Full text of The Monetary Versus Fiscal Policy Debate : Lessons from Two Decades, October 12, 1984

The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.

Preface and Acknowledgments

The selection of this conference topic provides an opportunity to honor
Homer Jones, director of research at the Federal Reserve Bank of St.
Louis from 1958 to 1971. Although some students of the monetary versus
fiscal policy debate may not recognize the name, Homer’s influence on the
topics discussed in this volume is substantial. As director of research,
Homer was instrumental in the development of the St. Louis Fed’s statistical publications that provide current information on the important
monetary and financial measures. His intellectual prodding of the
research staff to investigate various aspects of monetary theory and policy,
and his encouragement to report their findings in the Bank’s Review and
in scholarly journals helped to contribute to our understanding of the
effects of monetary policy actions on the economy. In appreciation for his
efforts, we gladly dedicate this conference volume to Homer Jones.
Many people associated with the Research and Public Information
Department at the Federal Reserve Bank of St. Louis contributed to the
planning of the conference upon which this volume is based. In particular, I would like to thank Dan Brennan, Melissa Daubach, and Linda
Moser for their assistance. I also would like to thank Spencer Carr of
Rowman & Allanheld for his help in the production of this book.

2
Monetary versus Fiscal Policy Effects:
A Review of the Debate
Bennett T. McCallum
CARNEGIE-MELLON UNIVERSITY
AND NATIONAL BUREAU OF ECONOMIC RESEARCH

1. Introduction
The “monetary versus fiscal policy” debate has not attracted much attention in recent years, and, in some ways, this is not to be regretted. It may
therefore be useful to begin this discussion with a quote from Phillip
Cagan (1978, 85—86):
No one who was not in touch with the economics profession in the 1940s
and 1950s can quite imagine the state of thinking then in the profession at
large on monetary theory and policy. The quantity of money was not considered important, indeed was hardly worth mentioning, for questions of
aggregate demand, unemployment, and even inflation... Textbooks in
basic economics and even in money and banking mentioned the quantity
theory of money, if at all, only to hold it up to ridicule. Those textbooks
produced an entire cohort of professional economists who became the
teachers of hordes of economics students. There were, of course, many
exceptions, most notably at the University of Chicago... But if you traveled among the profession at large, mention of the quantity of money elicited puzzled glances of disbelief or sly smiles of condescension.
As a former member of the horde of students that Cagan refers to, I
would modify his statement only by suggesting that the period in question
lasted longer than he indicates. For readers who find this claim hard to
The author is indebted to Carl Christ, Stanley Fischer, Robert King, Thomas
Mayer, Lawrence Meyer, and Allan Meltzer for helpful comments and to the
National Science Foundation (SES 84-08691) for financial assistance.

10

Bennett T McCallum

believe, I would suggest perusal of the contents of the American
Economic Association’s Readings in Business Cycles, published in 1965
(Gordon and Klein 1965). Another interesting record is provided by the
chapter on inflation in Ackley (1961), the standard, graduate macro text of
the l960s. Ackley begins with a brief description of the “classical school”
theory in which inflation depends primarily upon growth of the quantity
of money, but quickly moves on to other ideas.’
Today, of course, matters are different. There are few economists who
would label themselves “monetarists,” but most publishing macroeconomists hold views that would have been two standard deviations away from
the mean—in the monetarist direction—in 1965. As one piece of evidence
in support of that claim, I would point to the specifications (i.e., list of
variables) of vector autoregression (VAR) models currently being used for
forecasting and/or analysis. In particular, the VAR systems of Sims
(1980, 1982), Litterman (1982), Gordon and King (1982), Webb (1984),
and B. M. Friedman (1984) all include monetary aggregates but no fiscal
variables.
Both sides in the debate can of course claim victory: the monetary policy supporters for the reasons implicit in the foregoing comparison, and
the fiscal policy supporters by citing theory and evidence indicating that
fiscal actions are not without effect on aggregate demand. But the shift
has certainly been in favor of the former.
The purpose of this paper is to review developments bearing on this
debate over the last 20 years. To describe all of the significant items in
the literature would require a paper of inordinate length and one that
would be extremely dull for the participants at this conference, most of
whom are intimately familiar with much of the material. Fortunately, it
turns out to be unnecessary for me to provide an extensive treatment of
the main threads of the argument, for a careful review was published
fairly recently by Meyer and Rasche (1980). I will be able, then, to pass
quickly through the well-known items and arguments and focus my attention on items that (i) have been neglected, (ii) are fairly recent, or (iii) are
somewhat original. The organization of the paper is very simple: Section
2 is concerned with empirical studies and econometric points of interpretation, Section 3 discusses some of the main theoretical issues, and Section
4 provides some tentative conclusions and judgments.

2~Empirical Results and Econometric Issues
2.1 Single-equation results
In discussing the empirical analysis and related issues, it is convenient to
organize the discussion around results featured in a long series of articles
prepared by the research department of the Federal Reserve Bank of St.

Monetary versus Fiscal Policy Effects

11

Louis.2 Brief mention should be made, however, of the earlier paper by
Friedman and Meiselman (1963), which is not discussed by Meyer and
Rasche.
The central ingredient of the Friedman—Meiselman study was a comparison of simple correlations (based on annual U.S. data for 1897—1958
and subperiods) of consumption with money stock magnitudes, on the one
hand, and consumption with “autonomous” fiscal variables,3 on the other
hand. Friedman and Meiselman found that movements in consumption
were more highly correlated with monetary rather than fiscal variations.
Their own summary statement goes as follows:
The results are strikingly one-sided. Except for the early years of the Great
Depression, money. is more closely related to consumption than is autonomous expenditures.... This is so both for nominal values.., and for
“real” values.
It is true both for absolute values and for year-to-year or
quarter-to-quarter changes. Such correlation as there is between autonomous expenditures and consumption is in the main a disguised reflection
of the common effect of money on both.... One implication of the results
is that the critical variable for monetary policy is the stock of money, not
interest rates or investment expenditures. (1963, 166)
. .

..

.

Given the climate described by Cagan, the central role of the consumption function in Keynesian analysis, and some questionable methodology,
the Friedman—Meiselman study was welcomed by the profession about
like an unexpected slap in the face. Strongly critical studies were published by Hester (1964), Ando and Modigliani (1965), and DePrano and
Mayer (1965). Strongly worded replies and rejoinders of great length followed promptly.
It is clear that the Friedman—Meiselman approach was in fact open to
several methodological objections. Its critics emphasized the questionable
nature of the measure used for autonomous expenditures as well as the
delineation of sample subperiods. Especially troublesome was the focus
on contemporaneous relationships in single regression equations including
only one or two explanatory variables. Most researchers in macroeconomics believed, I would guess, that investigation of the issues under
discussion could be adequately carried out in the context of a fully
specified, simultaneous-equation, econometric model. The judgment of
Blinder and Solow (1974, 65), expressed a few years later, was that “all of
this was essentially pointless. The issue is simply not to be settled by
comparing goodness of fit of one-equation models that are far too primitive to represent any theory adequately.”
Thus the Friedman—Meiselman results were in the process of being
shrugged off when the first of the St. Louis studies—that of Andersen and
Jordan (1968)—appeared. As is very well known, that study featured a
least squares regression fit to quarterly U.S. data for 1952.1—1968.2 in

12

Bennett T McCallum

which the dependent variable was the change in nominal GNP, and the
explanatory variables were current and lagged values of changes in the
money stock (Ml or base), full-employment expenditures, and fullemployment tax receipts. The striking finding was that the sum of the
coefficients on the monetary variable was highly significant, whereas the
sum of the coefficients on both fiscal variables was insignificantly different
from zero.4
2.2. Econometric objections
Again, sharply critical objections were raised. The main lines of argument
were brought together by Blinder and Solow (1974) and nicely reviewed
by Meyer and Rasche (1980). My strategy here will be to focus on three
points emphasized by Blinder and Solow (see below) and on the evidence
concerning reliability of “reduced-form” procedures that was subsequently
described by Modigliani and Ando (1976). It will be convenient, though,
to begin by quickly mentioning the observation of Benjamin Friedman
(1977) that when data for the period 1970.1—1976.2 is added to the
Andersen—Carlson (1970) sample, the sum of the fiscal variable
coefficients becomes significantly positive. It was quickly shown by Carlson (1978), however, that this conclusion does not obtain when the variables are entered in rate-of-change form rather than as first differences.
In effect, Carlson’s suggested specification amount to
=

a + f3(L)1Xm~+ y(L)z~g~+ u~,

(1)

where y~,m , and g~are logarithms of nominal GNP, a money stock meas1
ure, and a fiscal variable, respectively; u~is a stochastic disturbance, and
$(L) and y(L) denote finite polynomials in the lag operator defined by
L~x = ~
so that $(L)t~m stands for a distributed lag such as f30z~m1
1
1
+ fl,~m _, +
+ $ki~m~k.Carlson’s specification not only has the
1
desirable feature of relating relative rather than absolute changes, but it
also leads to residuals that are more consistent with the standard assumption that u1 is a white-noise disturbance.5 Consequently, (1) seems preferable to the specification used previously, and this judgment leads to the
conclusion that the inclusion of data for years since 1969 does not reverse
the original finding that the sum of the
coefficients is insignificantly
different from zero. In what follows I will accordingly presume that (1) is
the relevant specification.
Let us now consider, then, the criticisms of Blinder and Solow (1974).
Those writers summarize their position very concisely as follows:
...

In summary, the Andersen—Jordan study errs for at least three reasons, any
one of which is sufficient to render their results meaningless. First, by omit-

Monetary versus Fiscal Policy Effects

13

ting all exogenous variables other than fiscal or monetary policy, they seriously misspecify the reduced-form equation for real [sic] output
Second, they use an incorrect measure of fiscal policy, which biases the
coefficient toward zero. Finally, and most damaging to their position, they
treat fiscal and monetary policies as exogenous when it is intuitively obvious
that the authorities are in some sense reacting to movements in the macroeconomy. (pp. 70—71)
.

.

These three difficulties correspond to those discussed by Meyer and
Rasche (1980, 56—63) and Batten and Thornton (1983, n. 2). We will take
them in turn, starting with the omission of exogenous variables. In this
case the contention is that the true specification is not (1) but
a + f3(L)i~m1 + y(L)z~g + 8(L)z~z~
+ u~,
(2)
1
where; is an additional variable6 that has an important influence on y~.
For the moment, let us suppose that z~mtand z~g~
are in fact exogenous, a
supposition that will be considered later in connection with the third
difficulty. Now Blinder and Solow refer to; as an exogenous variable.
But the conditions for true statistical exogeneity are extremely stringent:
z1 must be generated by a process that is independent of current and past
values of Yt (and thus ui). But it is hard to imagine any important
macroeconomic variable that is truly exogenous in that sense; even population growth and technical change probably respond (with lags) to
fluctuations in GNP.7 The point is better expressed, then, as suggesting
that (2) is applicable with ~z, interpreted as an endogenous variable that is
affected by ~Yt’ ~
and z~g~
only with a lag. Suppose that relation is
=

=

a + ~
0

+ a 1~m~,+ a
g~,+
2
3~

~t’

(3)

where ~ is a stochastic disturbance term. Then by substitution and rearrangement we have
=

[a + 8(L)a } + [fl(L) + a L6(L)]1~m~
0
2

(4)

+ [y(L) + a L~(L)]~g~
+ a,~(L)z~y~,
+ [u, + 8(L)~,],
3

which is not in the form of (2) because of the appearance of a,8(L)y~,.
That can be remedied, however, by moving the ~
term to the left side
and multiplying through by [1
La ,~(L)]’:
—

l~y

=

t

a’ + /3’(L)z~m~
+ y’(L)z~g~
+ u’s.

(5)

Here we have a relation of the form of (2), but (of course) with disturbances and parameters that are quite different:

14

Bennett T McCallum

=

[1

—

La,~(L)]’ [f3(L) + a L~(L)]
2

=

[1

—

La,8(L)]’ [y(L) + a L~(L)]
3

=

[1

—

La,8(L)]’ [u + i3(L)~~].
1

What we have in (5) is not a reduced-form equation but a final-form
equation for ~Yt (still assuming that ~mt and i~g1are truly exogenous).8
Thus we see that the coefficients in (5), which will be estimated by the
St. Louis procedure, are not the reduced-form coefficients in (1). The
estimated values will, under the assumption that ~(L) ~ 0 in (2) and that
(3) is a “stable” relationship, reflect indirect influences of i~.m,and L~g,on
~Yt by way of 1~z,. But the importance of that observation is quite
unclear; the coefficients in a reduced-form equation like (I) also reflect
influences that are “indirect.” The first problem mentioned by Blinder
and Solow hardly justifies, then, terming the St. Louis estimates “meaningless.” It merely implies that in interpreting the estimates it may be
necessary to keep in mind that they reflect “indirect” effects.9
It should be added that the foregoing argument presumes that Eq. (3) is
“stable”—that is, unchanging over the sample period. Such relationships
may in fact change significantly over real-time sample periods, for reasons
explained by Lucas (1976) or for other reasons. But precisely the same
must be said for (1); these two relationships are on the same footing in
that respect. If there is reason to believe that (1) is stable over a period,
there is no particular reason to expect (3) to be shifting.
Let us turn next to the second of the Blinder—Solow (1974) points—that
the St. Louis measures of fiscal policy may be “incorrect.” This means
that the St. Louis variable does not correspond to the “weighted standardized surplus” shown by Blinder and Solow (1974, 23, 33—34) to be the
measure that would appear in a reduced-form expression within the particular model that they use for illustrative purposes. But this is a conclusive criticism only to one who has some attachment to the particular
model in question,’°and then only if he believes that the effect of using
alternative measures would be large. Although Corrigan (1970) found
that a specific measure, which Blinder and Solow consider reasonably
attractive, led to estimates indicating a significant fiscal effect, it is my
impression that the variable measurement problem is not of overwhelming importance empirically. Or, to put it differently, if there is a fiscal
policy measure that carries a strongly significant sum of coefficients in an
equation of the St. Louis form, its existence has not been well publicized.” In any event, analysis of the policy-variable measurement issue is
undermined by the third problem discussed by Blinder and Solow, to
which we now turn.

Monetary versus Fiscal Policy Effects

15

The issue in this case results from the endogeneity of policy actions, a
topic that was investigated in detail by Goldfeld and Blinder (1972). The
Goldfeld—Blinder paper correctly emphasizes the distinction between exogeneity, in the sense of coming from outside the private sector of the
economy and exogeneity in the statistical sense. And they note that it is
almost inconceivable that either the monetary or the fiscal variables in the
St. Louis studies could be exogenous in the latter sense, for that would
imply that the authorities’ actions are not systematically influenced to any
significant extent by current or past macroeconomic conditions. If in fact
the authorities’ actions are so influenced, then (2) may not be a proper
reduced-form equation even if no variables are omitted. In particular, the
disturbance term u, will be correlated with regressor variables if policy
actions respond to current-quarter conditions, or if policy responds with a
lag but the disturbance in (1) is autocorrelated. Under such circumstances, least squares estimates of f3(L) and y(L) will of course be biased
and inconsistent.
This point is clearly correct in principle and could easily be of importance empirically.’2 The numbers reported in Goldfeld and Blinder’s
Table 8 suggest, however, that the downward bias should be about the
same for monetary variable coefficients as for fiscal variable coefficients,
except in the event that the fiscal authorities are extremely prompt and
accurate in their stabilization efforts. It seems unlikely that this would be
the case for the U.S. economy.
Nevertheless, the point is important enough to warrant continued discussion. Given that the policy authorities do respond to current or recent
macroeconomic conditions, the best way to proceed in estimating equations like (1) or (5) is to append policy-rule or reaction-function
specifications for the authorities and carry out simultaneous equation estimation, as recommended by Goldfeld and Blinder. But it is not necessary
to explicitly formulate equations descriptive of policy behavior; consistent
estimates of (5) can be obtained by estimating that equation in isolation
but using instrumental variables (IV) rather than ordinary least squares
(OLS) estimators. Comparisons of IV and OLS results should indicate
whether severe biases are in fact induàed by the reactive behavior of the
policy authorities.
The results of an extremely brief and tentative exploration of that type
are reported in Table 2.1. The first pair of numerical columns gives
coefficient and standard error values for an OLS regression of the form of
(5), using quarterly data on nominal GNP, the Ml money supply, and
nominal federal purchases of goods and services for 1954-80. Despite the
absence of polynomial constraints and the use of a different expenditure
variable, the results are much like those that have been featured in the St.
Louis studies. In particular, the Am1~ variables are strongly significant

16

Bennett T McCallum
Table 2.1
OLS vs. IV Estimates of St. Louis Equations
Sample Period: 1954.1 - 1980.4
Regressor
Constant
~m
1
L~m~
1
L~m~
2

L~.m~

3

L~m~
4

OLS Estimates
Coef.
(SE)

0.007

(0.002)

0.651
0.226
0.308
0.148
—0.240
0.099
0.026
0.005
— 0.050
— 0.038

(0.119)
(0.128)
(0.148)
(0.182)
(0.155)
(0.031)
(0.032)
(0.032)
(0.032)
(0.031)

IV Estimates*
Coef.
(SE)
0.009
(0.003)
0.167
(0.387)
0.615
(0.320)
0.191
(0.251)
0.401
(0.341)
— 0.297
(0.240)
(0.221)
—0.190
0.058
(0.079)
0.049
(0.055)
(0.048)
—0.025
—0.033
(0.044)

~g,
L~g,
1
~gt-2
L~g,
3
~
2
R
0,531
0.053
1.91
DW
1.90
SE
0.0080
0.0113
*Instruments are fitted values from OLS regressions of L~m~
and L~g~
on L~m~
1
~gt-i
~
rti, r~ and
, a constant. The dependent variable in each case
2
i5 ~

as a group with coefficients summing to 1.09, whereas the fiscal variables
enter less strongly and their coefficients sum to only 0.04.13
In the second pair of numerical columns are IV estimates for the same
period, with the instruments for 1Xm~and L\g~created by first-stage regressions of those variables on their lagged values ~m, ,
~
1
~gt— ’ plus r , and rt , with rt the 90-day treasury bill rate.
1
4
2 the assumption that r , and r~ affect polIndentification
is provided by
2
1
icy instrument settings in period t. Because there is no strong sign of serially correlated disturbances, these estimates should be consistent for the
parameters of (5) even in the presence of current-period responses of ~m,
and Ag~to values of ~Yt or other endogenous variables. As is readily
apparent, there are two major differences in these estimates as compared
to OLS. First, the coefficients attached to current-period values of ~
and ~g, are much smaller, with the latter turning negative. Second, the
standard errors are much larger, indicating a substantial reduction in the
reported accuracy of the coefficient estimators. In part, the latter is
induced by lower overall explanatory power, but in part it is also due to
the increased collinearity that arises when ~ and ~ are replaced by
constructed variables that are primarily linear combinations of lagged ~m
and ~g values.
It therefore appears from this experiment that the effects stressed by
Goldfeld and Blinder may indeed be of quantitative importance. Yet
even in these estimates it remains the case that the coefficients of
.

. ,

Monetary versus Fiscal Policy Effects

17

sum to approximately 1.0, whereas those on
sum4 to approximately zero. (The numbers are 1.08 and
-0.14, respectively.)’
Once it is recognized—as seems necessary—that policy instruments are
set in response to current or recent conditions, then the motivation for
using high unemployment values of fiscal variables is lessened or eliminated. Furthermore, it even becomes unclear why a distinction is drawn
between “discretionary” changes in expenditures or taxes and automatic
changes brought about as a result of the built-in stabilizers. It would be
appropriate to distinguish empirically between automatic and nonautomatic instrument changes if there were theoretical reasons to believe
that these would have different effects on nominal GNP, but the
econometric reasons apparently motivating Andersen and Jordan (1968)
do not seem applicable.’5 Consequently, it would appear that there
remains room for an empirical study that emphasizes the endogenouspolicy effects emphasized by Goldfeld and Blinder, and that also considers the impact on aggregate demand of tax changes brought about by the
built-in stabilizers of the U.S. tax system. A major reason why such a
study is still lacking is the difficulty in modeling policy behavior, together
with the absence of genuine exogenous variables.

2.3 The Modigliani—Ando study
Let us now turn to the rather interesting experiment conducted by Modigliani and Ando (1976), who hypothesize that (even ignoring
endogenous-policy issues) estimates of policy multipliers obtained by
means of the St. Louis approach are extremely unreliable. In support of
that hypothesis, Modigliani and Ando report that the St. Louis equation,
when estimated using artificial data generated by simulations of the MPS
model, provides very poor estimates of that model’s known multipliers—
estimates that are on the low side for fiscal multipliers and on the high
side for monetary multipliers. Although this finding does not literally
imply that the St. Louis estimates of the economy’s multipliers are
incorrect, it has been regarded by several reviewers’6 as tending to
discredit the St. Louis procedure and, thus, its results.
Reflection suggests, however, that this conclusion may not be warranted:
The reason for doubt involves the point made above—that there are virtually no macroeconomic variables that can appropriately be treated as
exogenous. If that point is correct, then the Modigliani—Ando experiment
will be misleading in the following way. Under the hypothesis at hand—
that there are no exogenous variables—the MPS model (which treats a
large number of variables as exogenous) incorrectly omits a large number
of behavioral relationships analogous to (3). Therefore the “true” multi-

18

Bennett T McCallum

plier values for the MPS model reported by Modigliani and Ando
correspond to the fl(L) and y(L) values in Eq. (2). But estimates of the
MPS multipliers obtained by the St. Louis approach correspond to the
/3’(L) and -y’(L) coefficients in (5), with a~values in (3) being whatever is
implied by the historical values of the variables treated (incorrectly) as
exogenous, the historical values of the policy instruments, and the simulated values of the model’s endogenous variables. In other words, the
comparison reported is between actual MPS values of the coefficients in
(2) and estimated values of the coefficients in (5). Since these coefficients
are truly different, under the hypothesis that there are no exogenous variables in reality, the discrepancy between estimated and true values does
not indicate unreliability of the estimation procedure.’7 This argument, it
should be said, seems to amount to an elaboration of the comment provided by Darby (1976).
2.4 Evidencefrom large models
One useful feature of the Meyer—Rasche survey is its compact tabulation
of policy multipliers for real GNP implied by seven prominent
macroeconometric models. Although there is considerable disagreement
among the other models concerning the magnitudes of fiscal and monetary multipliers, the St. Louis results do depart significantly from the average values of the other models. The government spending multiplier after
four quarters, for example, is 0.5 for the St. Louis model as compared
with a mean of 2.17 for the BEA, Brookings, Michigan, DRI, MPS, and
Wharton models. In the case of the monetary variable, the comparison is
not straightforward, because the St. Louis multipliers pertain to an Ml
variable, and the others pertain to unborrowed reserves. But the othermodel average about four quarters is 3.0, so the St. Louis value of 4.4 is
much greater in elasticity terms, the ratio of M 1 to unborrowed reserves
being about 10.
It will be noted that these multipliers are for real GNP, so their magnitudes depend upon the model’s precise specifications of dynamic Phillips
relationships. Since it is well known that there exist major disagreements
over the proper specification of this relationship, it is, in principle, not
surprising that the multipliers diverge. What one might hope for is some
agreement concerning nominal GNP multipliers. That, in any event, is
the topic under discussion in this paper—the relative impact of monetary
and fiscal actions on aggregate demand.
But the foregoing statement applies only in principle. In fact, the
predicted price level responses from any of the models in question are so
slow that four-quarter multipliers are essentially the same both for
nominal and real GNP. Thus we see that there is a considerable

Monetary versus Fiscal Policy Effects

19

discrepancy between St. Louis and other-model responses to a monetary
policy action, though less than in the case of a fiscal policy action.
Niehans (1978) and McNelis (1980) have suggested that the implications of non—St. Louis econometric models are actually much more
“monetarist” than most observers have recognized. Their argument starts
with the idea that the proper comparison of fiscal versus monetary policy
effects requires that an unreversed $1 billion/year increase in government
spending should be compared with a continuing sequence of $1 billion
increases, one per year, in the high-powered money stock. And when this
comparison is made, it is found that “for every model except BEA and
Wharton III.
the peak of the money multipliers must be many times as
high as the peak of the hypothetical fiscal multipliers, a typical ratio being
perhaps 15:1
The proposition that the quantity of money matters
much more than the way it is created is evidently common to both”
monetarist and nonmonetarist models (Niehans 1978, 253).
At first glance, this argument seems flawed. What is the point of comparing effects of changes in stock and flow variables? But then one realizes that it is entirely appropriate to compare an unreversed $1 billion
change in expenditures financed by bond sales with a similar expenditure
change financed by money issues—the bond or money issues going on
period after period with total tax receipts unchanged. And, clearly, the
second of these sequences is equivalent to the first plus a continuing
sequence of open-market bond purchases. Thus the bond-financed
expenditure increase and the sequence of open-market purchases are the
two constituent parts of a money-financed expenditure increase. If the
second constituent is much larger than the first, as Niehans and McNelis
claim, then their comparison would be both sensible and justified.
Continuing to reflect reminds one, however, that the simulation experiments actually carried out in the large-scale models are ones that hold
constant tax rates, not tax receipts. Thus, putative bond-financed expenditures and bond-financed money stock expansions are in fact tax
financed to a considerable extent. This would tend, since the models are
not Ricardian, to depress multiplier values. It is not obvious to me that
this tends to bias the results in favor of either type of policy action, but it
would be preferable to compare the effects of the following two experiments:
.

.

....

(i)

an unreversed expansion of government expenditures financed by
bond sales, with unchanged tax-receipt and money-stock paths.

(ii) an unreversed expansion of government expenditures financed by
(high-powered) money issue, with unchanged tax-receipt and
bond-stock paths.

20

Bennett T McCallum

I would not be surprised if the outcome of such an experiment were to
support the Niehans—McNelis position—which in turn supports that of St.
Louis to a considerable extent—but as far as I can tell it has not yet been
conducted.

2.5 Sims~VAR evidence
Before concluding this section, we should mention the argument recently
advanced by Sims (1980, 1982) to the effect that the impact of monetary
policy actions on GNP is extremely small. This argument stems from vector autoregression (VAR) results, obtained by Sims, that show that
money-stock innovations18 have very little explanatory power for U.S.
postwar output when an interest rate is included among the VAR system’s
variables. These results have been interpreted as indicating that monetary
policy actions have been an unimportant source of movements in real
GNP—which would be, given the apparent slowness of price level
responses, inconsistent with St. Louis-type results for nominal GNP as
well, In a brief analytical note, however, I have shown that this conclusion is not implied by the empirical results in question (McCallum 1983).
The point is that money-stock innovations do not necessarily reflect irregular components of monetary policy. Indeed, when the Fed uses an
interest rate as its operating instrument—as it has during most of the
postwar period—it is likely that its irregular actions will be better
represented by a VAR system’s interest rate innovations than by its
money-stock innovations.19 And in fact interest-rate innovations do contribute importantly to output movements in Sims’s results. Thus, it cannot
be concluded that the actions of the monetary authority are unimportant
for the explanation of output and nominal GNP movements.20

3. Theoretical Issues
Let us begin the theoretical discussion by reviewing the effects of monetary and fiscal actions on aggregate demand in a Ricardian economy—that
is one in which agents take account of the government budget restraint
(GBR) in making savings—consumption decisions for an effectively infinite
planning horizon.2’ Under such conditions the asset value of government
bonds held by the public is offset by the present value of extra future
taxes necessitated by the existence of these bonds, so the latter do not
constitute wealth to the private sector as a whole. Thus a one-period tax
reduction financed by bond sales has no effect on aggregate demand, for
the implied increase in future taxes just offsets the effect on wealth of the

Monetary versus Fiscal Policy Effects

21

reduction of current taxes. This Ricardian equivalence result, well known
from the work of Barro (1974, 1984), provides the basis for a comparative
analysis of monetary and fiscal policy effects.
The case of a tax reduction financed by money creation is quite
different, for the asset value of the additional money is not offset by extra
implied future taxes, of which there are none.22 The added nominal
wealth may be negated in real terms by inflation, but the latter comes
about as the result of an upward shift in aggregate demand, which is precisely the effect being claimed. Thus, the fact that inflation may, for some
purposes, be viewed as imposing a “tax” does not alter the validity of the
statement beginning this paragraph; a money-financed tax cut increases
nominal wealth and aggregate demand.
Consequently, since a money-financed tax cut (of~say, $X) and a bondfinanced tax increase (of $X) are together equivalent to an open-market
purchase, we see that an open-market purchase unambiguously increases
aggregate demand (in a Ricardian economy).23 This result clearly implies
that an increase in government spending must have a larger stimulative
effect on aggregate demand if it is money financed rather than bond
financed, for the difference in the two actions is precisely an open-market
purchase (or sequence of purchases if we are discussing an unreversed
increase in government purchases).24
It remains to be considered whether a bond- (or tax-) financed increase
in government purchases will have a nonzero effect on aggregate demand.
Effects of both temporary and permanent changes in government spending in a Ricardian world have been analyzed, in nonmathematical but
careful fashion, by Barro (1984, 309—312 and 316-320). In the case of a
temporary increase there is an increase in output and in the real rate of
interest. The latter translates, in the absence of inflation, into a rise in the
nominal rate. For portfolio balance with a constant money stock, nominal income must then rise, if the income elasticity of real money demand
is less than unity. To see this, let us write the money demand function as
m—p=a0+a,y—a2r,

(6)

where m, p. and y are in logarithmic terms, r is the interest rate, and all
parameters are positive. This equation may be rearranged to read
y+p=m—a0+(1—a,)y+a2r,

(7)

so if a~< I the increases in both y and r tend to increase y + p, the log
of nominal income. In the case of a permanent increase in government
spending, Barro’s analysis leads to no chnage in r but, again, to an
increase iny, so again (7) indicates a rise in y + p.25 Thus we see that an

22

Bennett T McCallum

increase in (real) government spending tends to induce an increase in
aggregate demand.
In summary, a Ricardian analysis suggests that (i) open-market purchases are expansionary, so (ii) money-financed spending increases or tax
cuts are more expansionary than bond-financed ones. Indeed, (iii) bondfinanced tax cuts have no effect on aggregate demand, but (iv) bondfinanced spending increases are expansionary.
Now it is, of course, not the case that Ricardian assumptions are
literally satisfied by the U.S. economy. But the Ricardian model may
nevertheless provide a useful first approximation to the workings of the
economy, a fruitful starting point for thinking about the effects of policy
actions. In particular, it would seem more appropriate to regard bonds as
contributing to private wealth not at all, than to regard them as doing so
fully. If each dollar of bonds functions macroeconomically as k dollars of
net private wealth—which is the way Patinkin (1965, 289) puts it—then the
value of k is (I would conjecture) much closer to 0.0 than to 1.0.26
The relevance of this observation is, of course, that most of the theoretical literature of monetary versus fiscal policy effects has presumed—
usually without discussion—that bonds constitute wealth fully (i.e., that
k = 1.0). Indeed, this literature abounds with paradoxical results that
obtain in large part because of wealth effects due to ongoing expansion or
contraction of government debt,27 Another source of paradox in this
literature—the exposition of which emphasizes the role of the GBR28—has
been the practice of focusing on “long-run” effects, with the latter somewhat misleadingly defined so as to require a balanced budget. The inappropriateness of this terminology is emphasized by occasional statements
concerning the comparative long-run effects of money- and bond-financed
deficit spending.29
In any event there is one contention with which I want to take issue
that appears in this literature: the notion that open-market purchases are
contractionary or that, as Blinder and Solow (1976, 500) put it, “the longrun effect of government spending is greater when deficits are bondfinanced than when they are money-financed.”30 The basis of this contention is, as explained in McCallum (1981, 136), that in models of the type
now under discussion either the contention is true or the economy is
dynamically unstable when deficits are bond financed (i.e., when paths of
spending, the money stock, and income tax rates are determined exogenously so that the stock of bonds must adjust to satisfy the GBR). Then
by assuming stability, the authors in question obtain the result. But, as is
argued in detail in McCallum (1981, 136—137), the appropriate conclusion
is rather that the economy is unstable3’ under these conditions, and
open-market purchases are expansionary even according to the models
and concepts in question.32 That conclusion is in no way refuted by the

Monetary versus Fiscal Policy Effects

23

observed stability of the U.S. economy, because neither the money stock
nor taxes are in fact managed in the way assumed by the setup in which
the analysis is conducted.

4. Conclusions
The substantive conclusions of the foregoing investigation/review are
fairly easy to discern and are not very dramatic. The clearest is that an
open-market increase in the money stock has a stimulative effect on
aggregate demand—a conclusion that, in turn, implies that a moneyfinanced increase in expenditures (or reduction in taxes) is more stimulative than a bond-financed increase. This conclusion is supported by
empirical results obtained both from St. Louis-style estimates and from
large-scale econometric models. Furthermore, the conclusion is also supported by theoretical analysis involving both Ricardian and non-Ricardian
assumptions.
In the case of pure fiscal policy actions—that is, bond-financed tax cuts
or bond-financed expenditure increases—the situation is not as clear. But
theory suggests that the latter should be at least as stimulative as the
former and most probably stimulative to a positive extent. The evidence
on these points is mixed but is not obviously inconsistent with the theoretical predictions.
With respect to the textbook issue concerning the relative (per dollar)
effect of pure monetary and fiscal actions, the evidence seems—in a rather
disorderly way—to support the notion that a sequence of $A open-market
purchases, on each period, will be considerably more stimulative than a
single but unreversed $A/period, bond-financed increase in expenditures.
It might be added, however, that it is unclear that any great importance
attaches to this last issue, at least from a policy perspective, provided that
each type of policy has nonnegligible effects. If the object of the debate is
to determine whether monetary or fiscal variables would serve better as
instruments to be manipulated for stabilization purposes, then attention
should be focused on the relative accuracy of the effects rather than on
the per dollar magnitudes. Adjustments in the instrument settings do not
themselves involve costs, in terms of destroyed resources, as is implicitly
suggested by the view that it is relative magnitudes that matter.33 Nor is
it entirely clear why, from a policy standpoint,34 so much attention has
been devoted to whether the sum of fiscal policy coefficients is zero when
individual coefficients are significantly nonzero. Only a minor quibble35
keeps me from sharing Fischer’s (1976) view that “if fiscal policy had
significant short-term effects
but no long-run effects it would be an
ideal stabilization tool.”
.

.

.

24

Bennett T McCallum

More generally, the essential issues concerning stabilization policy that
continue to divide macroeconomists are those concerned with, first, the
desirability of activist policy and, second, whether activist policy (if desirable) should be executed according to well-specified rules or in a periodby-period discretionary manner.36’37 As various writers have noted,38 neither of these issues is strongly dependent upon the outcome of the relative
magnitudes question. That does not imply, however, that the monetary
versus fiscal policy debate has been unenlightening. On the contrary, the
various ins and outs of the discussion have served valuably to enhance
knowledge and awareness of the central importance of monetary actions
as determinants of nominal income. That contribution is easy to belittle
or overlook, given today’s wide acceptance of that importance. But the
formation of today’s views—that is, the dramatic change away from the
situation described by Cagan—amounted to a major overhaul in the practice of macro and monetary economics, and this change was aided substantially by the monetary versus fiscal policy debate.

Notes
1. Another interesting sign of the items is provided by Michael Parkin’s first
notable publication (Lipsey and Parkin 1970), which centers on a “prototype
model” of the wage and price inflation process. The model’s two endogenous
variables are the money wage and the price level; its exogenous variables are
import prices, output/manhour, the unemployment rate, and a measure of trade
union aggressiveness. Unless I am mistaken, there is no mention of “money” in
the entire paper. This example will be appreciated most by those of us who know
and admire Parkin’s later work as a monetary economist.
2. The original paper was Andersen and Jordan (1968), soon followed by
Andersen and Carlson (1970). Other notable items in the series include Carlson
(1978), Hafer (1982), and Batten and Thornton (1983).
3. Friedman and Meiselman used M2 as their monetary variable and defined
autonomous expenditures as “net private domestic investment plus the government deficit on income and product account plus the net foreign balance” (1963)
184). They also calculated partial correlations and devoted some attention to
quarterly data.
4. An interesting predecessor of the Andersen—Jordan study, which probably
influenced the latter, is Brunner and Balbach (1959).
5. Carlson (1978) indicates that the relative change specification passes, and
the absolute change specification fails, tests for the absence of disturbance
heteroskedasticity.
6. Or variables. My discussion proceeds as if only one such variable were
omitted solely for convenience of exposition.
7. The impact of variables that follow smooth exponential trends is, of course,
picked up by the constant term. This was noted (for linear trends) by Andersen
and Jordan (1968).
8. That the disturbance in (5) is a complicated function of current and lagged
values of the disturbance in (1) does not necessarily imply that the former is serially correlated, for the properties of the latter are unknown.

Monetary versus Fiscal Policy Effects

25

9. This point was made by Andersen and Jordan (1968), Darby (1976), and
probably others, It will arise again in the discussion of the Modigliani—Ando
(1976) results.
10. It is a Keynesian multiplier model.
11. Meyer and Rasche (1980, 59) conclude their discussion of this measurement issue as follows: “However, the modifications
have not generally
resulted in dramatic changes in sample reduced-form equations.”
12. Potential empirical importance is suggested by Modigliani and Ando
(1976, 40—41).
13. Whether the magnitudes of these sums are important will be discussed in
Section 4.
14. Similar conclusions obtain when once- and twice-lagged values of the
unemployment rate for adult males are also used in the first-stage regressions,
although in this case the results are closer to those obtained by OLS.
15. For a recent review of the theory of automatic stabilizers, see Christiano
(1984).
16. These include Meyer and Rasche (1980, 62), Purvis (1980, 108—9), and
McCallum (1978, 322).
17. Variants of this argument would appear to apply to the other Modigliani—
Ando results.
18. A variable’s innovation is its one-period ahead prediction error when the
prediction is the orthogonal projection of the variable on all past values of the
variables included in the system under consideration.
19. I do not mean to claim that interest-rate innovations actually reflect only
policy surprises. They do in the formal model in McCallum (1983) but would not
if the Fed’s operating procedure were slightly different than that assumed. The
main point of the demonstration is that it is unreasonable to use money-stock
innovations as representative of monetary policy surprises.
20. This argument does not imply that interest rates are, in general, better
“indicators” than money-stock growth rates of monetary policy impulses; the relations mentioned in the text hold only for innovations.
21. Also needed, of course, is the assumption that taxes are lump sum in
nature—that is, have no major substitution effects. That finite-lived individuals
may have effectively infinite planning horizons was shown by Barro (1974).
22. See Patinkin (1965, 289). The reason money is willingly held despite the
.

.

