’s guarantee
guarantee of
expectaard's
of long—run
long-run price
price predictability,
predictability, inflationary
inflationary expectadomitions have become
become endogenous and volatile, and their movements
movements domi-
nate fluctuations in nominal
nominal interest rates.
rates.
It is these
these factors which
which
have led to
to the association of
of high interest
interest rates and weak currencies.
Both are the
the consequence of an
an adverse response
response of inflation
inflation expectaexpectations to undisciplined
undisciplined and expansionary
expansionary monetary
monetary policies, as Frenkel
has argued.
II believe that the forgoing considerations
considerations have two important
important imim-
conduct of domestic monetary policy in
in the United
plications for the conduct
States, both
both now
now and in
in the future.
States,
emFirst, though at long last an em-
phasis on controlling monetary aggregates is replacing
replacing an
an emphasis
emphasis on
on
interest rate targets
targets in
in the conduct
conduct of
be foolish to
of policy, it would be
believe that the battle
battle here has
has been finally won.
being fought.
Rather it
it is still
The advocates of
of controlling monetary aggregates have
inferalways based much of their case upon the difficulty of
of drawing infer-
ences from aa particular
particular value of the
the interest rate about whether
whether policy
policy
is “tight”
do so.
tight or “easy,”
easy,1' and will continue to do
11
11
11
The
The forqoing
forrioing analyanaly-
sis
sis surely helps to
to bolster
bolster their case,
case, for it
it shows that there is
is an
to the problems to which
which they have
important international dimension to
been pointing, a dimension
dimension that adds
adds weight to
to the argument against
using interest rates as
as aa policy indicator.
-287-
The second implication worth pointing
pointing out
out is not of
of such
such imedi—
immediThe
ate concern,
as important.
concern, but is surely just as
forgoing argument
argument
The forgoing
proposiamounts to
to presenting aa special case
case of the following general proposi-
tion:
the way in which monetary policy impinges upon the domestic
economy, and the way in
in which
which domestic monetary variables
variables should be
interpreted by the authorities depend critically
critically upon the
the state of
of the
international monetary system and the nature of the country's
country’s exchange
rate regime.
regime.
II believe that many of
of the United States’
States' current policy
apprecidifficulties have arisen from aa failure of the authorities
authorities to
to appreciinternational factors are of prime rather than
ate the fact that these international
in the design of
of policy.
secondary importance in
To put the matter
matter in its
its
it is not just the way
way in
in which United States policy
simplest terms, it
affects the rest of the
the world that varies with the exchange rate regime
and the conduct of policy in other countries; the way in which
which it
it afaffects the United States is
is also profoundly influenced by these matters.
imII will now turn to aa more specific
specific discussion of this point
point as
as it
it imupon the conduct of
pinges upcn
of policy
policy under the present regime.
There is no doubt about the nature of
of the current macro
macro policy
problem facing the United States:
it is how
how to
to reduce the inflation
inflation
rate without at
at the same time causing more of aa real
real contraction than
is absolutely necessary
necessary (however
(however much that
that might be).
It is
is also
also true
that there is
is aa wide consensus that getting the monetary
monetary expansion rate
"under
key role in tackling this problem.
problem.
“under control"
control” must play aa key
Debates
it comes to the question
question of how to implement such a policy,
arise when it
monetary expansion
expansion “under
"under control”
control" means in
of specifying what getting monetary
practice.
At
At one extreme
extreme are those who
who follow the lead
lead that (I an
am glad
-288-
to
learn from Neil Wallace) Sargent
Sargent and Wallace
Wallace (1975) never meant to
give.
They argue for aa rapid, pre-announced, monetary
monetary slowdown
slowdown which
will, by way of a by
by now well—known
well-known "rational
“rational expectations"
expectations” mechanism,
only
impinge mainly upon prices and will affect output and employment only
extent that the pre-announcement
to the extent
pre—announcement is not believed.
extreme are those like Modigliani (1977)
(1977) who believe
At the other extreme
that aa monetary
monetary contraction
contraction can be fine tuned, while in the middle
middle
stand those who would support aa gradualist
gradualist contractionary
contractionary policy
policy of the
type advocated
Allan Meltzer.
advocated at this conference by Allan
To
obTo aa foreign ob-
server, the striking characteristic of
of this
this United States
States policy
policy debate
is the way in which the openness of
of the United States economy and the
nature of the exchange
exchange rate regime are virtually
virtually ignored by
by all
all particparticipants.
Nevertheless, the theoretical
theoretical and empirical results presented
presented
by Frenkel at this conference, not to
to mention a good
good deal of work on
stabilization problems
problems in open economies that
that has
has been
been carried on
on mainmainly outside the United States, is extremely
extremely relevant
relevant to these issues.
ly
Two key questions
questions underlie
underlie current debates about stabilization
policy.
The first concerns the speed
speed with
with which the private sector
sector of
the economy can absorb information about policy and translate
translate that ininformation into
into price changes,
changes, and
and the second, analyzed by
by Lucas
Lucas (1976),
concerns the stability
stability over time of the mechanisms whereby information
is
is absorbed
absorbed and acted upon and the independence
independence or otherwise
otherwise between
those mechanisms and policy actions
actions themselves.
If
If one
one believes
believes that
information is
is absorbed
absorbed and acted upon quickly, then rapid monetary
monetary
contraction
anti—inflation policy.
contraction is an appropriate anti-inflation
policy.
If
If one believes
that reactions here
here are slow, but
but that their
their time path in
in the future
can be inferred reliably from past behavior
behavior then one will advocate fine
-289-
tuning.
tuning.
mechanism
AA slow but unstable, and hence hard to predict, mechanism
underpins the case for gradualism.
(May I note here in passing that II
(May
believe Meltzer’s
Meltzer's analysis of the case
case for gradualism, which
which II largely
accept, would be
be enhanced if he would lay
lay more stress upon the ~pp~~—
unpredictability of
of the lag
lag structure
structure of
of his
his model in any particular
particular ininstance, and less upon its drawn out and backward looking
se.)
looking nature per se.)
‘s empirical work shows that the foreign exchange
exchange market
Frenkel 's
that all available information, including
including
is efficient, in
in the sense that
information about
policy, is translated quickly
quickly into movements
movements of the
about policy,
exchange rate.
exchange rate is,
is, therefore, aa price
price that, other
other
The exchange
things equal (the
qualification is
and I will
will return
return to
it
i~jj!9~~q~j
(the qualification
is important
important and
to it
in
in a moment) adjusts rapidly to policy changes.
AA number
number of recent
papers
augof the
the aggregate
aggregate demand—expectations
demand-expectations augpapers have analyzed versions of
mented Phillips curve model, which underlies
underlies so
so much United
United States
policy debate, extended explicitly to incorporate a foreign sector.
