View original document

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

FEDERAL RESERVE BANK OF NEW YORK

251

T h e C o n trib u tio n s and L im itation s o f "M o n e t a r y ” A n a ly sis

By P a u l A. V o l c k e r
President, Federal Reserve Bank of New York
Remarks before the American Economic Association
and the American Finance Association in
Atlantic City, New Jersey, on Thursday, September 16,1976
The invitation to address this annual joint luncheon
of the American Economic and Finance Associations is
a special honor. But for one engaged in policymaking
it also presents a special challenge. He may like to think
of himself as a practical man but certainly not, as Keynes
once put it, the “slave of some defunct economist”. At
the same time, he can hardly regard himself as “quite
exempt from intellectual influences”.
In that spirit, I would like to take this opportunity to
consider some of the approaches and practices of central
banking in the light of modern economic analysis.
Now, I fully realize that neither central banks as a
genus nor the species Federal Reserve— nor even that
special variety known as the Federal Reserve Bank of
New York—have had a reputation of moving in the van­
guard of professional opinion. Nor would I apologize for
a certain intellectual conservatism. What we do must take
account of human attitudes and institutional settings that
necessarily change slowly. Sorting out what is true and
valid from what is fashionable is never easy, and we have
no laboratory apart from the American economy itself.
Yet, as one who spent almost twenty years outside the
Federal Reserve before returning last year, I can testify
directly about how much has changed over that period.
I learned my economics and my central banking in the
first full flush of the General Theory. Perhaps symbolic of
that influence, in the mid-1950’s it was still something
of a challenge to calculate a meaningful money supply
series from the mass of statistics issued by the Federal
Reserve, while markets hung on the latest release of data
on free reserves or bank loans. Today, the situation is
almost reversed. Our computers spew out “M’s” in seem­
ingly infinite variety and with great rapidity. Meanwhile,




analysis of the asset side of financial balance sheets seems
relegated pretty much to a few specialists— or to bank
examiners and the SEC.
We need not look to “defunct economists” to help
explain the change, but to a school of thought that is very
much alive and well!
I know one always treads on dangerous ground in using
a shorthand label to describe any school of economic
thought. I will therefore run a risk of oversimplification
and even injustice in characterizing some of the views of
the monetarists today— assessing the contributions and
limitations of that analysis. But I do not think there can
be much doubt that that school, for all the differences
within it, has helped bring a distinctly different flavor to
much macroeconomic policymaking and analysis in recent
years.
Certainly, it has helped bring a new focus on the rele­
vance of monetary policy: the proposition that the stock
of money does matter. To be sure, relatively few econo­
mists— and almost no central bankers— have ever openly
argued the opposite proposition. But implicitly or ex­
plicitly, there was a rather common assumption two or
three decades ago that, while the money supply did have
an effect on credit markets and interest rates, the induced
effects on the economy were not terribly powerful in most
situations. Changes in the supply of money moved us
along a rather elastic liquidity preference schedule, and
the investment demand function was thought to be rela­
tively insensitive to interest rates. We therefore need to
look at fiscal actions and to other exogenous forces as
the main determinants of economic activity.
As I shall suggest later, the idea that, at least in the
short run, the supply of money and interest rates are

252

MONTHLY REVIEW, OCTOBER 1976

related still seems relevant today. But the monetarists
have usefully emphasized the danger of confusion between
nominal and real rates and the role of price expectations.
They have forcefully made the case for the view that in
the long run velocity is not related to the stock of money
and that, in the same long run, an excess supply of money
contributes not to real income or wealth but simply to
inflation.
That latter point is, of course, one of the oldest propo­
sitions in the history of economic thought. But there is
no doubt that too often we have lost sight of it amid the
urgent search for solutions to immediate policy problems.
The further extensions of the idea— that the rate of
monetary expansion can have relatively little effect on
the real rate of interest over time and hence on the mix of
consumption and investment— are controversial in their
more extreme form. But certainly there is more awareness
today of the real limitations on the possibilities for manipu­
lating the mix of fiscal and monetary policies to achieve
our objectives.
More generally, while the insight is hardly confined
to monetarists, modern analysis has typically emphasized
the length and probable variability in the lags between
policy action and the effect on the economy. As a result,
there is less faith in our ability to make short-term
adjustments— to “fine tune” the economy.
These lessons have not been lost on central banks in
the United States or elsewhere. In shaping their policies
and policy pronouncements, monetary officials have pro­
vided tangible evidence of the new emphasis in the
greater prominence given the behavior of broad monetary
aggregates.
At the same time, central banks have long shared an
understandable human interest in wanting to hedge against
an uncertain future. They want to retain the ability to
respond flexibly to emerging developments, to probe ex­
perimentally with new policy measures, to test market
reactions, and to learn from those reactions before fully
committing themselves to new directions. Indeed, this
flexibility to act and react has long been considered a
great strength of monetary policy. Concern that a needed
degree of flexibility might be impaired accounted, I be­
lieve, for some initial reluctance by the Federal Reserve
in adopting the practice of publicly specifying explicit
goals or targets for monetary aggregates for any substan­
tial period of time ahead.
More than a year ago, however, responding to Con­
gressional intent, the practice of each quarter announcing
such targets a year ahead was adopted, always retaining
the right to change the targets in the light of emerging
developments.




From my viewpoint, this experiment in “practical
monetarism” has proved useful. It has assisted in commu­
nicating our intentions both to the political authorities and
to the marketplace. I suspect it has provided a focus for
more informed and constructive public debate. Indeed,
I am hopeful that, by clarifying the nature of the policy
choices and dilemmas and by more clearly relating today’s
decisions to a longer term horizon, the temptation to en­
gage in more purely political debate about policy choices
has been moderated.
But most important, I think, is the discipline it provides
for our own debate within the Federal Reserve. In my
experience, each of our short-term decisions has needed
to be justified and rationalized in our own minds against
our earlier and broader judgment about what growth in
money seems appropriate over a longer period. The pres­
sure to react, and the temptation to overreact, to each
new piece of information must be filtered through that
earlier judgment and longer perspective.
The Federal Reserve is not alone among central banks
in adopting that sort of approach. In that sense, we have
all been influenced by the monetarist debate. But a con­
sensus on the usefulness of that approach does not, of
course, imply consensus on the substance of policy: just
where the targets should be set, the circumstances under
which they might be changed or temporarily set aside,
and the degree of importance accorded other variables
including interest rates. Moreover, there is no general
agreement on which monetary aggregate is most relevant
— a matter of some importance since both the trend and
short-term fluctuations frequently diverge.
Policy is made up of a succession of short-run de­
cisions. In making those decisions we face the simple fact
that, whatever the stability in the relationship between
money and nominal income in the longer run, there is
considerable instability in the relationship over time hori­
zons relevant to policymakers. Certainly the relationships
between money, interest rates, and nominal income have
been unusual over the year or so since I rejoined the
Federal Reserve. Specifically, over the first year of an
economic recovery that has proved very close to the
average of postwar recoveries, the velocity of Mi grew
substantially more rapidly than history or most econo­
metric analysis would have suggested, taking account of
the stability of interest rates. Indeed, the phenomenon of
stable or even declining interest rates alone is highly
unusual during the first year of recovery.
Suppose an approach had been followed since the
spring of 1975 that sought to set aside judgment in favor
of the statistical rule book. Presumably a monetary target
would have been set significantly higher than the roughly

FEDERAL RESERVE BANK OF NEW YORK

5 percent growth that actually occurred in Mb assuming
of course a desire to achieve a similar pattern of growth.
Those who, in contrast to the monetary school, empha­
sized last year the desirability of roughly stable interest
rates to promote vigorous recovery have seen that objec­
tive materialize. But members of this group typically
grossly overestimated the monetary growth that would
prove consistent with that scenario. I can only conclude
that, in periods such as that we have just been through,
we need to be alert to possible shifts in the demand for
money. Movements in interest rates are an essential
source of information about those shifts in money demand
and other relevant developments. At times, they remain a
useful, if not uniquely useful, guide to appropriate policy.
Recognition of the broad relevance and desirability of
longer term monetary targets also has left unresolved
important tactical issues as to just how these targets should
be achieved. This is a matter vigorously debated by
monetary economists out of concern that the choice of
technique biases the result. I reveal no secret when I say
that the subject returns again and again in the discussions
of the Federal Open Market Committee (FOMC) and is
a major preoccupation of the work of the supporting staffs
at the Board and the Reserve Banks.
The Committee’s record of policy actions, now re­
leased about a month after each FOMC meeting, reflects
the results of discussion of the appropriate tactical
approach adopted by the Committee at each meeting.
(I might note in passing that the amount of information
provided in these records probably sets a standard among
the major central banks of the world, and represents a
degree of openness entirely unknown to a central banker
of an earlier generation.)
These policy records show that, while the precise ap­
proach varies with circumstances, recent practice typically
involves numerical “tolerance” ranges for key monetary
aggregates for the period immediately ahead. While in­
fluenced by the immediate economic circumstance, these
ranges are designed to be generally consistent with the
one-year targets, allowing for the short-term volatility
of the numbers and expectations about their near-term
behavior. A range is also established for the Federal funds
rate, taking into account the evidence we have about the
interest rate-money supply relationship. Then the Open
Market Account Manager has the job of providing re­
serves on a week-by-week basis at a rate that is expected
to produce a Federal funds rate (and related money
market conditions) within the given range, typically mov­
ing higher or lower within that range as the aggregates
appear to be relatively strong or relatively weak in terms
of the objectives for those magnitudes.




253

I have not found anyone in the Federal Reserve who
is wholly satisfied with this technique. To me, one problem
is that it has encouraged a high degree of sensitivity
throughout financial markets to even relatively small and
potentially transient movements in the Federal funds rate,
because this rate is felt by the market to reflect so heavily
official intentions. But the relevant question, as always,
is not whether the present technique is problem free, but
whether more satisfactory approaches can be devised.
No doubt, improvements are possible and will come. But
no one should be under the illusion that any tactical
change will end controversy that, in the last analysis,
stems more from different judgments about relevant policy
variables than about operating techniques.
The proposal is frequently made that the Federal
Reserve would be more successful in achieving desired
aggregates within relatively short periods if it simply
adopted a target path for bank reserves, the monetary
base, or some variant thereof. These reserve magnitudes
(at least those exclusive of member bank borrowings) are
more or less directly under our control, and they can be
related to the money supply by projections of the “money
multiplier”. Usually, the concept is that these targets
would then be adhered to, almost regardless of short-term
money market implications.
One technical question arises immediately. While I do
not pretend to econometric expertise, I do know that a
massive amount of research has been conducted in this
area. The apparent result is that the relationship between
money and reserve aggregates, particularly in the short
run, appears no more reliable than the relationship be­
tween interest rates and money. In either case— whether
one uses money market conditions or reserve measures
as the immediate tactical targets— one comes up against
two hard facts: first, the monetary aggregates are going
to be subject to considerable short-run uncertainty and,
second, changes in the week-to-week tactical targets will
have their impact on monetary aggregates only with a sig­
nificant (and uncertain) lag.
Let me be more explicit. When short-term tactical
objectives are couched in terms of money market con­
ditions, it is necessary to forecast what the demands for
the various categories of bank deposits and currency are
likely to be under given money market conditions. Alter­
natively, if the short-term tactical procedures are couched
in terms of some reserve aggregate, such as nonborrowed
reserves, it is necessary to forecast the reserve-deposit
multipliers for the various monetary aggregates— and of
course one must also successfully forecast and offset mar­
ket factors affecting reserves in order to hit the reserve
target.