.

absence of the interest payments is, of course, that it provides transaction-

facilitating services to its holders.
23. An open-market purchase leaves fewer bonds outstanding and so requires
smaller interest payments in the future by the government. Under Ricardian
assumptions it does not matter whether this reduction in payments is accompanied
by lower taxes or by bond growth.
24. That maintained deficits are possible under bond finance, as well as under
money finance, is demonstrated in McCallum (1984a).
25. This type of result can be shown to hold in the Sidrauski-type version of
the Ricardian model used by McCallum (1984a) as follows. As in Barro (1984),
let the consumer obtain utility from government-provided services and express
this by writing the within-period utility function for the representative household
as u(c1 + age, m,) with 0 < a < 1. The budget constraint is not changed, so
the household’s optimality conditions remain (3)—(lO) in McCallum (l984a, 128—
129). Consider alternative steady states with zero inflation. Combining (4) and
(5) then yields

26

Bennett T McCallum
flu (c + ag,m)
2

=

(1

—

/1)u (c + ag,m).
1

In this particular model an increase in g has no effect on the steady-state value of
k (see p. 129), SO c + g is unaffected, and an increase in g lowers z c + ag.
From the equation above we have
dm
dz

+ /3’($—l)u
21
11
u (1—f3)fl’—u
12
22
From the latter plus the conditions u~< 0, u~< 0, and u > 0, we find that
12
c/rn /dz > 0. So real money balances fall with a reduction in z coming from an
*

U

—

increase in g. But with a constant money stock, that implies an increase in the
price level and thus in aggregate demand.
26. This condition does not require that individuals’ planning horizons extend
beyond their own lifetimes. I would also conjecture that analysis using the
recently developed approach of Blanchard (1984) would support my main conclusions.
27. Some of the prominent items are included in volumes edited by Gordon
(1974) and Stein (1976). Also influential were Christ (1968), Brunner and Meltzer
(1972), and a series of papers by Blinder and Solow (1973, 1974, 1976). A recent
review, which shares the criticized presumption of the items reviewed, is Mayer
(1984).
28. It is worth noting that analysis that ignores the GBR is not thereby
discredited, as long as it does not pretend that time paths of money, bonds, spending, and taxes can all be specified arbitrarily. An analysis that specifies paths for
only three of these variables arbitrarily and ignores the fourth, may be perfectly
logical as long as it does not require an infeasible path for the fourth variable—for
example, a path along which the bond stock grows exponentially at a rate exceeding the growth rate of output by more than the rate of time preference (McCallum l984a). Failing to keep track of the fourth variable may lead to errors if the
model is non-Ricardian, because of induced shifts in behavioral relations. But if
it is Ricardian and the fourth variable is bonds or tax receipts, then such shifts
will not occur.
29. See, for example, Blinder and Solow (1976, 506).
30. Their contention is accepted by Mayer (1984).
31. In McCallum (1981) it is suggested that the dynamic instability in question
can be avoided if the rate of output growth exceeds the after-tax real rate of
return. But with an income tax, a maximizing analysis in a Ricardian model of
the type used in McCallum (1984a) indicates that the steady-state, after-tax real
rate of return will equal the rate of growth plus the rate of time preference. So
the condition mentioned in my earlier paper as an escape from instability cannot
hold in the vicinity of the steady state. I was thus wrong to quarrel, in my 1981
paper, with the first of the two “messages” suggested by Blinder and Solow (1976).
32. This conclusion seems to agree with that of Christ (1979, 533).
33. A similar point was mentioned by Meltzer (1969, 31) but only with respect
to monetary instruments. Robert King has suggested to me that there may be
resource costs associated with adjustments of tax schedules or government expenditures, that would tend to make money the better instrument.
34. The question of whether the cumulative effect of such actions is stimulative, contractionary, or neither may be of theoretical interest, since it bears on the
appropriateness of competing theories. It is not the case, however, that a value of
unity for the sum of the monetary policy coefficients is necessary for “long run”

Monetary versus Fiscal Policy Effects

27

monetary neutrality. For a recent discussion relating to that point, see McCallum

(l984b).
35. The quibble is that zero effects after the first period, rather than zero
long-run effects, would appear to be preferable.
36. The efficacy of various possible instruments and institutional arrangements

is also of importance.
37. The advantage of rule-like behavior of monetary policy has been articulated by Barro and Gordon (1983).

38. Among these are Sargent (1976), Modigliani (1977), and McCallum (1978).

References
Ackley, Gardner. 1961. Macroeconomic Theory. New York: Macmillan.
Andersen, Leonall C., and Jerry L. Jordan. 1968. Monetary and fiscal actions: A
test of their relative importance in economic stabilization. Federal Reserve Bank
of St. Louis Review 50: 11—24.
Andersen, Leonall C., and Keith M. Carlson. 1970. A monetarist model for
economic stabilization. Federal Reserve Bank of St. Louis Review 52: 7—25.
Ando, Albert, and Franco Modigliani. 1965. The relative stability of monetary
velocity and the investment multiplier. American Economic Review 50: 693—728.
Barro, Robert J. 1974. Are government bonds net wealth? Journal of Political
Economy 82: 1095—1117.
1984. Macroeconomics. New York: Wiley.
Barro, Robert J., and David B. Gordon. 1983. A positive theory of monetary poiicy in a natural rate model. Journal of Political Economy 91: 589—6 10.

_____

Batten, Dallas S., and Daniel L. Thornton. 1983. Polynomial distributed lags and
the St. Louis equation. Federal Reserve Bank of St. Louis Review 65: 13—25.
Blanchard, Olivier J. 1984. Debt, deficits, and finite horizons. NBER Working
Paper No. 1389.
Blinder, Alan S., and Robert M. Solow. 1973. Does fiscal policy still matter? Journal of Public Economics 2: 3 19—37.
1974. Analytical foundations of fiscal policy. In The Economics of Public
Finance. Washington: Brookings Institution.
1976. Does fiscal policy still matter? A reply. Journal of Monetary Economics 2: 501—10.
Brunner, Karl, and Anatol B. Balbach. 1959. An evaluation of two types of monetary theories. In Proceedings of the 34th Annual Conference of the Western
Economic Association.
Brunner, Karl, and Allan H. Meltzer. 1972. Money, debt, and economic activity.
Journal of Political Economy 80: 95 1—77.
Cagan, Phillip. 1978. Monetarism in historical perspective. In The Structure of
Monetarism, ed. T. Mayer. New York: Norton.

Carlson, Keith M. 1978. Does the St. Louis equation now believe in fiscal policy?
Federal Bank of St. Louis Review 60: 13—19
Christ, Carl F. 1968. A simple macroeconomic model with a government budget
restraint. Journal of Political Economy 76: 53—67.

______ 1979. On fiscal and monetary policies and the government budget restraint.
American Economic Review 69: 526-38.

Christiano, Lawrence J. 1984. A reexamination of the theory of automatic stabilizers. In Carnegie—Rochester Conference Series on Public Policy, Vol. 20, 147—206.

28

Bennett T McCallum

Corrigan, E. Gerald. 1970. The Measurement and importance of fiscal policy
changes. Federal Reserve Bank of New York Monthly Review 52: 133—45.
Darby, Michael. 1976. Comments. In Monetarism, ed. J. L. Stein. Amsterdam:
North-Holland.
De Prano, Michael, and Thomas Mayer. 1965. Tests of the relative importance of
autonomous expenditures and money. America Economic Review 55: 729—52.
Fischer, Stanley. 1976. Comments. In Monetarism, ed. J. L. Stein. Amsterdam:
North-Holland.
Friedman, Benjamin M. 1977. Even the St. Louis model now believes in fiscal policy. Journal of Money, Credit, and Banking 9: 365—67.

______ 1984. Money, credit, and interest rates in the business cycle.

NBER Work-

ing Paper No. 1482.

Friedman, Milton, and David Meiselman. 1963. The relative stability of monetary
velocity and the investment multiplier in the United States, 1897—1958. In Stabilization Policies. Englewood Cliffs: Prentice-Hall.
Goldfeld, Stephen M., and Alan S. Blinder. 1972. Some implications of
endogenous stabilization policy. Brookings Papers on Economic Activity No. 3:
585—640.

Gordon, Robert A., and Lawrence Klein, eds. 1965. Readings in Business Cycles.
Homewood, Ii.: Richard D. Irwin.
Gordon, Robert J., ed. 1974. Milton Friedman~rMonetary Framework. Chicago:
University of Chicago Press.
Gordon, Robert J., and Stephen King. 1982. The output costs of disinflation in
traditional and vector autoregression models. Brookings Papers on Economic
Activity No. 1: 205—42.
Hafer, R. W. 1982. The role of fiscal policy in the St. Louis equation. Federal
Reserve Bank of St. Louis Review 59: 17—22.
Hester, Donald D. 1964. Keynes and the quantity thebry: A comment on the
Friedman—Meiselman CMC paper. Review of Economics and Statistics 46: 364—
77.
Lipsey, Richard G., and J. M. Parkin. 1970. Incomes policy: A reappraisal.
Economica 37: 115—38.
Litterman, Robert. 1982. Specifying vector autoregressions for macroeconomic
forecasting. Working Paper No. 208, Federal Reserve Bank of Minneapolis.
Lucas, Robert E., Jr. 1976. Econometric policy evaluation: A critique. In
Carnegie—Rochester Conference Series in Public Policy, Vol. 1, K. Brunner and
A. H. Méltzer, eds. Amsterdam: North-Holland.
Mayer, Thomas.

1984.

The government budget constraint and standard

macrotheory. Journal of Monetary Economics 13: 371—79.
McCallum, Bennett T. 1981. Monetarist principles and the money stock growth
rule. American Economic Review 71: 134—38.

______ 1983.

A reconsideration of Sims’s evidence concerning monetarism.
Economics Letters 13: 167—71.
l984a. Are bond-financed deficits inflationary? A Ricardian Analysis. Journal of Political Economy 92: 123—35.
______ l984b. On low-frequency estimates of long-run relationships in
macroeconomics, Journal of Monetary Economics 14: 3—14.
______
1978. Book review. Journal of Monetary Economics 4: 32 1—24.
McNelis, Paul D. 1980. Irrepressible monetarists conclusions from a nonmonetarist model. Journal of Monetary Economics 6: 12 1—27.
Meltzer, Allan H. 1969. Money, intermediation, and growth. Journal of Economic
Literature 7: 27—57.

______

Monetary versus Fiscal Policy Effects

29

Meyer, Lawrence H., and Robert H. Rasche. 1980. Empirical evidence on the
effects of stabilization policy. In Stabilization Policies: Lessonsfrom the ‘70s and
Implicationsfor the ‘80s. Center for the Study of American Business.

Modigliani, Franco. 1977. Monetarist controversy or, should we forsake stabilization policies? American Economic Review 67: 1—19.
Modigliani, Franco, and Albert Ando. 1976. Impacts of fiscal actions on aggregate income and the monetarist controversy: Theory and evidence. In Monetarism, ed. J. L. Stein. Amsterdam: North-Holland.
Neihans, Jurg. 1978. The Theory of Money. Balitmore: Johns Hopkins Press.
Patinkin, Don. 1965. Money, Interest, and Prices. 2d ed. New York: Harper and

Row.
Purvis, Douglas D. 1980. Monetarism: A review. Canadian Journal of Economics
13: 96—122.

Sargent, Thomas J. 1976. A classical macroeconometric model for the United
States. Journal of Political Economy 84: 207—38.

Sims, Christopher A. 1980. Comparison of interwar and postwar business cycles:
Monetarism reconsidererd. American Economic Review 70: 250—57.
1982. Policy analysis with econometric models. Brookings Papers on
Economic Activity No. 1: 107—52.
Stein, Jerome L., ed. Monetarism. Amsterdam: North-Holland.
______

Webb, Roy H. 1984. Vector autoregressions as a tool for forecast evaluation.
Federal Reserve Bank of Richmond Economic Review 70: 3—11.

3
Fiscal Policy in Macro Theory:
A Survey and Evaluation
Karl Brunner
UNIVERSITY OF ROCHESTER

1. Background
Almost twenty years ago the “fiscalist issue” emerged as a major focus on
macroeconomic debates. Milton Friedman and David Meiselman initiated the discussion with an article eventually published in the volumes of
the “Monetary Commission.” The debate was subsequently joined in the
middle 1960s by Albert Ando, Michael De Prano, Donald Hester, Thomas
Mayer, and Franco Modigliani. Jerry Jordan and Leonall Andersen, with
other members of the research staff at the Federal Reserve Bank of St.
Louis and their critics, continued the discussion toward the end of the
1960s into the early years of the l970s.
Another round of discussions followed in the first half of the l970s.
The focus had somewhat changed, however. The Keynesian side
acknowledged real effects of monetary influences, and the monetarists
participating in the discussion recognized temporary real effects and permanent nominal effects of fiscal policy. Starting from this position, the
contribution by Carl Christ (1968), Blinder and Solow (1974), Brunner
(1976), and Brunner and Meltzer (1972a, l972b, 1976) addressed mainly
the feedback via the asset markets resulting from prevailing budgetary
policies. This work reemphasized the idea of a “crowding out” of private
capital formation associated with the financing of a budget deficit.
The appearance of “rational expectations” with the seminal work of
Robert Barro, Robert Lucas, and Thomas Sargent modified our approach
to monetary processes. It affected also the analysis of fiscal policy. The
basic thrust provided by the prevailing formation of “rational expectations” encouraged the revival of an idea originally pondered by Ricardo.
Rational expectations of agents expressing concern for future generations

34

Karl Brunner

destroys the significance of financial decisions in the budget process. The
financing of current expenditures with tax revenues or the sale of bonds
yields the same results under the circumstances. Deficit finance determines
future tax liabilities with a present value just matching the tax revenues
currently suspended. Wealth, position, and opportunities of agents remain
unchanged. Deficit finance affects, therefore, neither interest rates nor
aggregate demand for output. This aspect of fiscal policy thus offers no
wedge for influencing the aggregate evolution of the economy. This result
contrasts both with Keynesian and inherited monetarist analysis. The
ensuing discussions, however, uncovered processes linking tax policies
(even lump sum policies) with real effects operating independently of
direct portfolio and asset market effects. These processes are centered on
intergenerational wealth transfers associated with debt and tax policies.
The macroeconomic role of tax policies depends ultimately, so it appears,
on the assumption of “intergenerational self-interest” or on specific risks
and uncertainties.
An important role of tax policies apparently survives the emergence of
rational expectations and so does a role for expenditure policies. Total
expenditures and their structure still affect, in the context of this neoclassic analysis, consumption, investment, aggregate real demand for output,
the supply of output, and real rates of interest. But the detailed nature of
the mechanism differs radically from the Keynesian story. The government sector operates essentially as a production process absorbing products from the private sector as an input to produce an output. This output either competes with private consumption or contributes to the private
sector’s production process. This approach was originally suggested by
Martin Bailey (1971) but disregarded by aggregate analysis. Both Keynesian and monetarist arguments treated “government” as a sinkhole swallowing a portion of private sector output. The emphasis on “government”
as a production process operating with distortionary taxes changes the
macroeconomic focus of fiscal policy in important ways.
The momentum of academic discussions substantially changed in
retrospect the range of issues surrounding fiscal policy. The analytic evolutions and the resulting discussions modified many questions and
emphasized new dimensions. Academic discussions were also influenced
in recent years by political events and discussions in the public arena.
The emergence of a comparatively large and possibly “permanent” deficit
in the Federal budget motivated another round of discussions. There
appeared voices claiming that such deficits produce, in contrast to Keynesian arguments, negative short-run effects on output. Others emphasized
the long-run effects on normal growth. Many concentrated on linking
high interest rates with the prevailing large deficit. The inflationary
significance of the deficit was also considered. Some arguments seem to

Fiscal Policy in Macro Theory

35

recognize a direct link between deficits and inflation. More carefully
developed arguments emphasize the longer-run effect of persistent deficits
on monetary policy. It would appear that an anti-inflationary policy pursued in the context of a permanent deficit cannot persist. Fiscal policy
appears under the circumstances as the longer-run determinant of monetary policy and a crucial characteristic of the ultimately prevailing monetary regime. It follows that no reliable change in monetary regime is
really feasible without an associated change in the long-run fiscal regime.
This argument introduced a new focus and attention to fiscal policy.
The evolution of questions, issues, and analysis over the past nineteen
years since the “war between the radio stations” (FM, AM, DM) in the
American Economic Review (1965) justified in the judgment of the Conference organizers, and also in my judgment, an appraisal of our intellectual
positions. My paper is addressed to this task. It offers essentially a survey
over major strands of the discussion evolving over the past twenty years.
This survey remains somewhat selective even within its confined range of
macroeconomic issues. Neither does it cultivate a “neutral” account. It
involves interpretations and evaluations referring to aspects of arguments
advanced or to dimensions of the analysis that require more attention.
The first section covers the fiscalist—monetarist debate of the late sixties
and early seventies. It evaluates the empirical work bearing on the central
questions addressed at the time. The following section attends to the
range of issues raised by the neoclassical cum rational expectations
approach. Section 4 examines a number of problems recently associated
with transitory and permanent deficits. The last section assesses the
consequences of our intellectual position with respect to fiscal policymaking and the choice of fiscal regime.

2. The Fiscalist Issue
2.1. The fiscalist—monetarist debate
2.1.a. The evolving theme. The debate emerged gradually in the late
1950s and was fully focused at the time of the conference on monetary
theory organized by the National Bureau of Economic Research in Pittsburgh (1963). The intellectual state was conditioned at the time by the
critical response of an increasing number of economists to the prevailing
Keynesian analysis of fiscal and monetary policy. The core of Keynes’s
General Theory presents a real theory of a low-level output trap. It is supplemented by a real theory of business cycles. The central theme

36

Karl Brunner

emphasizes~the operation of two fundamental failures embedded in
economic organizations relying on markets for the social coordination of
activities. These market failures center on the stock market, as a guide for
investment activities, and the labor market. They severely damage the
ability of a market system to function as an instrument of social coordination. The peculiar characteristics of the stock market shape, so the story
goes, a price behavior randomly related to the social function of rational
resource allocation. These characteristics also suspend a reliable feedback
from the saving—consumption nexus to an investment decision. Persistent
mass unemployment suggested, moreover, to Keynes that potentially
beneficial transactions remained suspended. Such potentialities are
expressed by an excess of the labor suppliers’ marginal utility of the wage
product over the marginal disutility of labor. The market process
apparently fails to provide a sufficient range of coordinating mechanisms.
These basic failures embedded in the structure of the economic process
could not be offset by increased wage flexibility or monetary manipulations. Such endeavors would produce, at most, temporary deviations from
the low-level output trap (Meltzer 1981). A more powerful instrument
was required to move the economy out of such doldrums and push it
nearer to full employment. Fiscal policy seemed to offer the instrumental
opportunities needed for Keynes’s purpose. The underlying analysis suggested that fiscal policy could be shaped to influence directly aggregate
expenditures and to affect indirectly, via the multiplier, the level of consumption expenditures and total output and employment.
The message infiltrated the profession with some variations on the
theme. Alvin Hansen’s secular stagnation centered on the basic real
phenomenon subsequently formalized by the “Keynesian cross.” This
formulation reenforced the “fiscalist thesis” emerging from the Keynesian
analysis, a thesis that attributed dominant positive and normative
significance to fiscal policy as an instrument conditioning the level of output and employment. Sir John Hicks’s reinterpretation in terms of the
IS/LM (“islamic”) paradigm complicated the pattern somewhat with its
inclusion of a feedback via asset markets and portfolio adjustments. But
the assumption of accommodating monetary policy or the prevailing view
attributing low-interest elasticity to aggregate demand and high-interest
elasticity to money demand yielded a close approximation to the implications of the “Keynesian cross.” The professional literature of the late forties and fifties reveals this state very clearly. The article on monetary poiicy by Seltzer (1945) in the late 1940s effectively reflects the dominant
intellectual mood. The book Policies to Combat Depressions (1956) based
on a conference organized by the National Bureau of Economic Research
expresses the basic theme. It concentrated fully and only on fiscal policy.
The slowly evolving flood of textbooks conveyed the same message and so

Fiscal Policy in Macro Theory

37

did, to mention another example, Tinbergen’s book On the Theory of
Economic Policy (1952.)
The existing professional state must be clearly perceived in order to
understand the subsequent intellectual developments. Doubts and reservations bearing on the central underlying theme of market failure never
vanished entirely. The victorious sweep of the fiscalist thesis did not
silence some expressions of doubt, Clark Warburton pursued the classic
program of monetary analysis, offering a substantive alternative to the
Keynesian vision. Milton Friedman (1952) reenforced the questioning
with an examination of the comparative role of government expenditures
and monetary movements in three wartime experiences. The classic
research program rejected the market-failure approach introduced by
Keynes as a serious misinterpretation of market economics and of specific
events observed in the l930s. Substantial doubts about the role assigned
to fiscal policy by the Keynesian position was unavoidable under the circumstances. The evolution of monetarist ideas thus continues essentially
a classical program. These ideas reject the dominance of fiscal policy as a
determinant of both short- and long-run, aggregate, nominal or real
demands. At least some strands of the analysis acknowledge an effect of
fiscal policy on short- and long-term nominal demands, and a short-term
effect on real demand for output, but they deny any effect on long-term
aggregate real demand. However, all strands attribute to monetary shocks
substantial short- and long-term nominal effects and a definite short-term
real effect. This basic position was, however, sharpened with a specific
“impulse hypothesis” incorporated into monetarist analysis. This position
reversed the Keynesian thesis and assigned a comparatively dominant role
to monetary impulses within the general pattern described above. This
thesis of a comparative dominance was not advanced as an “ontological
proposition.” It was interpreted to reflect simultaneously the response
characteristics of the economic mechanism and the historical circumstances expressed by the relative magnitude and variability of fiscal
and monetary impulses. The dominant impulse hypothesis maintained by
the monetarist was thus quite sensitively dependent on the choice of
financial regime.
2. 1.b. The empirical work. The professional state characterizing the
earlier postwar period outlined in the previous paragraph needs to be
fully appreciated in order to assess the empirical work initiated in the late
1950s by Milton Friedman and David Meiselman. This article motivated
critical responses by Donald Hester (1964), Ando and Modigliani (1965),
and De Prano and Mayer (1965). Another round of empirical discussion
was unleashed by Leonall Andersen and Jerry Jordan (1968) with an article in the Review published by the Federal Reserve Bank of St. Louis.
Other members (Michael Keran and Keith Carlson) of the St. Louis Fed’s

38

Karl Brunner

staff extended the discussion and so did Keynesian critics from the
Board’s staff in Washington (DeLeeuw and Kalchbrenner 1969) and the
Brookings Institution (Goldfeld and Blinder 1972).
Harry Johnson (1971) asserted in the early 1970s that the empirical discussion of the l960s was fundamentally flawed and methodologically
inadequate. Tobin (1981) recently repeated Johnson’s assessment. The
two authors thus convey an impression that little, if anything at all, can be
learned from the first round of discussion about the role and significance
of fiscal policy. Neither one of the authors, however, provides any references or offers any arguments or clues supporting their contention. The
methodological objections raised at the time by Keynesians were
apparently accepted at face value without further examination. It is
noteworthy, therefore, that Keynesian critics justified their rather categorically formulated conclusions in terms of the same methodological procedure. An array of objections advanced with a categorical import frequently involved, moreover, nothing beyond invocations of possibly alternative, but unassessed, hypotheses. A retrospective appraisal from the
vantage point of our current state of discussion may be useful for our purposes.
Consider first the basic purpose of the empirical work. It was addressed
to a preliminary assessment of a wide and influential class of microeconomic hypotheses. An array of formulations filling textbooks and
conveyed to the academic world with a sense of empirical relevance
implied the dominant significance of fiscal policy and the irrelevance of
monetary policy in shaping the evolution of output and nominal magnitudes. These implications were also mirrored by many policy statements
supplied to the public arena. These broad statements about the nature of
the economic process were the subject of a first round of searching investigation.
The examination, guided by the dominant underlying theme, essentially
addressed a basic class of Keynesian theories. The class was determined
by the specification of the income variable y, the autonomous expenditure
variable A, and the linearity constraint. These specifications yielded the
induced magnitude defined by I = Y
A and the reduced forms
Y = a + f3A +
or I = y + 8A + v. A wide range of specific
theories formulated with any given definition of Y and A yields the
specific reduced forms. Changes in the definition of Y and A modify the
relevant class under consideration. The dependence of the definition of I
on the definitions of Y and A is crucial in this context. This requirement
was occasionally violated by some authors. This basic class was contrasted with a “quantity-theory” of the form Y = a + bM + ji or
I = c + dM + w. The “simple” K-class implies that /3 and 8 are
significantly positive, whereas b and d are zero. The “simple” version of
the quantity-theory asserts the opposite.
—

Fiscal Policy in Macro Theory

39

The investigation extended beyond the basic class of income—
expenditure theories. They included classes defined by the reduced forms
Y = a + f3A + -yM + or I = a + bA + cM + ji. These formulations subsume all standard versions constructed in accordance with the
“islamic” paradigm. The underlying theme to be assessed implies that /3
dominates y (or b dominates c) in significance.
The empirical assessment exploited both versions of the reduced form.
The I-form offered a useful check on the Y-form. The latter produces a
biased estimate of the A -coefficient due to the implicit correlations
between Y and A, which would yield an apparently significant coefficient
estimate for A even in the absence of any relevant systematic connection
between I and A.
A remarkable fact emerged once the investigation was under way. We
seemed to have been conditioned to assume that we understood what
Keynesian theory meant. But such understanding required that the profession agree on the specification of autonomous expenditures. The investigations, however, revealed a remarkable disarray and confusion. The
critical comments advanced by Friedman and Meiselman (1965, 73) in
this respect are still worth noting today:
We and our critics all used the same measures (for the money stock)
without much ado. The contrast with ‘autonomous expenditures’ could
hardly be sharper. Among us, we have produced more measures than there
are critics. We settled on one; AM on a different one, which is the sum of
two separable components; DM, after running ‘basic tests... on 20
different, but not unreasonable, definitions of autonomous expenditures’ settle on two, but also carry two others along for the ride; Hester came up with

four measures that only partly overlap the others. And all of us harbor serious doubts about the measures we settled on. However useful ‘autonomous
expenditures’ may be as a theoretical construct, it is still far from having
any generally accepted empirical counterpart.
Another passage in Friedman and Meiselman’s comments also deserves
our current attention. It cautions against the potential evaluation of
Keynesian theory by choosing larger portions of Y and A:
By our model, we in effect treated the income—expenditure theory as saying:
if you know from other sources what is going to happen to roughly onetenth of Y or N, then the multiplier analysis will tell you (or give you an
estimate of) what will happen to the other nine-tenths. AM converts the
model into one that says: if you know from other sources what is going to
happen to nearly half of Y or over one-third of N, then the multiplier
analysis will tell you what will happen to the other half of Y or two-thirds

of N. DM’s two models treat only slightly less of total income as autonomous. If AM and DM were to continue along this line of ‘improving’ the

40

Karl Brunner
model by having it predict a smaller and smaller percentage of income
more and more accurately they would soon arrive at the point where it is
predicting nothing—perfectly! In the old saw, with such friends, the

income—expenditure theory hardly needs any enemies.
These issues found little attention at the time. They did reveal, however,
that the sweep of Keynesian ideas, in spite of a vast literature and
influential textbooks, had not been translated into a useful empirical
theory. The strong assertions conveyed by the basic core of the income—
expenditure approach, which frequently spilled over into categorical policy statements, were thus shown to have little substantive foundation.
2.1. c. The nature of objections and critique. An evaluation of the
first-round discussion especially needs to consider the major objections
advanced that possibly influenced Johnson’s judgment, which was
repeated by Tobin. Ando and Modigliani (1965), for instance, objected
vehemently against the “single equation—single variable approach.” They
also criticized the use of “simple models” in lieu of “sophisticated
models.” Ando and Modigliani emphasized, in particular, at a later
round of discussions that the regressions applied in the investigations
under consideration were demonstrably inadequate to assess fiscalist propositions on the basis of observations controlled by a world conforming to
a specific, large, econometric model (1976). Closely associated with these
points was the accusation (or condemnation) that the procedure chosen
reflected a “reduced form methodological commitment.” This “simplistic” commitment was juxtaposed to the “sophistication” of a “structural”
model. Authors were also inclined to criticize on grounds of “grievous
misspecification.” Ando and Modigliani, moreover, declared categorically
that the examinations executed by Friedman and Meiselman are “basically irrelevant for the purpose of assessing the empirical uselessness of
the income—expenditure framework.” Disputes arose over the choice of
exogenous variables. The other party’s choices were naturally wrong,
most particularly if one’s own selection yielded approximately the desired
estimate. Lastly, one author (Hester) asserted that “theory or intuition”
was necessary to specify (correctly?) “the components of autonomous
expenditures.” This argument could, of course, be extended to the choice
of exogenous variables.
These objections involve two important issues and reflect the confusions
frequently emerging in the interpretation of the profession’s empirical
work. The first issue addressed by the critique emphasizes the “reduced
form methodology,” the “single-equation—single-variable” fallacy, or the
propensity for simple, in lieu of, (more realistic?) “sophisticated” models.
These objections essentially fail to recognize the rather specific and limited purpose of the investigations. This failure is especially visible in the
quote drawn from the comments made by Ando and Modigliani. The

Fiscal Policy in Macro Theory

41

“battery of tests” undertaken by Friedman and Meiselman and also the
subsequent work contributed by the staff of the Federal Reserve Bank of
St. Louis were not immediately directed to an assessment of the income—
expenditure framework. A direct assessment of such a framework is logically impossible. Such assessments pertain to specific hypotheses or
classes of hypotheses formulated in accordance with general criteria
characterizing the framework. An assessment of the framework emerges
ultimately from the cognitive fate of the hypotheses it generates. It would
thus be clearly understood that none of the critical investigations was
addressed to the framework. Previous passages emphasized that the
assessment was directed at particular classes of hypotheses yielding strong
propositions about the comparative significance of autonomous expenditures and at least potentially about fiscal policy. And once more, the
prevalent intellectual state with its underlying theme fully justified this
limited purpose—that is, limited relative to the general framework and its
possible translation into classes of hypotheses.
The other three objections associated with the first issue involve
different aspects or verbalizations of the same problem. The attribution of
a methodological legislation insisting, as a matter of principle, on a
reduced-form procedure reflects a pervasive misunderstanding about the
logic of the procedure characterized above. The statements under examination pertaining to relative dominance of A do not bear specifically on
any particular structure; that is, they do not characterize or single out in a
detailed fashion a particular hypothesis. They consistently describe a
whole class of hypotheses and actually pertain to properties of the class.
The most efficient procedure, under the circumstances, for systematic
assessment of such statements uses the reduced form as a test statement.
The properties of the reduced form reflect the properties of the class of
hypotheses under examination. Structural estimation of a single member
of the class is inefficient and essentially uninformative for the purpose.
No methodological legislation is thus involved.
The use of a single equation with a single independent variable should
now be clear. It was the appropriate choice for an assessment of the core
class. It did not represent a single equation model or a disposition to favor
simple, as against sophistical, models. The “single equation with single
variable” was the appropriate choice for an evaluation of a class of
hypotheses seriously presented in textbooks and class teachings. The
objection thus either misses the point or really tells us that all the chapters
and classes elaborating the Keynesian cross or widely used versions of the
“islamic” framework should be clearly labeled as irrelevant pastimes
without any use for the justificiation of any policy statements. This would
also involve a major separation from Keynes, Hansen, and an influential
literature controlling undergraduate teaching during the l950s and in our
hinterlands still today.

42

Karl Brunner

A second issue embedded in the discussion bears on a pervasive confusion between logical issues and psychological effects. Our discussions frequently suffer under a disposition to reject an argument or hypothesis simply on the grounds that one is capable of formulating an alternative.
Such ability offers, of course, no information about the cognitive status of
the hypothesis under consideration. Objections adducing “misspecifications” are thus, by themselves, an empty gesture. They may be interpreted, however, as defining implicitly a program of further research. But
the formulation of such a program possesses, by itself, no evidential value
with respect to the initial hypothesis.
Some authors found that Friedman and Meiselman or Andersen and
Jordan had committed some serious misspecifications. Ando and Modigliani developed, in particular, a “more sophisticated model” within the
general “income—expenditure framework.” They concluded from this construction that the evaluations made by Friedman and Meiselman were
useless and irrelevant. We note first that the “laborious battery of tests”
executed by Friedman and Meiselman and others is indeed irrelevant
with respect to a wide class of hypotheses subsumable under the general
income—expenditure paradigm. This critique would be properly
addressed to general conclusions about the role of fiscal policy drawn
from the empirical work actually executed. Friedman and Meiselman and
Andersen and Jordan carefully avoided such sweeping conclusions. Their
evaluations were made relative to a specific class of “Keynesian-type”
models and their significance is conditioned by this context. The evaluations would still be correct but irrelevant and useless if the class of
hypotheses under consideration was demonstrably neglected, disregarded,
and without any influence on the profession’s policy thinking. But the
latter condition hardly describes the professional situation of the earlier
postwar period.
Hester’s argument pertaining to the choice of autonomous magnitudes
deserved special attention in this context: he suggested that theory or
intuition determine this choice. But this choice determines the precise
nature of the empirical theory. There exists no such theory involving a
definite empirical context before this choice is made. The notion of a
“theory” guiding the choice alluded to by Hester refers at best to formal
structure linking variables with generic names. The admissible range of
semantic rules connecting variables via measurement procedures with
observations still includes many diverse possibilities. Whatever a priori
notions of “preempirical theory” or intuition we exploit for the
specification of crucial magnitudes yield, however, no evidence of the
empirical validity of the choice is made. Neither Hester’s notion of a
“theory” nor any intuition can judge the empirically relevant choices.
The suggestion fundamentally misconceives the nature of scientific pro-

Fiscal Policy in Macro Theory

43

cedures and confuses the context of search for an hypothesis with the context of evaluation on the basis of critical observation.
2.1.d. The exogeneity issue. Hester’s issue naturally generalizes to the
specification of exogenous variables. Objections addressed to the selection
or construction of exogenous magnitudes form a standard procedure of
mutual criticism. But once again the execution of a set of regressions with
more “desirable” coefficient patterns based on alternative choices of exogenous variables expresses a rival hypothesis but offers per se no relevant
evidence discriminating between the alternative hypotheses. Reliance on
the correlation coefficient must be carefully examined in these matters.
Such reliance is quite appropriate for the evaluation of the core class
addressed by Friedman and Meiselman. The hypothesis of comparative
dominance implies systematic differences among correlation coefficients.
We note in contrast that the eventual observation of higher correlation
under alternative choices of monetary and fiscal variables possesses no evidential value discriminating between the selections. Such value can only
exist relative to a specified class of hypotheses that implies statements
about correlation patterns. This would require restriction on all the
parameters, including a full description of stochastic properties under the
alternative definitions. The observation of a comparatively high sample
correlation and more “desirable” coefficient patterns under alternative
exogeneity specifications in the absence of an initially formulated
hypothesis therefore contains no discriminating cognitive significance. It
forms at best the initial step in the formulation of an hypothesis still to be
subsequently assessed against new data.
The exogeneity issue was further elaborated by Stephen M. Goldfeld
and Alan S. Blinder (1972). The two authors explored in great and careful detail the implications of endogenous stabilization actions bearing on
coefficient estimations in reduced forms. Their investigation is anchored
by the reduced-form regressions used by Andersen and Jordan and
expressed by Eq. (1). The symbols F and M refer to fiscal and monetary
y=k+aF+/3M+e

(1)

variables, and c is a random term. It is, of course, well known that any
correlation between F and M and c as a result of an endogenous policy
regime produces biased and inconsistent estimators of a and /3. Goldfeld
and Blinder pursue the matter well beyond this general statement and
impose some structure on the nature of the correlations. This structure
reflects a variety of assumptions about the nature of endogenous stabilization policies. The endogenous policy variables can be represented as a
sum of a nonstochastic magnitude, which can be neglected for our purposes, and a stochastic term

44

Karl Brunner

F=-~
a

~

/3

where ~F = ~ + UF, ~M =
+ Urn~the variance of is a2, the variance of
2 2
2 2
UF is y a and the variance of UM is 8 a . The u-terms reflect the forecasting ability of fiscal and monetary authorities exercised in the context of
their stabilizing endeavors. The parameters y and 8 represent the authorities’ forecasting ability. Lower values of and 8 mean higher forecasting
quality. Similar or coordinated forecasting may produce a correlation p
between UF and UM. These assumptions imply the following probability
limit for the ratios a/a and f3/f3 of OLS estimates of the reduced form (1)
to the true parameters
~—l+
7(p6—y)
/3
z~=y2+82—apy6+7282(l—p2)>0.
8(pa—8)

a

‘

The three forecasting parameters associated with endogenous policy reactions clearly determine the outcome. Almost any combination of biases
can be produced by shifting patterns of p. y, and 8. With p = 0, 8 large,
and small, fiscal policy may appear to be insignificant in reduced-form
regressions of Eq. (1). An excellent stabilization record of the fiscal
authorities supplemented with a poor record of monetary authorities
could explain the Andersen—Jordan results even against the background
of a substantial a-coefficient.
The authors’ general result is confirmed by elaborations of the argument. These elaborations include erroneous assumptions about the multipliers made by the authorities and lagged policy responses to serially
correlated c-disturbances. They extend to the case of “policy interaction”
where one authority’s response takes into account the other agency’s
behavior. The argument also covers a more general model with exogenous variables added to Eq. (1). All these cases enlarge the range for
potential bias beyond the initial parameters p,y,8.
Goldfeld and Blinder supplement their analytic investigation of statistical implications associated with endogenous policy responses with a
Monte Carlo study. They use an econometric model originally published
by Moroney and Mason to generate data sets on the variables incorporated. These data sets were applied to structural and reduced-form
estimations. Structural estimations proceeded with and without inclusion
of reaction functions describing the policymakers’ endogenous responses.
It would appear that structural estimates are comparatively little affected
by nonrecognition of endogenous policy responses. Direct estimation of
~‘

Fiscal Policy in Macro Theory

45

the system’s (exclusive of reaction functions) reduced form produces, in
contrast, seriously biased multiplier estimates.
We should hardly quibble with the correctness of the argument
advanced by Goldfeld and Blinder. Endogenous policy responses can
yield seriously biased estimates of reduced forms exemplified by Eq. (1).
The meaning of this result must be carefully assessed, however. The
authors usefully sharpen our awareness for some important qualifications
applicable to the empirical work which emerged in the context of the
“fiscalist debate.” The results hold for classes of hypotheses recognizing
an actual state of comparative exogeneity with respect to policy variables.
Goldfeld and Blinder offer no argument or evidence bearing on the crucial issues of whether, when and how “policy was endogenous.” The
correctness of their analysis does not, unfortunately, establish its
relevance. We still need to judge the occurrence and nature of
endogenous policy responses. It is important to understand, however, that
judgments based purely on standard regression analyses can seriously
mislead us. We also need to examine carefully the institutional arrangements and explore special events in the manner of Friedman and
Schwartz (1963). Institutional information may often tell us what regressions, if any, are advisable. A “mechanical” linking of policy variables
with a GNP gap hardly conforms to institutional situations prevailing over
most periods or in most countries. The data used in regression estimates
of Eq. (1) by various researchers covered a variety of historical episodes
and also different countries. One would conjecture that the existing institutional differences generate widely different endogenous responses. A
consistent pattern of statistical results over periods and countries deserve
more serious considerations under the circumstances, even with full acknowledgment of the argument advanced by Goldfeld and Blinder. It is
very doubtful for instance that past episodes of U.S. fiscal and monetary
history exhibit a pronounced and approximately uniform stabilization policy. The available information about the Fed’s strategy and tactical procedure suggests, moreover, the operation of a major random process
affecting monetary growth over quarterly periods. This situation may, for
all practical purposes, amount to an approximate exogeneity actually
emphasized by Goldfeld and Blinder. The political economy of fiscal policy also suggests substantial doubts about the relevant occurrence of an
endogenous fixed stabilization policy. “Stabilization” may be a useful
rhetoric device addressed to academics and the public arena. But fiscal
policy is dominated by other considerations associated with the incentive
structure characterizing the existing policy institutions. This incentive
structure implies that “fiscal policy” is dominated by redistributional
interests with little substantive attention to “stabilization” per se. We
should expect under the circumstances that regressions expressing fiscal

46

Karl Brunner

reaction functions remain poorly defined and unstable over time. We
may, of course, encounter situations justifying a substantial suspicion that
endogenous policy responses do occur. But the implications with respect
to estimates of Eq. (1) are obscure until we know more about the nature
of endogenous policy patterns. Information about occurrence only establishes that we need be cautious in interpreting the results. But Goldfeld
and Blinder still enriched our discussion. This contribution warms us
essentially to extend the empirical work to as many different periods and
situations involving different institutional arrangements as possible. This
research strategy seems more promising than an approach relying on
explicit statistical reaction functions of the usual kind embedded in a large
structural model.
2.i.e. Some problems of statistical theory. A retrospective evaluation
guided by an interest to learn from the previous discussions for our
present purposes cannot avoid an application of standards—criteria of procedures developed subsequent to the discussion under examination. Two
aspects bearing on statistical procedures and matters of statistical theory
require our attention. Most of the discussants used level data; Ando and
Modigliani and De Prano and Mayer used data only in this form. Friedman and Meiselman also explored first differences, whereas Andersen and
Jordan only used first differences. The proper choice between these alternative procedures poses a subtle but very important and widely neglected
issue even today. The choice depends sensitively on the error structure in
the formulated regression (Plosser and Schwert 1978) or the nonstationarity of the variables (Meese and Singleton 1982). Plosser and Schwert
compared the problems posed by over- and under-differencing of data in
regression analysis. They explore, in particular, the asymmetric effect of
over- and under-differencing on statistical inferences. They show that
over-differencing produces a regression with an error term controlled by
an MA (1) process with a unit coefficient, whereas under-differencing
yields a regression with a nonstationary random term. Over-differencing
still allows, under the circumstances, reliable estimation and inferences.
Under-differencing, in contrast, poses a serious problem. The sample distribution of the estimator does not possess finite moments. No inferences
are possible in this case.
The authors elaborate the general problem with the aid of several
examples bearing quite directly on our issue. The first example explores a
regression log y = a + /3 log m + c, with y representing nominal income
and m the money stock. The regression is estimated in level form, in levels modified with a time trend, with Cochrane—Orcutt adjustment, and
also in first and second differences. The last four procedures yield essentially the same estimate for /3. The first estimate derived from level data
is separated by more than two standard errors from the estimate obtained

Fiscal Policy in Macro Theory

47

on the basis of first differences. Most interesting is a comparison of the
variances computed for the error term in the regressions. Overdifferencing would imply that the variance associated with the first
difference is twice the variance of the residual in the level regression.
Under-differencing implies, in contrast,. that the residual variance in the
level regression exceeds the corresponding variance in the first difference.
This implication is confirmed by the estimates.
The authors contrast this case with the “quantity theory of sunspots”
expressed by the regression log y = a + /3 log s + c, where s measures
the accumulated sum of sunspots. The same five estimation procedures,
previously mentioned were carried out. The results based on level data
convey the impression of a significant relation with a substantial correlation and a unit multiplier. The first difference yields a radically different
result. The “multiplier” /3-coefficient collapses to nonsignificance with a
standard error almost equal to the coefficient estimate. The estimate
derived from the second difference collapses even further and is hardly
distinguishable from zero. The residual variance in the level regression is,
moreover, almost 13 times the residual variance in first differences.
Lastly, the authors examine the data (A and M—in logs, however) used
by Friedman and Meiselman. They compare the regressions log C =
a + /3 log M + c with log C = y + 8 log A + 3. The results are
remarkably different. The five estimation procedures yield the same
results for /3. The differences are not statistically significant when
evaluated in terms of the standard error of /3 computed from first
differences. Once again the residual variance of the level regression substantially exceeds this variance of the first differences. This pattern of
residual variances occurs even more emphatically in the case of autonomous expenditures. The contrast offered by this regression appears,
however, most particularly in the pattern of 6-estimates associated with
the five estimation procedures. The coefficient collapses from 1.08 in level
data to 0.14 in first differences and 0.09 in second differences. The
differences in estimates are statistically highly significant. All estimates
remain, however, significantly different from zero.
This analysis with the examples immediately related to our problem
clearly reveals the danger associated with under-differencing and the
misleading inferences obtainable under these circumstances. The results
developed by Plosser and Schwert suggest a definite strategy for empirical
investigations, at least in our range of problems. In the usual absence of
sufficient information about the error structure we need to estimate both
in terms of level data (unadjusted and adjusted for possible serial correlation in residuals), first differences, and, possibly, even second differences.
The resulting pattern of residual variances and coefficient estimates determines our evaluation. A residual variance of level regressions substan-

48

Karl Brunner

tially higher than the corresponding variance associated with first
differences suggests the relevant application of first differences. The
regression obtained from level data should, moreover, be considered seriously suspect whenever the estimated regression coefficients substantially
collapse for first and second differences. It follows thus that the results
presented by Ando and Modigliani or De Prano and Mayer yield little
information until further reevaluation.
A possibly more basic problem was raised by Meese and Singleton
(1982) and Wasserfallen (1985). The standard assumptions for regression
analysis are satisfied for stationary stochastic variables and nonstochastic
independent variables. The latter case hardly applies to a relevant
analysis of data cast up by social processes. But nonstationary stochastic
data pose a serious problem for estimation and inferences. It would
appear that consistent estimates of a regression require that the diagonal
terms of the covariance matrix of independent exogenous variables converge to infinity with the sample size. Alternatively, it seems sufficient
that the independent (and exogenous) vector variable be controlled by a
finite autoregressive process. These conditions offer, however, no basis for
inferences. More structure must be imposed in order to derive an asymptotic sampling distribution. The weakest condition on the moment matrix
of independent variables seems to have been formulated, according to
Meese and Singleton, by Grenander (1954): “These conditions preclude
exponential growth of any variable.. Borderline non-stationarity (i.e.,
unit root) is allowed if regressors are fixed or strictly exogenous.” Meese
and Singleton emphasize that an independent error with finite variance is
not a sufficient condition for asymptotic inferences. Quite generally, conditions on the regressors required for deriving inferences pose a troublesome issue. In the absence of good grounds supporting the relevant application of asymptotic distribution theory to inferences derived from nonstationary data, we may possibly obtain estimates but no judgment on
evaluation. This argument reenforces the conclusion obtained from
Plosser and Schwert’s investigation: It seems advisable in the case of
nonstationary data to derive the inferences from suitable transformation
into stationary series. The neglect of this problem lowers the relevance of
some empirical work presented in the “fiscalist” debate.
2. i.f Some general conditions. We should emphasize first that
Johnson’s indictment, recently repeated by Tobin after more than ten
years, simply has no foundation. It reflects a somewhat casual misunderstanding of the nature of the argument. Friedman and Meiselman explicitly cautioned the reader that this assessment was quite provisional. It
was also a definitely limited examination, and so was the work undertaken
at the Federal Reserve Bank of St. Louis. The limitation is defined by the
class of hypotheses implicitly addressed by the tests used. There do exist
. .