Though such
exchange rate
such work is most
most highly developed for fixed exchange
regimes——see, e.g., Laidler (1975), Jonson (1976), Jonson, Moses and
regimes--see,
Wymer
(l979)——some results are now availWymer (1976),
(1976), Bilson (1978),
(1978), Burton (1979)--some
avail-
able for aa flexible rate regime.
Thus Laidler
Laidler (1977)
(1977) shows, albeit in
an extremely primitive model
model with zero capital mobility, that even
where
errors are made
made about the domestic price
price level, perperwhere systematic
systematic errors
fect foresight about the exchange
exchange rate
rate is
is sufficient to
to guarantee
guarantee that
that
domestic monetary policy impinges solely upon domestic prices and not
at all on output.
orate
much more elabelabBurton (1979 and forthcoming), in aa much
model that does
does incorporate capital mobility,
mobility, aa variety of
stochastic shocks, and rational expectations, finds that
that the behavior
of the exchange rate is aa key source of information for agents and that
-290-
the more rapidly information
information about it
it is
is available to them,
them, the
the more
direct is the linkage between domestic monetary
monetary policy and domestic
prices.
One must be careful not to
to read too much in the way
way of policy imin—
plications from analytic
analytic exercises such as these.
Nevertheless, the
work that II have referred to does point to the
the conclusion that
that aa flexflexible
efficient market,
narket, imparts to an
ible exchange rate, determined in an efficient
an
economy an
an extra degree of price
price flexibility that it
it does not have
under aa fixed rate.
This in turn suggests that estimates of the output
output
that might be lost in the United States
States while
while bringing inflation under
under
been generated from data on the fixed exchange
exchange rate
control that have been
period are likely to be exaggerated, even if there is nothing else
wrong with the techniques used to
to derive them.
there is a very important qualification to be
be added to
However, there
all this.
The theoretical
theoretical results
results to which
which I have alluded are premised
premised
on
on the price level, and implicitly the money market, in the rest of the
theoretical experiment from
world remaining undisturbed during the theoretical
which they are derived.
To put the matter in terms of Frenkel’s
Frenkel's frameframe-
work, they apply to situations in which
which nothing happening abroad disdis-
turbs equilibrium in the market for foreign money,
money, or
or foreign assets in
in
exchange rate originate
originate in the
the
general, so that all disturbances to the exchange
money supply.
behavior of the domestic money
Why this is an
an important
important qualiquali-
fication is easily seen by considering Frenkel ‘s
's analysis and his
his emempirical results.
If aa foreign monetary contraction begins at
at the same
time
question tells us that,
that, given
given
tine as aa domestic one, the analysis in question
for the
of simplicity that the relative
relative sizes of these contraccontracthe sake of
nothing will happen to
to the exchange rate.
tions are appropriate, nothing
—291—
-291-
In
In
that case domestic money markets must be cleared by domestic output and
price
price level
level fluctuations without help from aa quickly adjusting foreign
exchange
exchange market.
Frenkel ‘s
's results on
on purchasing
purchasing power parity lend
weight
view that
that domestic
prices adjust
adjust slowly
disweight to
to the
the view
domestic prices
slowly to
to monetary
monetary disturbances.
Thus there
there is every reason
reason to suppose that in
in this
this case,
and in the short run, which may nevertheless persist
tine,
persist for aa long time,
much of
of the effect will be
be on output.
The implications of
of looking at Frenkel
Frenkel 's
empirical results
results on the
‘s empirical
efficiency of
of the foreign exchange
exchange market
market in the
the light of the macromodels II have cited in the preceding
preceding section may
may be
be summarized as
follows:
a single
single economy seeking to tackle an inflation
inflation problem
against the
the background
background of an otherwise tranquil world economy will
will find
that the existence
existence of an efficient market
market for foreign exchange
exchange under aa
flexible rate enhances the flexibility of
of domestic
domestic prices.
Such
Such an
an
economy will
will enjoy an easier transition to aa lower inflation
inflation rate
rate than
one would expect from studying closed economy models.
However,
However, if
if that
same economy is one among aa number
number faced with aa similar problem, then,
even with
with a flexible exchange rate, the pressures
pressures of domestic deflation
deflation
will, if other
simultaneously deflating, be concentrated
other countries are simultaneously
on domestic output.
In general
general,, the extent to which
which this
this happens in
any one country
country will vary with the conduct of
of policy abroad.
In the current state of
of knowledge, II do not believe we can say
any more than this, but II would
would claim that even
even this much is important
important
to know.
know.
Our consideration of
open economy aspect of
of the open
of stabilization
policy has, after all, led
led us to argue
argue that the lags with which informinformation will become available,
available, and hence aa basis for action, will vary
—292-292-
with the way
way in
in which policy
policy is
is conducted not only at home but also
also
abroad.
The length and variability of such
lags are,
are, therefore, in
in any
such lags
particular instance, going to be next to
to impossible
impossible for policymakers to
predict.
predict.
However
However such unpredictability is the very essence of the
the case
case
for gradualism.
The
The analysis we have been considering
considering does, therefore,
make an important contribution to
to the current U.S.
U.S. policy debate.
—293—
-293-
REFERENCES
REFERENCES
0. (1978), “A
Devaluation,” Canadian
Bilson, J. F. O.
"A Dynamic Model of Devaluation,"
Journal of
of Economics 11,
ll , 194—209.
194-209.
Boyer, R. S.
S. (1978), “Financial
"Financial Policies
Policies in
in an
an Open
Open Economy,”
Economy," Economica
Economica
45, 39-57.
“Expectations and
Economy,” unBurton, D.
D. A.
A. T.
T. (1979), "Expectations
and aa Small Open Economy,"
unpublished PhD thesis, University of Western Ontario.
(forthcoming), “Expectations
Flex"Expectations and aa Small Open Economy Under FlexRates,” Cana
Canadian
- - iible
bl e Exchange Rates,"
di an Journal of Economics.
Economics.
Jonson, P. D.
D. (1976), “Money
"Money and Economic Activity in the Open
Open Economy,
The
The United Kingdom 1880—1970,”
1880-1970," Journal of
of Political Economy 84,
84,
979-1012.
979-1012.
Jonson, P. D.,
“A Minimal Model of
D., E. R. Moses and C. R. Wymer (1976), "A
of
the Australian Economy,”
Economy," Reserve Bank of
of Australia Discussion
Paper 7601.
7601.
“Price and Output Fluctuations in an Open
Laidler, D.
D. E. W.
w. (1975), "Price
Economy,"
Essays on Money
Money and Inflation,
I nfl ati on, Manchester,
Manchester,
Economy,” Ch. 9 of Essays
Manchester
Manchester University Press and Chicago, University of Chicago
Press.
_____
“Epectations and the Behaviour
(1977), "Epectations
Behaviour of
of Prices and Output Under
- - FFlexible
l ex i bl e Exchange Rates,”
Rates," Economica
Economi ca NS 44,
44, November, 327—336.
327-336.
Lucas, R. E.