254

MONTHLY REVIEW, OCTOBER 1976

We have techniques to make the needed forecasts with
both the interest rate and reserve approaches. The trouble
is the forecast errors are large no matter what procedure
is used, particularly over periods of one to three months.
Indeed, unimpressive as they are, I am told some of the
correlations observed in historical data between reserve
measures and monetary measures would prove to be spuri­
ous under a regime of rigid reserve targeting.
These uncertainties are likely to make precise monetary
control elusive under any set of procedures. A common
characteristic of the two approaches is that the effect of
changes in either operating target— interest rates or re­
serves— on the various monetary aggregates takes time
to have its full impact, and the largest impact is not the
closest time horizon.
The relevancy of these twin problems of forecasting
errors and lags— whatever the tactical approach— is that
we must constantly balance the danger of w/iJerreacting
to deviations of the aggregates from target paths against
the danger of overreacting. Clearly, there are risks in not
responding to bulges or shortfalls in the money supply
relative to objectives. For example, if growth in the mone­
tary aggregates falls short of objectives, but the shortfall is
treated as a momentary aberration and no action is taken,
a cumulative shortfall may develop, making it harder to
retrace our steps. At times, the bulges or shortfalls may
reflect important underlying developments, such as an
unforeseen change in business activity that we would ig­
nore at our peril.
But the danger of overreacting to deviations in the
aggregates from targets is just as real. Statistically, there
is a high probability that any deviation from target— even
of considerable size— will prove temporary. Attempts to
respond immediately by shifting reserve availability and
allowing the money market abruptly to tighten or ease
could therefore easily result in whipsawing of the market.
More confusion than light might be thrown on our inten­
tions as short-term gyrations in open market operations
obscure any more sustained strategy.
The problem is not a negligible one if one thinks in
terms of a really substantial month-to-month smoothing.
Since only a relatively small fraction of the impact of a
given move in reserve availability or money market con­
ditions is reflected in th6 behavior of the monetary ag­
gregates in the short run, very large movements in reserves
and money market conditions might be needed to correct
short-term aberrations. Worse, the lagged effect of these
moves might then have to be offset by even larger move­
ments in the opposite direction in the subsequent period—
a process that could easily lead to a serious disruption of
the whole mechanism.




To take a recent example, it is not easy to contemplate
what degree of money market tightness might have been
needed to prevent the 15 percent rate of Mx growth that
emerged this past April— or the implication of that degree
of tightness for growth in subsequent months as lagged
effects continued to be felt. Similarly, one wonders if the
outright declines in Mx that have occurred in some indi­
vidual months could have been prevented consistent with
any positive Federal funds rate or, alternatively, through
any feasible injection of nonborrowed reserves within that
month.
I recognize that few, if any, still seriously push the
need or practicality of keeping monetary growth rates on
track month by month. The significance of these response
lags comes in a somewhat longer run context. But the
general proposition remains: there are risks in quickly
adjusting our tactical sights, and risks in delay, when the
aggregates move off course.
I know of no purely mechanical procedure to avoid
these risks— to ensure just the right degree of respon­
siveness to deviations from targets. Whether and how
much to respond will, I think, always be a difficult matter
of judgment and will not be helped much by choice of
tactical approach.
Obviously, the search for improved tactical techniques
will and should go on. Perhaps the continuing effort to
achieve better econometric models of the markets through
which open market policy operates will help, although I
must say frankly that the experience we have had does
not encourage me to expect any startling breakthroughs.
There may be alternative ways of formulating and pre­
senting longer term targets that would improve upon pres­
ent procedures. Even on the basis of what we know now,
we need to consider carefully ways in which reserve tar­
gets could be more extensively used as part of our tactical
procedures; indeed, the FOMC has done extensive work
on this issue in recent years.
As we immerse ourselves in these tactical questions,
however, we need to realize the larger question is not
tactical but substantive— how much weight to put on the
monetary aggregates as opposed to other considerations.
Concentration on the problems of chasing aggregate tar­
gets should not cause us to neglect their limitations.
I have already suggested that the normal relationships
between the aggregates and the economy can break down
over time horizons long enough to be highly significant
for policy formulation. There are also times when market
conditions may deserve attention in their own right. One
thinks immediately of those occasions when markets are
unusually disturbed to the point that a potential impact
on business sentiment and financial availabilities cannot

FEDERAL RESERVE BANK OF NEW YORK

be ignored. At other times, relatively small changes in
the apparent posture of the Federal Reserve may trigger
undesired expectations in the market out of proportion
to any presumed gain in tracking monetary targets. I
think, too, we have seen plenty of evidence of the poten­
tial sensitivity of international financial markets to interest
rate differentials—that floating exchange rates cannot by
themselves eliminate that dimension of policy concern.
More broadly, I think the intellectual emphasis on
monetary aggregates that developed through the 1960’s
threatened and, on some occasions, did go too far in
implying that credit markets, broadly defined, “don’t
count”—that they are never or seldom a source of dis­
turbance in the economy or a legitimate concern of policy.
Indeed, I suspect the relatively little attention directed
toward serious and systematic analysis of the role of credit
markets, toward the financial complexities of the economy
generally, and toward their disruptive potential is a com­
mon failing of most modern theorizing, regardless of the
intellectual starting point.
We have had many occasions in the 1970’s to pay the
closest possible attention to particular financial problems
and to the potential vulnerability of various credit mar­
kets. I would remind you of the recurrent concerns about
thrift institutions and the mortgage market, Penn Central
and commercial paper, Herstatt and the Euro-dollar mar­
ket, New York City and the municipal bond market, and
the rising level of commercial bank loan losses a year
ago. Some of these situations had in them the potential
for grave problems. Happily, they have been contained
and dealt with through a variety of techniques, more or
less of an ad hoc nature.
But is it sheer coincidence that so many of these prob­
lems have arisen in so short a period? And what is the
present significance of such phenomena as the shifting
proportions of debt and equity for the nature and strength
of our recovery, for the vulnerability of the economy to
inflation or to new shocks at home or abroad, and there­
fore for monetary and fiscal policy?
Perhaps answers to questions like these can be traced
back in some ultimate sense to the behavior of money.
But I doubt it. The explanation is much more likely to
be found in other phenomena, including changes in social
and economic attitudes stimulated by the earlier period
of relatively stable prosperity.
I would go further and raise a question about the prac­
tical policy implications of the central policy theme of
monetarism: that “inflation is always and everywhere a
purely monetary phenomena”.
I do not want to be misunderstood. Central bankers,
as custodians of a nation’s money, commonly share the




255

observation and intuition that pressures to increase the
money supply to serve some presumed short-term objec­
tive are a basic source of inflationary pressure. Certainly,
excessive monetary expansion is a sufficient condition for
inflation and, in the longer run, it is equally clear that
no important inflation can be sustained without money
rising substantially faster than real income (taking into
account trend velocity). There is always some rate of
monetary growth (perhaps zero) that will in principle
achieve price stability. But, in the world in which we live,
I do not think we can draw much comfort from those prin­
ciples as a full explanation of where we are and a guide
as to how to proceed.
Take, for instance, the period since inflation began to
accelerate after the mid-1960’s, with the rate reaching a
peak during 1974 unprecedented for peacetime. Do we
really have an adequate explanation of this development
in terms of an acceleration in the rate of monetary expan­
sion alone?
To be sure, there was in that period a faster rate of
money growth. The two events were not unrelated. But,
as a technical matter, it is also true that as we got into the
1970’s the money relationships were not stable, so that
monetarists did not succeed better than others in antici­
pating the full force of double-digit inflation.
Plainly, even over a period of years, the relationship
between money and inflation is complex and the statistical
association rather loose. We do not need to look far
to find other, and supplementary, explanations of price
developments in the 1970’s— the oil situation, some crop
failures, the spread of unions into some new areas, and
shortages in particular industries that ran up against ca­
pacity pressures before the economy as a whole reached
full employment.
We can theorize that such developments affect only
relative prices and need have no effect on the general
price level if monetary growth is held steady. But the
argument rests on the assumption of a highly flexible and
quickly reacting price system. If, to the contrary, relative
price adjustments in circumstances like these are typically
slow in coming and resisted, economists would agree that
monetary growth at a noninflationary rate would depress
the level of real activity. The question is which view better
fits observations of reality, and there seems to me a lot of
evidence that it is the latter.
More generally, we cannot avoid asking ourselves about
the nature of the economic, social, and political forces
and attitudes that seem to have aggravated the difficulties
of reconciling full employment with price stability.
It is hardly a satisfactory answer to say that central
banks in principle can always resist inflationary pressures

256

MONTHLY REVIEW, OCTOBER 1976

by simply refusing to provide enough money to finance
them. Set against persistent expansionary pressures, ag­
gressive wage demands, monopolistic or regulatory pat­
terns that resist downward price adjustments, and other
factors affecting cost levels, such an approach would
threaten chronic conflict with goals of growth and em­
ployment that must rank among the most important na­
tional objectives. In a democracy, the risk would not be
just to the political life of a particular government, but to
our way of government itself.
In this larger social and political setting, we should per­
haps think of central banks themselves as “endogenous”
to the system. A theory of chronic inflation that points
only to the money supply is not going to prove adequate
to understand— or deal with— inflation in today’s world.
The danger is that it may discourage the search for partic­
ular remedies for particular problems.
There is no doubt in my mind that we must persist
in finding an answer to our inflationary problems. We can
take satisfaction in the progress of the past year. The
current underlying rate of 6 percent or so is half that of
1974, and it has been maintained in a period of rather
vigorous recovery. It feels better, and it is better.
But perhaps the greater test lies ahead. I hear from
many directions the argument that individuals and in­
stitutions have pretty well adjusted to the current rate
of inflation. Further progress, it is said, may be difficult
without an unduly depressed economy. Perhaps, the argu­
ment goes, the better part of wisdom would be to live with
the current rate rather than to try to reduce it further,
aiming ultimately at the restoration of price stability.
Now, I recognize that it is possible to conceptualize
about fully anticipated inflation being equivalent in its
real effects to confidence in price stability. But I also
question whether our institutions or individuals are in fact
fully adjusted, or really can be expected to adjust, to the
current rate of price increases or to any sizable rate of
inflation. In any case, such an adjustment, once initially
made, would not help us to deal with those forces that
upset price equilibrium in the past. Indeed, I suspect the
job of dealing with these forces would be much more




difficult. The difference between a goal of, say, living with
6 percent or a goal of evolving toward stability seems to
me profound from a psychological point of view. Willing­
ness to settle for just so much inflation, but no more,
would simply lack creditability with the public at large, or
indeed, with policymakers themselves. Resistance to in­
creases in the name of short-term advantages could only be
weakened, and we would be off again. And I think we
have learned enough to see that, in those circumstances,
even our employment goals will fall by the wayside.
My theme today is simple. As we look back over the
evolution of thinking about monetary policy and macroeconomic policy generally over the postwar years, we can
see the dangers of overly simple and overly confident
views of the way the economic world works. Eventually,
simple doctrine comes up against complex and harsh
reality.
Back in the days when I was learning economics and
central banking, the General Theory had cast fresh light
on old problems. The intellectual contributions were
immense. But popularized, bowdlerized, and pressed to
extremes, it lost fashion for good reason.
The monetarists, emphasizing old truths in modern
clothing, have provided a large service in redressing the
balance. It is in pressing the point to an extreme that the
danger lies— the impression that only money matters and
that a fixed rate of reserve expansion can answer most
of the complicated problems of economic policy.
In a way, I suppose full confidence in a simple, unified
view of economic policy is a comforting thing, a kind
of security blanket in an uncertain world. But Alfred
North Whitehead, in a different context, once pointed to
the danger: “There are no whole truths; all truths are
half truths. It is trying to treat them as whole truths that
plays the devil.”
He overstated the case. The practical man cut adrift
from any sense of what is the greater truth— distinguish­
ing, if you will, the one-eighth truth from the seveneighth truth—will soon lose his way. But, in assessing
those truths, he can never afford to lose sight of the messy
reality of the world in which we live.