Fiscal Policy in Macro Theory

49

classes of hypotheses that cannot be subsumed under the assessments carried out. Friedman and Meiselman properly stressed therefore that their
results are “not decisive.” They are certainly not decisive with respect to
the general paradigm and the general idea of fiscal effects on aggregate
demand. But the tests were properly formulated and executed relative to
the class of hypotheses considered. A reservation should be entered, however, with respect to the use of under-differenced and nonstationary variables. We should also note the reservation advanced by Goldfeld and
Blinder. This reservation simply suggests some further examination in
order to take account of potential effects of endogenous policy reactions.
The discussion also brought forth a perennial problem confronting our
empirical work. The choice of exogenous variables forms an important
component in the construction of a hypothesis (or class of hypotheses).
The development of statistical analysis has sharpened somewhat our
understanding and offered approaches to this issue. One lesson we should
emphatically learn in this context emphasizes that we need to address
more careful attention to the admissible interpretation of our work. Our
imaginative invention of alternative specifications or constructions of exogenous magnitudes offers per se no rational grounds for the rejection of
other specifications and choices. We need either (more or less) direct evidence bearing on the exogeneity of the variables concerned or to depend
on the evaluation of the hypothesis as a whole. Correlation statements
may appear as relevant test statements in this context, provided, however,
as in the case of the core class examined by Friedman and Meiselman, the
hypothesis under consideration implies comparative correlation statements. Comparison of correlations in the absence of such definite implications is meaningless and without any evidential value.
The substantive content of the discussion contributed in retrospect to
some clarification. The hard Keynesian position dismissing monetary
conditions was unanimously discarded. The relevance of monetary conditions became generally recognized. Substantial issues remained, however,
in this range. Some Keynesians argued that the money stock or monetary
growth exerts “permanent” (long-lasting) real effects. Monetarists confine,
in contrast, (temporary) real effects to monetary acceleration (or decelerations). More important for our immediate purpose was the general recognition that fiscal policy did probably modify to some extent nominal
aggregate demand for output. Substantial differences concerning orders
of magnitude and persistence of real effects remained.
There occurred also a subtle but interesting shift within the Keynesian
paradigm. This shift modified the meaning of “fiscal dominance.” We
observe well into the l950s an argument assigning a steep slope to the IS
curve and a flat slope to the LM curve. This assignment was justified in
terms of a borrowing-cost interpretation for the interest elasticity of aggre-

50

Karl Brunner

gate demand (Brunner 1971). There clearly emerged a revision of this
position during the l960s. The relative slopes shifted sufficiently to offer
monetary conditions a significant leverage. The inherited sense of “fiscal
dominance” unavoidably disappeared. The appearance of the assignment
problem and the policy mix analysis reveals this change. But the modified
Keynesian analysis produced a new sense of “fiscal dominance” visible in
the “Economic Reports” of the Kennedy—Johnson administration. Both
monetary and fiscal policies were recognized to influence real magnitudes.
Both policies were thus in principle applicable to stabilization purposes.
Monetary policy was, however, judged to concentrate the social cost of
stabilization policies on a small segment of the economy (housing). Fiscal
policy, in contrast, spread these costs more “equitably.” Fiscal policy was
also judged to operate “directly,” in contrast to the “indirect” effects of
monetary policy; and, consequently, with shorter lags. These considerations determined that fiscal policy was proposed as the active component
of a stabilization program with monetary policy assigned an essentially
accommodating role defined in terms of an interest-rate strategy. This
position was sensitively conditioned by the underlying paradigm summarized by the IS/LM approach and its economic interpretation. This paradigm with its confining view about the nature of the “transmission
mechanism” remained, of course, on center stage in the dispute between
“Keynesians” and “monetarists.”
2.2. The asset market effects offiscalpolicy and the stability of the

system
The discussion covered in the previous paragraphs addressed the comparatively immediate output-market effects of fiscal policy. Neither
Keynesians nor monetarists had at this stage integrated the ramifications
of deficit financing via asset-market responses explicitly into their analysis.
Attention focused by the early l970s on the government’s budget constraint. The relation between fiscal policy and asset-market responses
generated by the mode of deficit financing and the resulting interaction
between asset markets and output markets became the subject of further
examination (Christ 1979; Silber 1970; Blinder and Solow 1974, 1976;
Infante and Stein 1976; Brunner and Meltzer 1972b, 1976; McCallum
1981). The participants in this discussion agreed that fiscal policy (including especially the effects of distortionary taxes and the structure of expenditures) affects actual output, normal output, price level, and real rates of
interest. There remained, however, substantial variations in the details of
the analysis and probably the order of magnitudes involved. The question pursued was addressed to the real and nominal consequences implicit
in fiscal decisions beyond the effects attributable to (global) expenditures
and taxes per Se.

FiscalPolicy in Macro Theory

51

My summary of the issue exploits a scheme used in various papers by
Brunner and Meltzer. The scheme involves an interaction between four
lines represented in Figure 3.1. The vertical line describes normal output.
Some strands of analysis recognized the dependence of the line’s position
on taxes and most particularly on longer-run portfolio adjustments
between government securities and real capital. The bbe line represents
the balanced budget equation. It describes thus the locus of price level
and output combinations (p,y) that satisfy a balanced budget. The position of the line depends on real government expenditures, the stock of
outstanding debt held by the public, and a tax parameter. The slope of
the line expresses the nonhomogeneity of the deficit function due to progressive taxation, yielding a “bracket creep.” Under proportional taxation
the bbe line would be vertical. The d-line presents a pseudodemand curve
to be understood in a semireduced sense. The line summarizes the locus
of all pairs (p,y) that satisfy simultaneously, for any given set of other
variables, output- and asset-market equations. Lastly, the s-line describes
the “structural” supply function. The position of this line moves with the
expected price level, the stock of real capital, and technological progress.
The initial position in the diagram shows a state of full stock and normal output equilibrium. The short-run flow equilibrium determined by
the interaction between pseudodemand and supply yields a state on the
normal output line that simultaneously produces a balanced budget. Now
p

Figure 3.1

52

Karl Brunner

consider an increase in real government expenditures (or a lower tax
parameter). This raises the balanced budget line to the position bbe , and
1
the pseudodemand line to d1 (see Figure 3.2). The fiscal stimulus thus
immediately raises output, price level, and interest rates. It also produces
a deficit expressed by the distance of the bbe line from point A, which
describes the new flow equilibrium. Suppose for the moment that the
deficit is financed with a new issue of government bonds. The resulting
increase in the stock of securities and interest rates pushes the bbe line
further upward along the (vertical) normal output line. The net effect of a
bond-financed deficit thus depends crucially on the interaction between
asset markets and output markets. This interaction determines the movement of the pseudodemand curve relative to the balanced budget line
induced by the fiscal action. The result depends, within the context of the
IS/LM framework, on the relative magnitude of the vertical shifts
imposed on the two curves. A comparatively larger upward shift of the IS
curve due to wealth effects induced by the bond issue raises the d-line,
whereas a comparatively larger wealth effect operating on the LM curve
lowers the line. The Brunner—Meltzer asset-market analysis yields a
somewhat more complex pattern modifying the wealth effect with substitution effects between financial and real assets. The absorption of a larger
stock of government securities by the asset markets unleashes offsetting
p

‘V

Figure 3.2

Fiscal Policy in Macro Theory

53

influences via interest rates on financial assets and via the asset price of
real assets. The net effect thus remains quite ambiguous without constraining order conditions. Neither argument thus yields, without specific
order constraints, a definite answer. Even a positive response of the d-line
is not sufficient, however, for the stability of the stock equilibrium. This
stability requires that the upward shift in the d-line caused by deficit
finance exceed the corresponding upward shift of the bbe-line. This conditions assures that the d-line eventually catches up with the bbe-line, and
the flow equilibrium produces a balanced budget.
Some variations in the analysis occur at this point. Some authors
emphasize the transitory nature of movements along the s-curve. The
latter will shift upward in response to adjustments in the expected price
level. The final state of equilibrium will thus tend to the normal output
line. But the conditions for stability appear under either the IS /LM or
the Brunner—Meltzer analysis quite unlikely. Instability associated with
insufficient nonnegative or even negative responses of the d-line appear
more likely in the context of both analytic arguments.
In contrast, consider the case of a deficit financed with new base
money. The d-line is definitely raised, whereas the bbe-line is not further
raised by the increase in the money stock associated with deficit finance.
The stability of the stock equilibrium is thus ensured under the circumstances. The d-line eventually intersects the bbe-line on the vertical.
Several aspects should be noted here. The total effect on the price level
(and on output in case the nature-rate hypothesis is rejected) is a multiple
of the “immediate” effect associated with the flow equilibrium response to
a fiscal impulse. The latter is described by the shift of the state point to
point A, whereas the final state of stock equilibrium is controlled by the
intersection of the budget line and the normal output line. The total
effect of a bond-financed deficit is even large in case the stability conditions is imposed. This follows from the fact that the bbe-line moves during the adjust process beyond the position determined by the initial rise in
real government expenditures. The total effect thus reflects in both cases
the financial repercussions of fiscal policy decisions.
The implications of the more likely unstable bond-financed deficit process deserve some more attention. Carl Christ (1979) emphasized that this
result poses a problem for the imposition of a monetary rule. Such a rule
confines the proportion of the deficit financed with base money below the
critical level, assuring stability of the stock adjustment process whenever
the deficit is sufficiently large. Brunner (1976) and, recently, McCallum
(1981) emphasized the (partial) alleviation of the problem produced by
economic growth. Such growth moves the flow equilibrium with the normal output line to the right and closes the gap between the state point
fixed by the flow equilibrium and a given budget line. For any given nor-

54

Karl Brunner

ma! growth rate there exists an upper limit on the growth of real debt
(and thus of the deficit) beyond which economic growth cannot produce
stability; that is, the movement of the bbe-line to the right exceeds the
growth-determined shift to the flow equilibrium. We should also consider
that the analysis implies a negative effect of bond-financed deficits on the
rate of norma! growth via the longer-run adjustments in the stock of real
capita!.
But deficits beyond the critical level appear substantially more likely
today than ten years ago. A low-level monetary growth would thus
induce with substantial probability the unstable process discussed above.
This process would raise real and nominal interest rates over time and
lower the normal rate of real growth. None of these consequences
induces within the economic system, according to either type of analysis,
any feedbacks eventually terminating the process. However, we need to
broaden our vision at this state and admit the interaction with the political process. The persistent increase in the real debt raises real interest
rates and the relative burden of interest payments expressed by the ratio
of interest payments to national income. This trend eventually induces
rejection of low-level monetary growth in order to moderate the growth in
real debt or even lower this stock with the aid of a higher price level. A
change in fiscal regime offers an alternative avenue. The crucial conclusion from this stability analysis suggests that a stable noninflationary
monetary regime is unlikely to persist in the absence of a fiscal regime
effectively containing the average deficit, Both monetarist and Keynesian
analysis developed at the time implied that the fiscal regime determines
the longer-run opportunities of monetary policy. Alternatively, proposals
for a monetary rule require supplementary proposal for a fiscal rule.

3. The “NeoClassical” Contribution
The last phases of the discussion summarized in the previous section overlapped with a new thrust in fiscal policy analysis. The “rational expectations revolution” also influenced, beyond the approach to monetary
analysis, the analysis of the government’s fiscal policy. A series of articles
by Robert Barro, beginning in 1978, introduced a “neoclassical vision”
into our discussion. The emerging analysis radically changed the
economic significance of deficits. Deficits were interpreted similar to the
deviations between current income and current consumption of private
households in the context of an intertemporal allocation. The public debt
and its behavior thus reflects the public sector’s intertemporal optimization conditioned by the pattern of permanent and transitory government

Fiscal Policy in Macro Theory

55

expenditures. But decisions to issue bonds in order to finance expenditures affect, in contrast to both Keynesian and monetarist analyses, neither real interest reates nor the price level nor, even temporarily, output
and employment. The stability issue discussed above does not exist under
the circumstances. Traditional notions of stabilization thus offer no
relevant motivation for public debt policy. This position bearing on
deficits and debt does not extend to government expenditures on goods
and services and “non-lump-sum” taxation. Budgetary policies expressed
by expenditure and tax decisions do exert temporary and permanent real
effects This analysis most especially developed by Barro (198 la) actu
ally involved aspects of government budgetary operations, which, even if
known in a general sense, were long neglected by macroanalysis. Standard macrotheory typically presented the public sector as a sinkhole for
goods and services produced by the private sector. Martin Bailey moved
beyond this “sinkhole theory” of the government sector and considered
the government’s supply of goods and services to the private sector. The
substitutive or complementary nature of this supply with respect to private
consumption and investment may substantially modify tbe traditional
results. Barro (1981b, l984a) revived Bailey’s neglected initiative and also
elaborated more carefully the impact of non-lump-sum taxation. He also
revived in this context some earlier work by Miller and Upton (1974).
The analysis addressed, beyond these aspects, different mechanisms yielding the real effects of budgetary decisions. It shares with monetary
analysis an emphasis on wealth and substitution effects, in contrast to the
Keynesian reliance on the income multiplier. Its emphasis on perpetual
market clearing relative to all ongoing shocks differentiates it from both
Keynesian and monetarist arguments.

3.1. The “Ricardian Theme”
Barro must be credited for having revived a theme originally considered
by Ricardo. The stability analysis examined above assumed that the
financial decision between bonds and taxes exert a permanent real effect.
Barro’s argument persuasively challenges this position. He emphasizes
that the traditional argument neglects to incorporate the future tax liability associated with current borrowing. A careful separation of issues
requires, in this context, the assumption of lump-sum taxes. This assumption permits us to isolate the possible effects attributable to financial decision as such without contamination with the real effects of distortionary
taxes. The forward-looking behavior emphasized by rational expectations
interprets deficit finance essentially as an intertemporal reallocation of
taxes. This implies that current deficits, expressed by a new sale of bonds,
correspond to a shift of taxes from the present to the future. The present

56

Karl Brunner

value of the forfeited (current) tax thus equals the present value of the
future tax liability. This equality holds under an important assumption
introduced by Barro and discussed subsequently. It holds, in particular,
whether the bonds issues are maturing at a specified recognized date,
according to a contingent state pattern (Chan 1983), according to a probability pattern over time, or, lastly, whether the bonds are perpetuities.
The “Ricardian argument” requires the formal apparatus of intertemporal budget constraints for both a representative household and government. A first simple argument confines the repayment period of debt to
the representative taxpayers economic horizon. This constraint can be
used to demonstrate that the present value of the future (expected)
government expenditures plus the inherited government debt are equal to
the present value of expected future taxes. Similarly, the household’s
budget constraint shows that the sum of present value of expected future
consumption and the present value of expected future taxes is equal to the
present value of expected future income plus inherited assets. Any temporal reallocation of taxes combined with an unchanged stream of
government expenditures produces an equality between current non-taxfinanced government expenditures and the present value of future
changes in tax liabilities. This result follows from the government’s
budget constraint. It implies in conjunction with the budget constraint of
the representative household that the household’s real opportunities are
invariant with respect to the deficit. The household’s wealth position
defined by its present value of consumption remains unaffected and so
does the present value of taxes and income. Households find the optimal
choice of consumption pattern unchanged under the circumstances. The
creation of a deficit thus cannot modify household consumption decisions.
Substitution effects induced by distorting incentives are, moreover,
excluded by the lump-sum character of taxes. The household’s prior
optimal consumption plan extends to the new situation, whatever the
representative household’s utility function may be. The deficit thus
induces neither an intertemporal substitution nor a scale effect on consumption. The current deficit is perfectly matched under the circumstances by an increase in household saving. The government sector’s
dissaving is thus fully offset by additional private saving. The economy’s
total saving is therefore independent of the government’s financial decision. This saving provides the necessary funds to absorb the new bonds
into the household’s portfolio. Optimal portfolio management determines
this link between deficits, household saving, and bond acquisitions as a
consequence of rational hedging. With portfolios optimally adjusted in
terms of risk—return combinations before the new deficit emerges, households will find it advisable to hedge the expected tax liability by acquiring
a corresponding amount of bonds. The optimal portfolio position will

Fiscal Policy in Macro Theory

57

thus be maintained. In terms of the Brunner—Meltzer asset-market model,
this result implies that the public’s net (stock) supply of government securities disappears in the credit-market equation. It follows that variations
in the stock of government securities exert no effects on asset markets.
Asset prices and interest rates will not be affected. It follows, moreover,
with respect to the stability analysis of Section 2 that the d-line remains
unaffected by the government’s debt finance.
The argument establishing the Ricardian thesis is crucially conditioned
by the two budget constraints. These constraints involve several important assumptions that decisively determine the conclusions. Barro
explored these assumptions in careful detail and argued persuasively their
a priori reasonableness as a good approximation to reality. Buiter and
Tobin (1978) and Tobin (1980) examined the underlying assumption with
a matching vigor and argued persuasively their a priori unreasonableness
as an approach to reality. There is no need to reproduce these excellent
discussions here in depth. My comments concentrate on three aspects
associated with the intertemporal budget constraints. These aspects bear
on the horizon of the representative household, the nature of capital
markets, and risks or uncertainties associated with deficit finance.
The simple fact of mortality combined with our knowledge of demographic structure and debt policies destroys the assumption that the
representative household’s horizon is at least as long as the repayment
period of debt associated with a given tax cut. An “infinitely” living
household would circumvent this problem for the Ricardian theme. Barro
provided a subtle and extensive analysis interpreting this assumption in
empirical terms. He introduced the notion of “operative intergenerational
bequests and transfers” to link finitely living generations within an
infinitely operating household. This is achieved analytically with an
ingenious device. The utility of each generation depends on the utility
level of the next generation. The implicit nesting of utility functions
yields the appearance of an infinitely living decision unit. This does not
imply that the consumption of future generations is equally weighted as
the present generation’s own consumption. Its implication denies, however, the dependence of consumption pattern on age characteristics of life
cycle theories. Optimal consumption choices of infinitely living households produce a preferred pattern of intergenerational transfers (in either
direction). Any intrusion by the government to modify this pattern via its
budgetary operations necessarily fails under the circumstances. Impositions of tax burdens on future generations via a bond-financed deficit
induces offsetting transfer of wealth from the present to the future generation financed with the additional current saving produced by the tax cut.
The voluntarily determined optimal pattern of intergenerational transfers
dominates the outcome and overrides or offsets the government’s

58

Karl Brunner

budgetary operations. The extension to an infinite budget constraint still
assures that any cut in current taxes is exactly balanced by the present
value of future tax liabilities. A shifting of the tax incidence beyond a
mortal man’s life span does not change his optimal intertemporal choice.
The initial choice prevailing before the intertemporal tax shift remains
optimal. This result implies that current tax cuts yield no wealth effects
on current consumption or other real variables.

3.2. The Ricardian Theme: Its qualification
3.2. a. Intergenerational altruism and intergenerational selfishness. The
invariance of intergenerational transfers with respect to budgetary operations is crucially conditioned by the assumption of “intergenerational
altruism.” This assumption justifies the infinite horizon. Consider, in
contrast the opposite assumption intergenerational selfishness com
bined with the life cycle hypothesis of consumption (Kotlikoff 1984) The
latter hypothesis implies that the marginal propensity to consume
increases with age. The assumption of “intergenerational selfishness”
reenforces this pattern as older generations plan no bequests under the
circumstances. Any transfer between generations modifies in this case
aggregate real consumption. In particular, a transfer from the younger to
the older generation raises aggregate consumption A shift in taxes from
the old to the young would produce this result. Replacing some current
taxes with bonds maturing dunng the young generation s lifetime after the
older generation’s death would accomplish the necessary shift. The result
emerges with even larger weight whenever the future tax liabilities fall on
unborn generations. “Intergenerational selfishness” thus assures that any
temporal reallocation associated with a deficit induces intergenerational
transfer not “washed out” within an infinite intertemporal budget constraint. The intergenerational transfers are not offset and produce real
consequences. They actually modify the relevant budget constraint for
each generation.
The difference between the two cases may be conveyed in the following
terms. Let the expression

A1~1+C1+T1= Y1+A1
denote the ith generation’s budget constraint, where A + indicates the
1 1
bequests transmitted to generation i + 1 by generation i C1 T1 and Y1
describe the present value (as of generation 0) of consumption, taxes, and
income for generations i over its lifetime. With intergenerational altruism
A reflects the optimal transfer decisions in response to changes in T and
0
1
the matching changes in T1 Thus follows the relevance of the infinite

Fiscal Policy in Macro Theory

59

budget constraint. Intergenerational selfishness breaks up this pattern. Its
strict and narrow application means that A~+ = 0. A change in T thus
1
0
invariably induces a change in C that exceeds, with the differential in
0
marginal propensities, the change in C1 in the initial period.
This analysis suffers, however, in comparison with the “intergenerational altruism” model from some immediate confrontation with reality.
We do observe that wealth is actually transmitted from one generation to
another. These transfers are excluded by the narrowly formulated
“selfishness model.” The occurrence of transfers is, however, not necessarily an expression of voluntary and planned transfers associated with
intergenerational altruism once we move beyond the context of perfect
information. Blanchard (1984) argued that an insurance scheme under
risk yields the same result of no bequests. But such insurance does either
not exist (specifically the one used by Blanchard) or it at most only used
or offered partially by agents. But bequests apparently do occur. The
context of risk could explain, however, the appearance of bequest without
the bequest motive as formalized by Barro. Agents face some probability
that the remaining life span exceeds the expected time. Risk averse
agents will therefore adjust the use of their wealth in view of this uncertainty. Their problem is similar to the asymmetric risks associated occasionally with inventory decision. The asymmetry is probably even more
acute in this case. Using up all resources before death exposes a person
to serious hazards. This is balanced by unused resources at death,
reflecting lowered consumption before death. The comparatively lower
level of consumption can be understood as a premium paid for selfinsurance against the hazards of early exhaustion of one’s resources.
Wealth will be held under the circumstances beyond the requirements of
the statistical expectations bearing on the remaining life span. It follows
that, on average, unused wealth will be transmitted at death to the next
generation without an “operative bequest motive.” Intergenerational
transfers operated by budget deficits would in this case still raise current
consumption. The observation of substantial intergenerational transmission of wealth is thus consistent with the denial of “Ricardian
equivalence.”
It should be noted that this argument disregards the potential role of
annuities to be purchased by the older generation. But even an actuarially fair supply of annuities would probably not completely replace the
holding of tradeable wealth as insurance for old age hazards expressed by
only partially insurable large expenses on health problems. The annuity
business, moreover, operates at a cost. A comparison of this cost with the
cost of self-insurance could be expected to have a margin of tradeable
wealth. But an entirely different argument, developed by Bernheim,
Shleifer, and Summers (1985), explicity recognizes the occurrence of

60

Karl Brunner

voluntary planned bequests and offers probably a more relevant critique
of the Ricardian thesis of debt neutrality. This analysis will be examined
in subsequent sections covering the empirical work.
3.2.b. Corner solutions. The alternative hypotheses yield very
different interpretations of social security. The Ricardian theme implies
that such arrangements are offset by correspondingly larger transfers from
the older to the next generation with no effect on current consumption.
The alternative, articulated by Feldstein (1982), implies that this intergenerational transfer raises current consumption. An interesting explanation
of the emergence of social security (Meltzer and Richard 1985) reenforces
the “non-Ricardian” hypothesis. Social security is interpreted as a substitute for the voluntary social arrangements made within the larger family.
Social evolution gradually eroded such arrangements and raised the older
generation’s control problem associated with the extraction of the support.
Social security emerged as a political solution to this problem. This
explanation cannot be reconciled with a Ricardian equivalence. Its basic
structure asserts, on the contrary, that the intergenerational link described
by Barro has been suspended. This interpretation suggests, moreover, the
occurrence of several important facts bearing on intergenerational
transfers. Such transfers do not occur solely in the form of a tradeable
wealth accumulated by the older generation and eventually inherited by
the younger generation. Voluntary intergenerational transfers may frequently involve an intertemporal exchange between older and younger
generations. The older generation invests initially in health, education,
and other dimensions of the younger generation with the expectation of
support (i.e., negative bequests) during old age. The extraction of support
does not proceed without control and transactions costs, however. The
hypothesis indicated above states basically that specifiable social changes
raised these costs and increased the likelihood of corner solutions when
unrestricted optimality yields negative bequests. Under the circumstances
the government’s intervention increases current real consumption. Lastly,
we note that this argument also suggests that a single representative
household does not adequately represent the actual diversity encountered
in matters of intergenerational transfers.
The representative household’s infinite intertemporal budget constraint
requires, in addition to the special intergenerational link, an assumption
about capital markets. This assumption removes the occurrence of
another solution. The representative household’s lending and borrowing
rates coincide with the rates available to the government’s operation.. A
violation of this conditions produces wealth effects that suspend the Ricardian equivalence. Chan (1983) explored this theme in the context of an
assumption approximating the problem in terms of liquidity constraints
imposed on a subset of households. A debt-financed deficit operates

Fiscal Policy in Macro Theory

61

under the circumstances to substitute government borrowing for constrained household borrowing. Households with no liquidity constraint or
a nonbinding one respond to the deficit and its finance in the “Ricardian
manner” with the offsetting behavior discussed above. Households
suffering binding liquidity constraints before the event experience, through
the intermediation of the government, a relaxation of the constraint.
They will react by shifting future consumption into current consumption.
The distribution of significant liquidity constraints or the distribution of
differential borrowing rates of interest between government and households determines the net effect of a debt-financed tax cut on households’
current consumption.
Some fundamental theoretical exploration about the conditions for
Barro’s “operative intergenerational bequests” qualifies the debt neutrality
result. Barro’s analysis is conditioned by the assumption that intergenerational transfers from old to young occur independently of debt. The
emergence of debt changes neither opportunities nor preferences in the
world described. No real consequences thus ensue. An interior solution
is simply maintained. Philippe Weil (1984) examined the problem with
great analytic care. His basic theme emphasizes the existence of a “deep
connection between the efficiency properties of overlapping generation
economies without bequest motive and the possible direction of intergenerational transfers.” There exist, in particular, overlapping generation
economies that “justify,” in terms of efficiency, intergenerational transfers
from young to old. Weil demonstrates that the existence of “operative
bequests” in a suitably defined steady state depends on the discount
applies by the old to the future utility of the young. This discount should
not exceed a benchmark determined by the gap between the interest rate
and economic growth in a diamond-type model. Weil concludes that a
“wide class of economies with a bequest motive” do not satisfy Barro’s
proposition about Ricardian debt neutrality. The limited significance of
this analysis should be clearly understood. It essentially establishes that
Barro’s result is nonvacuous and places necessary and sufficient conditions
on its occurrence. This purely analytic result cannot settle the crucial
empirical issue at stake.
3.2.c. Risk and uncertainty. Risk and uncertainty are essentially exorcised from the argument supporting the Ricardian theme. A firm link
connects the household’s benefits from the current tax cut with the future
tax burden. The government’s infinite intertemporal budget constraint
also removes all risk and uncertainty with regard to the government’s
budgetary operation. Agents can rest assured that future tax liabilities
guarantee the crucial equality of present values.
The pattern is substantially modified once we admit risk and uncertainty on two levels into the analysis. Barro (l98la) and Feldstein (1982)

62

Karl Brunner

already noted in passing some consequences associated with uncertainty
about the households’ future tax liability. Barro suggested that uncertainty raises the perception of the present value attached to future tax liabilities by risk-averse households. Thus a debt-financed deficit lowers
under uncertain future tax liabilities the representative household’s perception of its wealth position. Current private consumption thus declines
as a result of the deficit. The suggestive remark was developed by Chan
(1983) in some explicit analytic detail. The argument still adheres to a
state of lump-sum taxes with certainty concerning the future aggregate tax
liability. Households suffer, however, incomplete information and consequently experience some risk about the distribution of the global burden
among the taxpayers. The individual household’s future tax burden is
determined by a stochastic process. It follows that the individual’s share
of the current tax cut does not match his share of the future tax liability.
This suspension of the crucial link produced by a stochastic tax incidence
implies, under conventional restrictions on preferences, that a debtfinanced deficit lowers current consumption. The larger the uncertainty
about the incidence of the future tax liabilities, moreover, the larger is the
negative effect of a deficit on current consumption. The result reveals that
households hedge against the risk imposed on them by saving even more
than determined under certainty. The hedging response induces a substitution of saving at the expense of consumption.
A similar theme, but different in its conclusion, was recently developed
by Barsky, Mankiw, and Zeldes (1984). Risk-averse households encounter
in this case not an uncertain tax incidence but an uncertain future income.
An intertemporal reallocation of tax liabilities (with less now and more in
the future) lowers the degree of uncertainty bearing on future income.
The precautionary demand for savings declines and real consumption
increases under the circumstances. This result is, moreover, produced
under strict “Ricardian conditions.”
The analysis proceeds in the context of a two-period model characterized by three budget constraints. The two constraints describing the
household’s position appear as
(2)
C =(l+R)W +ji + ,
2
1
2 2

(3)

where C1 and C2 refer to consumption in periods 1 and 2, js and js + ~2
1
2
designate income in the two periods, W1 defines savings in period 1, and
R is the interest rate. The magnitudes js1 and js2 are nonstochastic,
whereas c is the stochastic component of future income. A temporal tax
shift is introduced by inserting Tpositively in Eq. (2) and (1 + R)T nega-

Fiscal Policy in Macro Theory

63

tively in Eq. (3) The signs are determined by the fact that any initial tax

burden is impounded into
budget constraints as

js

2

and

We thus rewrite the household’s

js .
2

(4)
C2=(l+R)W1—(l+R)T+js2+c.

(5)

The government needs to levy taxes in the second period in order to
finance the repayment of (1 + R)T. The required tax rate is thus determined by the relation
(1 + R)T

(6)

tjs .
2

=

The revenues from extra taxation imposed on expected income should
cover the repayment. The actual tax revenue is
t(ji +
2

2)

=

(1 + R)T + (1 + R)T~

The second period’s household constraint can thus be rewritten once more
as
C =(l
2

+R)W—(l

+R)T+ji +[l
2

—

(1 +R)T1~

(7)

Lastly, the authors postulate a three-times differentiable utility function in
C1 and C2 satisfying the conditions
(1 + R)U

222

—

U

> 0.

122

The first-order optimality conditions are immediately derived:
EU1(C1, C2)

=

(1 +. R)EU2(C1, C2).