E.,, Jr. (1976), in Karl Brunner and Allan H. Meltzer (eds.),
Labor Markets, Carnegie—Rochester
Carnegie-Rochester Public
The Phillips Curve and Labor
Policy
Journal_of
Policy Series supplement to the Journal
of Monetary Economics.
Economics.
“The Monetarist Controversy
Modigliani, F. (1977), "The
Controversy or Should we
we Forsake
Stabilization Policies?”
kierican Economic
1—19.
Stabilization
Policies?" American
Economic Review 67, March, 1-19.
“Rational Expectations and the
Sargent, T. J. and N. Wallace (1975), "Rational
Theory of Economic Policy,”
Policy," Research
Research Department, Federal
Federal Reserve
Bank of Minneapolis Studies in Monetary Economics 2.
Triffin, R. (1961), Gold and the Dollar Crisis, The Future ConvertiConverti~jjj~,
UniversitjTress.
bility, New Haven, Yale University
Press.
-294—
-294-
FLOATING EXCHANGE RATES IN THE 1970s:
1970s:
AA DISCUSSION OF
OF THE HELLER PAPER
PAPER
Geoffrey E. Wood
Heller’s paper
asserThe first section
section of
of Dr. Heller's
paper consists of four assertions about the consequences for the world economy of the move to
to
Floating
Floating exchange
exchange rates.
On
On the basis of these four assertions Dr.
leller
ieller proceeds to
to make recommendations
recommendations first for the future conduct of
of
J.S.
J.S. economic policy, and second for the future shape of the interintermtional
1ational monetary system.
In these comments
comments it will be
be argued first that his four asserassertions
of exchange
exchange rate
rate flexibility
tions on
on the
the consequences
consequences of
flexibility are
are at
at the
the least
least
nisleading and, in some cases, not supported
supported by any evidence
evidence at present
~vailable.It
,vailable. It will then
then be shown
shown that his policy
policy recommendations
recommendations for
the future of the international monetary system are based on
on misunder—
misunder;tanding
of exchange rate volatility and the reasons
;tanding both
both the causes
causes of
capital movements.
movements.
for international capital
The comments conclude with aa sumsum—
nary of what appear to be the true lessons
lessons of the
the floating exchange
exchange
rates
1970s.
rates experience
experience of the 7970s.
DR. HELLER’S
HELLER'S ASSERTIONS
ASSERTIONS
Dr. Heller asserts
“the operation
Or.
asserts that "the
operation of the
the flexible exchange
rate
system since
1971 has
significant increase
in costs
rate system
since 1977
has entailed
entailed aa significant
increase in
costs to
to
)r. Wood is aa member of the Centre for Banking and International
Dr.
Finance,
Finance, The City University, London, England. This is a revised and
?xpanded
version of
of comments
comments made
the conference
conference on
,xpanded version
made at
at the
on Dr.
Dr. H.
H. Robert
Robert
leller’s
,eller's paper. The author
author is indebted to several
several conference
conference partici—
partici,ants for remarks which have
have improved these coments.
comments.
Dants
—295—
-295-
the business sector.”
sector." The trouble with that
that statement
statement is
is that
that Dr.
Dr.
Heller does not make clear what comparison he
he is
is making when he
he says
says
costs have increased.
There has been a substantial increase in the
There
dispersion of inflation rates in the O.E.C.D.
O.E.C.D. (Organization for Economic
l970s as
compared to the
Cooperation and Development) area
area in
in the 1970s
as compared
1
1960s.1 Had exchange rates
rates remained pegged despite
despite this change, they
been kept so
so by
by an
an increasing proliferation of exchange
exchange
could only have been
controls to restrict capital movements and of
of tariffs and quotas to rerestrict trade,
trade, and by increasing volatility of
of national monetary
policies.
policies.
It is
is impossible to
to believe that these developments
developments would
would
not have imposed costs
costs on
on the business sector, and Dr. Heller
Heller certainly
does not demonstrate that these costs would
would be less
less than the costs imimposed by floating exchange rates.
rates.
Indeed, it should be
be pointed
pointed out that
that there is absolutely
absolutely no
no evievidence
dence in
in support
support of
of Dr.
Dr. Heller’s
Heller's view that
that floating
floating exchange
exchange rates
rates have
have
inhibited international trade.
trade.
This issue has
has been studied fairly exex-
tensively, and there is not one
one study which has found that
that floating
trade.
rates have had any dampening
dampening effect whatsoever on world trade.
But dede-
spite that, there may be some truth in
in this
this particular belief.
All studies so
so far undertaken have looked at the effect of
of the
exchange rate regime on
on international trade as aa whole.
Recent
1‘See
Wood and Nancy Aamon Ji
anakopl os, "Worldwide
see Geoffrey
Geoffrey E. Wood
Jianakoplos,
“Worldwide
Economic Expansion: Are Convoys or Locomotives the Answer?”
Answer?" Federal
Reserve Bank
Bank of
of St. Louis Review, July 1978.
—296—
-296-
theoretical
empiritheoretical work by Ronald
Ronald McKinnon,22 supported by
by forthcoming empirical work by Stephen Carse, John Williamson and the present author,33
cal work by Stephen Carse, John Williamson and the present author,
this is not appropriate.
suggest that this
AA substantial part
part of international trade is in primary
primary or
or semi—
semimanufactured goods.
sanufactured
goods.
The prices
prices of
of such goods are continually held
held
close together
together across countries by arbitrage.
Thus traders in
in such
goods are not affected by
by exchange
exchange rate fluctuations provided
provided that
that they
hold inventories equal to
to their indebtedness arising from trade
trade -- and
-—
the evidence is
is that to aa first approximation
approximation they do.
There is therethere-
fore no reason to expect trade in
in these
these goods to be in any way affected
by exchange rate changes, whether or not
not these changes are anticipated.
In contrast, manufactured
manufactured goods do not have their
their prices
prices quickly arMarbi4
traged
goods are
traged into
into equality
equality internationally.
internationally. 4 Traders
are
Traders in
in such
such goods
therefore exposed to exchange risk.
Tests for the effects of exchange
Tests
rate
rate fluctuations on
on trade should focus on these
these categories of goods,
goods,
rather than
than on trade as aa whole; looking at trade as aa whole may have
led to the concealing of the effect of
of exchange rate fluctuations on aa
sub-section of trade.
(This hypothesis is currently being explored by
by
the present author, but no
no results sufficiently
sufficiently reliable to report are
~t present available.)
,t
2Ronald McKinnon, Money
e, Oxford UniverMcKinnon,
in International Exchan
Exchange,
University Press, New
New York, 19
1979.
3Stephen
Geoffrey E.
E. Wood, Financing
stephen Carse, John Williamson, and Geoffrey
~c~!jjcfpj~iLftad~,
'ractices
in U.K. Foreign Trade, Cambridge University Press, Cambridge,
1980.
1980.
