Economic Heresies

By the same author

THE ACCUMULATION OF CAPITAL
COLLECTED ECONOMIC PAPERS (3 VOLS.)
ECONOMIC PHILOSOPHY
ECONOMICS: AN AWKWARD CORNER
THE ECONOMICS OF IMPERFECT COMPETITION
AN ESSAY ON MARXIAN ECONOMICS
ESSAYS IN THE THEORY OF ECONOMIC GROWTH
EXERCISES IN ECONOMIC ANALYSIS
FREEDOM AND NECESSITY
INTRODUCTION TO THE THEORY OF EMPLOYMENT
THE RATE OF INTEREST AND OTHER ESSAYS
ECONOMIC
HERESIES Some
Old-Fashioned Questions
in Economic Theory

JOAN ROBINSON

Palgrave Macmillan
 © 1971 by Basic Books, Inc.

All rights reserved. No part of this publication

may be reproduced or transmitted, in any form
or by any means, without permission.

First published in the Uni ted States of America 1971

First published in the Uni ted Kingdom 1971
First published in paper-covered edition 1972
First Edition 1971
Reprinted 1972
Published by
THE MACMILLAN PRESS LTD
London and Basingstoke
Associated companies in New York Toronto
Dublin Me/bourne Johannesburg and Madras

ISBN 978-0-333-12883-1 ISBN 978-1-349-81576-0 (eBook)

DOI 10.1007/978-1-349-81576-0

Tbe paperback edition of this book is sold subject to the

condition that it shall not, by way of trade or other-
wise, be lent, resold, hired out, or otherwise circulated
without the publisher's prior consent, in any form of
binding or cover other than that in which it is published
and without a similar condition including this condition
being imposed on the subsequent purchaser.
FOREWORD

While working on this book I have profited very much from

arguments and discussions with a number of colleagues and
pupils. I would like particularly to mention Professor A. Asi-
makopulos of McGill University, Mr. John Eatwell of Trinity
College, Cambridge, Professor Don Harris of Wisconsin Univer-
sity, Mr. Jan Kregel of the University of Bristol, and Dr. Amit
Bhaduri.
A large part of Chapter 6, Prices and Money, appears as an
article in the Journal of Money, Credit and Banking, November
1970, under the title "Quantity Theories Old and New."
Cambridge
December, 1970
INTRODUCTION

The orthodox doctrines of economics which were dominant in

the last quarter of the nineteenth century had a clear message.
They supported laisser faire, free trade, the gold standard, and
the universally advantageous effects of the pursuit of profit by
competitive private enterprise. This was acceptable to the au-
thorities in an expanding and flourishing capitalist world, espe-
cially to the authorities in England, l which was still felt to be
the dominant center and chief beneficiary of the system.
The arguments on which the economists' doctrines were
based, however, had little relevance to the problems on which
they pronounced. The structure of economic theory was a de-
ductive system based on a priori premises, such as that the be-
havior of individuals is governed by the principle of maximizing
utility; the argument was set up in terms of the effects of a dis-
placement from an already established equilibrium or in terms
of comparisons between two equilibrium positions, without any
discussion of the process of changing from one to another. For
instance, the case for free trade, which was a central part of
orthodox teaching, was based upon comparing the situation of
two countries, each in isolation in a stationary state, with given
resources fuIIy employed under the rules of perfect competition,
1 This was true of England, rather than of Great Britain or the United
Kingdom as a whole.

