YRJO JAHNSSON LECTURES

Asset Accumulation
and
Economic Activity
Reflections on
Contemporary Macroeconomic Theory

JAMES TOBIN

BASIL BLACKWELL - OXFORD

 I
Real Balance Effects Reconsidered

Keynes and Underemployment Equilibrium

Let us go back to the 1930s and to the theoretical con-

troversy ignited by Keynes’s iconoclastic General Theory,
published in 1936. Keynes claimed to have found under-
employment equilibrium. The word is equilibrium. For
Keynes was not content to assert the empirical possibility,
even the likelihood, of involuntary unemployment, indeed
of stubbornly persistent involuntary unemployment. That
was, after all, scarcely an extravagant observation after
seven years of world-wide depression and a decade of high
unemployment in Britain. As a matter of theoretical prin-
ciple, he went much further. The General Theory
ny iin itsis 0opening three ch r

“denies their existence, furthermore, in a competitive

1], Keynes, The General Theory of Employment, Interest and

Money, Macmillan, London, 1936.
2 Rear BALANce Errects RECONSIDERED

economy; Keynes does not say that the mechanisms fail

because of misguided government interventions in the price
system—minimum, wages and the like—or because of pri-
vate combinations in restraint of trade—trade unions and
industrial cartels. Instead he challenges orthodoxy on sacred
ground, its faith that competition will so adjust prices of
products and factors as to eliminate excess supplies, or
demands, in all markets. He does not say merely that this
process may take a very long time; he says that it does not
work at all. The challenge was deliberate and explicit. The
mistaken orthodoxy he described, somewhat imprecisely,
as “classical” theory, and its foremost exponent he iden-
tified as his own Cambridge friend, colleague, fellow-
student of Marshall, Professor A. C. Pigou. Pigou was not
slow to take up the challenge. i
Keynes’s argument for equilibrium with involuntary i

unemployment had two strands. The first was an explana-

tion why the price of labor—the money wage rate—did not
fall in the face of excess labor supply. The second was an
explanation why even if it did fall, as it should in a well-
behaved competitive market, the result would not be an
increase in employment. (Keynes hinted that the two points
might be related, in the sense that one reason for labor
resistance to wage deflation could be an intuitive apprecia~
tion of its futility.) Pigou’s response was devoted primarily
to the second point, and that is also the primary concern of
this lecture.
As for the first point, Keynes’s argument was that un-_

unemployed workers can bid for jobs against each other and
against employed workers. It is set by employers unilater-
ally or in concert with their employees, organized or unor-
ganized: in either event, the chief concern is the wage rate
 ReaL BALANCE ErrecTs RECONSIDERED 3

relative to wages in competing firms or in comparable

occupations and situations, and not the availability of
cheaper workers at the factory gates. The special position of
the inside work force, whether organized or not, derives
from its possession, individually and collectively, of firm~
specific skill and experience.
These observations, which I have liberally paraphrased,
Many of them are

wage levels—are eroded only very slowly by unem-

ployment. But they do not say that money wages will not
erode at all. The same economic climate that generates high
unemployment also impairs employers’ ability to pay high
and increasing money wages to their existing employees.
Layoffs, plant closings, bankruptcies, and threats of such
isa fr ployed workers with choice between...
wage conces The Great Depression, as
“well as the recent severe recession, provides numerous
examples. Thus Keynes's first point serves better to empha~
size the difficulty and slowness of melting frozen wage
levels or wage-increase patterns than to establish that they
never melt at all. Presumably on similar reasoning, Profes-
sor Pigou concentrated his counter-fire on the second point.