FEDERAL RESERVE BANK OF NEW YORK

257

T h e B u sin e ss Situation

Recent business statistics suggest that the economy is
continuing to move up, although it is still too early to
tell whether the pattern of slower growth evident since
last spring has ended. Consumer spending showed renewed
strength in August, as retail sales posted a sharp and
broadly based advance. At the same time, housing starts
and newly issued building permits increased substantially,
providing signs of prospective strengthening in residential
construction. Distinctly less encouraging, however, were
the August declines in durable goods orders and in the
index of leading indicators. Overall, while business activity
continues to exhibit upward momentum, the extent of the
strengthening, if any, is difficult to assess. Capital spending
is likely to play a key role in the months ahead, but the
outlook for such expenditures is unclear. On a more posi­
tive note, while the nationwide automobile strike has
entered into the fourth week, a tentative contract agree­
ment has been reached.
The latest price data have been mixed. At the con­
sumer level, overall price increases have been running
at around a 6 percent annual rate for the past several
months, but wholesale price increases accelerated in
September. Some other developments have been more
favorable for the price outlook. Most notable are
recent price developments in major metals industries.
Steel producers rescinded some previously announced
price hikes, and aluminum firms postponed price in­
creases. In addition, spot prices of raw industrial com­
modities, which are generally regarded as particularly
responsive to demand pressures, registered outright
declines in September for the second successive
month.
INDUSTRIAL PRODUCTION AND CAPACITY

Industrial production rose in August for the seventeenth
consecutive month, according to the Federal Reserve
Board’s index. Preliminary estimates show that output
of the nation’s mines, utilities, and manufacturing estab­
lishments rose 0.5 percent in that month, the same as the
increase recorded in July (see Chart I). Following the




sharp cyclical expansion in the early months of the
recovery, the rate of growth in production has slowed
noticeably in recent months. Nevertheless, output in
August was about 18 percent above its trough level
reached in March 1975. In large measure, some modera­
tion in production gains was to be expected as the recovery
matured. In fact, at this point in the recovery, the growth
of production is comparable to that in most previous post­
war recoveries. Nonetheless, the recent data have been
distorted by strikes in the rubber and coal industries, and
production in September was affected by the strike of the
United Auto Workers' union against the Ford Motor
Company, which began on September 15. Recent gains in
industrial output have been concentrated largely in durable

Chart !

INDUSTRIAL PRODUCTION
Seasonally adjusted
Percent

Percent

258

MONTHLY REVIEW, OCTOBER 1976

goods materials, business equipment, and construction
supplies. Production of consumer goods has been relatively
unchanged in recent months, no doubt reflecting a re­
sponse to the midsummer lull in consumer buying.
While the continued growth of industrial production has
raised the level of output close to its pre-recession mark,
concerns over the reemergence of widespread capacity
problems appear to have diminished. Virtually all mea­
sures of capacity utilization indicate that the economy is
operating well below its productive limits. The McGrawHill measure of manufacturing utilization stood at 77
percent in August, well below the 88 percent peak re­
corded in 1973. The Federal Reserve Board’s index of
capacity utilization in materials-producing industries,
which is generally regarded as a sensitive indicator of pos­
sible production bottlenecks, also is far below its peak.
The level of plant and equipment utilized in the materialsproducing sector stood at 81.5 percent in August, some
11.5 percentage points below the peak rate attained in
August 1973. Thus, the latest readings suggest that there
is ample capacity available to accommodate continued
growth of the economy. Moreover, the likelihood of con­
straints being reached in the near term has been further
reduced, of course, by the more moderate rate of growth
of real output experienced in the spring and summer.
M AN U FACTU R ER S' ORDERS A N D INVENTORIES

Durable goods manufacturers’ new orders fell 1.5
percent in August. The decline resulted mainly from a
substantial drop in new orders for nondefense capital
goods, which, in contrast, had posted an unusually large
increase in the previous month. Because month-to-month
changes in any economic series can be erratic, such large
offsetting movements are not particularly meaningful.
Because of concern over the behavior of capital spending
thus far in the recovery, however, the decline in capital
goods orders has raised, to some extent, renewed anxiety
over the outlook for this sector. While there can be little
doubt that this sharp decline diminishes somewhat the
otherwise generally brightening picture for business fixed
investment, its importance should not be exaggerated.
Aside from the August drop, nondefense capital goods
orders have increased continually this year. Bookings for
defense goods recovered in August, after dropping sharply
in July. The recent erratic movements in defense orders
might be related to the changeover by the Federal Govern­
ment to a new fiscal year.
Inventories in manufacturing and trade continued to
rise, by $1.6 billion in July, with nondurables manufac­
turing and durables retail trade accounting for most of




the increase. The accumulation began in January in re­
sponse to a strong pickup in sales and continued into the
summer, partly as a result of the recent sluggish growth
in final sales. Hence, the inventory-sales ratio has re­
bounded recently, although it has remained well below
the level of a year ago. As in July, manufacturers’ inven­
tories rose by more than $900 million in August, owing
mainly to the buildup in the nondurables sector.
The index of leading indicators fell 1.5 percent in
August, with the decline related to the recent sluggishness
in consumer spending and some areas of capital invest­
ment. Contributing heavily to the drop were the rise in the
manufacturing layoff rate, the decrease in contracts and
orders for new plant and equipment, and the shortening
of the average workweek in manufacturing. While the
August decline— the first in eighteen months— may con­
tribute to anxiety over the future course of the economy,
a one-month decrease is not sufficient to signal a stalling
of the recovery. Historically, the index has turned down,
sometimes for several months, even though the economy
subsequently continued to expand.
RETAIL SA LE S A N D PE R S O N A L INCOME

Following several months of lackluster performance,
retail sales rebounded strongly in August, rising 2.3
percent on a seasonally adjusted basis. After adjustment
for higher prices, this represented an increase of 1.9 per­
cent, pushing sales in constant-dollar terms slightly above
the previous peak registered in April. The August sales
gain was broadly based, with all major categories of retail­
ers—particularly auto dealers and department stores—
posting substantial gains. Some analysts have contended
that auto sales would have been even stronger except for
shortages of some popular models. It may also be, how­
ever, that in efforts to reduce stocks of existing models
dealers have offered generous price concessions and con­
sumers have responded by snapping up the available
models.
Since last April, retail sales have fluctuated irregularly,
with an increase in one month followed by a decrease in
the next. The failure of retail sales to break out of this
pattern is generally considered a major factor in the overall
slowing in the nation’s economic advance. A prolonged
pickup in sales depends on consumers’ willingness to
spend as well as on the growth of income and employ­
ment. In this regard, the latest surveys of consumers’ atti­
tudes suggest no change in their assessment of the econ­
omy or in their interest in buying big-ticket items. On the
other hand, the prospects for continued gains in personal
income appear good. After a strong surge in July due

259

FEDERAL RESERVE BANK OF NEW YORK

mainly to the midyear cost-of-living increase in social
security benefits, personal income posted a more modest
increase in August. Part of the slowdown was attributable
to the small advance in that month of average hourly
earnings in the private nonfarm economy. A decline in
farm income, as a result of a sharp drop in wholesale
prices of farm products during the month, also was im­
portant in moderating the personal income gain. From a
longer term perspective, personal income has grown at
better than a 9 percent annual rate over the first eight
months of the year, well above the experience of earlier
recovery periods.

Chart III

PLANT A ND EQUIPMENT SPENDING IN
RECESSION A ND RECOVERY
Index as a percentage of trough -q uarter l e v e l*
Percent_____________________________________________________________________ Percent

A v e ra g e o f four
previous cyclesy
/

\

Chart II

HOUSING STARTS
Seasonally adjusted a n n u a l rates
M illions of units

Source:

M illions of units

United States Department of Commerce, Bureau of the Census.




y

__

/
P lann ed

.........ill cycle

Q uarters after

Quarters before

trough

trough

Housing activity showed some encouraging signs in
August, with housing starts rising to a seasonally adjusted
annual rate of 1.54 million units (see Chart II). This
was the highest level of activity since April 1974 and put
the level of building activity some 22 percent above a
year earlier. Much of the August increase was in the multi­
family sector, where starts rose from about 260,000 to
nearly 350,000 units. Since one of the major factors
restraining the recovery in residential construction activity

J?
/ I

/

f

RESIDENTIAL CONSTRUCTION
AND CAPITAL SPENDING

*

1

-4

-3

1

-2

1

-1

1

0

1

1

1

1

1

1

2

3

4

5

6

7

Note: The National Bureau of Economic Research-dated recession troughs
occur in second quarter 1954, second quarter 1958, first quarter 1961,
and fourth quarter 1970. The trough quarter for the latest recession has
not yet been dated by the NBER. The first quarter of 1975 was selected
as the tentative trough quarter.
* F o r each reference cycle, the trough-quarter level equals 100 percent.
Source:

United States Department of Commerce, Bureau of Economic Analysis.

has been weakness in the multifamily sector, the pickup
in starts in this sector was particularly encouraging. Fur­
ther bolstering near-term prospects was a jump in permits
to build apartment complexes of five or more units.
Despite this tentative evidence of a pickup in the multi­
family sector, the level of activity remains well below
that of the peak years of 1972 and 1973. The single­
family sector, which has staged a stronger recovery,
exhibited renewed strength in August. Single-family starts
rose to 1.2 million units, and newly issued permits in­
creased to their highest level since early 1974.
According to the Commerce Department survey taken
in July and August, business spending on plant and equip­
ment is expected to rise 7.4 percent in 1976, little changed
from the 7.3 percent increase indicated in the previous
Commerce Department survey. However, while the yearover-year growth has changed only slightly, there has been
a substantial redistribution of the pattern of spending. As
a result, spending in the final quarter of this year is
expected to be substantially above that planned earlier.
While capital spending typically lags the general recovery,
thus far in the current business expansion capital spending
has been trailing unusually far behind (see Chart III).
Plant and equipment spending currently is only 3 percent
above the trough-quarter level. In contrast, at comparable

260

MONTHLY REVIEW, OCTOBER 1976

points in previous postwar recoveries, capital spending
had typically registered better than a 10 percent gain.
Since this comparison is in nominal terms, allowing for the
faster rate of inflation in the current recovery would fur­
ther underscore the recent sluggishness of capital spending.
Provided that the revised expenditure plans are realized,
the gap between the current experience and earlier recov­
eries will be reduced. Even more importantly, the pros­
pective revival in capital outlays could go a long way
toward quickening the overall pace of activity.
R E C E N T P R IC E D E V E L O P M E N T S

There has been little indication of a fundamental change
in the price situation. Consumer prices rose at a 6.3
percent annual rate in August, only slightly higher than
the increase posted over the preceding three months.
Retail food prices rose at a 4 percent rate, a somewhat
more rapid increase than the very modest advances of the
two prior months. With a large harvest in prospect, the




outlook for food prices in coming months continues to be
favorable, although the weather always adds an important
element of uncertainty. Consumer power and fuel prices
continued to increase at a rapid rate in August; on the
other hand, price increases for consumer services moder­
ated somewhat to a 6.6 percent annual rate.
Fragmentary information about prices of goods at the
initial stage of the production process suggests that some
further diminution in inflation may be forthcoming. There
have been reports of price concessions and increased dis­
counting on a number of basic raw materials. Additional
evidence on this was provided by the September survey of
the National Association of Purchasing Management,
which reported a continued decline in the proportion of
respondents facing higher prices. Moreover, steel manu­
facturers rescinded a previously announced price hike,
while aluminum producers postponed planned price in­
creases. In addition, spot commodity prices dropped con­
siderably in August and September, according to the
Bureau of Labor Statistics index.