4See
Williamson
“The British
1i ams on and Geoffrey E. Wood, "The
see e.g. John Wil
[nflation:
Inflation: Indigenous or Imported?”,
Imported?", American Economic Review,
September
,eptember 1976.
-297—
-297-
absence of confirming evidence,
evidence, Dr.
Or. Heller may be
So, despite the absence
correct in saying that trade has been inhibited by
by exchange
excrange rate fluctfluctuations"
uations.
But three noints
points shou·ld
emphasized.
Should be emphasized.
evidence to support his assertion.
evidence
provides no
First, he provides
Second, he should have compared
compared
Second,
what would have
have happened to trade under aa fixed rate system defended
defended
against the consequences of divergent
divergent inflation rates by
by proliferating
controls, with the effect of
of exchange rate fluctuations on trade.
Third, even if he
he is correct that exchange rate fluctuations inhibit
trade, it
it is far from clear that official exchange market intervention
intervention
is thereby justified.
His second major assertion is that flexible exchange rates have
“not brought about a climate for the conduct of
"not
of more
more effective
effective stabi—
stabi-
lization policies.”
policies." The
The only possible response
response to that is to ask why
on earth they should.
Under
Under a fixed exchange rate system,
system, the burden
by any country’s
country's government was
was in
of mistakes in stabilization policy by
part borne by the foreign
foreign sector.
Excess
Excess demand
demand was in part met by
foreign supply,
supply, while deficient
deficient home demand was in
in part
part offset by
demand from overseas, so long
long as the demand and supply imbalances were
at least partly
partly due to monetary policy.
is the
the
(An example of this is
United Kingdom experience in
l96Os; see Williamson
in the 1960s;
Williamson and
and Wood,
op.cit.)
Floating exchange
exchange rates, by eliminating
eliminating flows across the
close this
this safety valve; one should therefore
therefore expect
foreign exchanges, close
of stabilization
equal) that the performance of
(other things being equal)
stabilization
policies should deteriorate rather than improve under
under floating rates.
But
wHte very
But Dr. Heller did not write
very precisely
precisely at this point;
point; he
does not say exactly what
1;hat he
he means by the “climate
"climate for the conduct of
more effective
effective stabilization
stabilization policies.”
policies."
-298-
He may
may mean not the actual
achievement of such policies, but rather
rather how policymakers
policymakers have reresponded to divergences of the economy from its
its desired path.
is what he means, then
then he is pretty clearly wronq.
example.
If
If that
The U.K. is aa good
It
It was
was only after
after the collapse of sterling’s
sterling's foreign
foreign exchanoe
exchanoe
value in 1975 that the U.K. government took any serious measures
measures to end
the gradually
gradually accelerating
accelerating inflation of
of the
the previous twenty years.
Why
Why
they so responded can only be conjectured; but
but the explanation may be
be
that floating exchange
exchange rates bring home to the electorate
electorate the costs
costs of
inflationary
policies rather
inflationary policies
rather more
more quickly
quickly than
than did
did fixed
fixed rates,
rates, and
and
next election.
thus may
may influence their voting behavior at
at the next
Dr. Heller next claims that floating exchange
exchange rates have not
“decreased
the cost
cost of
[foreign exchange
exchange market]
of [foreign
market] intervention
intervention to
to cencen"decreased the
tral banks.”
banks."
Dr. Heller is really very careless in his use of
Or.
of the word
“cost.”
cost."
11
He never tells us what costs he
he has
has in mind
mind in
in the present
present inin-
stance.
It is certainly
certainly clear, however, that the amounts of interveninterven-
tion have been large,
large, and
and it is on this issue rather than
than the undefined
one raised by Dr.
Or. Heller that
that we next comment.
Why have exchange
exchange rates been so volatile? Where
Where have
have the private
stabilizing speculators
speculators been?
stabilizing
been?
these questions.
Dr. Heller
does not
Dr.
Heller does
not attempt
attempt to
to answer
answer
Fortunately, an answer has
has been provided by aa large
body of previous work.
rate volatility is, in large part,
Exchange rate
the consequence of volatile
volatile national monetary policies.
policies.
This has been
l97Os; it was al
also
192Os.
true not just in the
the 1970s;
so true in the 1920s.
The
The concon-
clusions of aa recently published paper by my colleague Roy Batchelor
summarize the evidence
evidence very
very well.
—299—
-299-
Stable inflation rates are
all
all that
that is required
required to
to keep the
trend in exchange
steady..., efficient exchanqe marexchange rates steady_...
markets should keep fluctuations around the
the trend within
within the
same
exsame margins as in the
the 1920s.
1920s. What
What is necessary
necessary for exchange rate stability is that monetary expansion
expansion be
be predictpredictable..
able
... 5
The
expressed in the quotation
The reason
reason for this is admirably expressed
quotation from
Gustav Cassel with which
which Jacob
Jacob Frenkel
Frenkel concludes
concludes the paper
paper he
he presented
at this conference.
The international valuation of aa currency
currency will, then,
then,
generally show aa tendency to anticipate movements, so
so to
speak, and become more an expression of
of the internal
internal value
that the currency is expected
to possess in aa few months,
6 to
or perhaps
year’ss time,
or
perhaps in
in aa year
time.6
1
The
The more volatile is aa nation’s
nation 1 s monetary
monetary policy,
policy, the more frefre-
expected future internal value of
of its currency change,
change,
quently will the expected
and so the more frequently will its
its exchange
exchange rate change.
The primary
primary
source of exchange rate volatility is therefore volatility
volatility in
in national
monetary
monetary policies.
Understanding that is
is central to drawing
drawing the
the corcor-
rect lessons for future policy of the exchange rate experience
experience of
of the
1970s.
of private
private
Understanding that also helps explain the absence of
stabilizing speculation; because of
of the volatility of national
national monetary
expolicies, speculators have had
had very
very little
little basis on which
which to form ex-
pectations of future exchange
exchange rates.
In this
this context, it is worth pointing out
out that
that (as
(as Jacob
Jacob Frenkel
Frenkel
shows) exchange
shows}
exchange rates have been no more
more volatile
volatile than
than prices in other
5Roy Batchelor, "Must
“Must Floating Exchange Rates be Unstable?"
Unstable?”
Annual Monetary
Monetary Review, Centre for Banking and
and International Finance,
Finance,
The City University,
University, London, England.
England.
6Gustav Cassel, ~
199, pp.
Money and foreinExchanges
Foreign Exchanges after 1919,
pp. 149149150, Macmillan,
Macmi 11 an, London, 1930.
-300-300-
asset markets, thus emphasizing the common cause
,sset
cause of such volatility.
volatility.
rurther,
stressed that
'urther, it
it must
must be
be stressed
that D.