2M. N. Baily, “Wages and Employment Under Uncertain

Demand,” Review of Economic Studies, January 1974.
, “On the Theory of Layoffs and Unemployment,”
Econometrica, July 1977.
C. Azariadis, “Implicit Contracts and Underemployment
Equilibria,” Journal of Political Economy, December 1975.
 4 REAL BALANCE ErFects RECONSIDERED

Deflation and Aggregate Demand

According to Keynes, there are theoretically conceivable

and empirically important circumstances in which reduc-__
tion of money wage rates would not succeed in increasing.
“aggregate demand for goods and services. Production and
employment ‘would remain unchanged. Prices would
be lower in the same proportion as wages. Real wages,
real profit margins, indeed all real variables, would be
unaffected. In short, the real equilibrium of the
economy—unemployment and all—is independent. ofthe
level of money wages and prices. It is of course not inde-
pendent of the real wage. Employers would offer more
employment at a lower real wage, and workers—whether
previously employed or unemployed—would be glad to
accept it. But _unless.a-reduction.of.the money wage would.
ww increase aggregate real demand, there is no
hich the mutual latent. willingness..to
“demand : and to supply more labor at.a lower real wage—or
possibly even at the same real wage—could be actualized.
This is the Keynesian impasse.
In fact, Keynes himself described one way out of the
impasse, a mechanism which considerably restricted the
generality which he seemed to be claiming for it in the first
part of his great book. This mechanism, sometimes called
the “Keynes effect,’ was the following: At lower money
wages, prices, and incomes, supply of money would be
larger in real volume or in Keynes’s own wage units. The
transactions demand for cash would be smaller; the excess
stock of cash would bid up the prices of interest-bearing
securities and lower interest rates. At lower interest rates
real investment would be higher. Thus aggregate demand,
further boosted by the multiplier, would expand output and
employment. Hence wage and price deflation is an equi-
 REAL BALANCE EFFECTS RECONSIDERED 5

valent—a bizarre and second-best equivalent in Keynes’s

view—to expansionary monetary policy. If one will work,
so will the other.
Here enters the famous liquidity trap, the situation in
which an increase in the real quantity of money, whether by
active central bank intervention or by deflation, will be
ineffectual. This is the situation in which interest rates
relevant to investment are already as low as they can go.
The absolute floor for nominal interest rates is zero, the
return on money itself. The effective floor, at which people
will be indifferent between holding money idle and buying
interest-bearing assets, might be, Keynes thought, a bit
above zero. Even short term government paper would
have to provide a minimal fractional gain relative to hoard-
ing money, to compensate for transactions costs, imper-
fect liquidity, and risk. Long term interest rates, which
could be regarded as an average of current and expected
short rates, would be held above zero by expectations, or
simply fears, that short rates will rise from rock bottom
in future.
Like money wage patterns, such stickiness might be
regarded as a disequilibrium phenomenon, in principle
transient even if in practice stubborn. But all Keynes really
needs anyway is the zero floor, combined with the possibi-
lity that the full employment equilibrium real interest
rate—the Wicksellian natural rate that equates full em-
ployment investment and saving—is below zero. That is a
possibility which, it seems, cannot be excluded by a priori
restrictions on technology and taste.

The Pigou Effect

Or can it be? Pigou did not regard the liquidity trap impasse
as particularly plausible. But he accepted it for the sake of
 6 Rear BaLaNce Errects RECONSIDERED

argument, and pointed out that the real value of the wealth
of the community wo ed by deflation. Money,
“and other assets denominated in money, are part of the
public’s wealth. At lower prices their purchasing power is
greater, while the N | in the form of
goods is ‘unchange . People save ‘to accumulate wealth to
provide for their consumption, or that of their heirs, in
future periods and contingencies. When the real value of
their existing assets is increased, these purposes are more
adequately satisfied and they will increase current con-
sumption at the expense of saving. This is the Pigou or “real
balance”’ effect?
Pigou’s first try misfired. Kalecki* reminded him in print
that the largest part of private holdings of monetary assets,
including bank deposits counted as money, had direct or
indirect counterpart in private debt. Deflation raised the
burden of the debts as much as the real value of the assets, As
“‘Pigou acknowledged, the correction left him with a much
smaller net base. One component is the part of the public’s
money stock supplied directly by the government: cur-
rency and coin and their equivalent in central bank de-

A possible second component is the public’s holdings of

non-monetary interest-bearing government. obligations.
“Whether ‘the: *, or any fraction of them, constitute net
wealth has been a matter of controversy at least since

5 A.C. Pigou, “The Classical Stationary State,” Economic Journal,

December 1943.
*M. Kalecki, “Professor Pigou on ‘The Stationary State’—A
Comment,” Economic Journal, April 1944,
A. C. Pigou, “Economic Progress in a Stable Environment,”
Economica, August 1947,
 REAL BALANCE EFFECTS RECONSIDERED 7