FEDERAL RESERVE BANK OF NEW YORK

261

T h e M o n e y and B o n d M a r k e t s in S e p te m b e r

Interest rates fluctuated in a narrow range during Sep­
tember following rather steady declines over the summer
months. After edging slightly lower in the three previous
statement weeks, rates adjusted upward subsequent to
the announcement on September 23 of a massive increase
in the money stock during the week ended September 15.
The one-week increase more than offset declines in earlier
weeks, and it dampened previous speculation among mar­
ket participants that the Federal Reserve System would
encourage slightly lower short-term interest rates. During
the statement week ended September 22, however, the in­
crease in the money supply for the previous week was sub­
stantially reversed by a large decline. Rates on new corpo­
rate bond issues dipped below the year’s previous lows in
April, and this development apparently prompted a num­
ber of corporate borrowers to schedule new financings.
As a result, the calendar of forthcoming issues grew as
the month progressed, further tempering market sentiment.
The Treasury continued with its program of fairly regu­
lar coupon offerings, and the two new issues were sold
during the month. The sluggish state of business loan
demand and the low rates available to borrowers in the
commercial paper market induced a number of major
banks to reduce the prime lending rate by V4 percentage
point to 63 percent. Early in October the 6% percent
A
rate became widespread, as most major banks joined in
the reduction.
Preliminary estimates indicate that in September the
narrowly defined money stock (Mt) showed a modest rise,
on balance, after substantial week-to-week fluctuation.
The broadly defined money stock (M2) continued to
reflect rapid growth in consumer-type time deposits and
rose at a relatively strong rate. Further declines in the
volume of large negotiable certificates of deposit (CDs)
outstanding held the bank credit proxy to a small increase.

(FOMC) specified that the Federal funds rate should be
held within a 5 to 5 Vi percent range during the period
until the next meeting on September 21. During that
interval, the funds rate remained around the 5lA percent
level that has prevailed since early in the summer (see
Chart I). During September as a whole, the effective rate
on Federal funds averaged 5.25 percent, down 4 basis
points from its average in August. At the end of Septem­
ber, the rate on 90- to 119-day commercial paper was
5V4 percent, Vs percentage point below the end-of-August
level. Rates on bankers’ acceptances also changed little
and closed the month at 5lA percent. Similarly, CDs
maturing in 90 days traded in the secondary market at
5.28 percent at the end of the period, a decline of 5 basis
points over the month. In view of the continued sluggish­
ness in business loan demand and the downward drift, of
commercial paper rates in recent months, a number of
large banks lowered their prime lending rate by lA per­
centage point to 63 percent, with most others joining in
4
this action in early October.
Federal Reserve open market operations in September
were, in part, concerned with offsetting the impact on
bank reserves of the September 15 tax date for corporate
and individual income taxes. As tax payments accumu­
late in the Treasury Tax and Loan Accounts at commer­
cial banks, the Treasury typically transfers such receipts
to its accounts at Federal Reserve Banks. These transfers
—Treasury calls— drained reserves from the banking sys­
tem in the last two statement weeks of September. Federal
Reserve open market operations supplied reserves and
offset these drains (see Table I).
The Federal Reserve adds reserves to the banking
system either through outright purchase of securities or
through repurchase agreements, under which it buys
securities from dealers who enter into agreements to buy
the same securities back at a later date. These agreements
have short-term maturities, usually from one to seven
THE M O N E Y M AR KET A N D THE
days. In the past, dealers had the option of withdrawing
MONETARY AGGREGATES
early from all repurchase agreements that had an initial
Interest rates on most money market instruments were maturity of more than one business day. When a contract
little changed on balance in September, At its meeting was terminated prior to maturity, it drained reserves from
on August 17, the Federal Open Market Committee the banking system and thus offset part of the reserve




262

MONTHLY REVIEW, OCTOBER 1976
Chart I

SELECTED INT£REST RATES
July-Septem ber 1976
Percent

N o te:

M O N E Y MARKET RATES

BOND MARKET YIELDS

Percent

D a ta a re shown fo r business days only.

M O N E Y MARKET RATES Q UOTED: Prime com m ercial loon rote at most m ajor banks;
o ffe rin g rates (quoted in terms of rate of discount) on 90- to 119-day prime com m ercial
p a p e r q u o ted by three of the five d ealers that re p o rt th eir rates, or the m idpoint of
the ran g e q u o ted if no consensus is a v a ila b le ; the effective ra te on F ed eral funds
(the rate most representative of the transactions executed); closing bid rates (quoted
in term s of ra te of discount) on newest o u tstan d in g three-m onth Treasury bills.
B O N D MARKET YIELDS QUOTED: Yields on new A a a -r a te d public utility bonds ore based
on prices a sked by u n d erw ritin g syndicates, adjusted to m ake them eq u iv a le n t to a

injection made by the Federal Reserve. Early terminations
of outstanding agreements were often large, and their
volume was not readily predictable. At the end of Sep­
tember, the Trading Desk of the Federal Reserve informed
dealers it would enter into four- and seven-day repurchase
agreements at the start of the October 6 statement week
and, in this operation, the seven-day contracts would not
include the right of early withdrawal. Similar fixed-term
contracts with a six-day maturity were made on October 1.
Preliminary estimates indicate that Mj—private de­
mand deposits adjusted plus currency outside commercial
banks— grew in September at a somewhat slower pace
than in the previous two months. The average level for
the four weeks ended September 22 stood at 3.0 percent




stan d ard A a a -ra te d bond of a t least tw enty years' m aturity; d a ily averag e s of
y ields on seasoned A a a -ra te d corporate bonds; d a ily a v e ra g e s of yield s on
lo ng -term G o v e rn m e n t securities (bonds due or calla b le in ten years or more)
and on G overn m en t securities due in three to five years, com puted on the basis
of closing bid prices; Thursday a v e rag es o f yield s on tw enty seasoned tw entyyeor ta x -e x e m p t bonds (carrying M oody's ratings of A a a , A a , A , and B aa).
Sources: Federal Reserve Bank of N e w York, Board of G overnors of the F e d e ra l
Reserve System, M oody's Investors Service, Inc., and The Bond Buyer.

(seasonally adjusted annual rate) above its average level
in the four statement weeks ended eight weeks earlier. At
its August meeting, the FOMC established a tolerance
range of 4 to 8 percent for growth of
over the two
months ended with September. As a consequence of
somewhat more rapid expansion of time and savings
deposits in September, the spread between the growth of
M2— Mi plus these deposits other than large negotiable
CDs— and M* widened. A wider spread between growth
of M2 and Mi is often associated with a decline in money
market yields relative to yields prevailing on time and
savings deposits, such as occurred during the June-August
period. Over the four statement weeks ended September 22,
the average for M2 was 9.9 percent at an annual rate above

263

FEDERAL RESERVE BANK OF NEW YORK

its average during the four statement weeks ended eight
weeks earlier. The FOMC tolerance range for M2 growth
over the two months ended in September was 7.5 to 11.5
percent. Comparing the averages for the four weeks ended
September 22 with similar averages fifty-two weeks earlier,
Mi and M2 have expanded 4.3 percent and 10.1 percent,
respectively (see Chart II). After posting a modest de­
cline in August, the bank credit proxy— total member
bank deposits subject to reserve requirements plus certain
nondeposit sources of funds—was little changed in Sep­
tember, as large negotiable CDs outstanding continued to
slide in response to weak demand for business loans.
THE G O V E R N M E N T SECURITIES M AR K ETS

Yields on United States Treasury securities declined
during most of September before partially retracing these
steps late in the month. Participants initially believed that
the pause in the economic recovery and the slow growth
of Mx would lead the monetary authorities to encourage
interest rates to edge lower. However, the sharp increase
in Mi announced after midmonth prompted some re­
visions in this outlook, and thereafter money market
interest rates were generally anticipated to remain stable
in the near term.
In the Treasury bill market, the extension of earlier
declines in bill rates was reflected in the auction of 52week bills on September 15. The average yield in this
auction was 5.56 percent, 7 basis points below the auction
four weeks earlier (see Table II) and the lowest average
rate since February 1975. The apparent tempering of
market sentiment after midmonth, however, resulted in a
reversal of generally declining rates at the auctions of
three- and six-month bills. At the last regular auction in
September, the average issuing rates on these bills were
5.07 percent and 5.33 percent, respectively, up 4 and 9
basis points from the previous week. No new cash was
raised in September by the Treasury on these short-term
issues. Rates on most bills ended September unchanged to
9 basis points below levels at the end of the previous
month.
The Treasury continued during September to raise
cash by auctioning coupon issues. On September 22, $2.5
billion of two-year notes was sold at an average yield of
6.30 percent, 37 basis points below the return on a similar
offering of two-year notes in the previous month. The
issue, which raised $819 million in new cash, attracted
broad-based bidding interest. On September 28, $2.5
billion of five-year notes was auctioned, all for new cash.
The average issuing rate at this auction was 7.08 percent,
down from 7.63 percent for a similar offering on June 29.




Table I
FACTORS TENDING TO INCREASE OR DECREASE
MEMBER BANK RESERVES, SEPTEMBER 1976
In millions of dollars; (+) denotes increase
and (—) decrease in excess reserves
Changes in daily averages

— week

ended
Net
changes

Factors
Sept.

1

Sept.

8

Sept.
15

Sept.
22

Sept.
29

“ M arket” factors

Member bank required reserves ..
Operating transactions
Federal Reserve float ..............
Treasury operations* ................
Gold and foreign account ........
Currency outside banks ..........
Other Federal Resesrve
liabilities and capital ...............
Total “market” factors ..........

—

67 + 465 — 263 + 121 — 925

_ 669

—2,421
— 698
—2,975
+
9
+1,361

—1,334
+ 29
—2,088
+ 20
+ 805

- 446 + 626 — 240 + 79 — 119
J—
+ 352 —3,936 +3,230 —6,175 —3,346

— 100
—2.003

591 —3,771 —3,466 +5,991 +3,493

+1,656

+ 230 + 74 —3,324 +3,154 +1,166
14 — 8 — 4 — 21 — 17
_.
_
23 _
— 18 — 3

+1,300
— 64
— 44

—
+
—
+

+ 499
— 37
+
2
+ 21
+
4
+ 295
+1,974
— 29

+
+
+
—
+

419
208
23
19
653

+3,471
— 138
-J-3,399
— 8
— 408

4-3,493
+1,017
+3,007
+ 5
— 297

—6,296
— 360
—5,542
+ 33
— 504

Direct Federal Reserve credit
transactions

Open market operations
Outright holdings:
Treasury securities ...................
Bankers’ acceptances ...............
Federal agency obligations . . .
Repurchase agreements:
Treasury securities .....................
Bankers' acceptances ................
Federal agency obligations . . .
Member bank borrowings ............
Seasonal borrowings f ....................
Other Federal Reserve assets? . .
Excess

reserves}§

.....................................