A straightforward manipulation of this condition yields the marginal propensity to consume:
—

3T

—

(1 + R)Cov[(1 + R)U
U12, c]
22
—js [EU
3(1 + R)EU12+(l + R)2EU22}
—

2

11

>

0

(8)

—

A reduction in current taxes, appearing as a position magnitude in the
constraint, thus raises current consumption. Tax policies induce, under
the circumstances, specified real effects in the economy. These effects
emerge even with the households’ definite perception that the present
value of its expected future tax liabilities equals the taxes foregone in the

64

Karl Brunner

present. The result is crucially conditioned by the positive covariance
term in Eq. (8). This result vanishes in the absence of the third-order
derivative condition imposed on the utility function. Once we accept this
condition the crucial aspect centers on the reduction in the variance of
second-period income produced by a nonvanishing tax rate
t = (1 + R)T/ji . The variance is actually reduced in the proportion
2
2
(1
t) .
The authors offer two distinct interpretations for the two-period model.
One confines the model to the lifetime of a single individual. Income
uncertainty bears in this case on an individual’s uncertainty within the life
cycle. The other interpretations introduce intergenerational relations.
The second period refers to the economic uncertainty attached to a
household’s descendants. The uncertainty of both cases is incisively
demonstrated by the authors. Extensive simulations proceeding under a
variety of assumptions offer some insights beyond the qualitative result in
Eq. (8) about the order of magnitude of tax effects on real consumption.
The issue addressed by Barsky, Mankiw, and Zeldes was already
analyzed by Chan (1983). The detailed formulation differs slightly, but
the conclusion is the same once the preference structure is properly restricted. Chan emphasizes, however, that the insurance scheme introduced
with the special tax policy arrangement is essentially independent of the
debt—tax mix problem.
The infinite intertemporal budget constraint of government also reflects
a crucial link suspending relevant risks and uncertainty. The equality of
present values of expenditures and taxes expresses agents’ certain
knowledge that current expenditures will eventually be covered by taxation. Suppose, however, that large debt-financed deficits persist for 10,
100, or 1000 years. Is it reasonable to assume an invariant certainty that
after x +1 years all will be unwound with appropriately increasing tax liabilities? The basic thrust of rational expectations would suggest that
agents learn. An experience of accumulating debt-financed deficits would
induce doubts and reservations about the relevance of infinite intertemporal budget constraints. The time-inconsistency problem diligently discussed with examples from tax and monetary policies should actually be
highly significant in this context (Baltensperger 1984). Suppose that
agents were exposed to a long series of deficits financed by issues of
interest-bearing debt. This experience induces some revisions in agents’
expectations. The probability assigned to possible defaults, particularly to
default by inflation, will rise under the circumstances. The increasing risk
associated with large and persistent deficits generates over time an anticipated purchasing-power risk attached to the government debt. This
purchasing-power risk modifies the Fisher equation with the appearance
of a specific covariance expression representing the purchasing-power risk.
—

Fiscal Policy in Macro Theory

65

The emergence of this risk term implies an increase both of the nominal
and the real rates of interest (Baghat and Wakeman 1983). This problem
will be reconsidered in a subsequent section attending to the long-run
interdependence of monetary and fiscal policies, an issue raised by Sargent and Wallace (1982).
3.2.d. The changing nature of the issue. The risk problem introduced
in the previous paragraph bears essentially on long-run aspects associated
with the cumulative effects of a long series of large deficits that erode the
relevance of the infinite intertemporal budget constraint expressed in
terms of ordinary taxes. This seems to be not the only aspect of risk and
uncertainty associated with the macroeconomic consequences of the
government’s budgetary process. Such risk and uncertainty have so far not
been integrated into macroeconomic analysis. Our most recent fiscal
experiences suggest that a potentially useful research program would
examine the impact of the uncertain amount and nature of tax liabilities
on the public’s balance-sheet risks and consequently on asset markets,
with further effects on consumption and investment. This investigation
moves us substantially away from the “Ricardian theme.” This theme,
restrained by the assumption of lump-sum taxes in order to isolate a possibly pure public-debt effect, can only admit uncertainty about the temporal distribution, the personal incidence of tax liabilities and future
income. The first type of uncertainty does not modify the Ricardian
theme, but the second and third violate the neutrality pattern of debtfinanced deficits. Risk and uncertainty surrounding tax liabilities have a
much wider field of operation once we introduce “non-lump-sum” taxes.
The uncertain incidence of future tax liabilities on nonhuman wealth,
human wealth, and consumption probably affects portfolio risks and thus
the required average return. The consequent adjustment in asset values
modifies consumption, saving and investment patterns. The importance
of debt-financed deficits may, in this context, not so much emerge because
of a direct effect of public debt on real interest rates in defiance of any
“Ricardian equivalence.” The deficits and the resulting increase in public
debt yield possibly their most important effects via a different channel.
These phenomena may be a signal of mounting uncertainty about future
magnitude and incidence of tax liabilities. Agents’ perceptions bearing on
the interpretation of observed deficits would operate as a crucial link in
the process. Balance-sheet risks and asset-market responses will vary with
the perceived duration of the deficit.
This theme also offers an avenue for an analysis of fiscal regimes in contrast to the usual analysis of fiscal policy actions and their effects on the
evolution of economic activity. Fiscal regimes could be differentiated
according to their respective risk patterns parameterized in a specific
mode. The procedure would follow the imaginative attempt made by

66

Karl Brunner

Stulz and Wasserfallen (1985) for the case of monetary regimes. The
investigation was motivated by the recognition of the comparative importance of the stochastic trend and the relatively modest significance of the
stationary component in economic fluctuations (Nelson and Plosser 1982).
The array of specific policy actions addressed to the cyclic component
remains, under the circumstances, confined to a smallish range of
influence. Stulz and Wasserfallen show that, in contrast to influences
exerted by specific actions, the influence of the regime~ characteristics
expressed by some risk parameters substantially contribute to global
economic evolution by conditioning the properties of the stochastic trend.
The uncertainty imposed by the regime affects the stochastic properties of
assets, their risks, and the portfolio risk with consequences on returns and
output behavior. These issues seem worthy of further exploration in a
program designed to integrate finance and macroeconomic analysis (Lucas
1984; Plosser 1984; Fischer and Merton 1984).
3.2.e. Some final remarks on the Ricardian theme. It is interesting to
reflect at this stage on the consequences of the discussion initiated with
Barro’s revival of the Ricardian theme. The neoclassic analysis rejects the
“conventional” position that the government’s financial decisions
expressed by the debt—tax mix induce real consequences over both the
short and long runs. The IS /LM framework concentrated on the wealth
effects of debt policies as the crucial condition of real effects. The
Brunner—Meltzer asset-market analysis, on the other hand, emphasized
the comparative small order of wealth effects associated with a pure-debt
policy even when future tax liabilities were disregarded. The wealth effect
measured as the vertical shift elasticity of the pseudodemand line in figure
3.1 occurs as a product of two terms. One term consists of components
with opposite signs reflecting offsetting influences produced by interacting
asset-markets. The other term is the ratio of government debt at market
value to total nonhuman wealth. The real consequences of debt policies
in the Brunner—Meltzer asset-market analysis were dominantly produced
by substitution and relative price effects produced by the shifting composition of assets. The discussion of the stability analysis revealed furthermore that this short-run, pure-debt effect, again evaluated by shifts of the
pseudodemand line, is at best very modest with respect to aggregate output and the price level. This portion of “conventional” analysis assigned
more significance to the long-run effects of a pure-debt policy centered on
the adjustment of the optimal stock of real capital and consequently the
position of the normal output line in Figure 3.1. Both mechanisms
stressed by conventional analysis became suspended in the neoclassic
analysis. Intergenerational altruism and optimal hedging removed both
wealth and substitution effects.
The discussion seems hardly to restore the “conventional” position. Its
most significant elements modifying the Ricardian theme do not suggest

Fiscal Policy in Macro Theory

67

much potent effects of debt—tax mix policies per se. Kotlikofl’s analysis
directs our attention neither to any wealth nor subsitution effects. This
analysis implies, based on the assumption of “intergenerational
selfishness,” that debt policies induce intergenerational wealth transfer
that modify current real assumption. Debt policy is, however, not a
necessary condition of such transfers. Such transfers may occur without
debt policy as a result of nontemporal tax shifts. But debt policies do
induce, on the other hand, the transfers described by Kotlikoff. This
analysis thus emphasizes the effect of debt policies on the composition of
aggregate real demand. Debt policies raise real consumption and lower
real investment in the short run and lower (comparatively) the capital
stock and normal output in the long run. But the mechanism involved
differs from those described by conventional analysis. We note, however,
that Kotlikofi’s analysis is consistent with the operation of a wealth and
substitution effect (via asset markets) induced by debt policy.
The consequences of risk and uncertainty induced by debt—financed
deficits, so clearly visible in the past five years, may also trace potentially
important transmission channels for debt policies. The papers by Baltensperger and Chan are somewhat suggestive in this respect. The issue
raised by Baltensperger suggests that an increasing risk of default by
inflation associated with a permanent, large, debt-financed deficit raises
the purchasing-power risk of government debt and consequently raises the
(gross) real rate of interest. An alternative mode of approaching the same
issue proceeds along lines suggested by Bomhoff (1983), Mascaro and
Meltzer (1983), and Evans (1984). Permanent and large debt-financed
deficits contribute to uncertainty about the course of monetary policy.
This uncertainty produces a risk premium embedded in interest rates and
raises real rates. And once we move beyond the realm of pure-debt policies and consider deficit policies in a world of distortionary taxes, the risk
problem appears, on a first impression, to magnify. But we still lack at
this stage an adequate analysis of portfolio risks induced by persistent
deficits and the associated uncertain course of tax policies. This analysis
would also extend to the effects of such portfolio risks on real consumption, investment, and real returns of assets. We may ultimately learn from
the work initiated by the “Ricardian discussion” that the “Ricardian
world” offers similar to the Modigliani—Miller theorem a useful benchmark for any analysis of our real problems. It may well be that the new
analysis gradually emerging yields insights into more significant mechanisms associated with debt-financed deficits than elaborated by “conventional analysis.”

3.3. Some recent empirical work on the Ricardian theme
Analytic arguments and counterarguments hardly settle the issue. They
may establish some presumptions with varying weights. An uncertain

68

Karl Brunner

incidence of future tax liabilities seems to be, for instance, a better
approximation to reality than the matching of the distributions of current
tax cuts and future tax liabilities. The fact of uncertainty appears clearly
more acceptable than its denial or the postulated perfect matching. The
approximate realization of such matching in the tax cuts effected on the
basis of the Kennedy or Reagan decisions would be, in my judgment,
quite astonishing. The matching would have to be reflected in a
corresponding matching of experienced tax reductions, additional savings,
and resulting acquisitions of government bonds. Simple institutional facts
(size of denomination relative to tax cut, access costs to capital markets)
distort the pattern and most likely prevent an approximate matching
between tax cuts and bond acquisitions, even with a maintained matching
between tax cuts and savings. But the partial distortion of the matching
need not be decisive per se. The nonmatching segment of taxpayers may
be concentrated toward the lower end of the income distribution involving
a smallish fraction of total tax cuts. Alternatively, these taxpayers may
invest the accrued savings in one form or another of indirect claims on
bonds. The financial intermediation involved in this case usually redistributes risk between the intermediary and the holder of its liabilities. The
more or less indirect claims on bonds are thus not equivalent to bonds. A
positive but smallish wealth effect could thus emerge. But an empirical
assessment of the Ricardian thesis along these lines seems very costly and
quite uncertain.
A study of another major implication bearing on bequest patterns may
be more promising. The Ricardian theme implies that any tax cut relative
to permanent government expenditures induces adjustments of bequest by
the older generations in order to maintain the utility level of the subsequent generation. A similar effect occurs with changes in social security
benefits (i.e., negative taxes) for the older generation. The studies actually
executed so far (Feldstein 1978; Barro 1978) yielded conflicting results. A
direct examination of bequest patterns linked to major tax-policy shifts
could add some information. The Ricardian theme implies, however, in
the context of our actual age distribution and conditional life expectation
a somewhat loose relation between relative tax cuts and additional
bequest. Some of the anticipated tax liabilities will still be borne by the
“older” agents. A tighter relation should prevail between relative tax cuts
and additional savings, however.
Assessments based on some of the crucial linkages emphasized by the
Ricardian hypothesis require a large amount of rather specific information
in order to produce approximately useful results. The operation of
liquidity constraints offers a good example. There is good evidence for
the relevant occurrence of such constraints for some segment under the
wealth distribution. We know from various studies that there exist social

Fiscal Policy in Macro Theory

69

groups which confront borrowing rates massively higher than the
government’s borrowing rates. But it is difficult and somewhat speculative
to assess the relative significance of this fact without detailed additional
cross-sectional data. Professional research quite sensibly attemped, under
the circumstances, another route. Early investigations by Tanner (1970)
and Kochin (1974) explored the implication bearing on the invariance of
consumption expenditures with respect to government debt or debtfinanced deficits. Buiter and Tobin (1978), among others, followed these
efforts. The net result of this early round was hardly conclusive, with some
diversity of results. Six more recent studies by Feldstein (1982), Kormendi (1983), Aschauer (1985), Plosser (1982), Boskin and Kotlikoff
(1985) and Beritheim, Shleifer, and Summers (1985) are selected for closer
examination.
3.3.a. Feldstein. Feldstein expresses the major implication of the
Ricardian thesis in terms of specific constraints on the coefficients in a
regression. This regression relates consumers’ expenditures (C)
C=$0+/31Y+fl2W+/33SSW+/34G+$5T
+ /36TR + /37D + ~
with the relevant variables under consideration, where Y is national
income, W wealth, SSW measures social security wealth, G expresses
government expenditures, T taxes, TR transfer payments, D total government debt, and ji is a random term. All variables are measured in real
terms per capita. The Ricardian hypothesis implies the following patterns:
/35=0,

/36=°’

/33=0

and

/37=—/32’

The last condition assures that an increase in measured wealth due to
government debt exerts no effect on real consumption. Feldstein considers, in addition, a special “fiscal impotence” hypothesis defined by the
four conditions listed plus /34 =
1. This hypothesis does not, however,
represent the neoclassical position developed specifically by Barro. This
problem will be discussed in a subsequent section. The rejection of
= —1 yields, in particular, no evidence bearing on the Ricardian
thesis.
Feldstein concludes an examination of 11 distinct estimations of the
regression equation with the judgment that government spending and
taxes “can have substantial effects on aggregate demand.” He also concludes that “each of the implications of the pre-Ricardian equivalence
hypothesis is contradicted by the data.” This strong and unambiguous
conclusion is somewhat puzzling when evaluated against the results
—

70

Karl Brunner

obtained from the regression analysis. The reader may judge for himself
with the aid of the following table. This table compares the frequency
among the regression results with which the standard error of the
coefficient estimate se exceeds or equals the coefficient estimate of ce, and
the number of cases with a coefficient estimate at least double the
coresponding standard error.
se > ce ce > 2se

f3~

7

136
/33

10

2
7

The last condition—that is, /3-,, =
/32—can only be judged on the basis of
three regressions. Two cases confirm the hypothesis, and one case provides negative evidence. This pattern is really rather mixed and hardly
offers a decisive rejection of the Ricardian thesis. The condition on
transfer payments (i e f~6= 0) emerges as the clearest but not particu
larly overwhelming rejection The other results bearing on the remaining
three conditions seem actually more supportive with respect to the Ricardian thesis.
Several issues associated with the regression analysis obscure the
interpretations and assessment of the results. The taxes used naturally
refer to “non-lump-sum” taxes with their specific incentives and disincentives Even a highly significant /3~< 0 would be difficult to interpret
under the circumstances. We also note that the error-structure problem
explored by Plosser and Schwert and the issues associated with potential
nonstationarity are neglected.
3.3.b Kormendi. Kormendi recently offered an interesting paper
exploring our subject. His discussion expands the role of government following the suggestions of Martin Bailey The government sector is essentially recognized as a production sector supplying consumption and
investment goods. It also operates with a “dissipation factor” representing
the social cost of government production. A “consolidated explanation”
of private-sector real consumption (excluding purchases of consumer durables) is developed according to the rationality concept of the neoclassical
position. This means that households’ information about their available
resources or opportunities fully reflects the underlying social reality
without any distortion of their perceptions. The resulting consumption
function is represented by the regression
—

PC,

=

a~+ a11Y, + a12Y,_1 + a2GS, + a3W, + a4TR, +

ii,,

(10)

where PC = personal consumption, Y = net national product, GS =
government spending on goods and services, W = private real wealth,

Fiscal Policy in Macro Theory

71

and TR = transfer payments. All variables are again in real terms. The
“consolidated hypothesis” implies that a11 > 0 < a3 and
1 < a2 < 0.
The magnitude of the latter coefficient reveals the nature of the government sector’s production process and of its output. Kormendi emphasizes,
moreover, that a4 > 0 can be reconciled with the consolidated hypothesis.
This pattern occurs in case transfer payments involve a redistribution
from social groups with lower marginal propensity to groups with higher
marginal propensity to consume.
The comparative robustness of the “consolidated explanation” yields, in
Kormendi’s judgment, some initial indications of the neoclassical
hypothesis. This robustness is evaluated in accordance with the procedure
developed by Plosser and Schwert (1978). The results are quite satisfactory. The estimates derived from the application of ordinary least squares
to level data, from generalized least squares, and from first differences
coincide very closely.
A second step of the examination enlarges the regression. This yields a
“nested specification” subsuming both the standard version and the consolidated explanation:
—

t~PC,= a0 + a11z~Y~
+ a12z~Y~_1+
a2~~xGS,

(11)

+ a3z~W,+ a4E~TR,+ a5~~.TX,
+ a6z~RE,+ a7t~GINT+ u,,
where TX = taxes, RE = retained earnings, and GINT = interest payments made by the government. The alternative hypotheses imply the
following patterns with respect to the crucial four coefficients:
standard version:
consolidated version:

0, a6 < 0, a7 > 0,
a6 = a7 = 0.
The results are quite unambiguous. The standard version is clearly
rejected. The estimates of a5, a6, and a do not differ significantly from
7
zero at standard levels. An F-test applied to the last three coefficients
confirms the consolidated version. It should also be noted that the estimates of a11, a12, a2, a3, and a4 coincide with the estimates obtained from
the examination of comparative robustness. The standard version implies,
moreover, that a5 =
a1 = (a11 + a 12)~This condition is also rejected.
The test of the net wealth position of government debt yields, in view of
the discussion of uncertain future tax liabilities, a remarkably interesting
result. The full discounting of future tax liabilities associated with current
debt finance implies that the coefficient for government debt D in an
extended regression including this variable must be zero. The standard
version would assign, in contrast, a positive coefficient. Estimation based
on a sample including the war years produces a highly significant negative
—

a2
a2

=

<

0,
0,

a5
a5

<
=

72

Karl Brunner

value for the debt coefficient. The exclusion of the war years still yields a
negative coefficient 1.6 standard errors from zero. The author’s interpretation of the result is worth quoting:
The real income stream deriving from government debt involves inflation
risk and some default risk to holders of the debt. The future tax stream
implied by the debt, on the other hand, involves that same inflation and
default risk plus considerable additional risk as to both its intertemporal
and cross-sectional incidence. Thus, in rationally assessing the future tax
consequences of government debt, the current certainty equivalent value of
the future taxes may exceed the current certainty equivalent value of the
income stream (which is simply the market value of the debt). In such a
case, the net wealth of the private sector is adversely affected by government
debt, implying a negative effect for z~GB,(and a positive effect for I~TX)on
private consumption.
3.3.c. Aschauer. Aschauer’s exploration also addresses, similarly to
Feldstein and Kormendi, the implications of the neoclassical thesis bearing on households’ consumption behavior. The present paper develops,
however, a difference procedure. The author investigates the Ricardian
equivalence proposition in the context of an intertemporal optimization
framework. A rather standard separable utility function is maximized
subject to a consolidated condition derived from combining the representative households and the government’s budget constraint. The integrated
constraint reflects the households’ full recognition of the real conditions
determined by the government’s fiscal operation. The argument of the
utility function refers to “effective consumption” C* = C + 8G, defined
as the sum of private-sector consumption plus a component of government spending G that contribute to the households’ consumption. The
parameter expresses the marginal rate of substitution between the two
components in effective consumption. With a quadratic utility function,
constraint maximization yields an Euler equation

e

EC,*

=

a + /3C,~1

(12)

.

The coefficients are determined by the parameters of the utility function
and the constraint. Upon translation into a stochastic context, the Euler
equation coincides with Hall’s (1981) formulation. The latter argued that
in the context of a life cycle model, /3 = 1 and that consumption moves
along a random walk.
Aschauer supplements the Euler equation with past values of the deficit
D and estimates the regression
C,

=

a + /3C,_1 + y1D,1 + 72D,_2 + y D,_ + y D,_ +
3
3
4
4

ji,.

(13)

Fiscal Policy in Macro Theory

73

This extension was motivated by the author’s concern to evaluate the
impact of fiscal policy. The neoclassical position incorporated in the
underlying optimization schemes implies that all the y-coefficients are
zero. An OLS estimation of the augmented regression yields a clear contradiction of the neoclassical thesis. The deficit variable contributes
significantly to the explanation of private consumption C. The sign pattern
is significantly different from zero. Aschauer argues that this result “may
be more apparent than real due to the fact that past taxes and deficits
may help to predict current government spending.” This information
content of past deficits combined with the substitution relation between
G and C conditions nonzero levels for the y-coefficients.
The author develops this idea in two steps. First he decomposes
effective consumption C* in the (stochastic) Euler equation
EC* = a + /3C,1 and obtains
C,

=

a + flC,1 + /38G,_1

—

8G,° +

U,,

(14)

where G,°= E,1 G,. Secondly, he introduces a forecasting equation for
G, in order to relate G,°with observations:
=

y + c(L) G,_

1

+ w(L) D,

1 + v,.

(15)

The last two expressions determine the system with the crucial properties
to be estimated and evaluated. This system consists of the forecasting
equation, Eq. (15), and Eq. (16):
C,

=

6 + $C,1 + i~(L)G,_1+ js(L)D,_1 + u,.

The derivation of the system implies some cross relations linking
and w. These cross relations offer the relevant test statements:

(16)
i~with ji

(They) restrict the way in which past government expenditures and past
government deficits may influence present consumption expenditure. In
particular, if the Ricardian equivalence proposition does not hold, past
values of the government deficit should have explanatory power for consumption expenditure apart from the role in forecasting government spending. Consequently, a finding that the data do not do violence (to the restrictions) yields some grounds on which to argue that to a first approximation,
the joint assumption of rational expectations and Ricardian equivalence
provides a plausible description.
The nature of the formulated hypothesis implies that the relation between
constrained and unconstrained estimates determines the crucial test information. The log-likelihood ratio provides thus the relevant test statistic.

74

Karl Brunner

It appears that the null hypothesis representing the neoclassical thesis cannot be rejected at a significance level lower than 24 percent. A test based
on a forecasting equation for G with longer lags yields a significance level
of 10 percent and another test imposing additional constraints shows a
significance level of 25 percent for the likelihood statistic.
3.3.d. Provisional comments. Kormendi and Aschauer manage to
demonstrate that the neoclassical position represented by the Ricardian
equivalence proposition cannot be so easily dismissed. The case for the
conventional alternative, which assigns significance to financial decisions,
is neither clear nor overwhelming. The evaluation centered on the patterns exhibited by consumption behavior so far remains unsettled and
open. Adyocates of the “conventional hypothesis” (like the author of this
paper) are forced to admit that the neoclassical position deserves serious
investigation. Further examination may affect both the conventional and
the neoclassical thesis. Neither position may survive unscathed. The
suggestions concerning the role of uncertainty and risk joined with
Kormendi’s estimate of the debt parameter may give a clue for possible
future work. Such work should, in particular, also attend to a serious gap
in the papers discussed above. The Ricardian proposition implies a tight
relation between relative tax cuts, additional savings and acquisitions of
government debt by households. According to this proposition the
induced additional savings do not spill over beyond bond acquisition to
consumer durables. The absence of any significant effect of the relevant
fiscal policy variables on real consumption expressed by nondurables, services, and use-value of durables offers only very partial evidence. The
regressions explored need be complemented with a similar regression
addressed to the investment in consumer durables and possibly other
assets typically held by households that are not equivalent to government
debt.
Some problems associated with Aschauer’s paper need to be noted. The
statistical work is based on level data. The error structure problem
emphasized by Plosser and Schwert may not be serious according to the
value of the DW statistic. The problem posed by potentially nonstationary data remains and is not clear in its implications. Some more attention, as exemplified by Kormendi, to these troublesome issues would seem
desirable. Aschauer’s procedure offers, moreover, no sharply focused
discriminating test between the neoclassical and the alternative thesis.
The Euler equation supplemented with lagged deficits [i.e., Eq. (13)]
hardly represents the “conventional thesis.” The weird pattern of
coefficients for D,_1 and D,_2 contradicts the conventional thesis. The
subsequent test based on Eq. (15) and (16) with their cross parameter properties thus offers at best a test of the neoclassical proposition against an
uninterpreted class represented by the augmented Euler equation excluding the conventional thesis.

Fiscal Policy in Macro Theory

75

3.3.e. Plosser. Plosser’s examination of the Ricardian thesis addresses
an entirely different dimension than the previous three papers. His paper
investigates the implication of the Ricardian hypothesis bearing on assetmarket patterns. A temporal reallocation of taxes expressed by a
corresponding accumulation or decumulation of debt does not affect asset
prices and interest rates under the Ricardian hypothesis. The basic idea is
implemented in the spirit of a neoclassic analysis. Market efficiency or
rational expectation is combined with the expectations hypothesis of the
term structure of interest rates. This analytic foundation implies a relation between the surprises in holding period yields of securities with
different maturities and corresponding surprises in fiscal or monetary
magnitudes:
H,+1

—

EH,+1

=

B[X,+I

—

EX,+1] + v,+1,

where H,+1 is a column vector of holding period returns from t to t + 1
of bonds of various maturities, E refers to the expectational magnitude, B
is a matrix conformable with the dimensions of the vectors H and X; the
latter vector contains the relevant fiscal and monetary variables. These
variables refer in this specific case to the debt monetized by the central
bank—the government debt held by the public and government purchases
of goods and services. v,~ denotes a random vector.
1
The neoclassical position implies that an unexpected increase in government purchases raises interest rates and thus lowers holding period
returns. This implication is compatible with Keynesian or monetarist
analysis. The underlying analysis, however, attributes the result in each
case to a somewhat different mechanism. The crucial difference between
the neoclassical thesis and the alternative positions, however, surrounds
the role of government debt. Surprises in this magnitude yield no consequences with respect to holding period returns under the Ricardian
hypothesis. They should produce, in contrast, negatively related consequences under the alternative hypotheses. Lastly, surprises in the monetized portion of the public debt yield, according to Keynesian and (older)
monetarist analysis, positively related surprises in holding period returns.
Neoclassical analysis is not inherently inconsistent with a nonvanishing
effect of monetary surprises. It remains, however, somewhat ambiguous
on this point without further specifications bearing on expectations and
shock structure. With some dominance of comparatively permanent
shocks in the monetary variable, its surprises convey a useful signal value
bearing on future inflation. This response mechanism would produce
negative reactions in holding period returns to positive monetary surprises.
These reactions would, moreover, increase with the maturity of the security under consideration.

76

Karl Brunner

In order to complete the analysis, two analytic building blocks are
added. The forecasting equation
X,~1= A(L)X, + u,~1

(17)

is introduced with A(L) designating a matrix of polynomials in the lag
operator L; u again refers to a random vector. This formulation advances
implicitly an hypothesis about the structure of the process generating the
observations of the X-vector. Lastly, an expression for EH,+1 is derived
from the rational expectation theory of the term structure:
EH,+1

=

R1, + 4~,

where R1, refers to a vector consisting only of the current one-period spot
rate, and 4 is a vector containing the marginal increments of liquidity
premia associated with different maturities beyond the spot rate. Plosser
also offers an alternative interpretation of 4 based on the Sharp—Lintner
capital-asset model. In this case 4 would be equal to a /3-vector multiplied by the difference between the expectation of a holding period yield
for a market basket and the certain current spot rate R1,.
The first step in the empirical examination evaluates the joint
hypothesis about market efficiency and the term structure. The implication that current surprises of holding period returns are independent of
past observations on money, debt, and government purchases is tested
with a suitable regression. The results are quite unambiguous and support the joint hypothesis. A second step in the procedure explores the statistical results bearing on the matrix of coefficients B in Eq. (16). Holding
period returns for four different maturities are investigated. In two cases
the coefficient estimate of the debt variable is less than its standard error.
One coefficient estimate is significant at the 10 percent level, and the
fourth coefficient estimate, slightly exceeding its standard error, occurs
with a significance level above 10 percent. All coefficient estimates associated with the debt variable are positive. The signs are thus inconsistent
with the “conventional analysis.” The significance levels confirm, on the
other hand, the null hypothesis expressed by the neoclassic thesis. Sign
and significance level together yield a clear rejection of the “conventional
hypothesis.”
The coefficient estimates of the monetary variable also support some
version of the neoclassic thesis. Their signs, with the exception of the
security with the shortest maturity, are, however, consistent with the “conventional analysis.” Only one out of four coefficients reaches a
significance level of at most 10 percent. The sign and the confirmation of
the null hypothesis obviously provide no support for the supplementary

Fiscal Policy in Macro Theory

77

hypothesis bearing on the signal effect of monetary surprises in the con
text of a specific shock structure mentioned above. The coefficient
estimates associated with government purchases provide some support for
the occurrence of an effect. All estimates exhibit the sign implied by the
hypothesis. Two of the four estimates are significant at the 5 percent level
and one at the 10 percent level The significance level for holding period
returns on the longest matunty nses beyond 10 percent We also note the
significant estimates for the constant parameter in the regression. This
confirms the occurrence of a liquidity premium or the operation of a nonvanishing /3-factor on government securities.
The interpretation of the test deserves some attention. This issue may
be usefully addressed with the aid of a somewhat more explicit characterization of the test procedure. The following pair of propositions control
the test:
F ~ [ME.EHTS’NCH ~ T],

F ~ [ME.EHTS. CH ~

—

T}.

F refers to a sentence presenting the hypothesis summarized in the forecasting equation ME denotes a sentence advancing the market efficiency
hypothesis, and EHTS denotes a hypothesis about the term structure of
interest rates; NCH and CH refer to the alternative hypothesis—that is,
the neoclassical and the conventional. Lastly, T refers to the test statement about the B coefficients under the null hypothesis. The crucial point
to be addressed here is the conditionality of the test relative to the
untested hypothesis contained in F The compelling force of any
confirmation or disconfirmation of T depends decisively on the assumption made about F. Suppose F is true and ME.EHTS and T are
confirmed. We can effectively discriminate under the circumstances and
conclude that NCH is confirmed and CH disconfirmed. But suppose that F
is false. The logical relations between the various sentences offer in this
case no grounds to discriminate between NCH and CH on the basis of
tests bearing on T. With the truth of F given, and ME~EHTSconfirmed
independently, the confirmation of T must disconfirm CH and can confirm
NCH in order to satisfy the truth of the whole conditional proposition.
Should F, on the other hand, be false, then the whole conditional is true
irrespective of our decision about NCH, CH, and T. The truth of the
conditional imposes under the circumstances no constraint on confirmation or disconfirmation of NCH or CH relative to any given decision
about T. The falsehood of F implies that the test is irrelevant and uninformative. It follows that in the absence of any information about F the
results of the test cannot be assigned substantial weight.
One procedure designed to raise the informativeness of the test involves
repetition with a vanety of different forecasting equations Plosser s paper
moves in this direction. The test is repeated with a forecasting equation

78

Karl Brunner

including current bond-market information. The results bearing on
government purchases appear somewhat sharpened. The other results are
essentially unchanged. We may conclude that Plosser’s paper does not
essentially modify the balance of evidence summarized above. It did,
however, usefully direct our attention to an alternative dimension—the
financial markets and the corresponding opportunities for systematic
future examinations of the neoclassic thesis.
3.3.f Bernheim—Shle(fer—Summers and Boskin—Kotlikoff Two papers
recently appeared in the professional circuit that appear to shift the balance of evidence somewhat against the hypothesis of debt neutrality.
Both papers move beyond the macro data applied to an examination of
consumers’ behavior. They exploit in very different ways some micropatterns systematically related to the relevant analysis. The first paper, by
Bernheim, Shleifer, and Summers (1985), seems particularly noteworthy.
It starts with the observation that most of the existing wealth was
transmitted by the prior generation. We noted in a previous paragraph
that an “accidental transfer” hypothesis without any bequest motive could
explain the occurrence of bequests in a non-Ricardian world. This explanation would probably hold even with an availability of annuities. But it
seems doubtful that this hypothesis can explain the order of magnitude of
the bequests. The crucial contribution made by BSS (i.e., the three
authors) emerges from the explicit construction of a detailed (“nonRicardian”) hypothesis explaining the occurrence of bequests that incorporates a bequest motive. This hypothesis, in spite of the acknowledged
bequest motive, yields radically different implications from Barro’s
hypothesis. The explicit specification of this alternative bequest hypothesis offers opportunities for new and richer critical evaluations with
untapped data.
The “exchange motivated bequests” hypothesis involves a simple idea
developed with a subtle analysis. The idea applies the basic REMM
(resourceful, evaluating, maximizing man) model to the interactions
between older and younger persons. The hypothesis thus rejects Barro’s
assumption that older persons are concerned about the future consumption opportunities of their descendants. Older generations accumulate
wealth in tradeable (i.e., bequeathable) form in order to purchase services
from the younger generation. The services bought by the prospect of
bequests occur essentially in the form of attention extended by the
younger to the older. An implicit exchange transaction of potential
wealth for current attention determines the evenutal wealth transfer from
older to younger.
The authors structure their analysis with two sets of utility functions
and a set of constraints describing wealth accumulations over time. The
utility functions represent the agents in the interacting groups. The
expression

Fiscal Policy in Macro Theory

~

/3ts

P(c, t) U,(C,, a,)

79

(18)

describes the utility function of the testator, where /3 is a discount factor,
P(s,t) is the probability of survival from the initial period s to period
and U, is the instantaneous utility as a function of consumption C, and
the attention vector a, = (a1, . . afl,). This attention vector describes the
degree of attention extended to the testator by each of the N members of
the set A of “credible” beneficiaries addressed by the testator. The latter’s
utility increases with C, and a,. He is also assumed to have a finite deterministic life span T so that P(s, T + 1) = 0.
The utility function of the N beneficiaries is given by
~

/3,~S

~

(19)

We need only comment here that BSS assume beneficiaries to live forever.
This assumption simplifies matters with no loss in relevant substance. U~
has a positive derivative with respect to C,~and a negative derivative with
respect to an,. Attention is valued quite differently by the two sides of the
exchange.
Expressions (20) and (21) introduce descriptions of wealth accumulations.
a.

w,~1 = (W~

C,

b.

B,=W,—C,—A,.

—

—

A,) (1 + r,) + A,(l + p,),

(20)

The first applies to testators, with W designating wealth, A the annuity
investment made, r the market rate of interest, p, the rate of return on
A, and B = ~
the total sum of bequests; b~is the component of B
going to beneficiary n. Formulation (a) holders in case the testator survives, and (b) holds in the case of death during the period.
=

(Wa,

—

C, + b~,I(t

—

1)) (1 + r,).

(21)

Expression (21) supplies the condition for beneficiary n. 1(t) = 1 in case
the testator dies during the period; otherwise it is zero.
The structure of the strategy game between testators and beneficiaries
has now been defined. The testator invests in bequeathable wealth in
order to induce attention from potential beneficiaries, and the latter compete with attention for bequests. BSS develop a clever and intricate
argument to determine a (Nash) solution to this game. A general sense of
the detailed argument will suffice here. The testator chooses an optimal
plan (C,*a, b~,. . . , bN) for consumption, attention, level, and distribution

80

Karl Brunner

of bequests. Optimization proceeds subject to the constraints, including a
feasibility condition. The latter imposes that the choice of (a,b1,. . . , bN)
must be confined to a range assuring to beneficiaries at least a utility
equal to nonparticipation in the game (i.e., a~= b~= 0). The
beneficiaries may be interpreted to play a subgame conditioned on the
vector (b~, . . . , b).
The testator can, of course, not impose his optimal
choice of attention on beneficiaries. He thus faces a problem of selecting
a bequest rule—that is, a vector function b(H,B), where H denotes the history of the game and B the total level of bequests—that induces
beneficiaries to supply a* voluntarily in their own interests. BSS show the
existence of a specific rule in situations with at least two beneficiaries.
This rule does produce a Nash equilibrium solution (a*, b*). This equilibrium implies that the testator fully appropriates the surplus utility
created by the exchange. The Ricardian equivalence theorem fails to be
satisfied under this “exchange motivated bequests” hypothesis. Bequests
do occur and they are motivated. They are, however, not motivated by
the future welfare of descendants but by the purchase of current attention.
A debt-financed deficit yields, in general, under the circumstances, no
offsetting intergenerational transfers and personal saving. Opportunities
are modified and real variables change.
An extensive empirical evaluation follows the analytic argument. The
hypothesis implies that parents influence their children’s behavior by
holding wealth in bequeathable forms. It implies, in particular, that contacts between parents and children within families with bequeathable
wealth are more extensive. The authors exploit the data from a longitudinal retirement history survey. They especially derive data on bequests b
and contacts. A normalized measure V of contacts per child is constructed with the raw data. An OLS regression of V on b yields the
“proper sign” but does not confirm the hypothesis. The authors trace this
negative result to potential endogeneity of b. The hypothesis does imply
that b and V are simultaneously and interactingly determined. A TSLS
procedure thus corresponds better to the structure of the hypothesis. The
results are dramatically different in this case. The null hypothesis of no
effect on b on V can be rejected with high confidence. A special test
assessing the exogeneity of b supports the choice of a TSLS procedure.
Exogeneity of b can also be rejected with substantial confidence.
The authors recognize that consistency of the empirical pattern with the
hypothesis cannot exclude other possible explanations. They evaluate a
variety of such alternatives in order to strengthen support for the
hypothesis advanced. They consider thus the possibility that influences
emanating from several omitted and personal dimensions are erroneously
attributed to b. The effect of b on V survives this examination. Some
alternative explanations of the observed correlation between V and b do

FiscalPolicy in Macro Theory

81

not distinguish between wealth in bequeathable and in nonbequeathable
forms. The hypothesis advanced assigns, in contrast, no significance to
the latter. The statistical results again confirm the hypothesis under
examination. The case for alternative hypotheses emphasizing the role of
the cost of contacts imposed on children or of the parents housing wealth
fails similarly when confronted with relevant data. Other implications are
also exploited in order to extend the range of relevant observations bearing on the hypothesis. Lastly, data on the comparatively low frequency of
privately purchased annuities or of gifts offer some useful information.
BSS discriminate, with their help, between the hypothesis considered, the
“accidental bequest” hypothesis discussed in a previous section, the
“bequests for their own sake” hypothesis, and the “intergenerational
altruism” hypothesis. The authors’ hypothesis also survives this last round
of assessments. A careful reader of this paper may agree that the wideranging and imaginative empirical evaluation establishes a serious case on
behalf of the hypothesis that bequests are a component of an exchange.
The relevant and pervasive occurrence of such transactions is, moreover,
inconsistent with the Ricardian thesis of debt neutrality.
A serious limitation of prior studies bearing on the Ricardian thesis is
their concentration on macro data of consumption patterns. BSS substantially enriched our assessment by exploring a wider range of implications
requiring micro data. Boskin and Kotlikoff (1985) pursue a similar course
in an investigation that required a massive labor and computational
input. Their examination addresses an important implication of the
Ricardian hypothesis expressed by the intergenerational altruism model.
Under this hypothesis consumption expenditures depend only on “collective resources” representing the real underlying situation. It implies, in
particular, that consumption does not depend on the age distribution of
the population. The analysis is based on a present value of family utility.
This formulation involves a series of instantaneous utility functions
specified for all descendants and their respective age groups. A description of wealth accumulation for both household and government serves as
a constraint on the optimal choice for an intertemporal consumption pattern of the “infinitely living” family. The first-order condition determines
an expression within the usual family of Euler equations relating consumption in adjacent periods. The problem, however, allows no tractable
analytic solution describing consumption as a function of predetermined
magnitudes. The authors thus pursue with great patience a different
course. They solve a finite-period approximation to the infinite optimization problem. The approximation is chosen so as to lower changes in the
optimal consumption pattern produced by extending the period to a negligible level. The data set used covers the period 1946—81. A dynamic programming approach is applied to compute the solutions for the relevant

82

Karl Brunner

sample period. It should be noted that this optimal pattern C is derived
in the context of stochastic uncertainty about future rates of return and
earnings. The derivation, moreover, depends on a specific utility function
and an age-weight assigned to age-group instantaneous utilities. The
authors actually investigate combinations of parameters (of instantaneous
utility and the discount factor) in order to find the selection that determines for C the closest fit (in terms of root mean squares) to the data.
The test is performed with a regression of actual consumption C on C
and five age groups expressed in terms of their respective income shares.
They hypothesis of intergenerational altruism,
(22)
implies that A0 = A. = 0 for i = 1, . . . , 5 and /3 = 1. The statistical
inference confirms the last condition but disconfirms the others. The pattern of A-coefficients seems consistent with some life cycle hypothesis.
The coefficient pattern estimated implies, in particular, that a redistribution of 10 percent of income from the younger to the older generation
would raise consumption by 1 percent and lower the net national savings
rate by 9 percent.
The authors’ preliminary capital and labor-intensive investigation
clearly disconfirms the Ricardian thesis couched in terms of an “intergenerational altruism” hypothesis. This disconfirmation must be substantially
qualified, however. The test does not uniquely address this crucial
hypothesis. It is mixed with an array of auxiliary hypotheses bearing on
the choice of utility function, age-weight assignments, the specification of
uncertainty, and other components. The disconfirmation could, of course,
apply to the set of auxiliary hypotheses. This comment does not lower the
value of a major piece of work developed by Boskin and Kotlikoff. Their
examination does reenforce the results obtained by BSS, and further
research involving variations in the choice of auxiliary hypotheses may
confirm the rejection of the Ricardian thesis.
3.4. A neoclassical analysis of government expenditures
The Ricardian theme does not imply irrelevance of fiscal policy. Expenditures and taxes remain potent instruments shaping output, employment,
and welfare. Barro (l98lb, 1984) also initiated in this field the neoclassical explorations. The general analysis of fiscal policy uses the same
market-clearing approach so extensively exploited in monetary analysis.
Barro’s discussion of fiscal policy moves incisively beyond the sinkhole
theory of the government’s operations. The private-sector output
acquired by the public sector forms the basis for the supply of public con-

Fiscal Policy in Macro Theory

83

sumption goods to households and productive input services to private
producers. “Public consumption” competes to some extent with private
consumption. A parameter, a, summarizes this fact. It expresses the marginal rate of substitution between public and private consumption. With
a = 1, “public consumption” and “private consumption” substitute one
for one. A vanishing a—that is, a = 0—reflects absence of any substitution between the two types of consumption. With 0 < a < 1 a unit
increase in “public consumption” lowers private consumption by less than
one unit.
The government’s supply of productive services raises an input available to private producers. Private-sector output thus expands in accordance with the marginal productivity of this input in private production. A
parameter /3 reflects this marginal productivity. The parameter describes
the increase in private-sector output produced by a unit increase in
government real expenditures. Government real expenditures both fully
reflect the public sector’s absorption of private output and its supply of
goods and services to the private sector. The government sector is made
to behave as if it contracted for goods and services produced by the
private sector that are immediately made available to private households
and producers. It is, moreover, assumed that the government’s input services do not affect the marginal product of labor and capital in the private
production process.
The system used to analyze the impact of government expenditures and
(distortionary) taxes is confined to some basic elements expressed by two
equations:
C(r, C,,

+ I(r,...) + C,

~.)