vii
viii / Introduction

with the situation in which they are trading in equilibrium, im-

ports being equal to exports, with everything else unchanged.
(Even then, the argument that neither country could improve its
position by protection contained a logical flaw; this scandal was
hushed up in academic teaching until it broke out again in the
1930s, when orthodox doctrine as well as the world economy
was in a state of disarray.) 2 The lack of correspondence be-
tween the assumptions of theory and the facts in reality did not
matter because the doctrines were acceptable; since the main
doctrine was laisser faire, no prescriptions for any positive policy
were required; there was no need to bother about studying situa-
tions to which policy might have to be applied. The economists
could go on happily categorizing empty boxes without feeling
any need to fill them with information.
After 1918, the situation of the British economy in the world
was drastically changed but the economists had not believed
themselves to be influenced by English national interests; their
doctrines had always been set up as universal truths; they were
now carried over into a situation where they were no longer ap-
propriate. There was one fact which was particularly awkward.
While the United States was enjoying the long boom of the
1920s, Great Britain was suffering from low profits and heavy
unemployment. Now, it was an axiom of the orthodox scheme,
inherited from the classics, that there cannot be unemployment
because of Say's Law. When a program was suggested in 1929
for public expenditure to relieve unemployment, it was answered
by the famous Treasury View,3 according to which there is a
fixed fund of saving available to finance investment. If the gov-
ernment borrows part to spend upon public works, there will be
an exactly equal reduction in foreign investment, so that un-

2 See A. Lerner, "The Diagrammatical Representation of Demand

Conditions in International Trade," Economica (August 1934).
3 See Memoranda on Certain Proposals Relating to Unemployment,
Cmd.3331.
 Introduction / ix

employment due to the reduction of the balance of trade would

more or less completely offset the increase in employment due
to public works. Soon the world slump set in. The total bank-
ruptcy of the orthodox theory became evident to all but its
professional devotees and the Keynesian Revolution emerged
from the ruins.
On the plane of the development of ideas, the main point of
the General Theory was that it broke out of the theological sys-
tem of orthodox axioms; Keynes was looking at the actual situa-
tion and trying to understand how an actual economy operates;
he brought the argument down from timeless stationary states
into the present, here and now, when the past cannot be changed
and the future cannot be known.
At the time it seemed like a revolution; a new day had dawned
in which economics was going to be a serious subject concerned
with serious problems. But the day soon clouded over. After
1945, Keynes' innovations had become orthodox in their turn;
now governments had to admit that they were concerned with
maintaining the level of employment; but in respect to economic
theory the old theology closed in again. Keynes himself began
the reconstruction of the orthodox scheme that he had shattered.
"But if our central controls succeed in establishing an aggregate
volume of output corresponding to full employment as nearly as
is practicable, the classical theory comes into its own again
from this point onwards. . . . It is in determining the volume,
not the direction of actual employment that the existing system
has broken down." 4 He had been too much occupied with im-
mediate problems to think very much about what the neoclassi-
cal theory (which he called classical) really entailed. In some
moods he found capitalism morally and aesthetically abhorrent
but his object was to save it from destroying itself; he did not

4 John Maynard Keynes, The General Theory of Employment, Interest

and Money (London: Macmillan, 1936), pp. 378-379.

A·
x / Introduction

press his criticism either of the system or of its apologists very

deep. In particular he did not distinguish between profitable in-
vestment and socially beneficial investment, and he was rather
averse to considering problems connected with the distribution
of income between families in an industrial nation (the prob-
lem of distribution of income in the world had not yet come
into fashion) .
A new orthodoxy was soon established by a simple device. A
substitute for Say's Law was provided by the assumption that
a well-managed Keynesian policy keeps investment running at
the level which absorbs the saving forthcoming at full employ-
ment. The rest of the doctrines of the neoclassics could then be
revived.
The neo-neoclassics, however, seem to have overlooked some
serious inconsistencies in the old scheme which made the new
synthesis unsatisfactory.
For instance, there is an inconsistency between the assump-
tion of a perfectly competitive market and the assumption that
every trader is maximizing his gain. A group of individuals, say
sellers of a particular commodity, can increase the gain for each
other by acting in concert. This was the flaw in the case for
free trade. Even within the strict assumptions of the argument,
it could be shown that either country can gain an advantage
from turning the terms of trade in its favor by restricting the
supply of its product and reducing its demand for the product of
the other. In the pursuit of self-interest, each country will try to
gain at the expense of the rest of the world. Free trade is not an
equilibrium position unless it is enforced by general agreement. 5
Another drawback of the neoclassical scheme was that it was
fully worked out only for a stationary state while the neo-neo-
classics wanted to make use of it to discuss the now fashionable