-

—

839
80
25
26
2
+
+ 271
- 294
+ 58

—3,226
— 496
— 115
—
47
—
2
— 152
—3,970
—
34

107
9
— 4
+ 15
— 2
+ 211
—3,240
— 10
—

—

+2,479
+ 318
+ 64
—
17
+
1
— 16
+5,958
— 217

+2,192
+ 70
+ 82
+ 44
+ 5
— 19
+3,520
+ 174

Daily average levels

M onthly
averagesll

Member bank:

Total reserves, including
vault cash|§ ..................................... 34,078 33,579 33,839 33,501 34,600
Required reserves ......................... 33,772 33,307 33,570 33,449 34,374
Excess reserves § ....... ....................
306
272
269
52
226
Total borrowings .........................
93
46
61
44
88
Seasonal borrowings t .................
32
30
28
34
29
Nonborrowed reserves ..................... 33,985 33,533 33,778 33,457 34,512
Net carry-over, excess or
deficit (—)fl .....................................
123
154
159
135
62
Note: Because of rounding, figures do not necessarily add to totals.
* Includes changes in Treasury currency and cash,
t Included in total member bank borrowings.
t Includes assets denominated in foreign currencies.
§ Adjusted to include waivers or penalities for reserve deficiencies in
accordance with the Regulation D change effective November 19, 1975.
II Average for five weeks ended September 29, 1976.
If Not included in data above.

33,919
33,694
225
66
54
33,853
127

264

MONTHLY REVIEW, OCTOBER 1976

Trading in seasoned coupon issues followed the general
market pattern, and the index of three- to five-year bonds
closed the month 10 basis points lower than its level at
the end of August. Similarly, the index of long-term bonds
fell 9 basis points over the month.
Yields on Federal agency issues moved modestly
lower in September. The Government National Mort­
gage Association raised $264.6 million through the
auction of IV2 percent modified pass-through securi­
ties on September 14. These securities, which have an
average life of about twelve years, were well received when
reoffered to yield 8.25 percent on a corporate bond equiv­
alent basis. The Farm Credit Banks redeemed $82 million
of maturing securities and refunded the remainder with
$1,419.3 million of new issues, consisting of $569.6
million of six-month Banks for Cooperatives bonds yield­
ing 5.60 percent and $849.7 million of nine-month Fed­
eral Intermediate Credit Bank bonds yielding 5.80 per­
cent. On September 22, the Federal National Mortgage
Association raised $400 million of new cash through the
sale of fifteen-year debentures at 7.80 percent. At the

GROW TH OF SELECTED MONEY STOCK MEASURES
Seasonally adjusted a n n u a l rates

Table II
AVERAGE ISSUING RATES
AT REGULAR TREASURY BILL AUCTIONS*
In percent
Weekly auction dates— September 1 9 7 6
M a tu rity
Sept.
3

Sept.
13

Sept.
20

5.087
5.333

5.099
5.309

5.02S
5.236

Sept.
27

5.072
5.325

Monthly auction dates— July-September 1 9 7 6

July

21

Aug.

IS

Sept.
15

Fifty-two weeks
‘ Interest rates on bills are quoted in terms of a 360-day year, with discounts from
par as the return on the face amount of the bills payable at maturity. Bond yield
equivalents, related to the amount actually invested, would be slightly higher.

end of the month, the Federal Land Banks offered $1.66
billion of bonds consisting of $450 million of fifteen-month
bonds priced to yield 6.10 percent, $858.2 million of
41
4-year bonds yielding 7.10 percent, and $350 million
of twenty-year securities with a return of 7.95 percent.
These issues generally encountered a good reception by
investors.

Percent

15

Ml

OTHER SECURITIES M AR KETS

From 13
w eeks e a rlie r

J

—
\

f

/TV

_

—

s '*
From 5 2

^
\

//*V \

u /

v

w eeks e a rlie r

III III

I I I 11

1 1 1 I I 1 M ! I J_

...

From 13
y ^ ^ v e e k s e a rlie r

M2

-

\

/

\
—\ S

From 52
\
w eeks e a rlie r

1 1 1 1 1 1 1 1 1 1 1 I I 1 1 1 L 1 ! JL1_
1975
Note:

.I..
1976

Growth rates are computed on the basis of four-week averages of daily

figures for periods ended in the statement week plotted, 13 weeks eonier and
52 weeks earlier. The latest statement week plotted is September 22, 1976.
M l = Currency plus adjusted demand deposits held by the public.
M2 = M l plus commercial bank time and savings deposits held by the public, less
negotiable certificates of deposit issued in denominations of $100,000 or more.
Source:

Board of Governors of the Federal Reserve System.




Yields on corporate and high-grade municipal bonds
extended the declines of the previous three months dur­
ing most of September. Increased supplies of new issues,
together with uncertainty over the future course of mone­
tary policy, however, put upward pressure on yields late
in the month. The number of new corporate offerings was
slim early in the period, but the calendar expanded
steadily as corporations were attracted by the lowest inter­
est costs in three years. Heavy new issue activity in the
tax-exempt market throughout the period generally en­
countered good investor demand.
The decline in yields in the corporate sector was ex­
emplified by the successful sale early in the month of a
$60 million offering of Aa-rated electric utility bonds
priced to yield 8.45 percent in thirty years. This return
was 22 basis points below that of a comparable issue,
which sold slowly when offered in July, and was the low­
est yield on an Aa-rated electric utility since March 1974.
The attractiveness of prevailing yields to corporate bor­
rowers brought forth a $100 million offering of Aaa-rated

FEDERAL RESERVE BANK OF NEW YORK

telephone debentures which had been postponed earlier in
the year. The bonds were priced to yield 8.00 percent,
25 basis points below a comparably rated telephone issue
offered the previous month. This was, moreover, the low­
est return on such an issue in nearly three years. The
yield was set in anticipation of further interest rate declines
and did not initially prove attractive to investors. At
the end of the month, the bonds were released from syndi­
cate price restrictions and subsequently traded in the mar­
ket at yield increases in excess of 10 basis points. Under­
writers priced $250 million of Aa-rated debentures of
Hydro-Quebec to yield 8.60 percent in thirty years. This
was 70 basis points below the yield on a similar offering
last June. It proved insufficient to attract investors, and
a large volume remained in dealer hands at the month
end.
Lower interest rates established this past summer in the
tax-exempt sector were little changed over the month of
September. For example, good receptions were afforded
both South Carolina’s $70 million of Aaa-rated bonds
scaled to yield from 3.0 percent in 1977 to 5.05 percent
in 1991 and Ohio’s $70 million of Aa-rated bonds with
yields ranging from 3.5 percent in 1978 to 6.2 percent
in 2002. The return in 1977 on the South Carolina issue
was 20 basis points below a comparably rated offering
marketed in June. However, New Jersey’s $75 million




265

offering, rated Aa/AAA (Moody’s/Standard & Poor’s),
sold slowly when priced to yield from 3.5 percent in 1978
to 5.8 percent in 1996. These yields were as much as 30
basis points below those on the state’s previous offering
in January. In one of the largest of such offerings, $700
million of Pennsylvania 8 ¥2 -month notes sold quickly at
a yield of 3,2 percent. Late in the month Texas issued
$40 million of Aaa-rated bonds, with yields ranging from
4.50 percent in 1985 to 5.30 percent in 1993.
The New York State Housing Finance Agency
(HFA) passed a key test in September, marking a
milestone on the road to renewed access for New York
State to the capital markets. The issue was a $149 mil­
lion package of State University of New York A/AArated construction bonds? consisting of $18 million of
serial bonds priced to yield from 5.50 percent in 1977 to
7.50 percent in 1986 and $131 million of term bonds
priced to return 8.50 percent in thirty years. The offering
provided for the retirement of some bonds issued earlier
in the year as part of a plan by New York State to keep
the HFA and several other state agencies out of default.
The size of the issue was originally $50 million but was
raised to $97 million subsequently. Demand for the new
bonds was so strong by the time it came to market that
the HFA increased the amount of the offering to $149
million.

266

MONTHLY REVIEW, OCTOBER 1976

A lte rn a tiv e D efin itio n s o f th e M o n e y S to ck
and th e D em an d fo r M o n e y

By L a u r e n c e H. M e y e r *
In the last five years, the monetary aggregates have
played an important role in the formulation and execution
of monetary policy. At the same time, there has been
growing concern that developments in financial practices
over the postwar period and innovations in financial
instruments and technology over the last few years have
reduced the importance of the narrowly defined money
stock (Mi) in the financial system and have blurred the
distinction between Mx and savings and time deposits at
both commercial banks and thrift institutions. The ratio
of Mi to total liquid assets has declined steadily over the
postwar period. More recently, the introduction of nego­
tiable order of withdrawal (NOW) accounts, checking
accounts at thrift institutions, expanded third-party pay­
ment privileges from savings accounts, telephone transfers
between savings and demand deposit accounts, and elec­
tronic funds transfer systems have made the association
of Mi with the means of payment increasingly less com­
pelling.
This paper is concerned with the problem of defining
the money stock in this changing financial environment.
In Part I, the five official measures of the money stock
are presented, the savings instruments included in the
broader money stock measures are briefly defined, and the
developments that may have altered the role of M! in the
financial system are discussed. Empirical evidence on the
definition of money is reported in Part II. The empirical
analysis is confined to Mu M2, and M3, the three money
stock measures for which the Federal Reserve System
currently sets growth ranges.
In determining the ranges for growth in the monetary

* The author wishes to thank Martin Mauro for excellent re­
search assistance, and Michael Hamburger, Arline Hoel, Kevin
Hurley, Fred Levin, and Gary Stern for valuable comments and
suggestions.




aggregates, the Federal Reserve System uses econometric
models, among other tools, to predict the paths of income,
prices, and employment associated with alternative rates of
monetary expansion. The precision with which the System
can predict the economy’s response to alternative mone­
tary growth rates depends, in some of these models at least,
on the precision with which it can estimate the demand for
money. And precision in estimating the demand for money
can be maximized by selecting the definition of money
with respect to which wealth owners exhibit the most
stable and systematic behavior. Hence, the focus of Part
II of this paper is on the relative predictive performance
of demand functions for alternative definitions of the
money stock. The empirical results reported in this section
suggest that there has been a perceptible deterioration in
the predictive performance of the Mx demand function
relative to the M2 and M3 demand functions in the 1970’s.
While the Mx definition permitted the most accurate pre­
diction of money demand over the full sample period in­
cluding both the 1960’s and 1970-75, the M2 and M3 equa­
tions yield more accurate predictions when the analysis is
confined to the 1970’s alone.
DEFINITIONS OF THE M O N E Y STOCK

There recently has been a proliferation in the number
of official definitions of the money stock.1 Until April
1971, the Federal Reserve Bulletin recognized only the
Mi definition of money in reporting financial data. At that
time, it began to report regularly three measures: Mi,
Mo, and M3 (the sum of M2 and deposits at mutual sav-

1 For the purposes of this study, measures of money regularly
reported in the Federal Reserve Bulletin are considered official.

FEDERAL RESERVE BANK OF NEW YORK
Table I
MEASURES OF THE MONETARY AGGREGATES
In billions o f dollars, June 1976
Money stock measures

Totals

Mi

(Private demand deposits adjusted + currency outside
commercial banks) .......................................................

303.1

m2

(M i + savings and time deposits at commercial banks
other than large negotiable certificates of deposit) ....