M

=

=

Y(r,G , G,...),
1 +
—+

L(r+i~,Y,G ,...).
—
+— 1

(23)
(24)

Equation (23) describes the market-clearing condition for the output
market. The various symbols have standard meanings: C = private real
consumption, I = private-sector real investment, G1 = current government real expenditures, G = permanent government real expenditures
anticipated for a horizon beyond the current period, Y = real income,
r = real rate of interest, M = money stock, P = price level, and i~ =
expected rate of inflation. The signs below the variables indicate the
direction of response of the dependent variable. The two distinct magnitudes for government real expenditures are introduced in order to analyze
the impact of transitory and permanent changes in government expenditures. This analysis proceeds initially under the assumption of lump-sum
taxes. Distortionary taxes are introduced at a later stage. The path of the

84

Karl Brunner

money stock together with the two government expenditure variables refer
to the exogeneous components in the analysis. This implies that ~ is also
held fixed in accordance with the path of M.
We consider first the impact of a transitory increase of C,. A simple
diagram (Figure 3.3) is used for this purpose. Line d represents the
demand for output occurring on the left side of Eq. (23). Line s marks
the supply located on the right of the same equation. An increase in C,
against a background of a constant permanent government expenditure
implies that some expected future expenditures must fall to offset the temporary rise of C,. Constancy of G means that the representative real
opportunities expressing the households real wealth remains unchanged.
These real opportunities are defined by the difference between the present
value of future real income and the present value of future government
real expenditures.
A constant ~ implies that an increase in C, induces no wealth effect on
demand or supply of output. This means, in particular, that z~G1> 0
shifts the demand and supply line in the graph for reasons other than
changes in basic real opportunities. The net effect of the impact on output and real interest rates can be easily determined by comparison of the
horizontal shifts of the two curves. The supply shifts to the right by an
amount /3 per unit change in C,, whereas demand shifts by an amount of
rr

S

V
Figure 3.3

Fiscal Policy in Macro Theory

85

1
a in the same direction. Assuming that a + /3 < 1, as Barro does,
the increase of G1 creates an excess demand of 1
a
/3 per unit
increase at the initial real rate. The market-clearing condition thus forces
a rise in the real rate of interest. The simultaneous shift of both curves to
the right assures, moreover, that output also rises. We note immediately
that we observe, in contrast to a Keynesian analysis, no multiplier effect.
The diagram suggests that the increase in output per unit increase in C, is
less than 1
a < 1. The economy’s response thus imposes a crowdingout effect on both private consumption and investment. Crowding out is,
however, avoided with /3 = 1 and a = 0.
The solution for r and Y can be inserted into the portfolio equation in
order to determine the response of the price level. Before we consider this
final result we need to examine the rationale of C, in money demand. C,
operates in the nature of an in-kind transfer. It is thus interpreted to
lower, relative to Y, the household’s monetary transactions. The increase
in G1 thus reenforces the effect of a higher real rate of interest on real balances. The net result thus depends on the comparative strength of the
income effect. The price level rises in response to a temporary increase in
government expenditures, provided that the effect exerted by r and G1 on
real balances dominates the income effect. With a dominant income effect,
prices fall.
A different pattern emerges under a permanent increase in government
expenditures. This requires a simultaneous rise in G1 and G. We need to
consider, therefore, the consequences of an increase in G. Once again we
evaluate the shifts in demand and supply of output. We note first that the
immediate effects induced by C, expressed by a on household_demand
and /3 on productio~ido not occur in this case. The increase in G with C,
constant provides no additional goods or services in the current period.
The current impact of ~ operates via wealth and labor effort. A unit,
increase in G lowers real opportunities of the private sector by one unit,
holding r constant. This is modified by future a and /3 effects. The first
effect provides additional resources to households, and the second raises
the future income stream. Additional modification results from a higher
labor supply produced by lower real wealth. This last modification raises
real income by an amount a. The net wealth decline is thus
a
/3
a. Barro postulates at this stage a one-to-one relation
between changes in real wealth and real consumption. This implies that
the average propensity to consume with respect to wealth prossesses a
suitable positive derivative with respect to real wealth. The increase in G
induces no direct effects on investment and C,. The total direct-demand
effect coincides therefore with the horizontal shift of the consumption
function—that is, —(1
a
/3
a). We combine this shift in aggregate demand due to ~ with the shift due to C, (which is 1
a) and
—

—

—

—

—

—

—

—

—

86

Karl Brunner

obtain /3 + a as the net shift in aggregate demand attributable to a permanent rise in government expenditures. This matches the positive horizontal shift on the supply side, which also equals /3 + a. It follows that
a permanent expansion of government expenditures raises output with no
effect on real rates of interest. The price level effect is algebraically larger
than the response produced by a transitory increase in government expenditures. But the price effect still remains ambiguous and conditioned by
the comparative strength of Y and C, in money demand. The dominance
of Y implies a fall in the price level. Lastly, similar to the case of transitory expenditure increases, there occurs no multiplier effect. The expansion of output /3 + a < 1 is less than the unit increase in G1 + G. The
resulting crowding out is, moreover, concentrated on private consumption.
The impact of tax policies is investigated in a similar vein. A flat-rate
tax on income net of depreciation and exemption is introduced. This tax
modifies the relevant margins for household and investor decisions. This
modification implies that the after-tax real rate of interest and the tax rate
appear as arguments in the consumption and supply functions. It follows
that an increase in the tax rate offset by expanded exemptions in order to
satisfy the government’s budget constraint lowers output and real interest
rate.
This analysis has been exposed by Barro to some empirical assessment.
One evaluation (1 984a) traces the broad contours of transitory expenditures. Such expenditures are essentially confined to war episodes and thus
emerge for all practical purposes in the nature of military expenditures.
The Vietnam episode can be disregarded, since the comparatively negligible magnitude of transitory spending (2 percent of GNP) is probably
swamped by other influences on real GNP. In contrast, the two world
wars and the Korean war episode are quite informative. Transitory spending loomed with substantial force. It is remarkable that aggregate output
responded with a fraction of around .55 to .60 to transitory expenditures
in all three cases. It is noteworthy, moreover, that crowding out operated
dominantly on real investment. These observations appear quite
consistent with the implication of Barro’s analysis.
A more thorough statistical investigation addressed the relation between
aggregate output on the one side and both transitory or permanent expenditures on the other side. The latter two variables were represented by
some ingeniously complicated measures, which probably avoid nonstationary characteristics. The nonstationary character of output is attended
by the inclusion of a time trend. The results seem remarkably consistent
with the implications bearing on the response pattern of output. Transitory expenditures induce a larger impact than permanent expenditures.
But the case for even a most modest multiplier effect for transitory spending remains quite weak. Still, larger permanent expenditures do lift aggre-

Fiscal Policy in Macro Theory

87

gate output permanently above the time trend, but at some cost of private
real expenditures.
The results may be sensitive to the range of auxiliary hypotheses guiding the measurement procedures. Further investigations along this
interesting approach need to be explored in order to deepen the assessment of a neoclassical fiscal analysis. One particular issue needs our
attention in this context. Nelson and Plosser (1982), followed by Stulz
and Wasserfallen (1985), demonstrated with extensive tests that the
hypothesis of a deterministic trend is overwhelmingly dominated by the
hypothesis of a stochastic trend. A stochastic trend usually collapses the
stationary component and radically modifies the results of regression
analysis applied to it. One suspects that substituting a stochastic trend for
the time trend and then regressing the resulting stationary component on
transitory and permanent expenditures would substantially alter the outcome.
Some comments should still be addressed to the analysis. The contrast
with Keynesian analysis should be noted first. Barro confirms that an
essentially price-theoretical approach (Brunner 1970) lowers the
significance of interacting multiplying flows. The latter moves into center
stage of the analysis once we proceed on the assumption of given price
level or nominal wages. A “non-market-clearing” analysis typically converts fiscal actions into a multiplier effect on output. This effect is produced by variations induced in aggregate demand encountering unresponsive prices. Barro’s neoclassical approach differs radically from the
Keynesian tradition. Fiscal policy is analyzed in the context of full information and market clearing. This context would prevent any real effects
of monetary shocks. This environment does not preclude temporary and
permanent real responses to fiscal actions. These responses, however,
depend crucially on the properties of output supply. The supply
responses fully determine the output effect of permanent expenditure
increases. We also note the emphasis on changing output composition
produced by fiscal action associated with the “crowding out” of private
consumption and investment. The difference in this respect between temporary and permanent fiscal action is also significant. This result cannot
be reproduced within the standard IS/LM framework.
There remains a question bearing on the interpretation of this analysis.
The full information market-clearing approach describes, in my judgment,
a benchmark forming a “gravitational center” of economic processes
approximating long-run aspects. Two important revisions would probably
move us somewhat closer to reality in some sense without sacrificing
Barro’s basic price-theoretical thrust. The prevalence of incomplete information needs to be recognized, and a more general concept of market
clearing needs to be used. This concept acknowledges that prices do not

88

Karl Brunner

reflect all ongoing shocks. They will rationally reflect perceived permanent shocks but not (perceived) transitory shocks. There will be a
market clearing under the circumstances relative to (perceived) permanent
shocks but not with respect to all shocks. Some shock-absorbing buffers
thus operate in the economy and distribute the output adjustment to
shocks over time. This pattern would probably produce a more substantial difference between the effects of temporary and permanent expenditure changes.
A representative work of neoclassical analysis in the new mode was
recently developed by Aschauer and Greenwood (1985). The analysis is
built in a dynamic choice-theoretic context. A representative household
optimizes over two periods subject to a constraint that incorporates tax
parameters affecting labor income in either period and a tax parameter
imposed on the second period’s capital income. Market-clearing conditions are added for each period. These conditions reveal on the supply
side the structure of production; that is, investment in the first period adds
to output in the second period. The first-order optimality condition
together with the market-clearing condition determines the system of
equations used to examine the impact of fiscal policy. A budget constraint for the government sector with a “Ricardian thrust” assures a consistent pattern of fiscal action. Moreover, the government provides, similar to Barro’s case, consumption goods to households and input services to
private producers. The households’ utility function depends, moreover,
on total consumption available to the household, which is the sum of
private consumption and the weighted government contribution with
weight less than unity. This weight again represents the marginal rate of
substitution between the two types of consumption.
The analytic structure is applied to an examination of tax policies, stabilization policy with the aid of adjustable tax parameters, expenditure
policies, and optimal taxation. An increase in the second period’s (flat)
tax rate on labor income induces intertemporal substitutions that raise the
first period’s labor supply and investment, but lower the second period’s
labor supply. The welfare effect of this tax increase depends crucially on
the occurrence of distortionary taxes. Welfare falls when first-period
income is untaxed. Welfare rises, however, in case first-period income is
already taxed. The new tax modifies to some extent the intertemporal distortion of the first tax.
The consequences of stabilization policy are explored under the
assumption that the production process in the second period is subjected
to random shocks. The larger their variance, the lower the representative
households’ welfare level. But a stabilization policy operating with a
state-contingent tax parameter actually lowers the welfare level. However,
uncertainty due to stabilization raises current consumption and decreases
current work effort, output, and investment.

Fiscal Policy in Macro Theory

89

A temporary (unexpected) increase in first-period government expenditures (expressed by a linear combination of consumption and productive
services) lowers welfare, provided the marginal product of government
productive services and the marginal rate of substitution between private
and public consumption is less than unity. Work effort in both periods
increases, whereas investment declines. Output rises in the first period
and falls in the second period.
An anticipated increase in the second-period government expenditures
raises first-period investment and work effort in both periods. The consequences of a permanent increase in government expenditures follow from
combining the prior two cases. Work effort and output rise in both
periods, and consumption declines.
Aschauer’s (1985) elaborate analysis produces results somewhat similar
to Barro’s investigation. The argument also proceeds on the basis of full
information and continuous full-market clearing. There is, however, no
money and, thus, no price level in the model. The choice-theoretic foundation prevents simply adding a money demand equation. Money would
have to be added to the utility function or embedded in a production or
exchange constraint (Brunner 1951). The usual homogeneity conditions
can, however, be expected to be satisfied. Aschauer’s results probably
carry over to a monetary economy. Finally, an examination of the
detailed structural knowledge required for stabilization and optimal tax
policy reveals the dubious relation between such analysis and actual policy issues. This aspect will be reconsidered in the last section of the
paper.

4. Deficits, Monetary Regimes, and Economic Activity
4.1. The endogenous states of the monetary regime:
Sargent and Wallace
The “stability problem” associated with deficit finance revealed an interrelation between fiscal and monetary regimes. This issue surfaced again in
recent years, but in a modified context. Thomas Sargent and Neil Wallace (1982) approached the interrelation between the two regimes, or the
financial coordination problem, with a concern directed to a very different
issue. They question the long-run survival of an anti-inflationary monetary regime when confronted with persistent deficits sufficiently large to
raise the real stock of government debt relative to real national product.
The problem may be explored with the aid of the government’s budget
constraint:

90

Karl Brunner
0

0

S+B=G+TR—TA+iS,

(25)

where S denotes the stock of publicly held debt, B is the monetary base, G
refers to nominal government expenditures on goods and services TR
designates transfer payments and TA designates tax revenues i should be
interpreted as the average interest rate on outstanding debt The budget
equation can be translated into the following expression:

~=a~+
—

(rr—n)+

(~+n)+

n—-~-+
y
+

~

p

~—n
y

ir—-~-

s

(26)

p
b—b

where s describes the ratio of real debt to real national income. b similarly represents the volume of base money per unit of nominal national
income (i e it is the reciprocal of base velocity) def consists of the basic
deficit ratio expressed as
(27)

~—fG+TR—TA

Y
with Y indicating nominal national product. The other symbols are rr =
real interest rate, n = normal rate of real growth, y = actual output, ir =
expected rate of inflation, and p is the price level.
Expression (26) may be considered as a differential equation in s. A
stable process requires that the bracketed expression associated with s on
the right side be negative. Actual real growth L~y/yand actual inflation
E~p
Ip sufficiently large would produce a negative sign But this state is
purely transitory Over the long run relevant for this investigation the
sign would be determined by rr
n, the relation between the real rate
and the normal growth rate. This relation appeared with a major role in
traditional growth theory. It implicitly occurs also in the analysis surrounding infinite intertemporal budget constraints. The arguments bearing on the Ricardian thesis require in particular that rr > n A real rate
rr exceeding the normal real growth n is a necessary condition for the
Ricardian thesis. Government debt would appear as net wealth under the
opposite inequality. However, the inequality rr < n raises subtle issues
about its relevant occurrence in a steady-state context A major problem
is the reconciliation of finite assets values with rr ~ n. Such reconciliation
could possibly be achieved in a model combining intergenerational
selfishness” with uncertainty of death (Blanchard 1984). This combination would determine a discount rate exceeding the real rate of interest.
—

Fiscal Policy in Macro Theory

91

The relevant long-run relations may now be written as
~—def+(rr—n)s—(~+n)b.

(28)

We note that a steady-state condition also requires that b = 0. This
expresses the fact that the price level is fully adjusted at any time to the
prevailing volume of the monetary base, and b is fully adjusted to the
ongoing inflation. Under the first state—that is, rr > n—consistent with
the Ricardian thesis the debt—deficit process is unstable For any initial
value ~ >0 the real debt ratio will persistently rise with the persistence of
deficits def and low inflation. According to the Ricardian thesis, such a
state does not persist. The temporal distribution of taxes implies that a
stream of positive deficits over the nearer future will definitely be offset by
higher taxes and negative deficits in the wider future. We should clearly
recognize here the structure of the argument. A predetermined path of
deficits def excluding “inflation taxes” proceeds in a “Ricardian World.”
The infinite intertemporal budget constraint thus imposes eventually an
increase in taxes. But the predetermined characters of def and, implicitly,
of ordinary taxes means that the inflation tax remains as the only possible
adjustment to satisfy the infinite budget constraint. It thus follows that a
noninflationary policy cannot be maintained over time in the context of a
permanent deficit that is sufficiently large.
The infinite budget constraint reveals the nature of the problem. The
single-period constraint
G1 + S11(l + i,~)

=

(29)

T~+ S, + B~—

can be assembled into an intertemporal expression in terms of real magnitudes per unit of real income
I

I

l+n
1=0

1 + rr

l+rr
(g~— t ) +
1

SØ

~ + n

l+n
=

1 + rr

1

~B
1

~

~,

(30)

where V is the velocity of the monetary base. Once the left side is fixed at
a substantial positive level, it follows that a very low growth of zXB/B =
over an initial segment must be followed by larger monetary growth in
the future.
The argument presented by Sargent and Wallace can be reconstructed
along the following lines, using expression (30) above. The stock of real
debt per unit of real income at time N satisfies
s(N)

=

—

s

0

+ [def—(i~-+ n)b}

exp(rr — n)N
rr — n

.

(31)

92

Karl Brunner

Under the assumption that def> (‘iT + n)b, the stock s grows monotonically with time N. Sargent and Wallace simply state that there exists ultimately a limit for s (somehow). The higher the levels reached by s(N) the
greater must be the subsequent inflation. This follows again from expression (29). Once an upper limit § is imposed with an unchanged d~1’,the
inflation rate must adjust in accordance with
=

-~

+ (rr

—

n)~

—

n.

(32)

The positive relation between ~ and i is immediately obvious. A persistent accumulation of the real debt burden can ultimately always be terminated by sufficiently high levels of inflation.
Expression (32) can be used with two different interpretations according
to the nature of the inequality between rr and n. With rr > n it describes
the inflation rate required to satisfy an imposed stock of relative real debt
s. Alternatively, with rr < n it determines the equilibrium stock s associated with any predetermined inflation rate; that is
=

def

—

(‘ii

+ n)b

(33)

In the first case fiscal and debt conditions dominate monetary policy, and
in the second case monetary policy dominates the debt position.
The difference between the two states involving the nature of the
adjustment to a constant real debt ratio s can be outlined in terms of Figure 3.4.

so
line 1

if

line 2
Figure 3.4

Fiscal Policy in Macro Theory

93

Line 1 represents the component def + (rr
n)s of and line 2 the
component
(ir + n)]b. Now consider the first state with rr> n and
an inflation rate ITØ below the inflation rate ~ required to hold s constant.
This means that the vertical distance between line I and the ir-axis
exceeds the vertical distance between the ir-axis and line 2. The rate of
change of s is positive under the circumstances, and line 1 drifts higher.
This drift continues so long as the basic deficit def is maintained and iro is
less than ~ [Eq. (32)]. In order to stablize the real debt ratio s, the
inflation rate must be raised to the required rate ~, which rises with s. The
higher line 1 was allowed to drift up, the further out to the right shifts ~.
In the second state characterized by rr < n, the real debt ratio s
adjusts, in contrast, to any predetermined level 1T~. Suppose the initial
condition is again described by the graph. The real debt ratio s rises
under the circumstances, and this lowers because rr
n < 0, line 1.
This process persists until the vertical distance between line 1 and the ‘iTaxis coincides with the vertical distance between the ir-axis and line 2.
Darby’s (1984) rebuttal of the problem addressed by Sargent and Wallace concentrates essentially on the stability of the process and some
observations supporting the required inequality. But the stability hardly
removes the central issue brought to our attention by Sargent and Wallace. The following table reveals the problem. We are shown the equilibrium real debt ratio associated with various basic deficit ratios under a
stable process.
—

~‘,

[—

—

def

‘ir=0,

s

.01
.40
.05
2.40
.1
4.90
b=.05, rr=.02,

n=.04

The assumptions made with respect to ~i, b, rr, and n are listed below the
table. A glance at the table indicates that deficits of the order experienced
or still expected in the United States and Europe would eventually produce, even in the context of a stable process, a massive increase in the real
debt ratio from the current U.S. level of about .35. We should be reminded that this massive relative expansion of the government debt would
occur in the context of a “non-Ricardian” world defined by rr < n.
Important real consequences emerge under the circumstances. Real
rates of interest rise, and output is shifted from investment to consumption. Normal growth will consequently decline. The negative difference
rr
n may thus disappear, and the economy may move into an unstable
debt accumulation process. The smaller (absolutely) the initial negative
difference rr
n and the larger the permanent deficit with the implicit
—

—

94

Karl Brunner

equilibrium real debt ratio, the greater is the likelihood of a change in the
sign of the crucial inequality.
The shift from rr
n < 0 to rr
n > 0 does not mean that the system necessarily assumes a “Ricardian property.” The change in inequality is certainly consistent with such a property. It need not occur, however, because the inequality forms just a necessary condition for the Ricardian equivalence to occur. Persistence of a “non-Ricardian” world under
the condition rr > n aggravates, of course, the real consequences outlined
above. These consequences especially induce the persistent rise of interest
rates and the real debt burden. They eventually determine responses in
the political arena that initiate an inflationary regime.
We should note that this argument is consistent with the monetarist and
Keynesian analysis of the financial consequences associated with a budget
deficit discussed in Section 2.3. The “monetarist” framework used 12
years ago for this purpose can easily be applied to the current problem.
This common argument contrasts, however, with the argument attributed
above to Sargent and Wallace in order to obtain the same result. The
infinite budget constraint is made to impose the eventual inflationary solution. The Sargent—Wallace analysis suffers, however, from a troublesome
indefiniteness associated with the arguments centered on infinite intertemporal budget constraints. The latter only conveys to agents that taxes must
be adjusted some time at an indeterminate future in order to obey the constraint. The economic or social mechanism eventually producing the shift
from debt finance to inflationary remains, moreover, quite obscure. A
“Ricardian world” does not seem to provide such a mechanism. Sargent
and Wallace suggest in passing the operation of limits of demand for
securities. But such “limits” would be reflected in rising real interest rates
and thus violate the “Ricardian pattern.” The implicit indefiniteness in
the world described by Sargent and Wallace means that agents are confronted with a substantial uncertainty with respect to timing, magnitude,
and variance of the inflation tax.
This uncertainty is not recognized by the “Ricardian model.” The
resulting purchasing power risk associated with nominal bonds produces a
risk premium represented by a covariance term (Baghat and Wakeman
1983) in the standard capital-asset pricing model. This risk premium is
added to the basic (risk-free) real rate of interest to form the effective real
rate. Considerations of uncertainty operating in the context of an infinite
intertemporal budget constraint thus move the analysis beyond a “Ricardian world.” The uncertain temporal reallocation of taxes produces real
consequences affecting agents’ real opportunities.
A cautionary note must be added. Once we abandon “nonlumping”
taxes, the relevant real rate rr should be net of taxes on interest payments
or receipts. Second, the relation between rr and n needs to be more care—

—

Fiscal Policy in Macro Theory

95

fully rephrased in order to avoid a relation between a risky return
expressed by n—most especially once we acknowledge the relevance of a
stochastic trend (Nelson and Plosser 1982; Stulz and Wasserfallen 1985),
and a possibly riskless return to represent rr.
It was argued above that the stability of the debt process—that is, the
sign of the difference rr
n—does not address our crucial problem raised
by Sargent and Wallace. The crucial problem is the long-run danger of
permanent inflation at a potentially massive scale.
The next table covering both states summarizes the long-run inflation
threat associated with persistent deficits. The inflation rate is computed
under the condition that the real debt
—

s=.33,b=.05
.06=rr>n=.03

andS=.5
.02=rr<n=.03

def

IT

def

iT

.01
.05
.10

39%
139%
239%

.01
.05
.10

10%
90%
190%

.06=rr>n=.03

.02=rr<n=.03

def

def

iT

.01
.05
.10

7%
87%
187%

.01
.05
.10

50%
150%
250%

ratio is held constant either at .33 or .5. We notice that, irrespective of the
stability conditions rr
n, the long-run inflation threat embedded in a
permanent basic deficit of 5 percent of gross national product would move
us to levels not yet experienced as a maintained phenomenon in the
United States or Europe.
—

4.2. The endogenous state of the monetary regime:
a choice-theoretic analysis
The absence of any motivating force in the account presented by Sargent
and Wallace explaining the eventual change in monetary regime directs
our attention to McCallum’s (1984) paper. The author adapts the model
introduced by Sidrauski for his purpose. A representative agent maximizes
the present value of current and future (instantaneous) utility over an
infinite horizon. Instantaneous utility depends on real consumption and
real money balances. Bonds, issued by the government, convey no direct
utility to the agent. They do occur, however, in the agents’ budget constraint. The agent is, moreover, visualized as a joint consuming-producing

96

Karl Brunner

unit. A simple production technology is thus incorporated in the budget
constraint. Maximization proceeds subject to the (infinite) set of these
constraints. The first-order conditions. yield restrictions on consumption c,
real balances m, the stock of real capital k, the Lagrangian variable, and
the usual inequality constraint bearing on the real rate of interest and the
agent’s bond holdings b. These conditions are supplemented with the
government sector’s budget constraint, which provides for finance of the
deficit by means of money creation and bond issues. Consolidation of
household and government constraint yields an income—expenditure statement in real terms.
A steady-state solution for all the relevant variables is easily satisfied
under mixed bond—money financing. The steady-state condition requires
that both money and bonds rise by equal percentages in this case. The
deficit is, moreover, defined in the sense of the basic deficit exclusive of
interest payments. Once the real deficit is fixed the associated percentage
rise in nominal money balances determines the inflation rate. The
optimality condition for real capital k yields the optimal stock by equating
its marginal product with the exogenous utility rate of time preference.
Insertion of the latter result into the consolidated constraint yields the
optimal rate of consumption. The remaining first-order conditions determine real balances, Lagrangian multiplier and equality between the real
rate of interest and the utility rate of time preference. The latter follows
from the occurrence of a positive bond stock. Lastly, the magnitude of
this stock is settled by the real version of the government’s budget
constraint.
The solution of the mixed finance carries through without a hitch and
also satisfies the transversatility conditions. Similarly, financing the deficit
with money only poses no problem. The argument proceeds as above.
This case implies, however, that the real rate of interest remains below the
utiiity rate of time preference. A problem arises, however, when the
deficit, as defined, is only financed by bond issues. The transversatile conditions are violated in this case. There thus exists no steady-state solution
for bond-financed permanent deficit. We also note that the real version
of the government budget constraint would imply a negative steady-state
stock of bonds that cannot be reconciled with the model.
McCallum demonstrates that the problem is quite sensitive to the
specification of the deficit to be held constant. The pattern changes once
we move beyond the basic deficit and includes the interest service on the
outstanding debt. The transversatility condition is not violated for the
stock of bonds once this extended deficit measure is held constant. This
reflects the fact that the growth of real debt per capita does not explode as
in the case of a constant “basic deficit” but actually converges to zero.
The difference in this case is that the persistent increase in interest pay-

Fiscal Policy in Macro Theory

97

ments due to the rise in the stock of bonds is matched by corresponding
increases in taxes. Last, McCallum establishes that a permanently
noninflationary bond-financed deficit can occur, provided the growth rate of
bonds is less than the utility rate of time preference. This implies that the
basic deficit is negative and converges with time toward zero. The
extended deficit remains, however, positive and converges to equality with
the government’s interest payments.
McCallum’s analysis qualifies somewhat the central proposition
advanced by Sargent and Wallace. A permanent deficit need not be
inflationary and need not impose a change in the monetary regime. It
depends partly on the nature of the policy rule specifying the permanent
deficit. But McCallum’s analysis still leaves unanswered a central issue in
the Sargent—Wallace argument. We learn that a noninflationary bondfinanced basic deficit is impossible. It is impossible because the steadystate system cannot produce that result. This answer does not attend to
the question of why, in reality, initial attempts at a noninflationary finance
of a basic deficit are eventually doomed. A similar problem involves the
other result, suspending the Sargent—Wallace proposition. The analysis
describes a world of certainty, perfect foresight, and nondistortionary
taxes. The real variables are not affected by noninflationary finance
under the circumstances. It was argued in an earlier section that the prevalence of distortionary taxes with uncertainty about the incidence of
future taxes very likely modifies substantially this picture of a bondfinanced permanent deficit. This case is reenforced by the remaining possibility that we live in a “non-Ricardian” world.
4.3. The empirical relevance of the issue
We still need to consider the relevance of the potential threat posed by a
permanent large deficit to an anti-inflationary regime. The analysis offers
by itself no evidence that an anti-inflationary policy executed in the
present against a background of a permanent deficit policy cannot be
maintained over the long run. A number of papers over the past years
examined this issue. They investigated, in particular, whether persistent
deficits eventually induced inflationary policies. Robert King and Charles
Plosser (1985) offer for our purposes an excellent example of this literature. They first introduce some historical background and discuss the
behavior characteristics of six different seignorage measures for the
United States. They find that seignorage averaged over the period considered about three-tenths of 1 percent of gross national product. The
various seignorage measures are also correlated for descriptive purposes
with a range of important macro variables. The authors note here a

98

Karl Brunner

correlation of .44 between the deficit and inflation over the postwar
period.
However, it is worthwhile to note that the correlation is essentially zero (.02)
for the 1929—1952 period. By way of contrast, there appears to be a positive
correlation between money creation and both real and nominal deficits in
the 1929—1952 period.
There is, in general, no evidence for a contemporaneous relation between
deficit and money creation in the United States. The underlying analysis
does not stress, however, any particular contemporaneous relation but
essentially an intertemporal dynamic relation. “The empirical strategy,
therefore, is to look for a dynamic relation between revenue from money
creation and deficits.” The crucial question is whether past deficits
explain subsequent revenues from money creation. The statistical results
show “that seignorage does not appear to make a significant contribution
to predicting any of the other government policy variables.” Even more
noteworthy is that deficits do not help predict seignorage.
The study extends attention beyond the United States. Among the
eight nations outside Latin America only Italy exhibits some significant
contemporaneous correlation between national income account deficits
and money creation. The four Latin American nations included show, in
contrast, some significant contemporaneous correlation between these
magnitudes. An investigation of dynamic intertemporal interrelation
confirms, however, the results obtained for the United States. Switzerland
forms an exception. This is somewhat surprising and makes, against the
background of Swiss financial institutions, little systematic sense.
The study demonstrates that we so far possess little systematic evidence
about the dynamic link between current fiscal policy and future monetary
policy. The problem may have barely confronted Western nations over
most of the postwar period. But the experience of Italy, Israel (Fischer
1984), and Argentina (Dutton 1971), and more casual observations from
other nations does suggest that we should not rely on the persistence of
the monetary regime observed over the sample period in a radically
different fiscal context. The dynamic link emphasized by Sargent and
Wallace hardly emerges in contexts of the modest deficits dominantly prevailing in the United States and some other nations until recently. The
issue has thus just been opened, and King and Plosser offer a useful starting point for future explorations of the subject. But such explorations
need to attend to a subtle but basic analytic issue. We encounter here a
generalization of the problem faced in the old fiscalist debate. It is anything but clear how the relevant fiscal variables need be specified. In particular, it is not obvious that the official administrative deficit, the national

Fiscal Policy in Macro Theory

99

income account deficit, or some real versions of these measures form the
relevant magnitude for the analysis of our problem. Kotlikoff (1984)
raised some searching questions about the standard measures of deficits.
These questions extend to the analysis of any possible relation between
budgetary operations and subsequent monetary regimes.
4.4. The deficit and economic activity
Traditional Keynesian and neo-Keynesian analysis suffered no doubt that
lower income taxes or higher government expenditures raise economic
activity. The presumed permanent inefficiency of the economy could be
expected to offer a leverage for fiscal policy to influence output and
employment. Recent experience with large deficits expected to persist
into an indefinite future generated in the public arena increasing doubts
about their beneficial effects. Analysis based on “non-Ricardian” assumptions emphasized, of course, for a long time the negative long-term effect
of permanent deficits. These effects operate via the asset market on capital accumulation and normal output. Public doubts concentrate, however,
for a variety of reasons on the short run effects associated with deficits.
Voices emerged that argued that deficits actually exert a contractionary
effect and lower output in the short run independent of further effects on
normal output growth. Academia could hardly stay behind the new
developments. The issue certainly deserves examination. Two papers are
selected in order to probe the argument of “contractionary deficits” or
“perverse fiscal policy.”
Feldstein (1984a) recently explored the possibility of a “contractionary
deficit.” His analysis revives an argument widely used by Keynesians in
the 1950s to explain in real terms a persistent inflationary drift. Prices are
supposed to be asymmetrically responsive to positive and negative
changes in demand. Allocative demand shocks are converted under the
circumstances into a persistent inflationary drift. Sectors experiencing rising real demands raise prices by a substantial margin, whereas sectors
exposed to shrinking real demand lower their price, at most, quite modestly. This idea is exploited by placing fiscal policy within an environment of specific interacting sectors. Feldstein initially presents a simple
argument along the following line. Suppose income taxes are reduced.
Consumption demand rises and investment falls. The latter follows from
higher interest rates. Consumption goods prices rise, whereas investment
goods prices remain unchanged. The increase in the price level against a
constant velocity and money stock necessarily lowers the level of output.
Lower taxes thus lower output.
This suggestive argument motivates a more general and explicit analysis
based on the following summary structure:

100

Karl Brunner
X

=

Ch + Cg + 4 +

‘g’

IT~(C)’C+ H~(I~.I
= v(r + ‘ir).M,
11
h+’g~

CCh+Cg,

(34)
(35)
(36)

where X = real income, C = real total consumption composed of household consumption Ch and government consumption Cg, I = real total
investment consisting of household investment 4 and government investment ‘g’ r = real rate of interest, H = expected rate of inflation, M =
money stock, ll~(C)= price level of consumption goods as a function of
C, [Ij(I) = price level of investment goods as a function of I, and v =
monetary velocity.
All variables in Eq. (34) occur in real terms. The simple addition of the
components C and I in order to sum up to real income will not hold, in
general, for a multisector specification with variable relative prices. Feldstein therefore postulates units for C and I that uniformly assure unit
prices in the initial position.
Expression (35) states a quantity equation. The left side represents
nominal income and the right side nominal demand. The structure is
completed by introducing a consumption function and an investment
function:
Ch

=

~

(X

—

T),

‘h =

~(r).

(37)

T refers to income taxes.
Expressions (34), (35), and (37) can be represented in a simple diagram.
Line 1 in Figure 3.5. describes Eq. (34), and line 2 Eq. (35). The signs of
the slopes are immediately established by inspection, and so are the directions of the shift induced by a reduction in taxes T. Both lines are shifted
up, increasing the real rate of interest. The effect on real income remains
ambiguous without additional constraints. Feldstein specifies three
conditions sufficient to produce a shift of the intersection point to the
northwest, as indicated in the graph. These conditions are (i) a lowinterest elasticity of monetary velocity, (ii) a large-interest elasticity of
investment, (iii) a high elasticity of H with respect to C.
The formal exercise certainly produces the desired perverse fiscal policy.
Its relevance remains, however, quite doubtful. The analysis appears
strangely regressive at this time. Important price-theoretical contributions
are disregarded, and we revert to a peculiarly ancient Keynesian world.
Analytically more serious is a basic logical flaw. Movements along line 1
and of shifts in its position hold prices constant at unity by construction.
Movements along line 2 and shifts in its position modify prices H~and
H~. Any shift of line L along 2 thus modifies prices under the assump-

Fiscal Policy in Macro Theory

101

line 2

x

Figure 3.5
tions of unchanged prices. The same contradiction appears whenever line
2 is shifted along line 1. A change in T (or Cg, Ig) is thus analyzed within
a framework exhibiting a flawed structure. This contradiction would seem
to remove all significance from this attempt to establish a presumption for
a “contractionary deficit.”
Consider now an alternative procedure to move into a new world of
fiscal policy analysis that reverses a long-accepted result of Keynesian
analysis. Mankiw and Summers (1984) explored this possibility in the
context of a slightly modified, but otherwise thoroughly standard, IS /LM
analysis.
Y=C(Y—T,r)+I(Y,r)+G,
M=L(C,I,G,r).

(38)
(39)

The letters designating the relevant magnitudes assume their usual meaning. The modification applies to the money demand in Eq. (39). Money
demand depends separately on the three expenditure components. Mankiw and Summers essentially concentrate on consumption C as the

102

Karl Brunner

relevant scale variable. The multiplier of income Y with respect to T is
immediately determined as
Cy[(Lc

where z~= —[1
—

—

—

Li)Ir

—

(40)

Lr]

C~ — Iy]’[Lj’Ir + L~’Cr + Lr]

[Ir + Cr]4Lc~Cy+ L .Iy]> 0.

1

The detailed structure of the positive denominator ~ need not concern us
here. It follows that the standard result holds according to the necessary
and sufficient condition
-~zZ0

Lr

iff Lc<Lj+-1-.

(41)

A necessary and sufficient condition for a “perverse response” can now be
given as

3T

> 0

iff c(L, C)>

c(L, I) + c(L, r)
c(I, r)

C
I

where c(y, x) denotes the elasticity of y with respect to x. The major portion of the paper offers a valiant attempt to present a case for the crucial
inequality yielding a positive tax multiplier. The discussion assumes that
L = Lg = 0.
1
The authors first assess with some rough calculations the crucial inequality. They assemble for this purpose available estimates and income
account data. The results confirm the condition. Similar calculations suggest, moreover, that the expenditure multiplier is little affected by the
change in specification. With a value less than one it is hardly a “multiplier,” however. The tax multiplier, in contrast, changes sign.
A survey of the literature indicates to the author the superiority of permanent income or wealth as a scale variable. The permanent-income
hypothesis of consumption suggests, under the circumstances, that consumption would offer a good proxy for the scale variable. The choice of
consumption is further supported in the eyes of the authors by the ownership distribution of money. Measures comparing the variability of velocity derived from a variety of scale variables, and the results obtained
from money demand regressions appear to offer more significant evidence
in this context. The first difference of a consumption velocity based on
Ml exhibits the lowest variability. Similarly, the regressions assign, in

Fiscal Policy in Macro Theory

103

general, a dominant weight to consumption compared to national income,
disposable income and final sales, or total private spending as a scale variable.
The reservations about the traditional Keynesian tax policy are
developed on the basis of a framework very close to those underlying the
original fiscalist debate and judged to be unsatisfactory at the time by
some Keynesian participants. Mankiw and Summers nevertheless raised a
relevant point that may be robust beyond the simple IS/LM formulation
used. An examination of the Brunner—Meltzer (1976) model reveal that
the impact of a tax cut on income from human wealth is substantially
attenuated by wealth and substitution effects on asset markets. A tax cut
on income from nonhuman wealth produces a stronger response. The two
tax cuts induce, moreover, opposite shifts between consumption and
investment. These distinctions are glossed over by Mankiw and Summers
and so are the substitution responses conditioned by distortionary taxes.
Their explorations could be interpreted to mean that the short-run stabilization function of tax policies should be recognized as a questionable
exercise. Comparatively much more important are long-run aspects of tax
policies shaping the use and development of an economy’s resources.
4.5.