D. Heller’s
Heller's belief
belief that
that “speculative
"speculative
ictivity may
,ctivity
may well accentuate rather
rather than
than reduce exchange rate fluctuafluctuations”
tions" is totally contradicted by evidence that
that there are no traces
traces of
;peculative
runs in the foreign exchange markets.77
;peculative “runs”
11
11
His last
assertion is
“fostered
His
last assertion
is that
that floating
floating exchange
exchange rates
rates have
have "fostered
:he decline
decline of the dollar as the world’s
world's leading currency.”
currency."
By this
this he
ieams that floating exchange
1eans
exchange rates have
have led to aa fall in the proportion
)f
assets in
of individuals
individuals and
if dollar-denominated
dollar-denominated assets
in the
the portfolios
portfolios of
and
:entral banks.
He is clearly right.
Portfolio
Portfolio diversification was to
iee expected as aa consequence of the move to floating rates, and it has
indeed happened.
But so what?
Why
Why is that
that undesirable?
Nowhere does
Jr.
Jr. Heller answer these questions.
U.S. POLICY
POLICY RECOMMENDATIONS
Turning first to his recommendations for the future conduct of
1.5.
J.S. policy, these are manifestly sensible
-—
they comprise recomendrecommend-
ng the announcement of
of intermediate monetary ranges targeted by base
:ontrol so
:antral
so as to ensure hitting
hitting them.
them.
The empirical and theoretical
iork on the causes of exchange rate volatility, which was referred to
,ark
?arlier, clearly
clearly indicates
such aa policy
exchange rates
•arlier,
indicates that
that such
policy would
would make
make exchange
rates
iuch less erratic in their movements, and such
1uch
such aa policy would also
also help
tabilize the U.S. economy
,tabilize
economy as aa whole.
77See for example Donald S. Kemp, “The U.S. Dollar in InternationDonald S.
"The
Dollar in Internationsee
tl,1 Markets,
mid-1970
mid—1976,” Federal
Bank of
Markets, mid1970 to
to mid-1976,"
Federa 1 Reserve
Reserve Bank
of St.
St. Louis
Louis
teview, August 1976.
-301-301-
INTERNATIONAL POLICY
POLICY RECOMMENDATIONS
RECOMMENDATIONS
INTERNATIONAL
Dr. Heller does not advise aa return to pegged exchange
exchange rates;
rates; he
recognizes that so long as national inflation
inflation rates are as
as diverse as
they
they currently
currently are such aa move would
would not be
be sustainable.
He does, howhow-
ever, encourage
encourage official intervention in
in the
the foreign exchange markets.
markets.
There are,
are, as Dr.
Dr. Heller recognizes, costs to such intervention-intervention—in particular,
particular, there may very well be an impact on domestic monetary
policy.
steady and predictable money growth
growth is
is the
the foundation of
Since steady
reasonably stable exchange rates, there are considerable risks that
central bank foreign exchange intervention would buy only short term
stability.
And what are the benefits of
of exchange
exchange rate stability
stability
achieved by official intervention in
in the foreign exchange markets?
markets?
What
What can justify
justify official intervention?
Central banks do not in general have any better
better knowledge than
does
the future course
course of economic
economic variables.
does the private sector of the
There can be
be occasions when
when they do
do have such knowledge -- because they
--
know their own intentions but have
have not published them, or because they
are privy to the otherwise
otherwise undisclosed intentions of
of aa foreign central
bank.
fluctuIn that case, intervention to prevent aa temporary
temporary market
market fluctu-
nay be justified
ation may
justified but
but such intervention is
is inferior to
to making
the confidential knowledge on which
which it is based.
public the
Making the
Making
central bank’s
bank's intentions
intentions public would help stabilize not just the
differing degrees, every
every other market.
foreign exchange
exchange market but, to differing
Publicity,
Publicity, therefore, clearly dominates intervention.
A second defense of occasional intervention nay
A
may exist if
if it
it is
is
catefound that fluctuating exchange
exchange rates do, indeed, inhibit certain
certain cate-
gories of trade.
If stable
stable national monetary
monetary policies are being
—302—
-302-
pursued, there may still appear to be aa case for intervention.
The
case would be that
that some of
of the benefits from exchange
exchange rate
rate stabilizastabilization accrue not as profits to speculators
speculators on the foreign exchanges, but
to traders in goods.
There would,
would, in other words, be aa divergence
between the private arid social
social benefits of stabilizing speculation,
with the social
social benefits outweighing the private ones,
ones, thus appearing
appearing
to
to justify intervention.
intervention.
second best.
best.
But here, too, exchange
exchange intervention is
As has emerged from the literature on protection, aa
direct subsidy paid to the affected
efficient means
affected sector is the most
most efficient
8
8
of
In
of assisting aa sector of an
an economy.
In the present
present case, intervening
in the exchange markets would mean that
that all traders
traders in international
money
noney markets, not only those in
in goods affected by fluctuating exchange
rates, were being assisted.
Here, too, then,
then, while exchange
exchange market
market
intervention may
may conceivably be justified
justified -- although
although the evidence
-—
Nhich
may justify
is not
justify it
it is
not yet
yet in
in -- again
again the
the policy
policy is
is aa second
second best
best
•hich may
——
one.
::me.
Two further possibilities remain.
remain.
An
An exchange
exchange rate
rate may be
be
changing very rapidly -- sterling in the three months to
:hanging
to July
July 1979 is
—-
an example.
This
This was imposing very rapid adjustment
adjustment costs on
on indusindus-
tries already required to respond to aa substantial
substantial change in
in the
oattern of comparative advantage.
,attern
If
If the authorities in
in such aa case
can slow
monetary control, then there
:an
slow the adjustment without loss of monetaryc2p~pj,then
are
ire benefits
benefits from
from their
their doing
doing so.
But
But the
the situations when
when they can
can do
do
8See
“Domestic Distortions,
see J. Bhagwati and V. K. Ranaswami,
Ramaswami, "Domestic
rariffs, and the
the Theory
Theory of the Optimum Subsidy,”
Subsidy," Journal of Political
conomy, February 1963, and Geoffrey E. Wood,
,conomy,
l✓ ood, “Senile
"Senile Industry ProtecProtection,11 Southern Economic
Economic Journal, January 1975.
1975.
tion,”
-303-
so are manifestly rare. The
U.K. was able to do so in that episode bebe-
cause aa large part of
of the inflow seemed
seemed to have resulted from aa desire
to buy just the kind of
of securities the
the U.
U. K. government would have
have had to
sell to sterilize the inflow, but the experience
experience of Germany in the
l960s and l970s
1960s
1970s shows that such
such episodes
episodes are unusual.
This
This case,
case, then,
then,
does constitute aa modest
modest defense of occasional intervention -- but the
——
circumstances are very special.
(And there will still
still be
be aa welfare
cost to the nation if the rate of
of return earned on international
reserves falls short of the rate paid
paid on
on foreign-owned national
national debt.)
The fourth, and last, defense
defense is
is when there is
is an increased
increased
demand on the part of
of non-residents to hold the money of some country
country -— —
not, it should be
be stressed,
stressed, assets denominated in
in that currency, but
the currency itself, including of course
course bank deposits.
deposits.