700.3*

m3

(M 2 +
deposits at mutual savings banks, savings
and loan association shares, and credit union
shares) .............................................................................

1,159.21*

m4

(M 2 +

large negotiable CDs) ...................................

770.9$

M5

(M 3 +

large negotiable CDs) ...................................

1,229.8$

* Tim e and savings deposits at commercial banks other than large negotiable
CDs (T ) were $397.3 billion,
t Savings deposits at thrift institutions (S) were $458.9 billion.
Large negotiable CDs were $70.6 billion.

t

ings banks and savings and loan association shares).
Beginning in April 1975, M3 was redefined to include
credit union shares and two additional measures of the
money stock were introduced: M4 (the sum of M2 and
large negotiable CDs) and M5 (the sum of M3 and large
negotiable CDs). Table I reports the magnitudes of the
five official measures and the savings components of the
broader measures as of June 1976.
SAVINGS AND TIME DEPOSIT COMPONENTS. As previously
noted, the broader money stock measures include, in addi­
tion to Mi, one or more of the three time deposit totals men­
tioned above. In this section the savings instruments in­
cluded in the broader definitions are described. Commercial
banks and thrift institutions issue a wide assortment of
savings instruments, differing in the maximum interest rate
that can be paid, in the term over which the account must
be held, and in the minimum denomination of the account.
Savings deposits or shares are mostly passbook accounts.
Although commercial banks and thrift institutions must
reserve the right to require at least thirty days’ written
notice before withdrawal, in practice withdrawals are
honored on demand. In effect, therefore, any amount may
be added to or withdrawn from a savings deposit account
at any time, making it particularly well suited for savers
whose deposits are in small amounts or whose needs for
withdrawals may be irregular or unpredictable. Time de­
posits, unlike savings deposits, explicitly specify a maturity
but may be redeemable prior to maturity with some sacri­
fice of interest. In addition, time deposits may require
some minimum denomination and are generally issued in




267

certificate rather than in passbook form.
During the 1950’s and much of the 1960’s, the savings
account was the major savings instrument issued by both
commercial banks and thrift institutions. However,
changes in regulations and competition for savings deposits
have resulted in the development of a wider assortment of
instruments. By the end of 1974, the value of time ac­
counts had increased to more than one third of the total
value of accounts in mutual savings banks, almost half of
the savings and time accounts (excluding large CDs) at
commercial banks, and more than half of the savings and
time accounts at savings and loan associations.
Several important regulatory changes affecting savings
deposits at commercial banks were initiated in 1975 to
permit commercial banks to compete more efficiently
with thrift institutions. Effective April 7, 1975, the Board
of Governors of the Federal Reserve System authorized
member banks to permit the use of the telephone to with­
draw funds from savings accounts or to transfer funds
from savings accounts. As of September 2, 1975, the
Board permitted member banks to offer a bill-paying
service through preauthorized transfers of funds from
customers’ savings accounts to pay their debts. Previously
only mortgage-related payments were permitted. These
two changes may have enhanced the substitutability of
savings deposits for demand deposits and, at the same
time, may have increased the distinction between savings
and time accounts. Effective November 10, 1975, the

Table II
CORPORATE SAVINGS DEPOSITS
AT COMMERCIAL BANKS*
In billions of dollars
Period

1975: October

.......................................................................

Total

0

November

.....................................................................

0.6

December

.....................................................................

1.5

......................................................................

2.6

1976: January
February
March

............. ............................................
.......................................................................

3.9
4.8

April ..............................................................................

5.3

May ................................................................................

5.9

6.0
♦These figures were derived by “ blowing up” the data from weekly reporting
banks and using the ratio of corporate savings deposits at all commer­
cial banks to those at weekly reporting banks based on a one-time
survey taken on January 7, 1976.

MONTHLY REVIEW, OCTOBER 1976

268

Chart I

RATIOS OF ALTERNATIVE MEASURES OF THE M O N E Y STOCK TO LIQUID ASSETS, 1 9 5 2 -7 5

1952
Note: Shaded areas represent periods of recession as defined by the N ational Bureau of Economic Research, except for the latest recession
which is tentatively judged to have ended in March 1975,
Source:

Board of G overnors of the Federal Reserve System.

Board amended the definition of savings deposits in Regu­
lations Q and D to permit corporations, partnerships, and
other profit-making organizations to maintain savings
deposits of up to $150,000 per depositor at member
banks. Estimates of the growth of corporate savings
deposits through June 1976 are reported in Table II.
These deposits increased to $6 billion dollars in seven and
one-half months. During late 1975 and early 1976, this
growth in corporate savings deposits may have contributed
to the sluggish growth in Mx.
A negotiable time certificate of deposit (CD) commits
the issuing bank to pay the amount deposited plus speci­
fied interest on a date specified. Because the instrument is
negotiable, it can be traded in the secondary market prior
to its stated maturity. Negotiable CDs typically are issued
in minimum denominations of $100,000 and are the most
volatile of the deposit measures. Negotiable CDs were
first introduced in 1961; prior to this date, M2 and M4
were identical as were M3 and M5.
r a t io s to t o t a l l iq u id a s s e t s .




ments in the ratio of each money stock measure to total
liquid assets over the period 1952-1 to 1975-IV. Total
liquid assets are defined here as M5 plus short-term Gov­
ernment securities, United States savings bonds, and
commercial paper. The secular trend in the money stock
measures points up a concern among those who favor the
broader money stock definitions. As a ratio of total liquid
assets,
has steadily declined over the period—from
almost half in 1952 to less than one quarter in 1975.
M2 has also declined as a percentage of total liquid
assets, although to a substantially smaller degree. The M2
ratio has declined from just over 60 percent in 1952 to
50 percent in 1975. The greater stability of the M2 ratio
reflects offsetting influences stemming from the declining
proportion of Mi in total liquid assets counterbalanced by
an increase in the proportion of savings and time deposits
at commercial banks. The share of the latter in total liquid
assets rose from 15 percent in 1952 to about 27 percent
in 1975. The proportion of total liquid assets held as
savings and time deposits at thrift institutions underwent
Chart I depicts move­ a similar increase, climbing from about 15 percent in

FEDERAL RESERVE BANK OF NEW YORK

1952 to about 30 percent in 1975. As a result of the
growth of savings and time deposits at commercial banks
and thrift institutions, the share of M3 in total liquid assets
increased from 75 percent in 1952 to 80 percent in 1975.
The behavior of M4 and M5 relative to total liquid assets
is also evident in Chart I. Over the period of its inde­
pendent existence, the ratio of M4 to total liquid assets
has changed little, with growth of CDs offsetting the
reduction of the M2 share. The ratio of M5 to total liquid
assets has increased from 80 percent in 1961 to 87
percent in 1975, reflecting the proportionate run-up in
both M3 and CDs.
Overall, Mx and M2 have declined as proportions of
total liquid assets, with the decline in the M2 ratio far less
dramatic than the halving of the ratio over the period.
On the other hand, the shares of M3 and M5 in liquid
assets have risen, while the M4 proportion has changed
very little. The pronounced decline in the ratio of M1 to
total liquid assets, however, is not necessarily evidence
that the usefulness of the M1 definition of money has
deteriorated.
could still be the aggregate with respect
to which wealth owners behave most systematically, and
the decrease in the ratio of M1 to total liquid assets simply
could reflect lower income and/or wealth elasticities of
demand for M1 compared with the savings components
included in the broader measures.
f in a n c ia l d e v e l o p m e n t s . Nevertheless, the financial in­
novations of the last several years seemingly have blurred
the distinction between savings and demand deposits
and may have weakened the close link between Mx and
the means of payment. Two important innovations were
the introduction of checking privileges on interest-bearing
savings accounts in New England and the spread of
noninterest-bearing checking facilities at thrift institutions
in states where checking privileges for savings accounts
are prohibited.
In January 1974, Congressional legislation became
effective authorizing all depository institutions in Massa­
chusetts and New Hampshire to issue NOW accounts
but prohibiting their introduction in other states. De­
pository institutions are permitted to pay a maximum
interest rate of 5 percent on NOW accounts; these
accounts can be issued only to individuals and nonprofit
organizations. In March 1976, Federal legislation which
sanctioned NOW accounts in the four other New England
states (Connecticut, Maine, Rhode Island, and Vermont)
became effective. The growth of NOW accounts is re­
ported in Table III.
Noninterest-bearing accounts with negotiable order of
withdrawal provisions, called payment order accounts,




269

were introduced by savings banks in New York in 1974.
In May 1976, New York State legislation which permitted
state-chartered thrift institutions to offer checking accounts
to individuals and nonprofit institutions became effective.
Of the eighteen states and territories (including Puerto
Rico) with mutual savings banks, ten currently permit
savings banks to offer checking accounts (Indiana, Dela­
ware, New York, New Jersey, Connecticut, Maine, Mary­
land, Oregon, Rhode Island, and Vermont). Federal
regulations prohibit Federally chartered savings and loan
associations from issuing checking-type deposits, but some
states have permitted state-chartered savings and loan
associations to offer either third-party payment orders
(Illinois) or demand deposit accounts (Connecticut,
Maine, Vermont, Rhode Island, and New York). Interestbearing check-like instruments, called “share drafts”, were
introduced by credit unions in October 1974.
Interest-bearing NOW accounts at commercial banks
are included in their savings deposit totals and therefore
are included in M2 but not in Mt. Interest-bearing NOW
accounts and noninterest-bearing checking accounts at sav­
ings and loan associations are included in M3 but
not in Mt or M2, and NOW accounts at mutual sav­
ings banks are included in M3 but not in M1 or M2.

Table III
NEGOTIABLE ORDER OF WITHDRAWAL (NOW) ACCOUNTS
IN NEW ENGLAND
In thousands o f dollars
i
Total of all
offering
institutions
in New
England

Month ended

..

1972: December

..........................

45,272

1973: December ..........................

|

1

Commercial (
banks
!

I

143,254

1974: December*

........................

312,576

1975: December ..........................

839,256

M utual
savings
banks

I Savings and
I
loan
1associations
1

45,272
143,254
65,249 1 213,661

33,666

386,560

93,756

358,940

1976: January ..............................

880,357

394,239

389,589

96,529

February ............................

942,779

435,080

406,217

101,482
112,196

.............................

1,091,004

543,456

435,352

April

M archf

.................................

1,206,880

627,525

456.166

123,189

May

...................................

1,324,030

738,586

458,288

127,156

1,415,712

804,328

476,112

135,272

June ..................................

* Congressional legislation enacted by the Congress which authorized all
deoository institutions in Massachusetts and New Hampshire to issue
N O W accounts became effective on January 1, 1974.
t Federal legislation which sanctioned N O W accounts in all depository in­
stitutions in Connecticut, Maine, Rhode Island, and Vermont became
effective on March 1, 1976.
Source:
(Research Department,
Federal Reserve Bank o f Boston), June 30, 1976.