Ricardo,® and the debate still rages today.° I will discuss it at

some length in the third lecture. The question is, whether .
taxpayers, anticipating that taxes will be levied to service a
Targer real public debt, regard themselves as poorer in the
same degree as the bondowners regard themselves.as.weal-
thier. Even if the government debt is washed out on this
account, the monetary base remains. So presumably do
those government obligations which in the liquidity trap
have become. the equivalent of money, bearing zero or
has observed, sufficient defla-
minimal interest. As Leontief
tion of money wages and prices can make it possible to
purchase the whole GNP with one dime.’ Presumably by
then the Wicksellian natural interest rate, equating full em-
ployment saving and investment, would be well in the
positive orthant and the Keynesian impasse would be
escaped.
Pigou relied on the response of consumption and saving
ced has been
by. Patinkin,®
to wealth; this tradition, reinfor
a4 m heorists. However, an argument with
parallel import
i could be made that increasing the real value
of monetary. wealth.is. favorable..to.investment...Portfolio
theory suggests that wealth-owners, finding themselves
not only with larger wealth but with a larger share ofit in

5D. Ricardo, The Principles of Political Economy and Taxation,

E. P. Dutton, New York, 1912, pp. 161, 198-9.
®R. Barro, “Are Government Bonds Net Wealth?”, Journal of
Political Economy, November/December 1974.
7 Quoted by P. Samuelson in ‘“‘A Brief Survey of Post-Keynesian
Developments,” Keynes’ General Theory: Reports of Three
Decades, Robert Lekachman, ed., St. Martin’s, New York,
1964, p. 333.
* D. Patinkin, “Price Flexibility and Full Employment,” Ameri-
can Economic Review, September 1948, and in Money, Interest and
Prices, 2nd ed., Harper and Row, New York, 1965.
 8 Reat BALANCE Errects RECONSIDERED

monetary form, will wish to shift towards goods or

equities. Thus they may lower the effective yields re-
quired of investments in consumers’ or producers’ durable
goods, relative to those available.on- money and Treasury...
bills ‘or bonds. This change in the structure of interest rates
‘could be favorable to investment even if rates on secure
liquid nominally denominated assets were stuck in the
trap.

think he did, busy as he was with practical affairs—that the

Pigou effect of deflation could be duplicated by fiscal pol-
icy, specifically by government spending or tax reduction
financed by printing money. That would not only provide
direct fiscal stimulus but also, like Pigou’s deflation, add
to the public’s wealth and specifically to its monetary
component.
In some modern minds, the question may arise how in
either case wealth is increased in any meaningful sense.
From a larger perspective, does.not.the wealth of a nation
consist of its real productive assets, human and nonhuman?
“These are what an outside observer, say in a space satellite
with a powerful telescope, would enumerate and value. (I
and he leave aside claims on the rest of the world. The
argument concerns a closed economy; Keynes did not con-
test the traditional view that deflation could work for a
small open economy with a fixed exchange rate.) How cana
nation make itself richer either by printing pieces of paper
or by iincreasing their value by charging each other fewer of
them in exchanges of goods and services? The answer, I
think, is in two parts. First, the social contrivance of com-
monly acceptable money, facilitating contemporaneous
and intertemporal exchange, is of social value to the nation
 REAL BALANCE EFrects RECONSIDERED 9

collectively as well as to holders individually. Its contribu.

tion to national wealth may well depend on its aggregate
amount. Second, in the particular situation under discus-
sion, it is only a breakdown—whether temporary or per-
manent—in internal arrangements which is responsible for
failure to use fully the real productive assets of the
economy. If additional monetary assets remedy the break-
down, the real national wealth they represent is the value of
the resulting stream of gains in production and consumption.