This does
does not
invariably constitute
constitute aa reason for supplying the currency; it may,
rather, often
often be
be an opportunity for reducing
reducing the inflation rate.
rate.
If,
however, inflation is at its desired rate, them
then the increased demand
for currency must be
be met by
by an increased
increased supply, and the simplest way
way
to be
be sure
sure of supplying the correct amount is to
to operate
operate on the
the foreign
exchange market.
But this is aa very special case indeed.
Summarizing then, the case for official intervention in the
foreign exchanges is
is very weak.
Recognizing that there can
can be substansubstan-
tial fluctuations of exchange rates about their equilibrium
equilibrium values
values does
not imply that these fluctuations should be corrected
corrected by
by official
official
intervention.
Dr. Heller is also concerned about the appropriate
appropriate reserve asset
for the international monetary system.
—304-
He believes that the currently
currently
evolving reserve asset system is inherently unstable, and that it
should be
be replaced by a single asset system, the
the asset being
being either
either the
U.S. dollar or aa somewhat modified SDR (Special Drawing
Drawing Rights).
It is
deal first
of aa dollar
It
is convenient
convenient to
to deal
first with
with his
his endorsement
endorsement of
dollar
standard. The weakness of such aa system was first diagnosed
diagnosed by Robert
Triffin.99 His diagnosis can be
be summarized very briefly as follows.
The reserve asset, the dollar, can be supplied only by the reserve
centre, the United States, running
running continual
continual deficits in its balance
confidence in
in the
of payments -- but that progressively undermines confidence
--
reserve asset which is being thus supplied.
supplied.
nally inconsistent.
Such
interSuch aa system is inter-
Dr. Heller provides
provides us with
with no reasons for
thinking Triffin to be wrong
--
indeed,
indeed, nowhere does
does he refer to
Triffin so
so his advocacy of
of a return to aa dollar standard cannot be
taken seriously.
The defect with his endorsement of
of an SDR-based
SOR-based system is that
under one set of circumstances
circumstances the scheme is unnecessary, while under
under
the alternative circumstances it
it will not work.
An international monemone-
tary system with all major
major currencies serving as
as reserve assets is
is not,
despite his belief
belief to the contrary,
contrary, inherently unstable.
Such aa system
will not
not be
be continually
continually destabilized by capital flows responding to
reasoninflation differentials -- so
so long as these differentials are reason--
ably stable and predictable.
And
And when these differentials are not
stable and
and predictable, there will be
be sudden and large movements of
9Robert Triffin, Gold and the Dollar Crisis, Yale University
Robert Triffin, Gold and the Dollar Crisis, Yale University
Press, New Haven, 1960.
—305—
funds from currency to currency whatever the official reserve asset of
the system may be.
be.
of the international monetary
Tinkering with the reserve asset of
system cannot substitute
substitute for stable
stable domestic monetary
monetary policies.
CONCLUSIONS
CONCLUSIONS
The lessons for the conduct of international monetary policy
policy
which have
have been
been provided
provided by
by the experience
experience of the 1970s can be stated
very
very briefly.
briefly.
Exchange
rates will
be volatile
so long
Exchange rates
will be
volatile so
long as
as national
national
monetary policies
policies are volatile.
volatile.
It is not clear what harm this exex-
change rate
rate volatility
volatility does, although
although the underlying monetary instainstability does cause considerable
Meltzer’s paper shows.
considerable harm as Alan Meltzer's
shows.
In
any
any event, the case for exchange market intervention to
to reduce this
this
volatility is very circumscribed
circumscribed indeed.
Nor can
can any case be
be made for trying to prevent
prevent portfolio
portfolio diversidiversification into
into aa range of reserve
reserve assets.
A multiple
A
multiple asset system
system will
will
be stable if
if national monetary policies
policies are stable,
stable, and if national
monetary policies are unstable then
then any international monetary system
system
will inevitably be unstable also.
The
l97Os experience of
The lesson of the
the 1970s
of floating rates, as of every
earlier floating exchange rate episode, is that the international monemonetary system will only be as stable as the set of
of national monetary
systems
systems which it
it links.
links.
-306-
l980s
MONETARY POLICY ISSUES FOR THE 198Os
Lawrence K. Roos
Roos
As one of
of the sponsors of this conference, it is aa special
pleasure to
to welcome
welcome all
all of
of you
you to
to the
the Federal
Federal Reserve Bank
Bank of St.
St. Louis.
Louis.
It is aa privilege,
privilege, as well, to have the
the opportunity
opportunity of joining
joining you
you in
might learn
learn from past experience
experience in
in planning monetary
pondering how we might
policy for the
the future.
In the
the tine
time allotted me, II would like
like to share with you some imimpressions of past policymaking that I, in my four years
years as president of
of
the Federal
Federal Reserve Bank of St. Louis have gained, and to explore with
with
you what II believe we
we might look
look forward to in the years ahead.
ahead.
Looking
l970s, I would be less
if II did
Looking back in the 197Os,
less than
than candid ·if
not admit
admit to
deep feelings
with the
not
to some
some deep
feelings of
of frustration
frustration with
the way
way in
in which
which
monetary policy has been
underbeen conducted, as
as well as to
to aa failure to
to under-
stand how policies which
which produced such
such adverse consequences managed to
persist for so extended aa period.
Perhaps the best
best way
way to express my
my
feelings is
is to
to focus
focus on
few fundamental
concepts which
feelings
on aa few
fundamental concepts
which have
have come
come to
to
doninate my own understanding of the impact of monetary policy on the
dominate
economy.
First, and foremost, is the concept that inflation
inflation is fundamenfundamentally aa monetary phenomenon.
phenomenon.
This is an extraordinarily appealing
appealing
notion to
to me, if for no
no other
other reason than its generality and sheer
sheer
Mr. Roos is President of the Federal
Federal Reserve Bank of St. Louis.
Louis.
—307—
-307-
simplicity. As Beryl Sprinkel recently noted:
“It doesn’t
"It
doesn't take a
genius to
to know
know that if you pump more and more money into
into the system,
you get inflation."
inflation.”
Now, I suppose that if Beryl is correct
correct that it
truly does not take aa genius to understand this, then there is still
hope that this
this concept will come to be
be widely accepted.
Unfortunately,
Unfortunately,
the time
time lag necessary for acceptance of this
this appears to
to exceed, by aa
considerable margin, the
the time lags with which changes in money affect
prices.
Because the full impact of changes
changes in the rate
rate of growth of
of money
on the inflation rate occurs over
over aa considerable period of
of time (esti(estimated
mated variously
variously from three to six years),
years), it
it is important that monetary
policymakers, as well as the general
general public, clearly understand that
11
11
the 11“core”
inflation rate,
core 11 inflation
rate, or “underlying”
underlying 11 inflation
inflation rate, or “basic”
basic 11
inflation rate (to mention just aa few of the terms that have been
inflation
attached to it)
it) is determined by the long-term
long-term trend rate of
of growth in
in
money after adjustments for changes
changes in money
money demand.