Monthly Statistical Release on Now Accounts

270

MONTHLY REVIEW, OCTOBER 1976

Checking accounts at mutual savings banks which do
not earn interest are not included in any of the monetary
aggregates. Accounts at credit unions against which share
drafts can be written are included in M3 but not in Mi
or M2. If NOW accounts and thrift checking accounts
become increasingly important, M1 may well include a
declining proportion of the stock of assets used as a means
of payment.
Overdraft banking also may have reduced the role of
Mx in the payments process. Overdraft banking permits
the user to secure a loan simply by writing a check in
excess of the current balance in his checking account. If
the overdraft is used to make a purchase and is subse­
quently offset by the transfer of funds from the purchaser’s
savings account, in effect final payment is made from the
savings account and Mt has played no role at all in carry­
ing out the exchange.2
v e l o c it y . In addition to these changes in financial prac­
tices, concern over the appropriate definition of money
has been heightened as a result of the behavior of Mi
velocity relative to that of the other monetary aggregates.
Velocity measures the relationship between the stock of
money and the flow of income or payments; more pre­
cisely, the income velocity of money is the ratio of income
to money, or the rate at which the money stock “turns
over” in income transactions during a period. Velocity
can be viewed as the link between the money stock and
spending. This can be seen by writing current-dollar in­
come (Y) as the product of velocity (V) and money (M ):
Y=VM
In Chart II, movements in the velocity measures corre­
sponding to the five monetary aggregates discussed in this
paper are depicted for the period 1952-1 to 1975-1V. The
only velocity measure with a pronounced trend over the
entire period is V 1. During the period, grew at a 2.9
percent compound annual rate. The trend rate was 3.2
percent over the period 1952-66, but then it slowed to
2.4 percent per year over the 1967-75 period.3 The veloc­
ity measure corresponding to M2 exhibits a more moderate
upward trend prior to 1962 and no significant trend there­
after. Over the period 1962 to 1975-1V, V2 remained in
the narrow range of 2.30 to 2.45. The V3, V4, and V5

2 Proposals to
banks with funds
3 However, the
roughly equal to

cover automatically overdrafts at commercial
from savings accounts are pending.
growth of Vi over the 1970-75 years alone was
that in the 1952-66 interval.




series (corresponding to Ms, M4, and M5, respectively)
also displayed less trend than the Vx series.
These differences in observed velocity do not necessarily
indicate that M1 velocity is more difficult to predict. Pre­
dictability and stability are not the same thing.4 The de­
terminants of velocity are precisely the forces that deter­
mine the demand for money; i.e., the value of V is the
outcome of portfolio decisions about how much money
wealth owners want to hold relative to income. Hence,
concern with velocity suggests the importance of focusing
on the relative performance of money demand functions
that employ alternative definitions of money.
E M PIRIC AL EVIDENCE
ON THE DEFINITION OF M O N E Y

The declining share of Mx in liquid assets, the marked
upward trend in its velocity, and recent innovations in
financial practices all raise questions about continued
emphasis of this aggregate. However, as noted above, the
issue hinges on the relative predictive performance of
demand functions for money using alternative definitions
of the money stock. From the perspective of stabilization
policy, accurate prediction of the demand for money is
important in part because it can affect the precision with
which the consequences of alternative monetary growth
rates can be forecasted. More specifically, in many models
one element in assessing accurately the implications of
alternative money growth targets for real income, employ­
ment, and prices is a reasonably stable money demand
function.
Reliance on the predictive performance of the money
demand function under alternative definitions of money
has been urged as a criterion for determining the defini­
tion of money by Friedman and Meiselman [4] and Fried­
man and Schwartz [5].5 The preferred measure of money
is the one exhibiting the smallest prediction error. The

4 Friedman [3] suggests that the best definition of money is the
one that yields the most easily predictable velocity. However, in
comparing Mi and M2, Friedman identifies stability of the velocity
series with its predictability and concludes that M2 should be
preferred to Mi, because M2 velocity (V 2) is relatively constant
while Mi velocity (VO exhibits a pronounced upward trend.
5 Friedman and Schwartz suggest that “the desideratum is a
monetary total whose real value bears a relatively stable relation
to a small number of variables that theoretical considerations lead
us to believe affect the real quantity of money demanded. . . .”
[5, pp. 139-40]. In their empirical work on the definition of
money, however, Friedman et. al. employed money income corre­
lations rather than explicit money demand functions.

FEDERAL RESERVE BANK OF NEW YORK

271

(2) Comparison of aggregate and disaggregated predic­
empirical analysis in this paper is confined to M1? M2,
and M3, the three money measures for which the Federal tions of M2 and M3 (Goldfeld [6]). For example, assume
Reserve System currently sets growth ranges.6 The empiri­ the choice is between Mx and M2. M2 predictions can be
cal approach employed below uses two separate objective made directly from an M2 equation (referred to as an
criteria—the first previously employed by Brunner and aggregate prediction) or by summing the predicted values
Meltzer [1] and the second introduced by Goldfeld [6]— of separate Mx and time deposit equations (referred to as
to assess the relative predictive performance of alternative a disaggregated prediction). If the disaggregated prediction
money demand functions and therefore to shed light on is superior to the aggregate prediction, this suggests that
“aggregation is inflicting some positive harm” [p. 594]
the appropriate definition of money:
(1) Comparison of percentage prediction errors for M1?and Mx is preferred to M2. On the other hand, if the ag­
M2, and M3 demand equations (Brunner and Meltzer [1]). gregate prediction outperforms the disaggregated, M2
would be preferred to M^ This second approach is
applied below to M2, M3, and the sum of savings and time
deposits at commercial banks and thrift institutions.
SPECIFICATION OF THE DEMAND FUNCTION. D e m a n d

funC-

6 For previous empirical research on the stability and predictive tions for M1? M2, and M3, time and savings deposits at
performance of the money demand function, see Brunner and commercial banks other than large negotiable CDs (T),
Meltzer [1], Laidler [8], Hamburger [7], and Goldfeld [61. Laidler concluded that the M2 function was more stable than the savings deposits at thrift institutions (S), and the sum of
T and S (TS) were estimated over the period 1952-11
Mi function, but the other studies favored the Mi definition.

C h a rt II

VELOCITY OF ALTERNATIVE MEASURES OF THE M O N E Y STOCK, 1 9 5 2 -7 5
V e lo c ity

V e lo c ity

6

Note:

Shaded areas represent periods of recession as defined by the N ation al Bureau of Economic Research, except for the latest recession

which is tentatively judged to have ended in March 1975.
Source:

Board o f G overnors of the Federal Reserve System.




272

MONTHLY REVIEW, OCTOBER 1976

to 1974-11.7 As in most standard formulations, the
demand for money is assumed to depend on income and/
or wealth and interest rates. The income version reflects
the transactions view of the demand for money.8 The
transactions view immediately suggests the
definition,
given that Mx appears to be the empirical counterpart
to the theoretical construct of the medium of exchange.
However, the increasing use of third-party payment
privileges from savings accounts, NOW accounts, and
checking accounts at thrift institutions suggests that the
empirical counterpart to the means of payment is no longer
as clear-cut as it once was.
The wealth version of the demand for money empha­
sizes the nonpecuniary return associated with holding
money. Money is viewed as a temporary abode of pur­
chasing power which bridges the gap between sales and
purchases of goods and services. Viewed in this fashion,
money includes the medium of exchange but may be
broader. The empirical counterpart is not so clearly
defined as for the means of payment, and the choice of
assets to include is, of necessity, an empirical issue.
Based on the general considerations just discussed, to
derive the money demand equation estimated in this
paper the long-run desired level of real money holdings
(m*) is first specified as a function of real income (y),
interest rates (r), and real net wealth (a):
m*=L(r,y,a)
Portfolio adjustment costs are assumed to induce wealth
owners to adjust actual money holdings (m) to the longrun desired level of money balances (m*) with a lag.
Using a stock-adjustment approach, the change in real
money balances between periods t and t-1 is related to the
discrepancy between desired money balances in t and
actual holdings in t-1, or
mt - m t-i = a (m*t - m t.1)
where a is the speed of adjustment. Following Goldfeld,
the basic form of the equations is logarithms in the levels

of the variables. Measures of money and time and savings
deposits, income, and wealth are deflated by the gross
national product (GNP) implicit price deflator. The com­
mercial paper rate and one rate on savings deposits are
used in each regression, except with M2 where only the
commercial paper rate is employed.9 The wealth variable
is net worth of households, constructed for use in the
MPS econometric model of the United States economy.
The basic form of the money demand equation tested is
as follows :
In mt = a+b In yt + c In at + d In RCPt
+ e In RSt + f In mt_i
where m is one of the various measures of the money sup­
ply or a component thereof deflated by the GNP price
deflator, y is real GNP, a is the real value of net worth of
households, RCP is the commercial paper rate, and RS is
one of the savings account rates.10 Three versions were esti­
mated for each aggregate depending on whether income
only, wealth only, or both income and wealth were in­
cluded as independent variables in the regression. Below
the results for the income versions of the Mi, M2, M3, T,
and TS equations are reported, along with the wealth
version of the S equation. More complete results can be
found in Meyer [9].
PERCENTAGE RMSE IN AND OUT OF SAMPLE. The mOSt
widely used criterion for comparing money demand func­
tions under alternative definitions of money involves
comparison of in-sample and out-of-sample predictive
performance. The in-sample results are based on the

0 The savings deposit rate was not statistically significant in the
Mo equation. The savings rate used in each equation is reported
in the following table. RSAV is a weighted average of rates on
savings and time deposits at commercial banks (R T D ), at savings
and loan associations, and at mutual savings banks; RTHR is a
weighted average of rates at savings and loan associations and
mutual savings banks. In both cases, the weights are the propor­
tions of the total stock of assets in the previous quarter:
Equation
Savin? rate
Mx ........................... RSAV
T
........................... RTD
S
RTHR
7 The end point of 1974-11 was selected because of indications
TS ........................... RSAV
that all equations exhibited structural shifts at or after this date.
M3 ........................... RSAV
8 According to this view, the essential property of money is
10 While this is a conventional specification of the money de­
that of a medium of exchange, facilitating purchases and sales of
goods and services. Money is held in portfolios between receipt mand function (see, for example, Goldfeld [6 ]), it is by no means
from sales and expenditures on goods and services because of the the only possibility. One characteristic of this equation which
transactions costs of moving between money and interest-earning should be pointed out is that it implies that nominal money hold­
assets. Income is included in the money demand function as a ings adjust to price disturbances within a quarter, while adjust­
measure of the volume of transactions, and interest rates reflect ment to changes in all other independent variables occurs with
a lag.
the opportunity cost associated with holding money.




273

FEDERAL RESERVE BANK OF NEW YORK
Table IV
PREDICTIVE PERFORMANCE OF M i, M2, AND M3 EQUATIONS

Table V
OUT-OF-SAMPLE PREDICTIONS OF M1# M2, AND Ms

RM SE as a percentage o f the mean

RM SE as a percentage o f the mean

In sample
Definition
Mi
m2
M3

..............................................................................
............................................................................
..........................................................

Out of sample

Definition

1962-69

1970-75

1952 to
1974-11

1962-75

0.42
0.47
0.44

0.96
0.96
1.23

Mi ...................................................................................
M2 ................................................................................
M3 ................................................................................

0.77
1.27
1.45

1.20
0.65
1.01

RM SE = Root mean squared error.