Irving Fisher on Deflation and Debt

Earlier in the same Great Depression another great econo-

mist, Irving Fisher, had reached a diagnosis precisely the
opposite of Pigou’s.® Fisher thought that reflation, not defla-
tion, was the remedy.He was struck by the increased bur-
d } at lower prices imposed on debtors—corporations,
proprietors,. home-owners, farmers. Debt squeezes,
‘defaults, and bankruptcies, he thought, intensified and
spread the slump in economic activity. He urged mea-
sures—monetary expansion, devaluation; marking up gold
prices—designed to restore commodity prices to pre-
Depression levels. For Fisher in 1932-3, more even than
Keynes in 1936, raising prices was a step indispensable _
fo recovery, not just an incidental byproduct of other
measures.
Irecall Fisher’s position not solely from Yale patriotism
but to bring our attention back to the casual “washing out”
of private debts and credits in the reckoning of the base for

° 1. Fisher, “The Debt Deflation Theory of Great Depressions,”

Econometrica, October 1933, p. 337.
, 100% Money, Adelphi, New York, 1936, especially
pp. 119-34.
 10 Rea BaLtance Errecrs RECONSIDERED

the Pigou effect. The gross amount of these “‘inside”’ assets

was and is orders of magnitude larger than the net amount
of the base. Aggregation would not matter if we could be
sure that the marginal propensities to spend from wealth
were the same for creditors and debtors. But if the spending
propensity were systematically greater for debtors, even by
a small amount, the Pigou effect would be swamped by this
Fisher effect.
There are indeed reasons for expecting, or at least for
suspecting, just that. The population is not distributed be-
tween debtors and creditors randomly. Debtors have bor-
rowed for good reasons, most of which indicate a high
marginal propensity to spend from wealth or from current
income or from any liquid resources they can command.
Typically their indebtedness is rationed by lenders, not just
because of market imperfection but because the borrower
has greater optimism about his own prospects and the value
of his collateral, or greater willingness to assume risk and to
die insolvent, than the lender regards as objectively and
prudently justified. Business borrowers typically have a
strong propensity to hold physical capital, producers’ dur-
able goods. Their desired portfolios contain more capital
than their net worth—they like to take risks with other
people’s money. Household debtors are frequently young
families acquiring homes and furnishings before they
earn incomes to pay for them outright; given the diffi-
culty of borrowing against future wages, they are
liquidity-constrained and have a high marginal -Propen-
sity to consume.
When nominal prices and wages are deflated, debt service
isa higher propo ors’ incomes, and the reduc
tion or elimination of their margins of equity disqualifies
them from further access to credit. Bankruptcies and
defaults do likewise, and transmit the distress of debtors to
their creditors, threatening the solvency and liquidity of
 REAL BALANCE EFFECTS RECONSIDERED 14

individual lenders and financial institutions. _Debtor

corporations, their ;
i
to restoration of financial structure above real investment.,
The declines ii real tharket value of their equities due to the
greater burden of their debts far surpass the gains to credi-
tors. These items in the Fisher scenario may well over-
shadow the positive effects of the increased real value of
creditors’ nominal assets.
These considerations do not apply solely to ancient his-
tory. Imagine the distress which would occur if debtors
who have borrowed in the 1970s in anticipation of con-
tinued inflation were suddenly to find themselves con-
fronted by price stability. Maybe Leontief is right that
sufficient deflation would make existing coins capable of
buying the whole GNP. It would also make existing debts
an astronomical multiple of the GNP.

Short and Long Run Effects of Prices on Aggregate Demand

To give Pigou and Fisher each his due, I am led to make a

distinction between the “long run” consequences of de-
flation and the “‘short run.” Perhaps the Pigou effect applies
to the first, the Fisher effect to the second. I am sure that
Pigou himself was conscious of this distinction, for he
entitled his final contribution to this subject ““The Classical
Stationary State.” His successors have been less careful,
confidently stressing real balance effects in short-run mac-
roeconomic analysis.
To understand the long run Pigou effect, we must use our
imagination and carry out a counter-historical ‘‘as if?’ ¢x-
periment, Imagine two alternative histories of the same
economy, over the same period of time. A common feature
of the two histories is the nominal value of the monetary
base at each point in time; this is the same in history | and
 12 REAL BALANCE EFFECTS RECONSIDERED

history I. At the outset and for an extended period there-

after both economies are in a liquidity trap and are suffering
unemployment. During this period nominal wages and
prices are 50% lower in history II than in history I, though
real wages, other relative prices, and real quantities are in
the beginning the same. The period lasts long enough for all
debts earlier contracted to mature and to be repaid or recon-
tracted; alternatively, all debts outstanding were contracted
with foreknowledge of the prices in each history. Pigou
}} would say that history If would show higher employment
and output than history I and would reach full employment .