The behavior of deficits

With the beginning of this decade, fiscal policy and deficits became a
major theme attracting public attention. Wall Street and the media
assigned increasing importance to the deficit. There emerged a widening
conviction that we entered a new era characterized by large permanent
deficits. This conviction essentially suggests that the nature of the process
determining the behavior of deficits in the United States substantially
changed by the end of the l970s. Such a change would reflect the emergence of a new pattern in the political process controlling the deficit. The
occurrence of a structural break in policy regime cannot be excluded a
priori. But our impressions of the past four years offer no relevant evidence in support of the thesis. In order to judge our situation we need a
theory about the process controlling the deficit. Guided by such a theory,
suitable tests yield some evidence on the thesis asserting the recent emergence of a new fiscal regime in the United States.
For most of the postwar period there was little reason to address this
issue. Macro analysis thus neglected to develop a theory about the
behavior of deficits. Barro’s (198la,b; 1984a,b) prolific contributions
attend in recent years also to this dimension of fiscal policy. The basic
idea guiding the research emphasizes that tax rates are essentially adjusted
to perceived permanent government expenditures. Their response to perceived temporary changes in government expenditures remains, under the

104

Karl Brunner

circumstances, quite modest. Tax revenues thus behave much more
smoothly over time than actual government expenditures. A first study by
Barro derives the tax-smoothing behavior from the government’s optimizing behavior addressed to the minimization of the excess burden imposed
by taxes on the economy. Bond-financed deficits associated with temporary bulges in government expenditures are thus designed to minimize
the social cost of taxation. The specific simple theory elaborated in
Barro’s first study (198la) actually implies that optimal tax policy sets a
constant tax rate over time. This view was recently contested by Feldstein
(1984b). This argument emphasizes that we need to compare the social
cost of changes in tax rates adjusted to finance transitory bulges in
government expenditures with the social costs of smaller but permanent
tax rate increases needed to finance the additional interest payments
resulting from the prior bond financing. An incorporation of plausible
parameter values into this analysis suggests that tax financing of temporary government expenditures is superior to debt financing, provided
the economy’s stock of real capital is smaller than its optimal magnitude.
Feldstein emphasizes, moreover, that in the case of permanent expenditure there exists essentially no financial choice. One way or the other
these expenditures are covered by taxes. This analysis seems to imply that
the optimal stock of public debt is zero.
Feldstein’s analysis obviously cannot provide the starting point for an
empirical analysis of deficits and public debt even if we accept its
relevance as a benchmark for a welfare analysis of budgetary operations.
Any positive interpretation of the essentially normative analysis is, moreover, immediately disconfirmed by the facts. Lastly; an evaluation of the
normative analysis as a benchmark for welfare statements lies beyond the
range addressed in this paper. But its normative relevance does not preclude that intertemporal tax smoothing may offer a good basis for an
analysis of deficit behavior. This smoothing pattern may be explained
very differently from Barro’s initial formulation that was questioned by
Feldstein. It thus seems quite appropriate that Barro simply postulates a
tax-smoothing pattern in his most recent examination. This hypothesis is
actually formulated in its most rigid form as a constant tax rate. The
basic idea is, however, consistent with some contingent flexibility of tax
rates over time.
The hypothesis to be tested is constituted by two building blocks. The
first introduces the tax-smoothing hypothesis, and the second combines it
with the specification of deficit in terms of the budget constraint. The
tax-smoothing hypothesis is applied to an intertemporal budget constraint
in order to determine the expected constant tax rate. This constraint is

jT0

00

J~ ‘r(t)y(t)e
00

g(t)e

“

dt + b(0)

=

—rt

dt,

(42)

Fiscal Policy in Macro Theory

105

where g(t) designates real government expenditures, b the inherited stock
of real debt, ‘r the tax rate, and y(t) the portion of income forming the tax
base. Equation (42) immediately implies the condition for a constant tax
rate:

T
fg(t)e~’
C 5 y(t)e_rt
dt +dtb(0)
(43)
0
This condition can be transformed
to yield a statement relating the tax
rate with permanent government expenditures g and permanent income
y* These permanent levels perceived at the intitial time t = 0 involve
two assumptions. The first specifies a trend growth rate n for g andy such
that the present value of g (or y) along the trend equals the present value
of the actual future course of g (or y). We thus obtain
—

*

J~g(0).e
00

5~g(t)e
00

(rn)

dt

=

—rt

dt

(44)

and a similar expression for y. Secondly, the permanent levels g (and y
are then defined by the condition
*

g *(Ø)

=

(r

—

n)

5~g(t)e

y (0)

=

(r

—

n)

f0 y(t)e_r~dt.

-r~ dt,

(45)

This condition establishes a relation between the present value of g(t) (or
y(t)) and its permanent level of g. A suitable replacement in expression
(42) yields the desired relation between the average tax rate
manent m~gnitudes
=

g* + (r

—

n)b(0)

i

and the per-

(46)

y
The second building block invokes the budget constraint expressed in
real terms
~=g+rb—ry,

(47)

where b represents the stock of real debt. Replacing T in (47) with the aid
of (46) and rearranging terms yields the basic relation

~=

dt

l_2~
y

~*+(r_n)b]+~_g*)+nb.

(48)

106

Karl Brunner

The first term describes a cyclical effect. The deficit increases during
recession and recedes during upswings. The effect of the business cycle
on the deficit increases, moreover, with g ~. The second term—that is,
(r
n)b—within the bracket is negligible and can be disregarded. The
second term indicates that temporary government expenditures affect the
deficit one for one. Lastly, the third term reflects the financing of interest
payments on outstanding debt made possible by the “normal growth” in
debt corresponding to the trend growth of the economy. This item unavoidably occurs once we assume a constant tax rate. Otherwise the relative interest burden would fall and tax rates could not be constant.
Equation (48) describes the growth of real debt as of any path of the
price level. It follows that multiplying both sides of Eq. (48) with the
price variable does not adequately render the growth of nominally valued
debt B. The expression irB multiplying the inflation rate with the stock B
must be added. We thus obtain
—

l—~

~=(n+IT)B+
dt

Pg*+P(g_g*).

y

(49)

One more term must be added, however. The structure of this last term
depends on whether the debt variable is measured at market or at par
value. In either case the dependent variable is modified over time whenever the average market rate of interest changes.
The underlying analysis ultimately yields the following regression:
B,

—

B,

1

PtYc

=

a0

B,

1

Pt_iYt_i

+ a1ir,
Pt_1Yt_1

(50)

+ a2 YVAR + a3 GVAR + a4 RVAR + U~.
YVAR refers either to (1
y/y *) (g */y,) or to an expression based on the
unemployment rate related to the first by a modified “Okun’s law.”
Either formulation reflects the cyclic influence on the deficit. GVAR
expresses the temporary government expenditures, and RVAR the adjustment required to reflect the effect of changes in interest rates on market or
par value of debt B. The hypothesis summarized by Eq. (50) implies that
—

=

n

and

a1

=

a

2

=

a

3

=

1.

(51)

An extensive range of empirical examinations substantially confirms the
component a1 = 1 of the total hypothesis. The confirmation holds for the
whole sample 1920—82, for the sample without the war, and subsamples
1920—40 or 1950—82. The coefficient a3 on temporary government expen-

Fiscal Policy in Macro Theory

107

ditures lies between .5 and .6 for the whole sample and drops to .22 whenever the war data are deleted. The results from the subsamples confirm
this pattern. The largest estimate is .16. Two estimates out of 12 are
(nonsignificantly) negative, and ten are smaller than their respective standard error. We observe here no confirmation of the hypothesis. All estimates of a3 are very significantly below unity. The “cyclic coefficient” a2,
in contrast, is uniformly positive and highly significant for all estimation
periods and choices of YVAR. The coefficient estimate, however, occurs
systematically on the high side. It clusters around 1.5. This result is not
consistent with the strict tax smoothing hypothesis. It reveals a measure
of cyclical flexibility in tax rates. Lastly, the estimate of a0 appears to be
persistently too low as an estimate of the trend growth in real GNP.
The regression analysis so far yields no evidence that the current decade
ushered in a new regime of budgetary processes. The observations on the
deficit cast up by the four years 1980—83 are quite compatible with the
patterns observed in past years. The estimated deficit traces the actual
value quite well, and the residuals remain in the usual range. The regression even overpredicts the deficit for three out of four years. Barro’s
results also imply that realization of the CBO’s baseline projections over
the next five years would indicate a break in the structure of the process
generating deficits. We would then actually have moved into a new policy regime. An unchanged policy regime would require substantial
adjustments in expenditures and tax revenues in order to stay below the
CBO projections.
The estimates of the coefficients associated with temporary government
expenditures confront us with a serious and frequently occurring problem
associated with the evaluation of hypotheses. Approximation errors in the
measurement of specific variables may burden the statistical work with an
uncertain interpretation. It is, in the present context, quite likely that the
measurement of temporary government expenditures contains a substantial and variable error margin. But we know, of course, that measurement
errors in the regressor bias the coefficient estimates toward zero. This
problem is worsened by the fact that the sample period 1920—82 contains
a single major observation for temporary government expenditures offered
by World War II. Small values and little variation in this variable over
most of the sample period combined with a potentially significant measurement error obscure the message of the data. A discriminating evaluation is not possible under the circumstances. It is very difficult to decide
whether the results bearing on a summarized above should be inter3
preted as a rejection of the hypothesis (a3 = 1) or can be reconciled with
the hypothesis on the basis of the data problems indicated. A possible
strategy to further explore the matter involves a search for data with more
“action” and experimenting with different approaches to the measurement

108

Karl Brunner

of theoretical entities as, for instance, in this case, the g variable. Barro
(1985) examined these opportunities with another paper exploiting British
data for the period 1730- 1918. These data exhibit much more variation
in temporary government expenditures produced by military spending.
The statistical investigation deletes the cyclic component. No reliable
data seem to be available for this magnitude. This omission may partly
explain the serial correlations of the regression residual. The regression
estimated was confined to the two terms
*

B,

—

B,

1

‘)tl’t

B,
=

a0

1

+ a1g, + w,,

(52)

where g~,denotes temporary government expenditures, and w a random
term. The first term in this formulation summarizes the first two terms of
Eq. (50). The hypothesis thus implies that a~is equal to the sum of trend
growth n plus the average inflation rate over the sample period 1755—
19 18. The estimate a0 = .021 coincides remarkably well with this sum.
The estimate for a1 moves, moreover, much closer to unity (i.e., a1 =.9)
than in the case of the United States. This estimate is two standard errors
from unity. The British data thus confirm the essential idea of the underlying hypothesis, but they do suggest some contingent tax adjustments.
The analysis of deficit behavior introduced a new dimension into
macroanalysis. The public discussion erupting in recent years speculating
about structural changes in our fiscal processes reveals the relevance of
this emerging research. Barro deserves some credit for raising the issue in
advance of the public arena’s attention. The exploration of the “Ricardian theme” eventually led to an examination of deficits and public debt.
We are hardly in a position at this stage to accept Barro’s hypothesis with
any sense of conviction. The evidence is still too unclear in this respect.
But neither can we reject it. So far, it is “the only game in town.”

5. Concluding Remarks
The professional discussion of fiscal policy has moved a long way over the
past 20 years. The “great debate of the l960s” centered on the role of
fiscal policy expressed by the effect of expenditures and taxes on national
income and economic activity. This discussion did modify the earlier position of the postwar period represented by the works referred to in Section
2. There seemed also to ensue by the early 1970s a consensus that fiscal
policy did probably produce permanent nominal effects and temporary
real macro effects. There remained substantial controversy about magnitude and temporal patterns of the consequences. There also persisted a

Fiscal Policy in Macro Theory

109

basic disagreement about the role of fiscal operations in a concept of stabilization policy. An activist exploitation of fiscal instruments for purposes
of short-run stabilization was typically advocated by the Keynesians. This
contrasts with a more classical conception advocated by scholars with a
monetarist and neoclassical view. This view rejects short-run manipulations and advocates policies addressed to the establishment of an institutional framework, not to sequences of fiscal actions. Such an “institutional policy” should provide an essentially confined, and thus reliably
predictable, pattern of fiscal operations. This issue will be considered in
the last paragraphs.
The temptation to sneer at the empirical work executed at the time
seems, on occasion, irresistible. It would indeed be pointless to reproduce
such investigations today. It would also express some measure of incompetence relative to the state of current economic and statistical analysis.
But this work had a function at the time and was not irrelevant with
respect to a strong thrust embedded in the Keynesian message. It did
contribute to modifying positions even when not explicitly acknowledged.
Modigliani’s presidential address to the American Economic Association
(1977) had a different flavor than the Keynesian stories of the l950s. We
note on the other side at least some monetarists with a more explicit
recognition of fiscal variables than before. An unpalatable fact found,
however, little resonance at the time. The central (non-price-theoretic)
emphasis assigned in Keynesian analysis to income—expenditure flows
made it quite sensitive to the choice of “autonomous expenditures.” This
choice seems anything but settled. The meaning of traditional Keynesian
theory, on an empirical level beyond formal classroom exercises, remains
obscure.
More recent discussions of fiscal impact on the macro behavior of the
economy may usefully lead us beyond this impasse. The market or coordination failure conception underlying the Keynesian approach naturally
produced an analysis downplaying the role of prices with an emphasis on
interacting income—expenditure flows. The classic tradition rejects the
basic failure conception motivating Keynesian analysis. A framework
emphasizing a system of (“multiplying”) interacting flows is thus replaced
with an essentially price-theoretical conception. This conception should
not be identified, however, with a (total) market-clearing analysis. The
price-theoretical approach developed by Barro includes, moreover, the
important distinction between anticipated and unanticipated fiscal events
or, most particularly, recognizes the differential impact of temporary and
permanent fiscal actions. The multiplier effect essentially vanishes in this
price-theoretical context. The consequences of fiscal events, moreover,
depend crucially on the government sector’s use of the goods and services
acquired from the private sector. Once we abandon the sinkhole theory,

110

Karl Brunner

new channels of influence are recognized. Barro’s analysis shows, for
instance, that the impact of government expenditures on goods and services depends significantly on their “supply-side effects.” This differs radically from the “demand-side dominance” of Keynesian analysis. This
analytic evolution, initially suggested by Martin Bailey, still needs further
elaboration. The operation of the government sector involves more than
a pure redirection of privately produced goods within the private sector.
It should be recognized as a production sector absorbing inputs converted
into an output. The difference in the incentive structure between private
and public sectors implies an important difference between the production
functions of the two sectors. We also need to consider to which extent
these consequences associated with the government sector as a production
sector bear more significantly on longer-term aspects of an economy.
The dimensions covered by fiscal policy were significantly extended
with the introduction of the Ricardian theme. The “Ricardian
equivalence theorem” defines a useful benchmark for our professional discussions. It obliges the advocates of a more or less “conventional thesis”
to specify the conditions responsible for the real consequences of the
government’s decisions how to finance expenditures. The possible recognition of a pure debt effect on real variables, explicitly acknowledged by
the “stability analysis” of the early 1970s, would hardly suffice today. The
discussion unleashed over the years directed our attention to new mechanisms or channels of influence conveying real impulses from the
government’s financial decisions. The association of deficit finance with
intergenerational transfers, the role of uncertainty and risk related with
future tax liabilities, and the condition of intergenerational transfers
linked with the function of bequests deserve critical exploration by the
profession. We may yet convince ourselves that the government’s “financial mix” does exert some real consequences but, to some extent, for reasons beyond the pure debt effect. We also note that the extension of fiscal
analysis summarized above increasingly directs our attention to a range of
influences modifying the division of total output between real consumption, household real capital, and real capital used in the production process. These channels condition normal output. Their operation may be
more important than the immediate impact on total output.
The emergence of an apparently persistent deficit was bound to attract
the profession’s interest. Attempts to justify some public concern that
“deficits are contractionary” are so far at best speculatively “interesting.”
More significant seems to be the attempt to explain the observable
behavior of deficits. A useful explanation would offer a criterion for judging the occurrence of changes in fiscal regime. The issue raised by Sargent and Wallace also deserves serious further attention. It involves a
basic question about the political role of monetary and fiscal institutions

Fiscal Policy in Macro Theory

111

and the relevant deficit measure guiding policy. The problem arises
whenever fiscal institutions dominate the monetary institutions over the
longer horizon. Political economy analysis seems to support this assumption. It follows under the circumstances that a breakdown of the “Victorian rule” (Buchanan 1985) ultimately determines an inflationary adjustment of monetary policy. The nature of the monetary regime ultimately
depends on the prevailing fiscal regime. A noninflationary monetary
regime thus requires for its survival a fiscal regime approximating a ‘Victorian rule.” But the analysis of this issue remained somewhat incomplete. The economic and political mechanism creating the accommodation
of monetary policy to a permanent deficit still requires some attention.
The empirical relevance is also unresolved at this time. There so far exists
little supportive evidence, but then we may have only entered the age of
permanent deficits.
Finally, a basic issue of political economy should be faced. The contrasting conceptions of fiscal policy offered by the Keynesian vision and
the classical tradition were characterized in a previous paragraph. The
issue cannot be left unattended with the easy escape into “ideology.”
There is more involved that deserves the careful attention of the political
economist.
Three substantive issues condition the policy conception: the basic
coordination failure of a market economy, the information problem confronting policymakers, and the characteristic operation of political institutions. Some Keynesians emphasize that the basic coordination failure of
market economies necessitates the intervention of the government in order
to offset this failure. Such intervention exploits to a large extent the
powers of fiscal policy. The coordination failure to be corrected with the
aid of fiscal policy involves, moreover, both a short-run and a long-run
dimension. The latter dominated the attention at the end of World War
II with the projections of oversaving and secular stagnation. This “structural coordination failure” justifies a permanent large deficit to offset
private oversaving. This portion of the Keynesian argument offers no
basis for activist manipulation of fiscal policy or for a large government
sector. Activist manipulation follows from an emphasis of a “dynamic
coordinatiOn failure” that produces inefficient fluctuations in output and
employment.
One strand of the issue thus depends on the substantive question of a
long-run (structural) coordination problem. The view advanced by
Keynesians in this respect forty years ago was thoroughly disconfirmed by
the end of the l950s. The issue remains, however, as we still encounter
assertions that an economy may be trapped within a set of multiple
“underemployment equilibria.” The other strand, represented by the
“dynamic coordination failure,” constituting the case for an activist fiscal

112

Karl Brunner

policy, involves three distinct substantive issues. We note first the idea
that fluctuations in output and employment are inherently inefficient. The
reviving interest in “real business cycle theories” warns us, however, that
economic fluctuations are not necessarily inefficient. The analysis
developed by Stulz and Wasserfallen (1985) demonstrates, moreover, that
economic fluctuations may reflect the characteristics of financial regimes.
But this cautionary note is really just a special case of the general information problem faced by policymakers. The activist argument implicitly
assumes that policymakers do possess reliable and detailed knowledge
about the dynamic properties of the economy. Such knowledge would
certainly allow the pursuit of an effective fiscal intervention. But such
knowledge, while necessary, is not a sufficient condition for socially successful fiscal activism. We still need to invoke a goodwill or publicinterest theory or benevolent dictator view of government. The case for
fiscal activism, at least for purposes of stabilization policy, thus involves
two important empirical assumptions bearing on required information
and the behavior of man in political contexts. The case for an “institutional policy” rests, in contrast, on the empirical proposition that the two
crucial conditions postulated by advocates of activism do not hold in reality. We lack the needed detailed and reliable knowledge about the
economy’s dynamic structure. The range of analytic results and empirical
positions covered in the survey demonstrates this state most explicitly.
The consequences of this information problem are reenforced by the fact
that self-interested behavior also permeates the political environment.
There is little evidence that political agencies operate according to a generally recognized social welfare function. Fiscal activism produces, under
the circumstances, more problems. We have no assurance that it will not
generate truly inefficient fluctuations. These issues associated with the
political economy of fiscal policy are wide open and far from settled at
this stage. We may yet achieve some cognitive progress in this field once
we recognize the substantive nature of the problems behind the ideological smoke.

References
Ando, A. and F. Modigliani. 1965. The relative stability of monetary velocity and
the investment multiplier. American Economic Review 55: 696—728.
1976. Impacts of fiscal actions on aggregate income and the monetarist
controversy: Theory and evidence. In Monetarism, ed. J.L. Stein. Amsterdam:
North-Holland.
Andersen, L.C. and J.L. Jordan. 1968. Monetary and fiscal actions: A test of their
relative importance in economic stabilization. Federal Reserve Bank of St. Louis
Review 50: 11—24.
Aschauer, D. 1985. Fiscal Policy and Aggregate Demand. American Economic
Review 75: 117—27.

______

Fiscal Policy in Macro Theory

113

Aschauer, D. and J. Greenwood. 1985. Macroeconomic effects of fiscal policy. In
Carnegie—Rochester Conference Series on Public Policy, Vol. 23. Forthcoming.
Baghat, S. and L.M. Wakeman. 1983. Non-diversifiable inflation risk and expected
treasury bill returns. Working paper, University of Rochester, Rochester, New
York.
Bailey, M.J. 1971. National Income and the Price Level. New York: McGraw-Hill.
Baltensperger, Ernst. 1984. The public debt: Limits and effects. Working paper,
University of Bern, Bern, Switzerland.
Barro, R.J. (1978). The impact of social security on private saving: Evidence from
the U.S. time series. American Enterprise Institute, pp. 1—36.
1981a. Money, Expectations, and Business Cycles: Essays in Macroeconomics. New York: Academic Press.
______ 198 lb. Output effects of government purchases. Journal of Political Econ
omy89: 1086—1121.
1984a. Macroeconomics. New York: Wiley.
1984b. U.S. deficits since World War I. Working paper, University of
Rochester, Rochester, New York.
1985. Government spending, interest rates, prices and budget deficits in the
U.K. 1730—1918. Working Paper No. 1, Rochester Center for Economic
Research, University of Rochester.
Barsky, RB., N.G. Mankiw, and S.P. Zeldes. 1984. Ricardian consumers with
Keynesian propensities. NBER Working Paper Series, No. 1400.
Bernheim, E.D., A. Schleifer, and L.H. Summers. 1985. Bequests as a means of
payment. Journal of Political Economy. Forthcoming.
Blanchard, Olivier. 1984. Debt, deficits and finite horizons. NBER Working Paper
Series.
Blinder, A.S. and R.M. Solow. 1974. The Economics of Public Finance. Washington, D.C.: Brookings Institute.
1976. Does fiscal policy still matter? A reply. Journal of Monetary Economics 2: 501—10.
Bomhoff, El. 1983. Monetary Uncertainty. Amsterdam: North-Holland.
Boskin, M.J. and L.J. Kotlikoff. 1985. Public debt and U.S. saving: A new test of
the neutrality hypothesis. Carnegie—Rochester Conference Series on Public Policy, Vol. 23. Forthcoming.
Brunner, K. 1951. Inconsistency and indeterminacy in classical economics.
Econometrica, 19: 152— 73.
1970. Em Neuformulierung der Quantitatstheorie des Geldes. Kredit und
Kapital. 1—30.
1971. Survey of selected issues in monetary theory. Schweitzerische
ZeutschrUi fur Volkswirtschafl und Statistik 107 (1).
1976. Inflation, money and the role of fiscal arrangements: An analytical
framework for the inflation problem. In The New Inflation and Monetary Policy,
ed. M. Monti. New York: MacMillan.
Brunner, K. and A.H. Meltzer. 1972a. Money, debt and economic activity: An
alternative approach. Journal of Political Economy 80: 951—77.
l972b. A monetarist framework for aggregate analysis. In Proceedings of
the First Konstanzer Seminar on Monetary Theory and Monetary Policy, ed. K.
Brunner. Berlin: Drunker and Humblot.
1976. An aggregative theory for a closed economy, and Reply. Monetarism: The principal issues: Areas of argument and the work remaining. In Monetarism, ed. J.L. Stein. Amsterdam: North-Holland.
______

-

______
______

______

______

______

______

______

______

______

114

Karl Brunner

Buchanan, J. 1985. Budgetary bias and post-Keynesian politics. In Growth of Government, eds. A. Lindbeck and J. Myrman.
Buiter, W. and J. Tobin. 1978. Debt Neutrality: A Brief Review of Doctrine and
Evidence. Manuscript.
Carlson, K. 1967. The federal budget and economic stabilization. Federal Reserve
Bank of St. Louis Review 49: 5—12.
Chan, L.K.C. 1983. Uncertainty and the neutrality of government financing decisions. Journal of Monetary Economics 11: 351—72.
Christ, CF. 1968. A simple macroeconomic model with a budget constraint. Journal of Political Economy 76: 53—67.
1979. On fiscal and monetary policies and the government budget constraint. American Economic Review 69: 526—38.
Darby, M.R. 1984. Some pleasant monetarist arithmetic. Federal Reserve Bank of
Minneapolis Quarterly Review 8: 15—20.
DeLeeuw, F. and J. Kalchbrenner. 1969. Monetary and fiscal actions: A test of
their relative importance in economic stabilization—A comment, Federal Reserve
Bank of St. Louis Review 51: 6—11.
DePrano, M. and T. Mayer. 1965. Tests of the relative importance of autonomous
expenditures and money. American Economic Review 55: 729—52.
Dutton, D. S. 1971. A model of self-generating inflation: The Argentine case.
Journal of Money, Credit, and Banking 3: 245—62.
Evans, P. 1984. The effects on output of money growth and interest rate volatility
in the United States. Journal of Political Economy 92: 204—22.
Feldstein, M. 1978. The impact of social security on private saving: Evidence from
the U.S. time series: A Reply. American Enterprise Institute, 37—47.
1982. Government deficits and aggregate demand. Journal of Monetary
Economics 9: 1—20.
1984a. Can an increased budget deficit be contractionary? NBER Working
Paper Series.
1984b. Debt and taxes in the theory of public finance. NBER Working
Paper Series.
Fischer, 5. 1984. The economy of Israel. In Monetary and Fiscal Policies and Their
Applications. In Carnegie—Rochester Conference Series on Public Policy, Vol. 20,
7—52.
Fischer, S. and R.C. Merton. 1984. Macroeconomics and finance: The role of the
stock market. In Carnegie—Rochester Conference Series on Public Policy, Vol.
_____

_____

_____

______

21, 57—108.

Friedman, M. 1952. Price, income, and monetary changes in three wartime
periods. American Economic Review 42: 612—25.
Friedman, M. and D. Meiselman. 1963. The relative stability of monetary velocity
and the investment multiplier in the United States, 1897—1958. In Stabilization
Policies. Englewood Cliffs, N.J.: Prentice-Hall.
1965. Reply to Ando and Modigliani and to DePrano and Mayer. American Economic Review 55: 753—58.
Friedman, M. and A. Schwartz. 1963. A Monetary History of the United States
1867—1960. Princeton: Princeton University Press.
Goldfeld, SM. and AS. Blinder. 1972. Some implications of endogenous stabilization policy. Brookings Papers on Economic Activity 3: 585—644.
Grenander, V. 1954. On the estimation of regression coefficients in the case of an
autocorrelated disturbance. Annals of Mathematical Statistics 25: 252—72.
Hall, RE. 1978. Stochastic implications of the life cycle—permanent income
hypothesis: Theory and evidence. Journal of Political Economy 86: 97 1—87.
_____

Fiscal Policy in Macro Theory

115

1981. Intertemporal substitution in consumption. NBER Working Paper
No. 720.
Hester, D.E. 1964. Keynes and the quantity theory: A comment on the Friedman—Meiselman CMC paper. Review of Economics and Statistics 45: 364—68.
Infante, E.F. and J.L. Stein. 1976. Does fiscal policy matter? Journal of Monetary
Economics 2: 473—500.
Johnson, H.G. 1971. The Keynesian revolution and the monetary counterrevolution. American Economic Review 61: 1—14.
Keran, M. 1976. Monetary policy, balance of payments, and business cycles.
Federal Reserve Bank of St. Louis Review 49: 7—17.
King, R.G. and CI. Plosser. 1985. Money, deficits and inflation. In Understanding
Monetary Regimes. Carnegie—Rochester Series on Public Policy, Vol. 22.
Kochin, L. 1974. Are future taxes anticipated by customers? Journal of Money,
Credit, and Banking 6: 385—94.
Kormendi, R.C. 1983. Government debt, government spending and private sector
behavior. American Economic Review 73: 954—1010.
Kotlikoff, Li. 1984. Taxation and savings: A neo-classical perspective. Journal of
Economic Literature 22: 4.
Lucas, R.E. 1984. Money in a theory of finance. In Essays on Macroeconomic
Implications of Financial and Labor Markets and Political Process. Carnegie—
Rochester Conference Series on Public Policy, Vol. 21, 9—45.
Mankiw, N.G. and L.H. Summers. 1984. Are tax cuts really expansionary? NBER
Working Paper Series.
Mascaro, A. and A.H. Meltzer. 1983. Long- and short-term interest rates in a risky
world. Journal of Monetary Economics 12: 485—5 18.
McCallum, B.T. 1981. Monetarist principals and the money stock growth rule.
_____

American Economic Review 71: 134—38.
______

1984. Are bond-financed deficits inflationary? A Ricardian analysis. Jour-

nal of Political Economy 92: 123—35.

Meese, R.A. and K.J. Singleton. 1982. On unit roots and the empirical modelling
of exchange rates. Journal of Finance 57: 1029—35.
Meltzer, A.H. 1981. Keynes’ general theory: A different perspective. Journal of
Economic Literature 19: 34—64.
Meltzer, A.H. and S.F. Richard. 1985. Why the social security system is in crisis.
Unpublished manuscript.
Miller, M.H. and C.W. Upton. 1974. Macroeconomics: A Neoclassical Introduction.
Homewood, Ill.: Richard D. Irwin.
Modigliani, Fl. 1977. The monetarist controversy or, should we foresake stabilization policies? American Economic Review 67: 1—19.

Nelson, CR. and CI. Plosser. 1982. Trends and random walks in macroeconomic
time series. Journal of Monetary Economics 10: 139—62.
Plosser, CI. 1982. Government financing decisions and asset returns. Journal of
Monetary Economics 9: 3 15—52.
______
1984. Money in a theory of finance. In Essays on Macroeconomic Implications of Financial and Labor Markets and Political Processes. Carnegie—
Rochester Conference Series on Public Policy, Vol. 21, 47—55.

Plosser, CI. and G.W. Schwert. 1978. Money, income and sunspots: Measuring
economic relationships and the effects of differencing. Journal of Monetary
Economics 4: 637—60.
Policies to Combat Depressions. 1956. A Conference of the Universities—National
Bureau Committee for Economic Research. Princeton: Princeton University
Press.

116

Karl Brunner

Sargent, T.J. and N. Wallace. 1982. Some unpleasant monetarist arithmetic.
Federal Reserve Bank of Minneapolis Quarterly Review 6: 1-17.
Seltzer, L.H. 1945. Is a rise in interest rates desirable or inevitable? American
Economic Review 35.

Silber, W.L. 1970. Fiscal policy in IS-LM analysis: A correction. Journal of
Money, Credit, and Banking 2: 46 1—72.
Stulz, R.M. and W. Wasserfallen. 1985. Macroeconomic time series, business
cycles and macroeconomic policies. In Carnegie—Rochester Conference Series on
Public Policy, Vol. 22.
Supplement to Review of Economics and Statistics. 1963. A Conference of the
Universities—National Bureau Committee for Economic Research. Conference
on Monetary Theory.
Tanner, i.E. 1970. Empirical evidence on the short-run real balance effect in
Canada. Journal of Money, Credit, and Banking 2: 473—85.
Tinbergen, J. 1952. On the Theory of Economic Policy. Amsterdam: NorthHolland.
Tobin, J. 1980. Asset Accumulation and Economic Activity: Reflections on Contemporary Macroeconomic Theory. Chicago: University of Chicago Press.
1981. The monetarist counter-revolution today—An appraisal. Economic
______

Journal9l: 29—42.

Wasserfallen, W. 1985. Makro okonomische Unterschuger mit Rationalen Erwartungen: Empirische Analysen fur die Schweitz. Habilitation thesis accepted by
University of Bern. To be published.
Weil, P. 1984. Love thy children: Reflections on the Barro debt neutrality
theorem. Harvard University Working Paper.

4
Ruminations on
Karl Brunner’s Reflections
Alan S. Blinder
PRINCETON UNIVERSITY

1. Introduction
Karl Brunner’s mammoth paper is a wide-ranging and detailed survey
and evaluation of many issues that are tied in one way or another to the
fiscal versus monetary policy debate. It is the kind of paper that
overwhelms a discussant by its size and scope. So, to keep my task
manageable, I will be selective and not try to touch every base that Karl
touches. Still, that will leave me with plenty to do.
The organizers of this conference must have known that the only way
to get me to raise my output above its natural rate was to hit me with a
series of unanticipated shocks. This they did—with Karl’s help. Originally, I was asked to be a discussant of a paper on the evolution of monetarism written by the man who invented the word. That sounded interesting. Then came the first unanticipated shock: I was told that the paper
would really be about fiscal policy. Somehow that made me anticipate a
paper full of Brunner—Meltzer type models with emphasis on asset substitutability and the financial aspects of fiscal operations. When the paper
finally arrived (just a few days ago) I received my second unanticipated
shock. I never imagined that Karl would try to resurrect the old AM/FM
debate, today, in the age of VCRs, digital recordings, and cable TV. But
he did!
Ironically, it was just this week that I lectured on the monetaristKeynesian debate in my graduate course at Princeton. In almost three
hours of lecturing, I never once mentioned Friedman and Meiselman,
Ando and Modigliani, Andersen and Jordan, Blinder and Solow, Goldfeld
and Blinder, or any of the other parties to this debate. Apparently, Karl
thinks that my graduate students were shortchanged So I d like to

118

Alan S. Blinder

redress that first. Then I will turn to issues pertaining to the government
budget constraint and to the Reagan—Barro equivalence theorem, where, I
am happy to say, our disagreements are quite minor.

2. The Intellectual Setting
But first a brief word about the intellectual setting for this debate. Like
McCallum (see Chapter 2), Karl reminisces about the bad old days in
which Neanderthal Keynesians roamed the land, spreading the false word
that money does not matter. McCallum even dates the Neanderthal
period as lasting at least until 1965.
Funny, but I don’t remember it that way at all. Maybe I’m just too
young. But I started studying economics in 1963—the year the
Friedman—Meiselman study was published—and grew up thinking that
money mattered quite a lot, even though I was exposed to one Keynesian
teacher after another and never saw a live monetarist until Leonall
Andersen gave a guest lecture at MIT in 1970. My first college textbook
was Paul Samuelson’s fifth edition published in 1961, and therefore written in 1960 or 1959, which by 1963 had been widely imitated. As I
remember, my young, impressionable mind got the strong impression that
money and monetary policy mattered quite a bit.
But Karl and Ben induced me to check my memory. So I dusted off
my old Samuelson. Let me read you a few quotations. In Chapter 15,
immediately after dismissing the crude quantity theory, Samuelson
remarked:
“Few people are still alive who subscribe to the crude quantity theory, but
we should not use its inadequacies to damn the whole idea that money can
have important effects on macroeconomic magnitudes.... The next few
chapters will show how monetary policy does have an important influence
on the total of spending.” (p. 315)
Then on the very next page, he explained that a “sophisticated quantity
theorist” does not believe that velocity is constant but claims instead that
controlling the money supply will help to control national income.
According to the Samuelson of 1961, “this is in agreement with almost
any modern theory of income determination” (p. 316). The last sentence
of the chapter entices students to read the next two chapters (on banking
and central banking) with the words: “So from every point of view, the
discussions in the ensuing chapters
are of tremendous importance.”
(p. 316)
. .

.

Ruminations

119

In Chapter 18 Samuelson began the “Synthesis of Monetary Analysis
and Income Analysis” with the words “Monetary analysis is seen to fit in
well with the modern theory of income determination; the stage is set for
stabilization policy—central bank monetary policies and government fiscal
policies and government fiscal policies.” (p. 366) (Notice who got first billing!) He then proceeded to outline the standard “Keynesian” transmission mechanism by means of interest rates and investment.
My question is a simple and rhetorical one: Can anyone reading this
book have come away with the ideas that money is unimportant and
monetary policy is impotent?

3. Simple Correlations and Reduced Forms
Up until a few years ago, I used to tell Princeton freshmen the story of
how the Neanderthal Keynesians, with their stone-age view that money
doesn’t matter, were vanquished by the Cro-Magnon monetarists, with
their equally silly view that fiscal policy doesn’t matter. When I did so, I
always put the story in the past tense, on the assumption that the issue
was dead and buried. The tone was clearly: Thank God we don’t argue
about that any more.
Now, Karl wants us to exhume the body. Although I’m not sure I have
penetrated his methodological discussion, he seems to defend simple
reduced forms, or even simpler correlation coefficients, as the “right” way
to test one broad class of hypotheses against another. As he puts it:
The “single equation with single variable” was the appropriate choice for an
evaluation of a class of hypotheses seriously presented in textbooks and
class teachings [p. 41]. Reliance on the correlation coefficient. is quite
appropriate for the evaluation of the core-class addressed by Friedman and
Meiselman [p. 43].
. .

Let me try to explain what I think Karl means, using as my example
the simplest version of the St. Louis equation. Then I’ll say why I think
he is wrong. Suppose that true model of the economy is Eq. (1) of Karl’s
paper:
Y~=k+aF +bM +e~,
1
1

(1)

where e includes quite a lot of things, some of which are at least partly
forecastable. If F, M, and e are all orthogonal random variables, then the
variance of Y is

120

Alan S. Blinder
Var( Y)

=

a2Var(F) + b2Var(M) + Var(e).

(2)

I think Karl wants to say that the “core-class” hypothesis of the CroMagnon monetarists was that b2Var(M) is much bigger than a2Var(F), so
that movements of M dominate movements of Y. Conversely, the core of
the Neanderthal Keynesian view is that a2Var(F) is much larger than
b2Var(M). With orthogonal data the simple correlation coefficients provide the data needed to discriminate between these two hypotheses, since
in this model r(Y, F) is proportional to aa(F~,and r(Y, M) is proportional
to ba(M). This decomposition of Var(Y) even makes sense, because a
high r( Y, M) and a low r( Y, F) would mean that monetary impulses
dominate the movement of Y, and vice versa.
At some level I have sympathy with Karl’s methodological point of
view, since I do think that an economic model must consist both of a set
of equations and a judgment about the nature of the dominant stochastic
disturbances. But my sympathy is only skin deep.
One reason is trivial and obvious. If F and M covary in the data, the
clean decomposition in (2) cannot be done. There is a covariance term
that Cro-Magnons can attribute to monetary policy and Neanderthals can
attribute to fiscal policy. Nothing in the data will adjudicate this dispute,
which is more teleological than logical.
Another problem arises when policy is set purposefully. At the risk of
some duplication, let me state that point as simply as possible, even
though Karl has discussed it at length, modernizing it as I do to account
for rational expectations. Goldfeld and I (1972) suggested that M and F
are not whimsical random variables but may instead be deliberately manipulated to offset changes in e. If monetary policy fully offsets the
expected value of e, then
M1

=

—(l/b)E1_ie~,

(3)

and (1) becomes
=

k + aF~+ (e,

—

E1_ 1ev).

(4)

If forecasts are pretty good, the innovation would have small variance,
r(Y, F) would be high, and regression (1) would yield a zero coefficient for
M. Hence, Neanderthal Keynesians would be judged correct by Karl’s
criterion. Both F-M and A-J would have become fiscalists. And all
because monetary policy was so effective.
On the other hand, if fiscal policy did the stabilizing and monetary policy was random, the corresponding equations would be

Ruminations
F~= —(1/a)E11e1,
=

k + bM1 + (et

—

E~_ie1).

121
(5)
(6)

Now V and F would be uncorrelated, whereas V and M are highly correlated, and a regression of the form (1) would assign a zero coefficient to
fiscal policy. These are just the findings of F-M and A-J. Karl would
judge the Cro-Magnon monetarists correct because fiscal stabilization was
so effective.
But neither conclusion makes sense; neither finding implies that either
policy tool is powerless to influence GNP. In (6) we have no way to estimate a and in (4) we have no way to estimate b. In either case, purposeful policy reactions deny the econometrician the information he needs to
estimate one of the multipliers.
Surely we all know by now that neither a nor b is zero, that both monetary and fiscal policy are from time to time used purposefully, and that
many variables are omitted from (1). So why argue about which of two
silly hypotheses is the sillier? I’d rather see the fossils of the FM /AM
debate left in the grave.

4, The Govermnent Budget Constraint and
Fiscal—Monetary Interactions
I have much more favorable things to say about Karl’s excellent discussion of the government budget constraint and the issues it raises. In fact,
I think we see eye to eye almost totally—which makes both of us, I think,
disagree with McCallum. By the way, I think that when Blinder and
Solow agree so closely with Brunner and Meltzer, that’s worth noting.
Maybe we have hit on some deep truth! My capsule summary of the
debate would differ in only minor ways from Karl’s. It goes like this.
The paper Solow and I published in 1973 dealt with a simple case of
fixed prices andfixed tax rates. It pointed out and explained the paradoxical result that a rise in government spending is more expansionary in the
long run if the ensuing deficits are covered by issuing bonds than if they
are covered by printing money—provided the system is stable under both
financial policies. More importantly, perhaps, it showed that the likelihood of instability is far greater under bond financing than under money
financing.
Notice that this is bad news for monetarism—not as theory but as
policy—because the “bond-financed” case is essentially the monetarist policy rule of steady (in this case, zero) money growth.