In fact, because long-term
long-term changes
changes in velocity
velocity have, roughly,
roughly, had
equal and offsetting impacts to that for changes in
in output, the
the core
inflation rate
rate is essentially
essentially equal
equal to
to the
the trend
trend growth in
in money.
money.
BeBe-
cause the trend rate of growth
growth of
of money
money has
has approximated
approximated seven
seven percent,
11
the current “monetary—induced”
monetary-induced 11 rate
rate of inflation is
is about seven perper-
cent.
To put it somewhat differently, had there been
been no oil
oil shocks
shocks or
non—monetary indOced
other exogenous non-monetary
indUced impacts
impacts on prices,
prices, we
we would
would nevernever-
theless be
be currently faced with an inflation
inflation rate of about seven perpercent due solely to the growth in money
money that has emerged from past monemonetary policy
tary
policy decisions.
decisions.
-308-
AA careful understanding of the difference between
between the actual ininflation rate
rate and the monetary-induced
monetary-induced or core
core rate of inflation
inflation is
crucial for the proper conduct of monetary policy.
Only the monetary—
monetary-
induced rate
rate of inflation should concern monetary policymakers; it
it is
the only component of
of inflation that they can influence.
influence.
Exogenous
shocks such as those caused by
by higher energy prices or crop
crop failures
will
will always contribute
contribute to
to the
the current measured inflation
inflation rate,
rate, but
but
their impact is transitory.
Attention paid to these
these exogenous influinflu-
ences on prices
prices must never divert monetary
monetary policymakers
policymakers from focusing
their actions
actions toward controlling,
controlling, and reducing, the monetary-induced
monetary-induced
rate of inflation.
A second key concept that has guided my understanding
A
understanding of
of the imim-
pact of monetary policy is that abrupt and substantial changes in the
pact
growth of money, if sufficiently
sufficiently prolonged,
prolonged, have dramatic
dramatic and usually
unfortunate consequences for the economy.
Unusually rapid growth in
in
money, if sustained for several quarters, while having some
some positive
positive
effects on employment and output for aa short time, will
will ultimately
ultimately and
inevitably increase the monetary-induced rate of
of inflation.
inflation.
Similarly,
Similarly,
unusually slow growth in money, if sustained for several
several quarters,
quarters, will
unusually
result in reduced growth in
in output
output and employment -- perhaps, even
even aa
--
recession,
recession, and ultimately
ultimately reduce the monetary—induced
monetary-induced rate of
of inflation.
inflation.
Careful understanding of the short-run consequences
consequences of
of sharp
proper conduct
conduct
fluctuations in the growth of money is crucial for the proper
of monetary policy.
of
To avoid undesirable
undesirable results such
such as
as recession or
or
To
an
an over—heating
over-heating of
of the
the economy, monetary
monetary policymakers
policymakers must avoid policy
policy
capricious changes in the growth of
actions that result in sudden or capricious
-309-
money. They
They should,
should, instead,
instead, conduct
conduct policy
policy In
in such a way
way that
that changes
changes
in
systematic and
and gradual
gradual.
in the growth of money are systematic
A third concept that has guided my understanding is that the
A
growth of money can best be
on the behavior
be controlled, not by focusing on
of interest rates, but by
by controlling the growth of the monetary
monetary base.
Since
Since the
the Federal
Federal Reserve controls the largest component
component of
of the monemane-
tary base
tary
——
credit -- growth of
of the monetary
monetary base is
Federal Reserve credit
-—
directly and completely in the hands of the Federal Reserve.
Similarly,
Similarly,
is considerable
considerable evidence that
that the multiplier
multiplier linking the monetary
there is
base to
to the money
money stock is sufficiently stable and
and predictable
predictable to
assure aa reasonably close relationship between growth of
of the base
base and
growth of money
money over
over all but
but the shortest-term
shortest-term period.
Consequently,
the
the lesson
lesson for policymakers is that, if control of the growth of money
is to be
be a crucial part
part of monetary policy, desired money growth rates
should be
be linked directly in
in the policy process
process to the qrowth of
of the
monetary base.
base.
Finally, aa fourth concept which has enabled me to understand the
impact of monetary
monetary policy on
on the
the economy is that economic markets,
markets,
especially the financial and foreign exchange markets,
markets, are reasonably
rational and efficient.
efficient.
Thus, increased rates of money growth tend to
of the dollar on
produce higher interest rates and to lower the value of
international exchange markets as soon as the financial market
market partici
participants,
pants, who seem to
to be
be well aware of the association between money
come to expect increased future inflation rates.
growth and inflation, come
It follows that, while so-called “tighter”
"tighter" monetary policy may immediimmediately produce higher interest rates, the same result occurs with
“looser”
looser monetary policy in the longer time
time span.
span.
11
11
-310-310-
Interest rate
movements per se are unreliable guides to policy. This is especially
true
true when we consider that interest rates,
rates, which
which represent
represent the price of
of
credit, are also affected by aa host of
of non—nonetary
non-monetary influences.
the above concepts is especially complex and certainly
certainly
Now, none of the
none is likely to
to be either
either new or controversial to
to most
most of you.
HowHow-
assessing the likely
ever, they do provide an analytical framework for assessing
results
results of monetary policy actions.
actions.
It is this basis of analysis that
has
ny frustration in viewing what has happened
has led
led to
to my
happened over the past
four years.
years.
No one, who
who believes as
as II do that the most
most significant
component of inflation is
is monetary, could
could have failed to
to have
have been
been conconcomponent
cerned
cerned with growth in money that accelerated from five percent over
over the
period
1/73 to 111/76 to eight percent
percent from 111/76 to 111/78,
111/78,
period from 1/73
core rate of
of inflation.
guaranteeing a significant increase in the core
No
No
one, who believes as II do that drastic changes in the qrowth of money
money
one,
produce undesirable economic
economic consequences, could have failed to
to be con
concerned when the
the money stock, having grown at the rate of eight percent
for two years, suddenly dropped to aa less than two percent growth for
the period from September
September 1978
1978 to
to May 1979, virtually
virtually assuring aa major
economic
economic slowdown.
~nd,
And, certainly, no
no one,
one, who believes as II do that
financial markets are rational
rational and efficient, could fail to be
be disdistightness’
turbed by
by the
the current
current expressions of concern with alleged "tightness"
of
of monetary
monetary policy, as judged by the ‘high’
"high" levels of
of nominal interest
interest
rates.
Money growth at rates approachinq
approaching 10 percent and an inflation
inflation
rate of close to
to 10
10 percent
percent are certainly not reflections of tightness
tightness.
Certainly
Certainly the financial
financial and foreign exchange market participants
participants have
not been fooled; witness the behavior of
of interest
interest rates and the value
of the dollar over the last few months.
months.