Unexpected ‘price reduction in a single history—low

prices following high prices sequentially—is an entirely
different matter. Contracts made when prices were higher
and expected to be higher remain in force. Fisher’s observa-
tions apply. It would take a very long time before such
contracts were worked off, and even then the economy
would not be the same as if they had never existed.

Prices and Output in Short-ruan Macro Models

At this point, 1 would like to make a mildly technical

digression. If, as I suspect, Fisher was very likely right about
the short run effects of movements of price level on aggre-
gate real demand, what does this imply for short-run macro
models? Does it mean that the IS curve in r-Y space shifts
upward as the price level rises in cyclical expansion? That
the IS curve in r-p space fora given level of output, e.g. the
full employment level, is upward sloping? Before drawing
these conclusions we should remember a short-run price
level effect of much greater importance now than in Fisher’s.
t This is the progressive structure of taxation relative to.
nominal, incomes.and profits; during a short run before
 Rea BALANCE Eprecrs RECONSIDERED 13

degislative adjustment, the fact that taxes area larger shareof _

lincome at higher prices works in. Pigou’s direc-.
eee even if the Fisher effect is stronger, and the
suswers to the questions above about the relation of IS loci
so price levels are affirmative, this does not mean that the IS
and LM loci jointly yield a positive rather than negative
ociation of Y and p for given settings of policy. The
_ Keynes effectthe fact that a given nominal supply of
-@oney is smaller in real value the higher the price
Eevel—still works in the conventional direction. Particu-
larly at high levels of output and interest, far from the
Squidity trap, it may well dominate any direct price level
effects on wealth and spending. The curvature of the liquid-
preference schedule, then, contributes asymmetry to the
_ Stuation. Altogether, aggregate demand couldbe positively _
zelated to price level at low levels of output and interest
autes but negatively related closer to full employment. This
does open some possibility of multiple equilibria.
The situation can be analyzed with the help of some
grams. In Figures 1-3, LM and IS loci are drawn for
gen values of the nominal money stock M and of real
wational product Y. They represent combinations of inter~
#sz rate and price level consistent on the one hand with
smonetary balance, and on the other hand, with balance in
gemand and supply of goods. Figure 1 shows the conven-
tonal story, including the Pigou effect. The LM locus is
s=pward sloping: the real stock of moncy is lower when the
eice level is higher, and it takes a higher interest rate to
ace the public to handle transactions with smaller cash
aoldings. The IS locus is downward sloping: according to
“Se conventional Pigou effect, a higher price level means
tess demand for goods and services, and it takes a lower
aaterest rate to offset this effect. As the diagram also says,
“spansionary monetary policy ‘M” shifts LM to the right,
sehile expansionary fiscal policy “FP” shifts IS upwards. An
 14 Rear BALANCE Errects REC
ONSIDERED

FIG. 2 reverse Pigou effect, case |

(given M, Y)

FIG, 3 reverse Pigou effect, case [1

(given M. ¥)
 Rear Barance Errects RECONSIDERED 15

increase in Y shifts LM up (more demand for money at any

given P) and shifts IS down (assuming the marginal pro-
pensity to spend is less than 1). The effects on p and rare also
summarized in the accompanying table.
Figures 2 and 3 show a reverse Pigou effect or Fisher
effect. The [S locus is upward sloping, because the spending
by debtors encouraged by higher p exceeds the spending by
creditors deterred. In Figure 3 the effect is so great that IS is
steeper than LM, and the comparative statics of monetary
and fiscal policy are reversed.
Figures 4 and 5 are special cases of Figures 1 and 3, in
r

\
\
\
EY
\
\
LM
N
xs
Y

p
FIG. 4 conventional Pigou effect (given r, Y)

J 1S
/

f
A y
o
a
LM
a“
“ a“

FIG. 5 reverse Pigou effect (given r, Y)