RM SE = Root mean squared error.

percentage root mean squared errors (RMSEs)11 over
the period 1952-11 to 1974-11; the out-of-sample results
are based on four-quarter dynamic simulations over the
period 1962-1 to 1975- IV.12
Table IV reports the RMSE as a percentage of the
mean for the Mx, M2, and M3 equations for both insample and out-of-sample predictions. The results of this
comparison favor the Mi definition in that the Mx equa­
tion yields the best predictive performance both in and
out of sample. The prediction error for Mx is 0.42 percent
in the in-sample results, compared with 0.47 percent for
M2 and 0.44 percent for M3. In the out-of-sample results,
the percentage prediction error is 0.96 percent for both
Mx and M2 and 1.23 percent for M3.13

These results appear to provide support for the Mx def­
inition. To determine whether or not there has been any
deterioration in the relative predictive performance of Mi,
compared with M2 and M3, the period used for out-ofsample tests was broken into two subperiods: 1962-69 and
1970-75. The basic equation described above, using alter­
native definitions of money, was simulated dynamically
for these periods. The results are reported in Table V.
During the earlier 1962-69 subperiod, the Mx defini­
tion yields the smallest percentage prediction error (col­
umn 2 of Table V ). During the 1970’s, however, the M2
definition yields the smallest prediction error and Mi
yields the largest error (column 3 of Table V ).14 The sub­
period results suggest, therefore, that the performance of
the Mi equation relative to the M2 and M3 equations has
deteriorated in the last several years.
AGGREGATE VS. DISAGGREGATED PREDICTIONS OF M2 AND

The second empirical criterion for the definition of
money, discussed above, involves a comparison of aggre­
gate and disaggregated predictions of the broader mea­
sures of the money stock. For example, if M2 is the
appropriate definition of money, it should be possible to
predict movements in M2 more accurately using an ex­
plicit demand function for M2 rather than making predic­
tions from separate demand functions for its components,
Mj and T. Similarly, predictions of M3 based on an aggre­
gate Ms function should be compared with predictions
derived from separate equations for its components.
This criterion is applied to the results of dynamic outof-sample simulations over the 1962-75 period. For the
m 8.

11 The RMSE is defined as the square root of the sum ot
squared errors divided by the number of forecasts. The percentage
RMSE is computed by deflating the RMSE by the mean of
the actual values of the variable being predicted. Again, while
comparison of RMSEs is a conventional procedure, it is not the
only way to distinguish between these aggregates. Use of the
RMSE means that large errors are given a great deal of weight.
12 In a dynamic simulation, the predicted value of the money
supply is used in the following period as the value of the lagged
dependent variable. A regression is initially estimated over the
period 1952-11 to 196i-IV. To determine the predicted value of
money in 1962-1, actual values are substituted for income, wealth,
and interest rates in 1962-1 and for the value of money in 1961-IV.
To determine the predicted value of money in the next quarter,
actual values of income, wealth, and interest rates in 1962-11 and
the predicted value of money in 1962-1 are substituted in the
equation. The same procedure is followed to yield predicted values
in 1962-III and 1962-IV. At this point, four quarters are added
to the sample period, the regression is rerun through 1962-IV,
and predicted values are determined for 1963. The procedure is
continued through 1975.
13 The four-quarter out-of-sample dynamic predictions are
extended through 1975 to increase the number of observations
available to discriminate between aggregates.




14 This result could depend heavily on the deterioration in the
Mi demand function evident since mid-1974. On this subject, see
Enzler, Johnson, and Paulus [2].

274

MONTHLY REVIEW, OCTOBER 1976
Table VI
AGGREGATE VS. DISAGGREGATED PREDICTIONS
Four-quarter post-sample predictions
Root mean squared errors
1970-75

1962-69
Definition
Aggregate

Disaggregated

Aggregate

Disaggregated

M 2 .....................

5.02

4.98

3.14

3.66

M 3

...............................

8.87

7.06

7.93

8.27

TS

.....................

7.63

6.84

7.75

8.20

entire period, the out-of-sample results provide some
superficial support for M2 over Mi in that the disaggregated
M2 prediction error exceeds the aggregate prediction
error. However, there is also evidence suggesting that
T and S should be treated as a single aggregate and, if
this is the case, the appropriate comparison is between
Mi and M3 and not between Mx and M2. The disag­
gregated prediction of M3 dominates the aggregate pre­
diction, suggesting a preference for Mx over M3.
Subperiod results for the post-sample prediction period
are reported in Table VI. The disaggregated predictions
are consistently superior in the 1960’s, although the dif­
ference with regard to M2 is small.15 The results for the
subperiod of the 1970’s, however, are just the reverse.
The aggregate equations consistently yield the better pre­
dictions. Taken together, these results seem to support
the deterioration in the relative performance of the Mi
demand function in the 1970’s that was evident in the
results based on percentage prediction errors for Mi, M2,
and M3. And, since these results also suggest that T and S
should be treated as an aggregate, M3 rather than M2
may be the preferred definition.

predicted depends on identifying those asset categories
with respect to which the private sector exhibits sys­
tematic behavior and, in part, it depends on estimating
demand functions for those categories.
A number of recent financial developments have raised
concern about the continued usefulness of the narrow
definition of the money stock. The secular decline in the
ratio of Mx to total liquid assets may suggest a diminished
role for Mi in the financial system. The greater stability
of the velocity measures corresponding to the broader
money stock measures has been cited in support of broad­
ening the definition of money. Yet neither of these secular
trends represents direct evidence about the appropriate
definition of money.
To provide direct evidence, this paper compared the
predictive performance of demand functions for the Mi,
M2, and M3 definitions of money. The evidence based on
the full sample and full post-sample periods indicated that
the private sector behaved in a more systematic fashion
with respect to the M1 variable than with respect to the
broader measures. When the period was broken into the
1960’s and the 1970’s, however, evidence of deterioration
in the predictive performance of the Mi equation relative
to the M2 and M3 equations was uncovered. While the Ml
definition yielded the best results during the 1960’s, the
broader money stock measures were generally superior
in the 1970’s. Consistent with this, moreover, was the
superiority in the 1970’s of the aggregate predictions of M2
and M3 in comparison with the disaggregated predictions.
REFERENCES

fl] Brunner, K., and Meltzer, A. “Predicting Velocity: Implica­
tions for Theory and Policy”. Journal of Finance (May 1963).
[2] Enzler, J., Johnson, L., and Paulus, J. “Some Problems of
Money Demand”. Brookings Papers on Economic Activity
(Washington, D.C.: The Brookings Institution, 1976:1).
[3] Friedman, M. “How Much Monetary Growth?” Morgan Guar­
anty Survey (February 1973).
[4] Friedman, M., and Meiselman, D. “The Relative Stability of
CONCLUSION
Monetary Velocity and the Investment Multiplier in the
United States, 1897-1958”. Stabilization Policies (Englewood
Cliffs, New Jersey: Prentice-Hall, Inc., 1963).
The ability of the Federal Reserve System to predict
and
A. Monetary Statistics of the
the response of the economy to monetary policy may de­ [5] Friedman, M., (NewSchwartz, National Bureau of Economic
United States
York:
Research, 1970).
pend, in part, on the precision with which it can predict the
[61 Goldfeld, S. “The Demand for Money Revisited”. Brookings
portfolio behavior of the private sector. In turn, the
Papers on Economic Activity (Washington, D.C.: The Brook­
precision with which private portfolio behavior can be
ings Institution, 1973:3).
[7] Hamburger, M. “The Demand for Money by Households,
Money Substitutes, and Monetary Policy”. Journal of Political
Economy (December 1966).
[81 Laidler, D. “Some Evidence on the Demand for Money”.
Journal of Political Economy (February 1966).
15 Similarly, for the period 1961-71, Goldfeld found that dis­ [9] Meyer, L. “Alternative Definitions of the Money Stock and
aggregated predictions of M2 outperformed the aggregate results
the Demand for Money”. (Research Paper 7614, Federal
(see [6, pp. 592-95]).
Reserve Bank of New York, June 1976).




NEW YORK

1976

JANUARY

(Rec. 1 /1 5 /7 6 )

P r i o r i t i e s f o r th e I n t e r n a t i o n a l Monetary System: An Address by P a u l A. V o lc k er
P e t r o - D o l l a r s , LDC’ s and I n t e r n a t i o n a l Banks: An A ddress by R ic h a rd A. Debs
The B u siness S i t u a t i o n
The Money and Bond M arkets in December
Index f o r th e Year 1975
FEBRUARY (Rec. 2 /1 7 /7 6 )
Remarks b e f o r e t h e New York S t a t e Bankers A s s o c i a t i o n by P a u l A. V o lcker
The B usiness S i t u a t i o n
The Money and Bond M arkets i n J a n u a r y
T rad in g i n B a n k e r s 1 A cce p ta n ces
MARCH (Rec. 3 / 1 1 /7 6 )
I s s u e s in th e F in a n c in g o f C o r p o r a te Tender O f f e r s : S ta te m e n t by R ich a rd A. Debs
T re a su ry and F e d e r a l R eserve F o r e i g n Exchange O p e r a t i o n s , By Holmes and P a rd e e
The B usiness S i t u a t i o n
Money and Bond M arkets i n F e b ru a ry
APRIL
The B usiness S i t u a t i o n
th e Money and Bond M arkets i n March
Banking S t r u c t u r e i n New York S t a t e :

P r o g r e s s and P r o s p e c t s

MAY (Rec. 5 / 1 4 /7 6 )
Our Changing F i n a n c i a l System an Address by R ic h a rd A. Debs
o f Monetary Control
The B usiness S i t u a t i o n
The Money and Bond M arkets i n A p r i l
The Strategy

JUNE

(R ec. 6 /1 1 /7 6 )

S u s t a i n i n g t h e B u s i n e s s E x p a n s i o n — An A d d r e s s b y P a u l A. V o l c k e r
The B u s i n e s s S i t u a t i o n
The Money a n d B ond M a r k e t s i n May
E v a lu a tio n th e L eading I n d i c a t o r s
T re a s u ry and F e d e r a l R e s e rv e F o re ig n E xchange O p e r a t i o n s : I n t e r i m
R e p o r t b y A l a n R. H o l m e s a n d S c o t t E. P a r d e e




JULY

(Rec.

7/14/76)

The B u s in e s s S i t u a t i o n
T h e Money a n d Bond M a r k e t s i n J u n e
A P r o b a b i l i s t i c A p p ro a c h to E a r l y W a rn in g o f C h an g es in Bank
F in a n c ia l C on d itio n
AUGUST

(R ec. 8 /1 2 /7 6 )

The B u s in e s s S i t u a t i o n
T h e Money a n d Bond M a r k e t s i n J u l y
The W eakness of B u s in e s s L oans in th e C u r r e n t R eco v ery




SEPTEMBER

(R ec. 9 /1 3 /7 6 ')

T r e a s u r y and F e d e r a l R e s e r v e F o r e i g n E x c h a n g e O p e r a t i o n s , by A la
H o lm e s a n d S c o t t E. P a r d e e
The B u s in e s s S i t u a t i o n
T h e M oney a n d B ond M a r k e t s i n A u g u s t
OCTOBER (Rec. 1 0 - / 8 / 7 6 )
The C o n t r i b u t i o n s and L i m i t a t i o n s o f "M onetary” A n a l y s i s : An A ddress by Pau
The B u s in e s s S i t u a t i o n
The Money and Bond M arkets i n September
A l t e r n a t i v e D e f i n i t i o n s o f th e Money S to c k and th e Demand f o r Money
QUARTERLY REVIEW
DECEMBER (Rec. 1 2 /2 1 /7 6 )
The New York C ity B udget: Anatomy o f a F i s c a l C r i s i s
M easuring C a p a c ity U t i l i z a t i o n i n M a - u f a c tu r in g
The B u sin e ss S i t u a t i o n
C u r re n t Developments
The Changing C om position o f th e Loabor F orce
The F i n a n c i a l M arkets
C u r re n t Deveopments
I n t e r e s t r a t e B e h av io r i n th e C u rre n t Economic Recovery
T re a s u ry and F e d e r a l R eserve F o r e ig n Exchange O p e r a tio n s