—311—
-311-
But my frustration is not confined only to the unfortunate conconsequences of past
past monetary policy actions. It
It also lies with the monemoneobtary policymaking process itself that produced the results we have ob-
served throughout the l970s.
1970s.
Time and time again,
again, II have
have observed
observed the
Reserve’s interest rate target while
achievement of the Federal
Federal Reserve's
while money
“desired” target
growth was permitted
permitted to
to wander at will outside its !ldesired
target
0
ranges.
As II noted in
in an earlier
earlier discussion in London last
last June, the
monthly “betting
odds” during
"betting odds"
during the past four and aa half years
years have been
been
only
only about one in two that
that Ml
Ml would remain inside its
its target range.
Moreover, there is little
little doubt that the conduct of
of monetary
monetary policy, by
of interest rates,
rates, has produced aa procyclical
focusing on stabilization of
pattern in the growth in money.
That
That pattern has tended to exacerbate
cyclical movements and exogenous
the impact of
of cyclical
exogenous shocks
shocks on the economy.
But,
But, again, none of this is especially new to you.
Many
Many of
of you
have
have contributed
contributed over
over the
the past decade
decade to
to studies critical
critical of both
both the
monetary policymaking process and
and policy
policy consequences.
I, too,
too, have
been convinced, both
exarguments to which II have been
been exboth by the economic arguments
posed, and by
by aa first-hand view of
of the disappointing
disappointing results of
of the
policies pursued, that only aa major
major change in
in the formulation of monemone-
tary policy -- away
away from concentrating on stabilization
stabilization of interest
—-
rates and towards focusing on the monetary
monetary base
—-
would enhance the
prospects
successfully achieving
achieving the results we desire
desire from
from monetary
monetary
prospects of successfully
policy.
The announcement
announcement by
by Chairman Volcker on
on Saturday, October 7, that
the Federal Reserve is changing its procedures of monetary
monetary policymaking
policymaking
controlling growth of
of the reserve
reserve aggregates
to place more emphasis on controlling
while
while permitting interest rates to fluctuate freely, represents a gaint
—312—
-312-
step In
in correcting past mistakes. There is no doubt In
in my mind that if
this new approach is effectively implemented
implemented in the upcoming months and
years, we can
can achieve control
control over the
the growth of money
money and,
and, conseconse‘basic” rate of inflation.
quently, control over the
the "basic"
Similarly, we can
avOid the
avoid
the adverse real sector impacts that have resulted from uninunintended drastic short-run
short—rum fluctuations in the growth of money around its
its
longer—run trend rate of
longer-run
of growth.
Finally, once the financial market
participants are convinced
convinced that we have indeed seized control over the
the
growth of
of money and intend to bring about the gradual reduction in
in
money
money growth necessary
necessary to reduce the
the core inflation
inflation rate, II believe
that we will see an end to the
the surges
surges in interest rates amd
and declines in
the value of the dollar
dollar which
which have proved so
so troubling
troubling in
in the past.
Thus, as you
enthusiyou may have inferred
inferred from my
my comments, II am
am enthusiastic and encouraged about the change in the policymaking process that
has occurred.
However, my euphoria is restrained by
by aa realization that
several problems still remain if this change in policy is to produce
results.
the hoped—for
hoped-for results.
To assure maximum effect from the Fed’s
Fed's new
policy the following steps must be
be taken:
taken:
1)
1)
Instead of placing
placing sole emphasis on controlling
controlling the growth
of
policymakers should focus also
of non-borrowed reserves, policymakers
on
on growth in the monetary
monetary base and
and total
total reserves.
reserves.
There
There
just too many
many slips
slips twixt growth in non-borrowed
reare just
non—borrowed reserves and growth in
in money.
2)
Policy emphasis must be firmly and fundamentally rediredirected from concern about movements
movements in the Federal funds
rate to concentration
concentration on
on growth in
in the monetary base
base and,
and,
hence, the money
money stock.
stock.
The substance
substance of policy must
must go
-313—313—
beyond merely widening the permtsslble
permissible range of movements
in
in the Federal
Federal funds rate.
rate.
For, if
if widened Fed
Fed funds
funds
rate constraints remain even
even remotely binding,
binding, monetary
control cannot succeed.
3)
The new method must be given
given adequate
adequate time
time to prove itit-
self.
The success of the
the new monetary control proceproce-
dure cannot be reasonably evaluated by observing money
stock behavior over aa short time span.
Not
Not even the most
most
ardent academic
academic advocate of
of base targeting
targeting asserts that
that
precise money
money control is possible over
over aa period of
of six
six
months or less.
less.
At the very least,
least, aa one year testing
testing
period is necessary for any comparison between
between previous
methods and the current
current one.
Moreover, no one should
should
inflation to dissipate in a matter of months.
expect inflation
InIn-
flation has
has been generated over
over aa period of
of 15
15 years and
cannot be eliminated overnight.
It
It would be
be tragic if
this new approach to policymaking were
were to
to be
be tried and
abandoned after aa short time because of false expectaexpectati
ons.
tions.
4)
Finally, and perhaps most importantly in the short run,
the procedures for implementation of
of the new policy, the
rules of the game, must be
be clearly enunciated to the
public.
As we have observed
observed during the first week after
announcement of the new
new approach, the lack
lack of clearly
clearly
announcement
articulated rules
rules produced
produced aa near panic in
in financial
financial marmar-
kets.
There
There is no reason
reason to shroud policy in secrecy
and to
to keep
keep markets guessing.
-314-
While
While surprises might
might have
had some value in policies directed
mardirected toward money
money market
counterproductive when
ket stabilization, surprises are counterproductive
when
monetary aggregates become the target.
Above all,
all, the attention of
of policymakers must be
be focused on the
longer—run impacts of policy.
longer-run
Unfortunately, as
as Arthur
Arthur Burns noted in
his
Per Jacobsson
Jacobsson Lecture,
Lecture, the
“anguish of
of central
central banking’
his Per
the "anguish
banking" has
has often
often
cone
come from the
the short—term
short-term political
political pressures
pressures on
on monetary authorities
--
pressures
for—bad-and—for—worse, the monetary authorities
pressures to which, for-bad-and-for-worse,
have all too often succumbed.
What is needed more than ever
ever before is a steady hand on the
tiller of monetary policy.
subNot only will the Fed’s
Fed's new
new policy be
be sub-
jected
jected to critical analysis by those who traditionally
traditionally have
have doubted
doubted the
feasibility of
of monetary control; the
the very credibility
credibility of this country’s
country's
central bank is at stake.
II trust that we will have the wisdom
wisdom to
to
implement our policy effectively, the openmindedness
openmindedness to judge our
our propro-
the courage to resist whatever pressures might arise
arise
gress fairly and the
to retreat
retreat from the historic
historic step
step we have
have taken.
taken.
—315—
-315-
@FedFRASER