122

Alan S. Blinder

Though the details of our model left much to be desired, I think these
basic findings have proven to be very robust. Brunner and Meltzer and
Tobin and Buiter established parallel results in full-employment models
with variable prices; Buiter, Pyle and Turnovsky, and others allowed both
prices and output to vary. Other extensions opened the economy, allowed
more assets, and so on.
None of these earlier contributions, however, dealt in a satisfactory way
with rational expectations. Actually, I think it is impossible to do that,
except in a totally arbitrary way. Let me explain why.
One problem posed by rational expectations is the Reagan—Barro
equivalence theorem. If bonds are just congealed future taxes, then the
wealth effects that lie at the heart of this analysis disappear. There is no
stability issue because bond financing is just tax financing. And the
relevant choice is between money finance and tax finance, not between
money and bonds. Karl discusses this extensively and well. But let me
defer it for the moment, for I have something else in mind.
In conjunction with any sort of forward-looking expectations, the
government budget constraint sets up dynamic constraints across policy
choices. To take a not very hypothetical example, suppose the current
government raises spending and cuts tax rates, thereby opening up a
deficit. Current and future governments are thereby obligated to do some
combination of
(a)
(b)
(c)
(d)

raising tax rates
cutting spending
printing money
floating more bonds.

(This latter possibility can last forever only if the conditions for stability
under bond finance hold; and they may not.) It is rational for people to
know this, therefore, and to expect some combination of these events
sometime in the future. But, as Karl points out, who knows when? And
who knows which ones the government will choose?
In Barro’s hands, rational expectations is interpreted to mean that tax
cuts covered by debt today must lead to future tax increases of equal
present value. In that case, under a host of other assumptions (see
below), we get non-Ricardian equivalence, and current bond-financed tax
cuts affect nothing. But that’s only one of several possibilities.
Sargent and Wallace (1981) assume that if the economy is unstable
under bond financing, the government will ultimately have to resort to
money creation. Since rational expectations in a frictionless world
effectively telescope the future back to the present, they conclude that
tight money might be inflationary. But that’s only a second possibility.
President Reagan and his crowd had a different form of rational expectations in mind. They argued that taking away the tax revenue today was

Ruminations

123

the way to get spending down tomorrow. So far that idea has not worked
very well. But who can say it was not a “rational” expectation, or that it
was less rational than Barro’s or Sargent and Wallace’s.
My view on this issue is very similar to Karl’s: who knows what or
when? If a government opts for bond financing of deficits, which seems
to be the default option (pardon the pun!), and it discovers that it has
thereby put the economy on an unstable path, something will definitely
have to give. The economy will not zoom off to either positive or negative
infinity. Something will happen. But what? Sargent and Wallace gave us
one possibility; another is that the economy will get a new government;
yet another is that institutional changes will take place, altering the structure of the model.
Karl gets this analysis just right, I think. He is also right, in my judgment, to point to the tremendous uncertainty that this must cause in
people’s minds. When an individual has very diffuse priorities over what
long-run government policy will be, it strikes me as plausible that his
point estimates of future policy variables may have weak effects on his
current decisions—which is just the opposite of what Barro and Sargent
and Wallace assume. If this is so, then expectational issues, although
deep and weighty, may not be of great empirical importance. I, of course,
do not pretend to know that this is the case. I merely raise it as a possibility.
Another problem stems from diversity of expectations across individuals. If Barro thinks that current deficits will eventually lead to massive
future taxes, if Sargent thinks they will eventually lead to a huge amount
of money creation, and if Reagan thinks they will lead to huge future cuts
in spending, and the economy will be stable under pure bond financing,
then the economy may not converge to any rational expectations equilibrium at all, as Phelps and others have pointed out.
4.a. Monetization
I’d like to say one thing about the empirical aspects of the fiscal—monetary
interaction. Following King and Plosser, Karl states that “There is.
no
evidence for a contemporaneous relation between deficit and money creation in the U.S.A.” (p. 98) That’s what I used to think. But I found otherwise in a paper in this conference series two years ago. Let me try to
reconcile the two views.
King and Plosser found no zero-order correlation between deficits and
money growth over the 1953—82 period. I did several things differently. I
used fiscal-year data and included off-budget items to get a more accurate
measure of the budget deficit. I used bank reserves rather than the
money supply to look directly at the monetization decision rather than at
the money multiplier. And I took care to make the dimensions of van. .

124

Alan S. Blinder

ables and their alignment in time consistent with the government budget
constraint. Nonetheless, I also found no zero-order correlation between
deficits and changes in bank reserves over the 1949—81 fiscal years.
But something quite different, and quite surprising, emerged when I
allowed the monetization decision to depend on lagged inflation and
lagged growth of real federal purchases: a strong and quite robust empirical relationship between deficits and changes in bank reserves emerged.
At least over the 1961—81 period, the deficit was a significant determinant
of monetization; but the fraction of the deficit that was monetized fell as
either inflation or the growth of federal purchases rose. This empirical
regularity survived most of my attempts to get rid of it, including adding
other variables, shortening the sample period, differencing (a la Plosser
and Schwert), and a Chow test for coefficient shifts.
In the end I changed my beliefs. I now think there is reasonably good
evidence that larger deficits typically cause (in the Hume sense, not the
Granger sense) faster growth of bank reserves in the United States. However, I should point out that the estimated fraction of the deficit that is
monetized is never very large, and it gets negative when lagged inflation is
high.

5, Non-Ricardian Equivalence
Before discussing Karl’s discussion of Ricardian equivalence, I’d like to
say a word about truth-in-naming. As we all should know by now, David
Ricardo thoroughly discussed what is now called the Ricardian
equivalence proposition—and rejected it, That does not mean that the
equivalence proposition is false; Ricardo probably also believed in bleeding. But it does mean that we should stop calling it Ricardian. I propose
that we call it the Reagan—Barro equivalence proposition.
When it comes to the substance of the issue, I feel relatively comfortable with Karl’s discussion. I would have changed the emphasis in some
places: for example, I attach more importance to the problem of corner
solutions at zero bequests. Especially in a growing economy in which
children are, on average, better off than their parents and in which
bequests in human form are substantial, I suspect that many optimizers
would like to leave a negative bequest but cannot. A tax cut or a rise in
social security benefits now, balanced by future taxes on our children; is
one way to accomplish this.
Karl correctly characterizes the empirical evidence on the equivalence
theorem as quite mixed. And he expresses some surprise that the data are
not more decisive, because he finds the Barro—Reagan proposition

Ruminations

125

implausible on a priori grounds. I am also surprised. The Barro—Reagan
view has proven much harder to reject than I thought. The evidence
adduced to date really is too mixed and, as previously noted, too ideologically correlated to resolve the issue.
In fact, a recent paper by Benjamin, Kochin, and Meador suggests that
many of the empirical tests of equivalence may be entirely beside the
point. Their basic argument is consistent with Barro’s work on optimal
debt policy and can be summarized as follows.
Tax rates are not arbitrary but are purposefully manipulated by government to minimize deadweight loss over time. As Barro pointed out, the
optimizing government will want to smooth tax rates relative to expenditures. Under certainty, the optimal tax rate is constant through time. But
when there is uncertainty, the optimal tax rate will evolve as a random
walk, following the current estimate of the present value of expenditures.
One implication of this kind of optimizing behavior, pointed out by
Benjamin et al., is that the rational expectations consumption function
becomes Keynesian. This is easy to see. The rational expectations consumption function appealed to by Barro is farsighted and forwardlooking:
V

°°

C,

=

k

~

‘

T

°°

~

s=o (l+r)s

t

—

s=0

‘+5

(l+r)5

.

(7)

In the Barro—Reagan story, any arbitrary change in current T, is balanced by changes in the opposite direction in some future ,T,+5’s of equal
present value. Hence, the tax term does not change, and neither does C,.
But if T, is a random walk, then current T (not current spending) is the
best estimator of any future T,~5. So a rise in T, will be interpreted by
consumers as indicating a rise in the permanent levels of government
spending and taxation. They will therefore reduce their consumption
accordingly. So current taxation will have strong Keynesian effects on
current consumption.
Note the strong parallels between this argument and the one I made
earlier about St. Louis equations. Both econometric procedures make
sense if the government policy instruments are set whimsically. But both
can give seriously misleading results if the government acts purposefully.

6. Conclusion
In sum, I disagree most emphatically with Karl’s attempt to resurrect and
legitimize the old reduced form approach, but I agree with most of what

126

Alan £ Blinder

he says about the government budget constraint and the non-Ricardian
equivalence theorem. Two out ofthree is not bad!
SinceKarlislabeledamonetaristandlamlabeledaKeynesian,and

since neither one of us shuns our label, maybe this wide-ranging agreement suggests that the labels are obsolete and possibly even dy4lmctlonal.
I, for one, would be happy to declare the monetarist—Keynesian debate
over today~right here in St Louis. What an appropriate place!

References
Barro, R. 11974. Are government bonds net wealth? Journal ofPolitical Economy
82: 1095-1117.
Benjamin, D. K., L Kochin, and M. Meador. n.d. Observational equivalence of
rational and irrational consumers if taxation is efficient Unpublished

manuscript
Blinder, A. S. and R. M. Solow. 1973. Does fiscal policy matter? Journal ofPublic
EconomIcs 2: 319—37.
Brunner, K. and £ H. Meltzer. 1976. An aggregative theory for a closed economy. In Monetarism, ed., 1 Stein. Amsterdam: North-Holland.
Buiter, W. 1976. Capacity constraints, government financing and the short-run and
long-run effects offiscal policy. Unpublished manuscript.
Friedman, M. and D. Meiselman. 1963. The relative stability of monetary velocity
and the investment multiplier in the United States, l897-l95& In Stabilization
Policies. Englewood Cliffs, NJ.: Prentice-Hall.
Ooldfeld, S. Pt and A. S. Blinder. 1972. Some implications ofendogenous stabilization policy. Brookings Papers on Economic ActIvity 3:585-MO.
King, R. 0. and C. L Plosser. 1984. Money, deficits and inflation. Presented at

Carnegie-Rochester Conference Series on Public Policy.

Pyle1 D. IL and S. J. Turnousky. 1976. The dynamics of government policy in an
inflationary economy: An ‘intermediate run’ analysis. Journal of Mone~Credit
and Banlclng8: 411-37.
Samuelson, P. A. 1961. EconomIcs: An Introductory Analysis. New York:

McGraw-HilL
Sargent, .t J. and N. Wallace. 1981. Some unpleasant monetarist arithmetic.
Federal Reserve Bank ofMinneapolis Quarterly RevIew 6:1-17.
Tobin, J. and W. Buiter. 1976. Long run effects of fiscal and monetary policy on
aggregate demand. In Monetartcm, ed. J. Stein. Amsterdam: North-Holland.

.5
Comment on Karl Brunner’s
“Fiscal Policy in Macro Theory:
A Survey and Evaluation”
Robert J. Gordon
NORTHWESTERN UNIVERSITY
AND NATIONAL BUREAU OF ECONOMIC RESEARCH

1. Introduction
It is a great pleasure for me to discuss the essay by Karl Brunner; this
continues a tradition of discussing each other’s work that dates back at
least a decade. Much of Brunner’s massive survey is admirable, reflecting
a deep and careful analysis of central issues in fiscal policy, and the way
events have changed perceptions of fiscal policy. The comprehensive and
up-to-date reference list adds to the value of the contribution. Because of
its length, I cannot delve into every issue raised by Brunner, but I will
concentrate on three areas where, I believe, his analysis needs to be
qualified and supplemented: (a) the intellectual history of the monetary—
fiscal debate; (b) the interpretation of St. Louis equations; and (c) Ricardian equivalence and other theoretical issues.

2. The Interaction of Events and Ideas
Brunner drags up from his dusty shelves of old journals the infamous
“battle of the radio stations (AM—FM)” and defends what has long been
dismissed—an evaluation of alternative viewpoints with simple correlations
between spending and money on the one hand and autonomous spending
on the other. The essence of Brunner’s defense is that “the ‘single equation with single variable’ was the appropriate choice for an evaluation of a
class of hypotheses seriously presented in textbooks and class teachings.”
Most spectators of the AM—FM debate found the single-variable framework unappealing because they could find no example of any influential

128

Robert J. Gordon

economist at that time who believed that “only fiscal policy matters,” and
even if such an economist did exist, he never would have expected to find
a stable and constant coefficient of total spending on autonomous spending because the output multiplier was a variable, not a constant, depending on, among other things, a host of changing tax rates. The Andersen—
Jordan St. Louis contribution was taken more seriously precisely because
its point of departure was a two-variable test of monetary and fiscal policy
together.

Brunner begins with intellectual history and then delves into
econometric details. Let me begin with my own version of the intellectual
history that emphasizes the influence of events (rather than journal
debates) on the evolution of ideas. This account is sprinkled with a few
quotes to indicate what people actually believed in the 1950s and early
1960s. We then turn to an interpretation of the empirical issues.
The central paradigm of macroeconomics as it emerged from World War
II was indeed the Keynesian multiplier theory and its endorsement of an
activist fiscal policy to overcome the inherent instability of private investment. Monetary theory lurked in the shadows, discredited at least temporarily as a result of a major event that dominated early postwar ideas—

namely the juxtaposition between early 1938 and late 1940 of a weak
economic recovery, explosive monetary growth, and a short-term interest
rate that was rapid and constant between early 1938 and late 1941, the
economy’s recovery floundered until military spending began in earnest in
late 1940, after which real GNP suddenly jumped by almost 20 percent in
a single year. This chronology ingrained a deep-seated belief in the
potency of fiscal policy and the “pushing on a string” analogy for monetary policy.
But money was not ignored totally in the late 1940s, and many economists took note of the fact that the quantity of nominal money had tripled
between 1940 and 1945. Contemporary accounts displayed a curious
inconsistency, with a monetary expansion viewed as impotent but a monetary contraction viewed as too dangerously potent to risk, as in Lawrence
Seltzer’s (1945) remark that “there is great risk that the deflationary effects
of a radical rise in interest rates might be so severe as to throw the whole
economy into a crushing depression” (p. 844).
As for the teaching of undergraduates, I have never managed to obtain
Samuelson’s 1948 first edition, but I do have the 1951 second edition,
which, I hasten to add, was not the edition that I used in college but was
obtained at a used book sale. Samuelson in 1951, more than a decade
before the AM—FM debate, does not reveal himself as a hard-line “only
fiscal policy matters” guy. Instead, his treatment reflects the uncomfortable asymmetry of early postwar Keynesian ideas. There are over 25
index entries for money and monetary policy and another 20 for the

Comment

129

interest rate. On page 342 are listed as effects of a $1 billion open-market

purchase:
the general easing of interest rates and the increased availiability of credit to
would-be investors,
the upward shift in the earlier chapters’ investmentincome schedule resulting from the lowered rate of interest,
the primary
and secondary increases in income resulting from the increased flow of
investment,
and the increased stock of buildings, equipment, and inventories that will later result from the cumulation of a high rate of investment.
.

. .

. . .

..

.

In typical late 1940s style, this account is immediately followed by three
qualifications that make monetary policy “at best a supplement to other
stabilization policies, such as fiscal policy.” The three qualifications are
these: (1) “Changes in the amount of money may have very weak effects
on the rate of interest if rates are already very low.” (2) “Even if there are
some changes in the rate of interest, the rate of investment spending may
turn out to be relatively little affected by changes in interest rates. The
prospects for invesement may depend much more on the depressed state

of business.” (3) “The Central Banker may be unwilling to push monetary
policy very far.” Clearly the first two qualifications reflected events of the
late 1930s, and the third the Fed’s pegging of interest rates in this preAccord edition of the textbook. The late 1940s asymmetry is implicit in
the assumption that the problem of monetary policy is pumping up a
depressed economy rather than slowing down an overheated one.
Over the following decade there was a gradual but continuous shift of
opinion toward an increased role for monetary policy, marked by
mileposts including the Patman Committee Inquiry, the negative reaction
of many economists to the downgrading of money in the Radcliffe report,
and the influence of the monetary research of Milton Friedman, his students, and others. The growing belief in the importance of money can be
traced to several episodes in the first postwar decade, Those who believed
that the large outstanding stock of public debt prevented effective monetary action and required the pegging of interest rates either lost credibility
or changed their opinions when the higher interest rates that followed the
Treasury—Fed Accord failed to have any disastrous consequences for debt
management or the economy’s performance in general. The relative mildness of the 1954 recession was due partly to countercyclical monetary policy and helped to lessen the belief that monetary policy was only effective
in countering inflation and suffered from an asymmetric impotence in
dealing with slack demand. The continued acceleration of inflation
despite rising interest rates in 1956—57 tempered the belief that monetary
policy had unique curative powers to combat inflation. By 1962 Harry
Johnson was able to observe that “the wheel has come full circle, and pre-

130

Robert J. Gordon

vailing opinion has returned to the characteristic l920s view that monetary policy is probably more effective in checking deflation than in checking inflation.” Although Johnson may have been ahead of his time,

influenced as he was by monetary research at the University of Chicago,
nevertheless his account provides a picture far from Brunner’s, with the

“hard-line fiscalists” hard to find.
Turning now to more contemporary quotes—I seem to have saved my
final exam in Economics 1 at Harvard, taken in May 1959—I find interesting evidence to support the idea of “fiscal dominance,” but I find no evidence at all of the influence of Brunner’s “hard-line fiscalists” believing
“money doesn’t matter.” Fiscal dominance is reflected in the fact that the
first half of the exam consisted of two questions: one hour on fiscal policy
and a half hour on monetary policy. Inspection of the four-part fiscal
policy question makes one scoff in retrospect at the idea of regressing
aggregate spending on autonomous spending, because 1959 Harvard
undergraduates were supposed to know that the multiplier effect of a
change in government spending depended on whether or not the spending
was financed by increased taxes, and they were given the option of concluding that an increase in spending accompanied by an equal increase in
taxes might raise unemployment or leave it unchanged. The half-hour
monetary question reflected not a ritual belief that “money doesn’t
matter,” but rather the same old asymmetry. To quote the question, “It is
frequently argued that monetary policy is effective in controlling inflation,
but less successful in fighting unemployment. Trace the mechanism
through which the tools of monetary policy operate under alternative cyclical conditions, and comment on their effectiveness.”
Returning to the influence of events, the AM—FM debate coincided
with the heyday of activist fiscal policy, dubbed the “new economics.” By
then, changes in government spending were recognized to involve gestation lags and to have allocative side effects, and so the central policy tool
had become changes in income tax rates, which of course involves
changes in the spending multiplier rather than a stable multiplier, as in
Brunner’s caricature of fiscalism. Although the consensus policy paradigm of 1965 did not neglect monetary policy nor deny that monetary
tightness could interfere with the pace of economic expansion, monetary
policy was basically kept in the background and relegated to the role of
maintaining a low and stable level of long-term interest rates to foster the
goal of stimulating long-term economic growth.
This policy framework collapsed with amazing speed after 1967 as the

result of the interaction of events and economic writings. My graduate
school classmates and I were acutely aware of the timing of this turn in

the intellectual tide, as we began our first teaching jobs in the fall of 1967
and almost immediately found our graduate school education incapable

Comment

of explaining the evolution of the economy.

131

The most important

ingredient in this revolution was the Friedman—Phelps “natural rate
hypothesis,” the role of which is well known and not our subject today.
More relevant was the blow struck by Andersen and Jordan in 1968.
Although activist advocates eventually regrouped and presented convincing evidence of fatal statistical flaws in the St. Louis procedure, particularly the contribution of Goldfeld and Blinder, their disarray lasted long
enough to partially discredit fiscal activism. To add to the overall indictment of fiscal policy provided by the St. Louis equation, Robert Eisner in
1969 made an important attack on the efficacy of the temporary tax
changes favored by mid-1960s policy activists. Using the framework of
Friedman’s permanent income hypothesis of consumption, Eisner argued
that a temporary income tax cut or surcharge would fail to alter permanent income and thus would have a lower spending multiplier.
Further, the lag in the effect of fiscal policy might be long and/or
unpredictable, with the length of the lag depending on the public’s subjective assessment of the likelihood that the tax change soon would be
reversed.
These academic criticisms of the activist case might not have been so
persuasive if they had not been accompanied by supporting events. The
dramatic drop in the personal saving rate in late 1968 and the failure of
spending growth to slow appreciably in response to the temporary tax surcharge was consistent both with the St. Louis claim that monetary multipliers had previously been underestimated and fiscal multipliers overestimated and with the Eisner critique. Blinder’s retrospective econometric
evidence of this period shows that temporary tax changes are not completely ineffective, but their multiplier impact may be as little as one-half
of tax changes regarded as permanent, and the effect on consumption of
any tax change may take several years to occur.

3. Empirical Issues in St. Louis Equations
The empirical issues involved in the AM—FM debate and subsequent St.
Louis equation are so well known that little time need be spent reviewing
them. The St. Louis equation represented an advance over Friedman and
Meiselman in three main dimensions: testing the effects of monetary and
fiscal policy in the same equation, using full employment instead of actual
government spending and revenues, and expressing variables in first
differences. However, the St. Louis reduced form was vulnerable to the
central criticism that coefficients of both the monetary and fiscal policy
variables were biased if there were any correlation between either policy

132

Robert J. Gordon

variable and the error term in the equation, representing the whole panoply of omitted demand and supply shocks that drove changes in aggregate
demand.
The general case for this point was best expressed by Goldfeld and
Blinder, and once their case was stated, everyone understood the argument that the monetary policy coefficients were biased upward since dur
ing all of the ongmal Andersen—Jordan sample period the Fed was acting
to stabilize interest rates rather than money thus creating a passive posi
tive response of money to any demand shocks in either the money or
commodity markets And there was no surprise when Ando and Modigli
am reported their experiment that when estimated to artificial data gen
erated by the MPS model, the Andersen—Jordan technique substantially
overstated monetary effects and understated fiscal effects. But this still left
open the source of the fiscal bias. If a downward bias on fiscal policy
coefficients in the St. Louis equation occurs because active fiscal policy
has been pursued within the sample period thus creating a negative
correlation between government spending and the error term what were
these episodes when fiscal activism was so effective? In two published
comments (1971, 1976), I pointed to the set of events in the Eisenhower
administration that led to this result.
Most important, there was a huge negative correlation between the
decline in defense spending that took place between 1953 and 1956, and
the (dare I say) autonomous bursts of automobile spending associated
with new models in 1955 and export spending partly associated with the
Suez crisis in 1956. (With reference to McCallum’s paper in this volume,
it is important to note that this negative relationship displays a positive
serial correlation extending over two years.) I recall Paul Samuelson’s
injunction to us fledgling graduate students in the mid-1960s that he
would flunk anyone who produced an econometric explanation of the
high level of auto sales in 1955. I later told the story, to explain the
Andersen—Jordan result in terms of efficacious fiscal policy, that
“President Eisenhower had decided to stop the Korean war in 1953
because he could see the 1955 auto boom and 1956 Suez crisis coming,
and he wanted to get defense spending out of the way to avoid overheating the economy.” What is not facetious is the remarkable record of the
Eisenhower administration in the 1958 recession in creating a time path
for nondefense government purchases that rose as the economy fell and
fell as the economy recovered. In my comment (1976) I showed, by alternatively including and excluding a proxy for autonomous spending from a
St. Louis equation, that in the Eisenhower period fiscal coefficients were
low and downward biased, in the Nixon—Ford period they were high and
upward biased as a result of procyclical fiscal policy, and in the
Kennedy—Johnson era they were in between. The original St. Louis equa-

Comment

133

tion was dominated by the Eisenhower sample period and by the negative
correlation between the post-Korean decline in defense spending and the
mid- 1950s business expansion.
Viewing this whole literature from the mid-l980s, we find naive the
entire literature on autonomous spending because (as McCallum s
paper suggests) nothing is truly autonomous. Recent papers have more
fruitfully viewed business cycles as being generated by innovations in
both financial and real variables, where “innovation” is defined as the
error in an equation that relates the variable in question to its own past
values and the past values of everything else. In this context I have
recently completed a research project that reexamines the behavior of
household and business investments in newly created quarterly data
extending back to 1919 [see Gordon and Veitch (1984)] Strong evidence
is provided to support both sides of the AM—FM debate, for innovations
in the money supply have a substantial influence on both household and
business investments, but there is still room for a major impact on the
business cycle of autonomous innovations in structures investments (both
residential and nonresidential).

4. Ricardian Equivalence and Other Issues
The rest of Brunner’s paper is more satisfactory. There is a sensible discussion of “intergenerational altruism” and “intergenerational selfishness”
in the context of the Barro—Ricardo equivalence theory. As a matter of
historical record, I wish that Brunner had cited Patinkin’s incorporation
into macroeconomic analysis of a variable proportion (k) of outstanding
government bonds treated as net private wealth. Patinkin’s treatment anticipated many of the implications of Barro’s analysis without taking any
position over whether k is at an extreme value of zero, as assumed by
Barro, or unity, as assumed in some traditional Keynesian analysis.
Brunner recognizes that “the context of risk could explain.., the
appearance of bequest without a bequest motive [as] formalized by
Barro.” Risk, however, is just one of the reasons why I was never convinced by the Barro logic, however dutifully I continue to teach it in the
graduate school classroom. As one without children and likely to leave a
substantial bequest, it immediately became evident that there is a more
important reason than risk to explain why individuals often leave bequests
without any necessary altruism for future generations. After all, we are
supposed to be able to insure ourselves against risk by buying annuities.
But a more important additional set of factors—high transactions costs and
inconvenience, as well as imperfect capital markets—make it almost

134

Robert J. Gordon

impossible for a well-off person to “go out” with a zero net worth. There
is no rental market for the type of house I live in, so in order to buy an
annuity with all my assets, I would have to move. Renting a car is expensive, and renting my personal library of books and journals would be
impossible. People like me are likely to behave according to a permanent
income theory of the flow of consumption services and to leave whatever
assets are necessary to maintain that flow of services to worthy charities.
Since my heirs are likely to be nonprofit and nontaxable organizations,
there is simply no present value of future tax liabilities to consider, and
the Barro theorem falls to the ground. The Reagan tax cuts financed by
deficit spending have made me feel good, and I have spent some of the
proceeds.
Of course, I have not spent all of the proceeds, because simultaneously
there have been substantial increases in Keogh and IRA ceilings that have
induced me to save more as well. This tug of war between conflicting
incentives bears on the empirical evidence “on the Ricardian theme”
reviewed at such length in Brunner’s paper. Reduced-form equations can
be useful, and I have estimated plenty of them in my work on inflation
and, more recently, on investment. But the equations of Feldstein—
Kormendi type, summarized by Brunner, seem unlikely to provide any
reliable evidence of the issues at hand. First, the inclusion of government
spending and tax revenues as explanatory variables in a consumption
equation runs afoul of the Goldfeld—Blinder critique for the same reasons
as does the St. Louis equation. Second, the tax schedule is progressive,
and if people in different tax brackets have different propensities to consume, the schedule relating total consumption to total tax revenue will be
nonlinear. Third, the lags that Blinder found between changes in taxes
and changes in spending are neglected. Fourth, no distinction is made
between temporary and permanent tax changes. Fifth, tax law changes
that alter disposable income, like a neutral surcharge, can have totally
different effects than legislative changes that twist the incentives to consume and to save, as in my IRA—Keogh example. Surely Brunner is aware
of all this, so I wonder why he takes all this empirical work so seriously.

5. Concluding Remarks
Bunner’s paper treats numerous other issues that do not appear to require
comment here. I would enter only two qualifications. First, Brunner’s long
and sensitive discussion of the Sargent—Wallace deficit analysis is slightly
marred from his uncriticial analysis of the Mankiw—Summers paper on
consumption and money demand. There are at least three problems with
that paper that shed doubt on its credibility: (a) the implausible assump-

Comment

135

tion that the responsiveness of money demand to changes in investment is
zero; (b) the lack of evidence for the postwar period of any difference in
the fit of GNP and consumption when entered into a money demand
equation with a flexible lag structure [see Gordon (1984)]; and (c) the
awkward fact that a larger share of demand deposits is held by business
firms than by households. Mankiw and Summers do not make a convincing case that would support contractionary effects of tax cuts.
The second qualification refers to Barro’s tax smoothing hypothesis on
the behavior of deficits. Although somewhat skeptical of the empirical
robustness of Barro’s approach, Brunner calls it “the only game in town.”
Yet a paper by Barro (1984) that Brunner does not cite was dismissed at a
recent conference as being unsuccessful both on theoretical and empirical
grounds. As one discussant asked, “How is it that policymakers in Washington figured out the Ramsey optimal tax rule 50 years before public
finance economists?” Consider social security, for which tax smoothing
means full actuarial funding of the expected program of benefits. And the
theory fails empirically, because the coefficient in a debt change equation
on temporary government spending is zero rather than unity as required
by the theory, and the World War II years have to be thrown out.
A concluding comment is that, given the length of Brunner’s present
paper, the attention given to reduced-form empirical evidence on various
issues seems excessive. Other, perhaps more interesting, issues regarding
fiscal policy might have been covered instead. He might have reconsidered the effects of changes in the monetary—fiscal policy mix on real
interest rates and the exchange rate, with and without Ricardian
equivalence. Was the shift in the policy mix with the ensuing appreciation
of the dollar the real key to the extent of disinflation; if so, what are the
theoretical arguments for reversing the mix to avoid future deficits or for
maintaining the mix to hold the benefits of disinflation? What are the
implications of tax reform for macro theory, particularly a shift to base
broadening with lower marginal rates, or a shift to a broad-base progressive consumption tax? What are the implications of current large government deficits for the public choice idea that the best way to reduce
government spending is to reduce tax rates—this just hasn’t happened
when defense spending is included. Overall, I cannot help feeling that the
empirical work examined is too flimsy to merit so much attention from
this fine theorist, and I cannot help wishing that a paper on fiscal policy in
macro theory had contained more about theory.

References
Barro, Robert J. 1984. The behavior of U. S. deficits. NBER Working Paper 1309.
March. In The American Business Cycle: Continuity and Change, ed. RJ. Gordon. Chicago: University of Chicago Press. Forthcoming.

136

Robert J. Gordon

Eisner, Robert. 1969. Fiscal and monetary policy reconsidered. American
Economic Review 59: 897—905.
Gordon, Robert J. 1971. Notes on money, income, and Gramlich. Journal of
Money, Credit, and Banking 3: 533—45.
1976. Perspectives on monetarism. In Monetarism, ed. J. Stein, pp. 52—66.
Amsterdam: North—Holland.
______ 1984. The 1981—82 velocity decline: A structural shift in income or money
demand? In Conference on Monetary Targeting and Velocity, pp. 67—99. Federal
Reserve Bank of San Francisco.
______ed. n.d. The American Business Cycle: Continuity and Change. University of
Chicago Press. Forthcoming.
Gordon, Robert J. and John M. Veitch. 1984. Fixed investment in the American
business cycle, 1919—83. In The American Business Cycle: Continuity and
Change, ed. R.J. Gordon. Chicago: University of Chicago Press. Forthcoming.
Johnson, Harry G. 1962. Monetary theory and policy. American Economic Review
52: 335—84.
Patinkin, Don. 1965. Money, Interest, and Prices, 2d ed. New York: Harper &
Row.
Seltzer, Lawrence H. 1945. Is a rise in interest rates desirable or inevitable? American Economic Review 35: 83 1—50.
______

6
Can Policy Activism Succeed?
A Public Choice Perspective
James M. Buchanan
CENTER FOR STUDY OF PUBLIC CHOICE
GEORGE MASON UNIVERSITY

1. Introduction
The question posed in the title assigned to me presupposes the existence
of an ordering of options along some scale of presumably agreed-on preferredness or desirability. Only if this presupposition is made does it
become appropriate to ask whether or not politics, as it operates, can be
expected to select the most preferred option on the ordering, or, less
ambitiously, to select, on average, options that would allow the pattern or
sequence of “choices” to be adjudged “successful.” The generalized
public-choice answer to the question, given the required presupposition, is
reasonably straightforward, and it is essentially that of classical political
economy. Those who make political decisions can be expected to choose
in accordance with agreed-on or “public interest” norms only if the institutional structure is such as to make these norms coincident with those of
“private interest.” The public chooser, whether Voter, aspiring or elected
politician, or bureaucrat, is no different in this role than in other roles,
and if incentives are such that the coincidence of interest is absent, there
will be no “successful” political ordering over the feasible options. I shall
return to the possible coincidence of interest following Section 2.
The more fundamental question to be asked, however, involves the
appropriateness of the required presupposition—that concerning the possibility of any meaningful ordering of policy options, quite independently
of any problems of implementation. This question has been obscured
rather than clarified by those economists who resort to “social welfare
functions.” These functions impose a totally artificial and meaningless
ordering on “social states” without offering any assistance toward facilitat-

140

James M Buchanan

ing choice from among the set of options feasibly available to the public
chooser. Section 2 examines this fundamental question in the context of
the issues that prompted the assigned title.

2. Is It Possible to Define an Ordering of Policy Options Along
an Agreed-on “Success” Scalar?
In this section I propose to ignore totally all problems of policy
implementation—all public choice problems, if you will. For simplicity,
assume the existence of a genuinely benevolent despot, who sincerely
seeks to do that which is “best” for all of those who are members of the
political—economic—social community. How can we describe the utility
function of this despot? It is easy, of course, to list several desired endstates. Full employment, stable and predictable value in the monetary
unit, high and sustainable rates of economic growth, stable international
order—these may be mutually agreed-on objectives for policy action. But
there may be conflict among the separate objectives (to raise a topic of
much debate—discussion of the 1950s that has been relatively neglected in
the 1980s). How are we to model the trade-offs among the objectives
within the utility function of the benevolent despot, if indeed such
conflicts should arise?
I presume that the despot can act so as to influence macroeconomic
variables in the economy; I leave possible rational expectations feedbacks
to the other paper in this session. But how “should” the despot act, and,
in this model, how “will” he act? There is no definitive answer to these
questions until and unless the utility function is defined more fully.
There is, of course, an empty response to the question posed in the title
to this section. Clearly, if the despot can, by our presumption, influence
macroeconomic variables by policy action, then, by some criterion of his
own, he can be “successful.” But presumably we seek to employ a more
objective criterion for success, one that can at least conceptually be
observed by others than the despot himself.
For simplicity, let us assume that the despot is concerned only about
domestic employment and monetary stability; we ignore all nondomestic
considerations, and we put aside problems of growth. Further, let us restrict attention to standard macropolicy tools. The despot here is assumed
to be unable, at least in the time frame of the policy under consideration,
to modify the structural features of the economy. With these
simplifications, we can go further and specify the objective function more
precisely. Let us assume that the despot seeks to guarantee that level of
employment that is consistent with stability in the value of the monetary
unit, given the institutional structure of the economy. The objective
reduces to a single price level target.

Can PolicyActivism Succeed?

141

Even in this highly restricted setting, which is by no means that which

might command consensus as a normative posture, the despot cannot simply “choose” the ultimate end objective from an available set of options.
That is to say, “stability in the value of the monetary unit” cannot be
selected as if from off a policy shelf. The despot is further restricted by
the tools of policy available, which in this setting are those of the familiar
fiscal (budgetary) and monetary instruments. Nominal demand can be
increased, directly or indirectly, or reduced, directly or indirectly, by the
use of fiscal—monetary tools, either separately or in some mix. Even if we
ignore, as indicated, the expectational-induced feedbacks generated by
resort to any instrument, there remains the task of predicting accurately
the relationship between the instrument, economic structure, and ultimate
objective. The structural features of the economy are not invariant over
time, and a policy thrust that might be successful under one set of conditions, say in t0, may fail, say, in t , because of structural shifts. At best,
1
therefore, the truly benevolent despot can only be partially successful,
even given the most clearly defined target for policy.

3. Monolithic and Nonbenevolent Despot
The presumption of benevolence on the part of political agents is not, of
course, acceptable within a public-choice perspective. It is precisely this
presumption that has been a central focus of the overall public-choice critique of the theory of economic policy. Political agents must be presumed
to maximize personal utilities in a behavioral model that is invariant, as
between public and private roles or capacities. The structure of decision
making may, however, affect utility-maximizing behavior through shifts in
the effective constraints on choice.
In this section, I shall discuss briefly the simplest possible decision
structure, one in which political decisions are lodged within a single
monolithic authority (in the limit in one person) which (who) is not
directly accountable to or subject to constituency pressures, whether or
not these be explicitly “democratic” (electoral) in nature. In this model, it
is evident, quite apart from any historical record, that the despot will find
if advantageous to resort to money creation over and beyond any amount
that might characterize the “ideal” behavior of the benevolent counterpart
considered above. This result emerges, quite simply, because incentive
effects must be taken into account, and the despot, even if totally immune
from constituency pressures, must reckon with individual adjustments to
alternative revenue-generating instruments. Through a policy of revenuemaximizing inflation, defined in a dynamic sense, the despot can extract
the full value of monetary structure (that is, the value differential between
a monetary structure and a barter structure))

142

James M Buchanan

The amount of revenue that may be potentially raised through money
creation is, of course, finite. And the totally uncontrolled despot may seek
to utilize the taxing and debt-issue power over and beyond the
inflationary revenue limits. The precise features of the despot’s policy mix
wifi depend, in part, on his time horizon in relation to the behavioral
reactions of the population. These features need not be examined in
detail here. It is sufficient, for my purposes, to conclude that the monolithic despot will be successful only in terms of his own criteria, and that
by any of the more familiar criteria for policy success, the failure would
be manifest.

4. Monolithic and Nonbenevolent Agent Subject to
Electoral Constraints
The analysis becomes more complex once we introduce electoral feedback
constraints on the behavior of the monolithic political agent. Assume now
that decision authority remains concentrated, but that the holder of this
authority is subject to potential electoral replacement at designated
periodic intervals. In this model the “governor” cannot expect to use his
authority for personal enrichment for any extended period. Under some
conditions, simple wealth-maximizing strategy might involve revenuemaximizing exploitation during the period of office, with no attention to
possible reelection. In other conditions, the wealth-maximizing strategy
might involve the effort to remain in office, in which case, short-run revenue maximization via inflation, debt creation, and taxation will be mitigated. If the agent is modeled as a simple revenue maximizer, it seems
unlikely that his pattern of behavior would be adjudged “successful” by
external criteria under either of these circumstances.
The more interesting model is one in which the agent is motivated by
other considerations than wealth, the simplest model being that in which
political position is itself the single maximand. The agent’s behavior will,
in this case, be constrained by expectations of electoral support. The question then becomes one of determining to what extent voters, generally, or
in a required winning coalition, wifi support or oppose patterns of policy
outcomes that might be deemed “successful” by external criteria. Given
the postulated motivation here, the agent will base behavior strictly on
constituency response.
Consider this question in the terms introduced earlier, that of a unique
objective of monetary stability. Will a sufficiently large voting constituency support a regime that seeks only this policy objective? This question may be examined in the calculus of the individual voter or potential
voter,

Can PolicyActivism Succeed?