 16 Rear BaLaNce Errects RECONSIDERED

which the Keynes effect is eliminated. This would apply in

the liquidity trap, or in the event of an accommodative
monetary policy that pegged the interest rate. Note that in
Figure 5, as in Figure 3, p and Y are positively rather than
negatively associated.
Figures 1-5 have two uses. A direct application is to the
classical flexible price world with real output Y supply-
determined. Reversing the Pigou effect alters some tradi-
tional comparative static results. An indirect application is
to the Keynesian world of demand-determined Y, This is
done in Figures 6 and 7. The first panels of Figures 6 and 7
r

LM
(3) (2) Py) (9)

ye?
P3 ys

P2
Ww
PL

Po

FIG. 6 Y
 REAL BALANCE Errects RECONSIDERED

LM
(Pg) (P3) (P2) (Py) (o)

FIG. 7

show IS and LM loci in traditional (Y, r) space. Each locus

assumes M and p, but a family of curves are drawn for
various values of p, po < pi < pz... The various IS/LM
solutions, numbered in the panel, are translated into the
second panel, where they form the Y® relations. In Figure 6
this has the expected downward slope, implying for
 18 Reav BALANCE Ersects RECONSIDERED

example that a reduction of money wages (depicted as a

downward shift of YS) would raise employment and out-
put. But in Figure 7, the Pigou effect reversal makes YP
change directions, rendering ambiguous the consequences
of a money wage reduction (compare shifts of Y{ and Y$)
and opening the possibility of dual equilibrium (Y§).

Conclusion
for Theory and Policy

There is another important difference between the two

cases earlier distinguished, the “‘as if” comparison
of alter-
native price levels and the sequential deflation. Deflation
in
real time—unless engineered by governmental fiat rather
than by markets—may generate expectations of furthe
r
deflation. Now expected deflation increases the demand for
money, making it more attractive relative to other assets,
particularly to goods and equities in goods. This effect
counters the price level effect and may be stronger. If
so,
deflation does not correct the initial deficiency in aggreg
ate
demand that triggered it. Then deflation has no stopping
point. The symmetrical case is hyper-inflation, in which
the
velocity of money rises astronomically.
Both Keynes and Pigou were aware of this problem as
a
practical matter but did not regard it as a part of their
theoretical game. They were too purist. Recall the central
issue: Does the market economy, unassisted by govern
~
ment policy, possess effective mechanisms for elimin
ating
general excess supply of labor and productive capaci
ty?
This question applies to real time and to sequential
pro-
cesses. Therefore the static long run “Pigou effect”
does not
entitle anyone to give a positive answer."

“J. Tobin, “Keynesian Models of Recession and Depres

sion,”
American Economic Review, Papers and Proceedings,
May 1975,
 Rear BaLance Errecrs RECONSIDERED 19

This does not mean that Keynes wins the more abstract
battle of theoretical principle. He did not’show the existence
of an excess-supply equilibrium, at least not in the meaning
af the magic word equilibrium in the classical, or neoclassi-
zal, economics he was criticizing. In that meaning, equilib-
rium is a stationary state, and a state in which expectations
are fulfilled. A sequence in which wages and prices are
falling, and in which debts are embarrassing debtors who
never anticipated prevailing wages and prices, is not such a
state, Pigou succeeds in restricting “equilibrium” to situa-
tions in which markets clear, and Keynes’s proposed
equilibrium with involuntary unemployment does not
qualify.
But why should Keynes care about such semantics? His His, v-e"
important message was that Pigou’s equilibrium may not
be globally stable, that even ifitis stable, disequilibrium can
be protracted and stubborn. The forces which lower money
wages and prices are slow and weak, and those which
translate deflation or disinflation into greater real demand
are uncertain. As Keynes also knew, protracted under-
production and under-utilization severely damages the
marginal efficiency of capital. In mild and short-lived reces-
sions investment is buoyed by belief that high employment
and prosperity are the long-term norm. Once this confi-
dence is destroyed, as contemporary events again demon-
strate, it is terribly difficult to revive it. The practical moral
is that active policy, along with market response, is part of
the social mechanism for maintenance or restoration of
equilibrium.