143

Two separate difficulties arise. The first involves the absence of individual voter responsibility for electoral outcomes in large number constit
uencies. Even if the individual knows that the agent elected is fully
responsive to the electoral process because he knows that his own voting
choice will rarely if ever be decisive, the individual may not vote And if
he does vote he has little or no incentive to become informed about the
alternatives. And if he votes, and even if he is reasonably well informed,
there is little or no incentive for him to vote his “interests” rather than his
whims Hence there is only a remote linkage between what might be
defined by the observing external “expert” as the “interest” of the voters
and the support that is given to a prospective political agent who promises
these externally defined interests This difficulty alone suggests that poli
tical agents cannot be held responsible by the electoral process nearly
to the extent that is suggested by naive models of electoral feedback.
A second difficulty emerges even when the first is totally ignored. Even
if all individuals are somehow motivated to vote and to do so in terms of
their well-considered interests, these interests will not be identical for all
voters. There are differentials among persons in the relative benefits and
costs of any macropolicy action. Even the ideally responsive political
agent will meet only the demands of the relevant coalition of voters, as
determined by the precise voting rules.
Consider a single political agent who must satisfy a simple majority of
constituency voters. If voters’ interests in the employment—inflation
trade-off can be presumed to be single peaked, the political agent’s
optimal strategy requires satisfying the median voter. It seems likely that
this median voter will tend to be myopic in his behavior in the electoral
process. He will place an unduly high value on the short-term benefits of
enhancing employment relative to the long-term, and possibly permanent,
costs of inflation. He will do so because, as a currently decisive voter, he
can insure the capture of some benefits in the immediate future. By foregoing such short-term benefits in a “rational” consideration of the longterm costs, the currently decisive voter cannot guarantee against the
incurrence of such long-term costs in future periods. This asymmetrical
result follows from the potential shiftability of majority voting coalitions.
A subsequent period may allow a different median voter or coalition of
voters to emerge as dominant—a decisive voter or group that may choose
to inflate from strictly short-term considerations. To the extent that this
takes place, all of the initial benefits of policy prudence may be offset. In
the recognition of this prospect, why should the decisive voter or coalition
of voters in the initial period exhibit nonmyopic “rationality” in the sense
indicated?2
The ultimate answer to the assigned question is clear in this highly
simplified model for “democratic” politics. Policy activism cannot be successful if the criterion of success is long-term monetary stability, a en-

144

James M Buchanan

tenon that seems most likely to emerge consensually in a constitutional
process of deliberation.3

5 Nonmonohthic and Nonbenevolent Agents in a Political
Structure Subject to Varying Electoral Constraints
The political models examined in sections 3 and 4 were oversimplified in
the assumption that authority was placed in a single agent or agency As
we approach reality it is necessary to recognize that policy making
authority is likely to be divided among several agents or agencies who
(which) may be subjected to quite different electoral controls or constraints and hence potentially affected by differing electoral pressures
For example fiscal or budgetary policies may be made in a wholly
different process institutionally, from monetary policy and even within
the institutional structure of budgetary policy, authonty may be divided
between executive and legislative branches of government, subjected to
varying electoral constraints as defined by such things as breadth of con
stituencies length of terms of office voting structure within agency (in
legislatures and committees) legally defined responsibilities and so on
The direction of difference in effects between this more realistic political
model and the monolithic model previously examined seems evident. To
the extent that policy-making authority is divided, the proclivity toward
response to short-term pressures is increased. Any array of results along
the success criterion indicated would indicate that the divided-authority
model ranks well below its monolithic counterpart.

6. Nonbenevolent but Monolithic Agent Divorced from Direct
Electoral Constraints but Subject to Legal—Constitutional
Rules against Personal Enrichment
If there is little or no basis for expecting political agents to express
benevolence in their policy behavior, and if, as suggested, the standard
“democratic” controls will not themselves insure patterns of outcomes that
meet reasonable criteria of success, alternative institutional structures must
be analyzed. Consider, first, a model in which decision-making authority
is lodged in a single agent or agency and one that is specifically divorced
from the electoral process—an agent or agency that does not face continual electoral checks. To prevent that potential for excess under the
model discussed in section 3 however suppose that the agent or members

Can Policy Activism Succeed?

145

of the agency are placed within enforcible legal—constitutional limits with
reference to his or their personal or private enrichment, either directly or
indirectly. That is to say, the agent or members of the agency cannot use
the money creation and /or taxing power to finance their own private consumption needs or accumulation (e.g., Swiss bank accounts) desires.
Beyond this restriction, however, we shall assume that the agent or
members of the agency is (are) not limited in behavior except in the
overall and general mandate to carry out “good” macroeconomic policy.
This model can, of course, be recognized as one that is closely analogous to the monetary authority of the Federal Reserve Board in the
United States. Some elements of the model discussed in section 3—that of
the nonconstrained despot—describe the existing structure, and, more
importantly, some political controls are exercised; but, for my purposes,
the existing monetary authority fits the model reasonably well.
The problem becomes one of predicting the behavior of such an agent
and of assessing this behavior in terms of the success criterion introduced.
Neither economic nor public-choice analysis is capable of being of much
assistance in this respect. To make a prediction, one must get inside the
utility function of the agent (or of those who participate in agency decisions). In particular, it would be necessary to know something about the
internal rate of time preference that will characterize behavior. If~as we
have assumed, demand-enhancing action is known to generate short-term
benefits at the expense of long-term costs, the behavior of the monopolistic and discretionary agent in making this trade-off will depend strictly on
his own, private, rate of time preference, as expressed “for” the community. That is to say, under the conditions indicated, the agent will not, personally, secure the benefits or suffer the costs. By definition, the agent is
not responsible, in the sense of a reward—penalty calculus.
This absence of responsibility itself suggests that the behavior of the
discretionary agent is likely to be less carefully considered, to be based on
less information, and hence to be more erratic than would be the case
under some alternative reward—penalty structure. The model further suggests that the agent here is more likely to be responsive to the passing
whims of intellectual-media “fashion” than might be the case in the presence of some residual claimancy status. To the extent that the agent is at
all responsive to interest-group pressures, such response seems likely to be
biased toward those groups seeking near-term benefits and biased against
those groups that might be concerned about long-term costs, if for no
other reason than the difference in temporal dimension itself. Organized
pressures for the promotion of short-term benefits exist while there may
be no offsetting organization of long-term interests. This bias might well
be exaggerated if the agent or agency is assigned functions that cause the
development of relationships with particular functional groups in the

146

James M Buchanan

policy (e.g., banking and finance). In sum, although there is really no
satisfactory predictive model for behavior of the genuinely discretionary
agent or agency, there are plausibly acceptable reasons to suggest that
policy failures will tend to take the directions indicated in the discussion
here.
Viewed in this perspective, and in application to the Federal Reserve
agency in the United States, and perhaps notably after the removal of
international monetary constraints, there should have been no surprise
that the behavior exhibited has been highly erratic. Any other pattern
would indeed have required more explanation than that which has been
observed. From both analysis and observation the ultimate answer to the
question concerning “successful” policy activism in this model, as in the
others examined, must be negative.

7. Nonbenevolent and Monolithic Agent Divorced from
Electoral Constraints but Subject to Legal-Constitutional
Rules Against Personal Enrichment but Also to
Constitutional Rules That Direct Policy Action
The generally negative answer to the question posed in the title prompts
examination of still other institutional structures that do not involve
attempts at “policy activism,” as such, but which, instead, embody sets of
predictable and directed policy actions in accordance with constitutionally
specified rules. In familiar terminology, if “policy activism,” when applied
in a setting of discretionary authority, must fail to meet the success criterion, can a setting of rules do better? It would be inappropriate to discuss at length the relative advantages of alternative regimes or sets of
rules. But it is clear that almost any well-defined set of rules would eliminate most of the incentive and motivational sources for the failure of discretionary agency models as previously discussed.
In a very real sense there is no agency problem in an effectively operating rule-ordered regime. A fiscal—monetary authority, charged with the
actual implementation of policy, but only in the carrying out of specified
rules, defined either in terms of means or objectives, cannot itself be
judged on other than purely administrative criteria of success or failure.
More ultimate criteria must now be applied to the alternative sets of rules,
with success or failure accordingly assigned. And working models of such
alternative sets might be analyzed, just as the models of a discretionary
agency have been analyzed here. But there seems to be a closer relationship between the rules that might be selected and the success criterion
adopted than there is between the latter and the pronounced goals of a
discretionary agency.

Can PolicyActivism Succeed?

147

The potential for success of rule-guided macropolicy depends, in large
part, on the absence of policy activism, not only for the removal of the
potential for self-interested behavior on the part of discretionary agents,
but also for the built-in predictability of such action that is inherent in the
notion of rules, as such. The relative advantages of rule-guided policy
over agency discretion could be treated at length, but this effort would
carry me well beyond my assignment in this paper.

8. Fiscal Policy and Monetary Policy
There are two distinct policy instruments, or sets of instruments, in both
the familiar textbook terminology and, indeed, in the overall subject of
this conference: fiscal policy instruments and monetary policy instruments. To this point I have made no distinction between these two sets,
and I have avoided altogether any discussion of relative efficacy as well as
relative vulnerability to the sorts of influences on behavior that are
emphasized in a public-choice approach. It is time to explore some of the
differences that are directly relevant to the arguments that I have
advanced.
Fiscal policy involves budgetary manipulation and, hence, a necessary
linkage between any macropolicy objectives and the whole process of
public-sector allocation. Given this necessary linkage, and given the
institutional—political history, it seems totally unreal to suggest that any
shift of authority over fiscal policy would be delegated to either discretionary or even to rule-bound authority. It seems highly unlikely that
fiscal policy, in any sense, would be removed from the ordinary procedures of democratic decision making, with divided legislative and
executive responsibilities and roles in its overall formulation. It becomes
unrealistic in the extreme to presume that we, in the United States, would
transfer to an agency immune from electoral constraints any authority to
manipulate either side of the budget in accordance with rules or intentions to improve macroeconomic performance. Decisions on tax rates,
spending rates, and, in consequence, deficits and borrowing requirements,
are likely to remain within the responsibility of “democratic” determination, with the predicted result that any meaningful success criterion will
fail to be satisfied. There will be a bias toward “easy budgets,” with
higher-than-desired deficits, to the extent that any considerations of
macroeconomic policy enter the policy argument.4
Given this predicted bias, and quite apart from any consideration as to
the independent efficacy of budgetary policy in effectuating desired
results, any genuine hope for “success” in macroeconomic policy must
involve a reduction or removal of budgetary manipulation from the

148

James M Buchanan

potentially usable kit of tools.5 If “fiscal policy” can be isolated so as to
insure that its operation does not make the task of monetary management more difficult, a major step toward genuine reform will have been
made. It is in this context that the argument for a constitutional rule
requiring budget balance becomes important in macroeconomic policy
discussion.
If fiscal policy is so isolated, the task of policy action is left to the
monetary agency or regime. A monetary agency can be made effective if
the discretion of the agent is limited by the imposition of legally binding
and enforcible rules for policy actions. These rules may take on any one
of several forms, and it would be out of place to discuss these alternatives
in detail here. The monetary agency can be directed to act on the defined
monetary aggregates so as to insure prespecified quantity targets (as in
some Friedman-like growth rule). Or the authority might be directed to
act so as to achieve a specifically defined outcome target, such as the
maintenance of stability in the value of the monetary unit. In either case
the structure of the rules must be such as to invoke penalities for the
failure of the authorities to act in accordance with the declared norms.
Some allowance for within-threshold departures from targeted objectives
would, of course, be necessary.
But only with some such feedbacks in place can the persons in positions
of responsibility as monetary agents be expected to perform so as to
further the success criterion that is implicit in the imposition of the rules.
It seems at least conceptually possible to build in a workable reward—
penalty structure for the compensation and employment of rule-bound
monetary agents. And, in the limiting case, such a reward—penalty structure, appropriately related to the achievement of the desired policy target,
may obviate the need for explicit definition of a rule for policy action.
For example, if the compensations of all employees of the monetary
authority should be indexed so as to insure personal penalty from any
departures from monetary stability, perhaps nothing more need be
required by way of rules. (Such a scheme might involve the maintenance
of fixed nominal salary levels against inflation, and double indexing of
salaries against deflation, or some more sophisticated formulae.)
If no incentive—motivational structure is deemed to be institutionally
and politically feasible, under the operation of any fiat money regime, the
argument for more basic regime shift in the direction of an automatic or
self-correcting system based on some commodity base is substantially
strengthened. The relative advantage of all such systems lies in their
incorporation of market-like incentives to generate behavior that will tend
to generate at least long-term stability in the value of the monetary unit.

Can Policy Activism Succeed?

149

9. Conclusion
In this discussion, as elsewhere, the primary implication of public-choice
theory is that institutional—constitutional change or reform is required to
achieve ultimate success in macroeconomic policy. There is relatively little
to be gained by advancing arguments for “better informed” and “more
public-spirited” agents, to be instructed by increasingly sophisticated
“economic consultants” who are abreast of the frontiers of the “new science.” All such effort will do little more than provide employment for
those who are involved. It is the political economy ofpolicy that must be
reformed. Until and unless this step is taken, observed patterns of policy
outcomes will continue to reflect accurately the existing political economy
within which these outcomes are produced. And we shall continue to
have conferences and discussions about the failures of “policy activism.”

Notes
1. For further elaboration and analysis, see Geoffrey Brennan and James
Buchanan (1980) Chap. 6; and (1981).
2. For further elaboration of the analysis, see Geoffrey Brennan and James
Buchanan (forthcoming), Chaps. 5 and 6.
3. I shall not develop the argument in support of the contractarian—
constitutional criterion for measuring policy success or failure. Let me say only
that such a criterion must be used unless we are willing to introduce external and
nonindividualistic standards of evaluation.
A more controversial position is the one that suggests that the monetary stability criterion would, indeed, be the one that would emerge from the ideally constructed constitutional setting. I shall not develop the argument in support of this
position, although I think it can be plausibly made.
4. For an early statement of this point, see Buchanan (1962). For a more
extended discussion, see James M. Buchanan and Richard E. Wagner (1977,
1978).
5. Keynes and the Keynesians must bear a heavy~responsibility for destroying
the set of classical precepts for fiscal prudence that had operated to keep the
natural proclivities of politicians in bounds. By offering what could be interpreted
as plausible excuses for fiscal profligacy, modern politicians have, for several
decades, been able to act out their natural urges, with the results that we now
observe. For further discussion see Buchanan (1984).

References
Brennan, Geoffrey, and James M. Buchanan. 1980. The Power to Tax. Cambridge:
Cambridge University Press.

150

_____

James M Buchanan

1981. Monopoly in Money and Inflation. London: Institute of Economic
Affairs.
______
nd. The Reason of Rules. Cambridge: Cambridge University Press,
forthcoming.
Buchanan, James M. 1962. Easy Budgets and Tight Money. Lloyds Bank Review.
64: 17—30.
Victorian Budgetary Norms, Keynesian Advocacy and Modern Fiscal
Politics. Prepared for Nobel Symposium on Governmental Growth, Stockholm,
Sweden, August 1984. Center for Study of Public Choice Working Paper No.
4—02.
Buchanan, James M., and Richard E. Wagner. 1977. Democracy in Deficit. New
York: Academic Press.
______ eds. 1978. Fiscal Responsibility in Constitutional Democracy. Boston: Martinus Nijhoff.

______

7
An Appeal for Rationality in the
Policy Activism Debate
John B. Taylor
STANFORD UNIVERSITY

My assignment for this paper is to provide an up-to-date review of the
rational expectations debate about whether activist monetary and fiscal
policies can improve macroeconomic performance. Preparing a review is
particularly difficult at the present time, because we do not seem to be
having much of a debate over policy activism. Looking back over the past
five years since I prepared a similar review paper for this conference series
[see Taylor (1980)], it now seems to me that the debate about policy
effectiveness that raged between rational expectationists and other
macroeconomists during the l970s essentially ended in the early 1980s.
Since then, only a few analytical or empirical studies of alternative policy
proposals have been conducted, and more importantly little effort has
been made to reach agreement among the various proposers, or even delineate specific reasons for disagreement. Clearly a renewal of discussion
on these important issues is in order.
Rather than provide a detailed review of an old debate, this paper
presents a case and outlines a framework for a new debate about policy.
It argues that a good framework for debate is the rational expectations
approach to policy evaluation that emerged from the policy ineffectiveness
debate of the 1970s, but which has been used far too rarely to study other
activist policy issues. The paper includes an outline of the essential aspects
of a rational expectations approach to the policy activism question.

1. The End of the Policy Ineffectiveness Debate
Rational expectations first became a big factor in the policy activism
debate in the early 1970s when Thomas Sargent and Neil Wallace wrote
their famous policy ineffectiveness paper They showed—using an elemen

152

John B. Taylor

tary macroeconomic example based on Robert Lucas’s then new model of
the Phillips curve—that an active monetary policy could not be effective in
stabilizing fluctuations in output and employment Hence a monetarist
constant-growth-rate rule for the money supply, such as the one proposed
by Milton Friedman years before, was optimal: it could not be improved
on by an activist or Keynesian countercyclical stabilization policy.
The Sargent—Wallace paper unsurprisingly ignited a great policy
debate The paper was soon followed up by demonstrations of empirical
support for the Lucas model by Sargent and Robert Barro and extensions
of the Lucas model by Barro and others. Almost all Keynesian
macroeconomists eventually joined in to register their disagreements. The
policy ineffectiveness debate raged for much of the 1970s and completely
replaced the monetary—fiscal policy debate among economists in most
universities.
The early rational expectations proponents of the policy ineffectiveness
view were quite explicit about their analytical framework and the assump
tions that formed the underpinnings of their conclusions. For this reason,
in my view the debate had a relatively high degree of rationality com
pared to many debates about economics. It focused on specific issues of
disagreement In a relatively short period of time bogus or irrelevant
issues had been cast aside and the central reasons for disagreement had
been isolated. Empirical tests of the crucial informational assumptions
underlying the theory also came surprisingly quickly
There is little doubt that the excitement surrounding this policy debate
was responsible for stimulating the great interest in rational expectations
shown by many young macroeconomists dunng the l970s The debate
also stimulated thinking about alternatives to Lucas’s theory of the Phillips curve—alternatives based on contracts and staggered wage setting with
rational expectations in which the policy effectiveness property did not
hold. Econometric techniques were improved in order to go beyond the
simple Sargent—Wallace type model and evaluate policy in large and pos
sibly nonlinear models with rational expectations. Much as the
monetary—fiscal policy debate, which was ignited by Milton Friedman’s
original monetarist proposals, generated empirical and theoretical research
that improved our understanding of macroeconomic fluctuations, the policy ineffectiveness debate had similar positive fallout.
However beneficial, the policy ineffectiveness debate of the l970s is
now over. There is general agreement that it is the market-clearing
assumptions, rather than the rational expectations assumptions, of the
Lucas model that are responsible for the policy results; contract models
with rational expectations introduced by Edmund Phelps, Stanley Fischer,
myself~and others imply that policies that react to the state of the economy can improve macroeconomic performance. These contract models are

An Appealfor Rationality

153

now as much a part of rational expectations as the market-clearing
models. In his textbook Michael Parkin (1984) has accordingly divided up
the rational expectations school into two parts: the “new classical” school
and the “new Keynesian” school. There also seems to be general agreement that the empirical support for the Lucas new classical model is
weaker than the early Barro and Sargent studies showed. There is also
general agreement that the new Keynesian models with rational expectations need some bolstering of their microeconomic foundations. Of course,
others might characterize these areas of agreement somewhat differently.
[See the survey by McCallum (1980) or a more recent one by myself
(1985) for details and references.]

2. The Current Deadlock
One might have expected (as I did) that when the controversy over the
policy ineffectiveness issue became resolved, rational expectations
researchers in macroeconomics would then turn to other important,
though perhaps less exotic, issues in the policy activism debate. Although
there was agreement that the constant-growth-rate rule is not necessarily
optimal, there still was relatively little agreement or even discussion about
what a better rule might look like. Reflecting on the policy effectiveness
debate, Stanley Fischer (1980, 226) noted, “After all, we do not know the
optimal activist policy.” There are many other issues to be resolved: how
would we implement an activist policy rule if that would improve
macroeconomic performance? Karl Brunner (1981) has raised questions
about this practical issue. Can one deal in practice with the serious problem of lags and uncertainty in the effect of policy that the proponents of
constant-growth-rate rules emphasize? Milton Friedman (1984) still feels
that this is the fundamental problem with activist policy:
slow, steady, monetary growth. That is not a necessary implication of
monetarist theory. A believer in monetarist theory still can favor an activist
monetary policy as a way to offset other changes in the economy..
[however] the monetary authorities have typically made matters worse.
they
have been a source of uncertainty and instability in the economy.
. .

. . .

These are important areas of controversy in the policy activism debate
that have not been resolved and about which there is little consensus or
agreement. Yet serious research and evaluation of alternative policy rules
using the rational expectations techniques that proved useful in resolving
the earlier issues (or, for that matter, using any other analytical frame-

154

John B. Taylor

work) is not underway at anything like the scale of research that we saw
in the policy ineffectiveness debate. The policy activism debate has not
moved in the direction that one would have thought.
I am not sure why this is so. Perhaps the apparent political success of
supply-side economics discouraged those who thought scientific research
in economics could have a hearing among policymakers. Perhaps the
costly 1980—82 disinflation disillusioned some enthusiasts of the rational
expectations assumption, though I do not feel it should have. Perhaps
constant talk of budget deficits has made it difficult to concentrate on discussions about policy activism or made one feel terribly impractical in
searching for long-run policy reforms.
Of course, policy talk has not stopped, and there have been interesting
proposals for new, and not so new, policy rules to replace the monetarist
constant-growth-rate rule: price rules, nominal GNP rules, interest-rate
rules, gold-standard rules. Indeed there is now more talk and proposing
than ever. The problem, in my view, is that there has been little attempt
to evaluate these proposals within a theoretical or empirical framework
that is specific enough to be criticized, debated, and eventually used to
resolve disagreements.
This no-debate situation is troublesome at a time when there is a clear
need for some consensus among macroeconomists. Lester Thurow (1983,
xv) expresses what is probably a commonly felt view: “The current intellectual disarray among economists is matched only by a parallel time of
confusion during the early days of the Great Depression.” The old Keynesian consensus is clearly gone, but nothing has yet replaced it. The lack of
such a consensus leaves the economy vulnerable to economic policy
actions based on little theoretical or empirical support. In his recent book
on policy Herbert Stein (1984, 324) expresses the situation more passionately but no less accurately:
Although there is much talk about economic policy there is no debate. People say what they have always believed, or what they find it convenient to
say, but there is no confrontation of the arguments. There is no effort to
find the sources of disagreement or to reach agreement, perhaps because the
participants think that the effort to change minds and reach agreement is
hopeless. Talk about economic policy has become only a way of rallying
one’s own troops.

3. What Is the Rational Expectations Approach?
If the rational expectations approach is to provide a suitable framework
for debating policy, it is necessary to have a general understanding of the
approach. Despite numerous conferences and survey papers, there is still

An Appealfor Rationality

155

great confusion—especially among noneconomists and economists outside
universities—about what the rational expectations approach to policy is.
Consider, therefore, the following five general principles that I think summarize the rational expectations approach to macroeconomics.
First, people areforward-looking, and their future expectations can be
modeled reasonably accurately by assuming that they have learned the
basic statistical regularities of the business cycle, and they use this
information to make unbiased (but not error-free) forecasts.

This, of course, is just the Muth definition of rational expectations
applied to macroeconomics. It seems like a reasonable assumption for
macroeconomic applications because many features of economic fluctuations are recurrent from one business cycle to another; there are established statistical regularities. Since business cycles have been observed
for hundreds of years, it makes sense to assume that people have become
familiar with them. Such a forward-looking unbiased forecasting assumption would not be reasonable for new unprecedented events for which
there is no experience.
Second, macroeconomic policy should be stipulated and evaluated as a

rule, rather than as one-time changes in the policy instruments.
Because people are assumed to be forward-looking, their expectations
of future policy actions affect their current behavior and the state of the
economy. Hence, in order to evaluate the effect of policy on the economy,
we need to specify not only current policy changes but also future policy
changes. In other words we need to specify a contingency plan that
describes how policy will react to future events. Such a contingency is
nothing more than a rule for policy. Of course, the contingency plan
could specify a constant-growth-rate rule for the money supply, but more
generally there will be some reaction from the state of the economy.
The rational expectations approach forces one to think about policy as
a rule or a strategy. Once you are working with a rational expectations
model, you soon realize that you have little choice but to specify policy as
a rule. My own experience is that I have naturally specified policy rules in
rational expectations policy evaluation studies without much thought
about it one way or the other. This practical reason for thinking about
policy as a rule does not seem to have been mentioned in the early discussions of rules versus discretion, but it does support the case for rules over
discretion.
Note that the focus on rules does not mean that the effect of one-shot
changes in policy should never be calculated with a rational expectations

156

John B. Taylor

model. Such a calculation can be a useful thought experiment to help
understand the workings of the model, but it is, of course, necessary to
specify whether the change is anticipated or unanticipated, as well as
whether it is temporary or permanent.
The famous critique of econometric policy evaluation put forth by
Robert Lucas (1976) is the technical side of this principle. Lucas showed
that traditional econometric models would give incorrect answers to policy
evaluation problems if expectations were forward-looking and there was a
change in the policy rule. Since these traditional models were based on
adaptive backward looking expectations their parameters would change
when the policy rule changed. This was the negative part of the critique
and has clearly made policy analysts wary of using the traditional models
But there was also a positive side The Lucas critique provided a general
framework for modifying the traditional models by stipulating policy as a
rule it is possible to calculate by how much the parameters of the traditional models would change. Much technical econometric research by
Thomas Sargent Lars Hansen and others has been devoted to developing
such a framework.
Christopher Sims (1982) has recently argued that the focus of the
rational expectations approach on alternative policy rules is irrelevant He
argues that we rarely get big changes in rules anyway so that we might as
well use reduced forms or conventional econometric models for policy. It
is true that there is a utopian flavor to the rational expectations approach.
The search is for big policy reforms that would improve economic welfare
over a long period of time The reforms would probably require changes
in the policy-making institutions or the creation of new institutions. Such
reforms are, by their very nature, rare. But they do occur. The creation of
the Federal Reserve System, the departure from the gold standard, and
the shift to floating exchange rates are all examples. These reforms seem
to have had substantial effects on the economy. A careful analysis of the
effects of future policy reforms therefore seems quite relevant.
Third, in order to get the benefits of a particular policy rule, it is
necessary to establish a commitment to that rule.

As was first pointed out by Finn Kydland and Edward Prescott (1977),
dynamic models with rational expectations can lead to problems of time
inconsistency. They discovered this problem while attempting to compute
optimal policy along the lines suggested by Lucas In a dynamic model of
investment and in a Phillips curve model, they found that once policymakers began on an optimal policy there was incentive in future
periods for the policymakers to change the plan—to be inconsistent. Policymakers could make things better by being inconsistent This was true

An Appealfor Rationality

157

even if the welfare function of policymakers was identical to that of people in the economy and did not change over time. However, by being
inconsistent the policymakers would be likely to lose credibility; people
would begin to assume that the policymakers would change, and this
would lead to a new policy-making equilibrium that was generally inferior
to the original policy plan of the policymakers. The implication is that to
prevent this inferior outcome it is better to maintain a firm commitment
to a policy rule.
There is a nice macroeconomic analogy to the macroeconomic time
inconsistency problem: patent laws. By promising a patent to inventors,
the patent laws stimulate inventive activity. Once a particular invention
has been made, however, it is tempting to break the commitment and not
give a patent. A policymaker who had the discretion to award patents
each year would indeed be tempted not to do so. By holding back the
patent, we avoid the economic inefficiencies of a monopoly. Fortunately,
reneging on patents does not occur in practice because it is so clear that
future inventive activity would suffer. As a result, we have patent laws
that limit such discretion. The time inconsistency research suggests that
discretion should be limited for similar reasons in macroeconomic policy.
It should be emphasized that evaluating policy as a rule does not prevent
time inconsistency. There still may be temptation to change the rule. The
commitment to the rule is the important feature of this third principle.
The previous two principles together imply that a rational expectations
analysis of “activist” policies is actually an analysis of policy rules with
feedback from the state of the economy to the policy instruments. There
is a big distinction between “discretionary” and “activist” policies. Those
in favor of discretionary policy disagree with the whole concept of a ruleof-the-game approach, whether the rule is a feedback rule or a constant
setting for the policy instruments; discretionary policy is formulated on a
case-by-case and year-by-year basis with no attempt to commit or even
talk about future policy decisions in advance. Activist and constantgrowth-rate policy rules have much more in common with each other than
do activist policy rules and discretionary policy. Both types of policy rules
involve commitments and lead to the type of policy analysis suggested by
the rational expectations approach.
The next two principles are related to the types of economic models
typically considered by rational expectations economists and to the factors
they consider in determining whether a policy is good or not. On these
two principles there is more variety among the different departments of
the rational expectations school than there is on the first three principles.
Fourth, the economy is basically stable; after a shock the economy will

eventually return to its normal trend paths ofoutput and employment.

158

John B. Taylor

However~because of rigidities in the structure of economy, not in

expectation formation, this return may be slow.
Formal rational expectations models of economic fluctuations are usually dynamic systems continually disturbed by stochastic shocks. After
each shock the economy has a tendency to return to a normal or natural
growing level of output and employment, although there may be
overshooting or a temporary cumulative movement away from normal. A
smooth return is never observed in practice, however, because new shocks
are always hitting the system. Since the economy is viewed as always
being buffeted around by shocks, rational expectations economists must
calculate a “stochastic equilibrium” rather than a “deterministic equilibrium” to describe the behavior of the economy. The combination of the
stochastic shocks and the dynamics of rational expectations models is
capable of mimicking the actual behavior of business cycles surprisingly
well. The properties of the stochastic equilibrium are much like the actual
behavior of business cycles.
The shocks can be due to many factors, but they usually have been
portfolio preference shocks, productivity shocks, or price shocks. The
dynamics are due to many possible rigidities in the economy, but price—
wage rigidities and slow adjustment of capital (including inventories) have
been the most important empirically.
Because of these rigidities, the impact of a shock to the economy takes
time to sort itself out. Suppose, for example, that there is a shift in money
velocity with people demanding to hold more money at any level of
income and interest rates. Eventually the price level will fall so that the
real supply of money is effectively increased, but if there are wage and
price rigidities, this adjustment will take time: first the increase in money
demand will cause an increase in interest rates; the higher level of interest
rates will in turn depress demand for durables and have repercusions
throughout the economy; depressed demand conditions will then begin to
put downward pressure on prices; the fall in prices then will begin to raise
the real supply of money; these prices will continue until the economy is
back to its natural level of output and employment. The whole process
could take more than a year.
Combined with these structural rigidities is the supposition that expectations are not restrained by similar rigidities. A shock can change expectations of inflation, exchange rates, and other variables overnight, even
though there are rigidities that cause the economy to take additional time
to fully adjust to the shock. The expectations take account of the structural rigidities, since these are part of the model. This combination of
rigidities in the economy with perfectly flexible expectations is an essential
feature of most rational expectations models. There has been a tendency

An Appealfor Rationality

159

to get expectations assumptions mixed up with structural assumptions
about how markets work. Hence, the comment that expectations might be
rational in flexible auction markets but not in sticky labor markets is frequently heard. These two types of assumptions should be usefully
separated. (Again recall that rational expectations are meant to apply to
recurrent events, not to unprecendented events. In response to a new
event or a new policy rule, slow adjustment of expectations would be
likely.)
There has been much research on price and wage rigidities in rational
expectations models. The important general feature of this research is that
prices and wages have a forward-looking feature, whether they are sticky
or not. When workers and firms set wages and prices, they look ahead to
the period during which the prices or wages will be in effect—to demand
conditions, to the wages of other workers, and so on. This means that
expectations of future policy actions will affect wage and price decisions, a
property that is quite unlike Keynesian models of wage and price rigidities. The view that the economy will eventually return to normal—
however slowly—after a shock is also inconsistent with the Keynesian view
of permanent underemployment equilibria.
Fifth, the objective of macroeconomic policy is to reduce the size (or

the duration) of the fluctuations of output, employment, and inflation
from normal or desired levels after shocks hit the economy. The objective is to be achieved over a long period of time that will, in general,
include a large number of business cycle experiences. Future business
cycle fluctuations are not viewed as less important than the current
one.

By responding to economic shocks in a systematic fashion, economic
policy can offset their impact or influence the speed at which the economy
returns to normal. It thus can change the size of the fluctuations. How this
should be done is a main subject of disagreement among proponents of
different policy rules.
From a technical viewpoint the disagreement can be addressed by
inserting alternative policy rules into a rational expectations model and
calculating how each rule affects the variance of output, employment, and
inflation in the stochastic equilibrium that describes the business cycle
fluctuations. We want to choose a policy that provides the best economic
performance as approximated by this stochastic equilibrium. One simple
criterion is the minimization of the variance of output and inflation. Since
in many models with price and wage rigidities there will be a trade-off
between the reduction of output and inflation variability, it will usually be
necessary to stipulate a welfare or loss function that reflects certain value

160

John B. Taylor

judgments Frequently one policy will so dominate another than the par
ticular welfare weights do not matter much, however.
Despite the need to make such value judgments the rational expectations approach is fairly specific about what the objectives of policy should
be. Changing the natural or normal levels of output and employment is
not the direct objective of stabilization policy from a rational expectations
perspective; of course, it is possible that reduced variability of output or
inflation could raise the secular growth rate of the economy or reduce the
natural rate of unemployment. As a first approximation, these normal 1evels are not influenced by macroeconomic policy. The secular growth rate
of the economy is influenced by tax policy and by the mix between fiscal
and monetary policies. But it is the average setting of these instruments
rather than their cyclical variations that is most important for long-term
growth.
The average rate of inflation can obviously be influenced by monetary
policy, and it is important to choose a target rate that maximizes
economic welfare. The objective of macroeconomic policy, however, is to
keep the inflation rate close to this target rate; that is, to minimize fluctuations around the target, regardless of what the actual value of the target is.
Alternatively if a zero inflation target is appropriate the objective of po1
icy is to keep the price level near some target; the specific target value
itself is much less important.

4. Some Proposals for Activist Policy Rules
Although there have been too few analytical or empirical investigations of
activist policy alternatives to monetarist rules, there clearly have been
some. Mention of a few here may serve as a departure for discussion. I
restrict myself to proposals being investigated by two of my
macroeconomist colleagues at Stanford, Robert Hall (1984) and Ronald
McKinnon (1984), as well as myself. Since the proposals are not exactly
alike, there is room for discussion, and the examples are obviously not
offered here as the final word.

4.1. An elastic price rule
Hall (1984) considers a policy rule in which the Fed manipulates its policy
instruments in order to keep the deviations of the price level from its target level equal to eight times the deviation of the unemployment rate
from its target level. The figure eight is chosen as an example; more generally, the exact number would be chosen after public discussion. The

An Appealfor Rationality

161

objective of the proposal is to stabilize fluctuations in both unemployment
and the aggregate price The level of unemployment rate is taken as given
and equal to the natural rate Equivalently the policy attempts to estab
lish an aggregate demand curve (in price output space) that is steeper
than a monetary rule or a nominal GNP rule and thereby less tolerant of
output fluctuations than a monetarist rule.
Clearly this proposal fits into the policy evaluation framework outlined
in the previous section. The emphasis is on cyclical fluctuations over a
long period of time the target level of unemployment is assumed to be
unaffected by policy and the policy evaluated is a rule Discretion is not
completely eliminated, however, because the Fed must decide the
appropriate instrument setting to achieve the rule but Hall does consider
the problems of stating the rule in terms of magnitudes that the Fed does
not directly control The policy is evaluated by using a dynamic stochastic
framework like that described under principle 4.
4.2. An exchange rate rule
McKinnon (1984) has been investigating an activist policy rule in which
the Fed increases the growth rate of the money supply whenever the dol
lar exchange rate appreciates (relative to some target) against other hard
currencies particularly the mark and the yen A lower exchange rate calls
for a reduction of the money growth rate. Similarly the rule calls for the
Bundesbank and the Bank of Japan to increase the growth rates of their
money supplies whenever their exchange rates appreciate In this sense
the rule involves policy coordination between the countries. There is also
coordination in maintaining agreement on the long-run trend path of the
world money supply or at least the group money supply for the United
States, Germany and Japan.
This exchange rate rule is designed to offset portfolio preference shocks
that McKinnon views as arising partly via currency substitution between
countries, and his analytical framework is directed toward such shocks.
Although rational expectations is not entered explicitly into this framework, the quick movement of forward-looking exchange rate expectations
in the face of rigidities elsewhere in the econom)’ is one of the motivations
behind focusing policy on the exchange rate.
4.3. An activist money-supply rule
In my own research I have investigated the properties of an activist
money-supply rule that reacts to the state of the economy. Although a
complex optimal rule was calculated for a particular rational expectations
model using the overall approach outlined here, that rule turned out to be

162

John B. Taylor

remarkably similar to a simpler rule in which the growth of the money
supply is increased whenever real output is below its trend growth target;
and by a little bit more when output is falling relative to its trend growth
target. As a close approximation, the rule involved no accommodation of
the money supply to inflation shocks. Hence, monetary policy has a stabilization role but no accommodation role. According to this framework this
specific ‘activist rule would work better than a monetarist rule.
An alternative to this proposal would have the stabilization role of
monetary policy given over to a fiscal policy rule similar to the automatic
stabilizers. This would make the monetary authorities responsible only for
maintaining a fixed money-growth rate, which could reduce temptation to
accommodate inflation. It would also prevent monetary policy from having international repercussions when attempting to react to domestic poiicy disturbances. However, allocation of all stabilization policy to a fiscal
policy rule might require some explicit attempt to deal with interestsensitive investment demand.

5.

Concluding Remarks

My aim here has been to present a rational expectations framework
within which a number of issues in the policy activism controversy might
be fruitfully discussed and debated. The hope is that such a framework
might bring more rationality to a debate that now seems to be in a slump.
The framework involves a number of specific features that I think are reasonable and on which there might be some agreement, but it is by no
means a straitjacket. It leaves plenty of interesting modeling questions
open to the investigator of a particular problem.

References
Brunner, Karl. 1981. The case against policy activism. Lloyds Bank Review No.
139.
Fischer, Stanley. 1980. On activist monetary policy with rational expectations. In
Rational Expectations and Economic Policy, ed. S. Fischer. Chicago: University
of Chicago Press.
Friedman, Milton. 1984. Has monetarism failed? Manhattan Report 4: No. 3. New
York: Manhattan Institute for Policy Research.
Hall, Robert E. 1984. Monetary strategy with an elastic price standard. Federal
Reserve Bank of Kansas City Conference on Monetary Policy.
Kydland, Finn, and Edward C. Prescott. 1977. Rules rather than discretion: The
inconsistency of optimal plans. Journal of Political Economy 85: 473—91.
Lucas, Robert E., Jr. 1976. Econometric policy evaluation: A critique. Carnegie—
Rochester Conference Series of Public Policy, Vol. 1. Amsterdam: NorthHolland.

AnAppealfor Rationality

163

1980. Rules, discretion, and the role of the economic advisor. In Rational
Expectations and Economic Pollcy~ed. S. Fischer. Chicago: University of Cliicago Press.
McCallum, Bennett t 1980. Rational expectations and macroeconomic stabiliza.
tion policy: An overview. Journal ofMoney~Owdlt and BankIng 12: 716—46.
McKinnon, Ronald L 1984. An International Standardfor Monetary Stabilitatlon.
Washington, D.C.: The Institute for International Edonomics. (Distributed by

M.I.t Press.)

Parkin, MichaeL 1984. MacroeconomIcs Englewood Cliffs, NJ.: Prentice-Hall.
Sims, Christopher. 1982. Policy making with econometric models. Brookings
Papers on Economic ActIvity 1.
Stein, Herbert. 1983. P?esldentlal Economics: The Making ofEconomic Policyfrom
Roosevelt to Reagan andBeyond. New York: Simon and Schuster.
Taylor, John B. 1980. Recent developments in the theory of stabilization policy.
Stablitatlon Policy: Lessonsfor the 197()’s and Implications for the 1980c, ed. L.
Meyer. Federal Reserve Bank of St. Louis.
1981. Stabilization, accommodation, and monetary rules. American
Economic RevIew 70: 145-49.
___n.d. Rational expectations models in macroeconomics. In Frontiers In
Economlcs~S. K. Arrow and S. Honkapojha. Oxford: Basil Blackwell. Forth-

coming.
Thurow, Lester. 1983. Dangerous Currents: The State of Economics. New York:
Random House.