5 Cf. T. Scitovsky, "A Reconsideration of the Theory of Tariffs,"

Review of Economic Studies (Summer 1942).
 Introduction / xi

concept of continuous growth. A hitherto stationary economy

cannot begin to grow without going through a drastic transfor-
mation-for instance, its investment industries, which have
been operating at a level just to keep their stock of equipment
intact, must be expanded sufficiently to allow for net investment.
Moreover, if the impulse to grow arises from a desire of in-
dividual households to save, how is the initial Keynesian slump
to be overcome? Marshall wanted to discuss growth but he was
daunted by the problem of adapting the formal theory to deal
with it. "In fact we are here verging on the high theme of eco-
nomic progress; and here therefore it is especially needful to
remember that economic problems are imperfectly presented
when they. are treated as problems of statical equilibrium, and
not of organic growth." 6 The neo-neoclassicals plunged in with-
out any such hesitation.
The most serious problem concerned the concept of "factors
of production." One view was expounded by Wicksteed: "We
must regard every kind and quality of labour that can be dis-
tinguished from other kinds and qualities as a separate factor;
and in the same way every kind of land will be taken as a
separate factor. Still more important is it to insist that instead
of speaking of so many £ worth of capital we shall speak of so
many ploughs, so many tons of manure, and so many horses, or
foot-pounds of 'power.' Each of these may be scheduled in its
own unit, and when this has been done the enumeration of the
factors of production may be regarded as complete." 7 This
point of view underlies the Walrasian scheme.
From another point of view factors of production are treated

6 A. Marshall, Principles of Economics (London: Macmillan), p. 461.

Except where otherwise stated, references to Marshall's Principles are to the

eighth edition.
7 P. H. Wicksteed, An Essay on the Co-ordination of the Laws of
Distribution (London: Macmillan, 1894), p. 33. Reprinted in Scarce
Tracts series, London School of Economics.
xii / Introduction

in the broad categories of Ricardo-land, labor, and capital.

The nature of capital was always a source of anxiety and trou-
ble. Marshall divided the factors of production into land, labor,
and waiting, and he regarded the real cost of production (as op-
posed to rent of natural resources) as composed of the efforts of
work and the sacrifice of waiting. Now, waiting consists of own-
ing property and refraining from selling out and spending the
proceeds. "That surplus benefit which a person gets in the long
run by postponing enjoyment, and which is measured by the
rate of interest (subject as we have seen to certain conditions),
is the reward of waiting. He may have obtained the de facto
possession of property by inheritance or by any other means,
moral or immoral, legal or illegal. But if, having the power to
consume that property in immediate gratifications, he chooses to
put it in such a form as to afford him deferred gratifications, then
any superiority there may be in deferred gratification over those
immediate ones is the reward of his waiting. When he lends out
the wealth on a secure loan the net payment which he receive!;
for the use of the wealth may be regarded as affording a numeri-
cal measure of that reward." 8
Thus the total stock of capital of an economy at any moment
has two aspects: it is a great collection of various kinds of equip-
ment, stocks, and work in progress, and it is a sum of wealth.
There is a third aspect of capital which mediates between the
other two, that is, capital as finance. An entrepreneur may own
wealth or borrow from rentiers. The command of finance per-
mits him to employ labor; wages are paid out week by week in
advance of the production of salable goods; when the goods are
sold (if all has gone according to plan) the initial finance is
recovered with a profit. If the delay is, say, six months, a loan

8 Marshall, Principles, 1st ed., 1890, p. 614. Here Marshall clearly

regards waiting as simply owning capital. In later editions a similar
passage is applied to waiting in the sense of saving (8th ed., 1920, p.
233), and the argument becomes extremely obscure.
 Introduction / xiii

equal to half one year's wage bill provides a wage fund which
permits him to employ labor indefinitely, as long as the same
conditions hold. He also advances raw materials, power, and
other running costs which will be recovered from sales over the
course of some months and he advances equipment, the finance
of which is to be recovered over the course of some years. Thus
finance is the link between capital as physical means of produc-
tion and capital as wealth.
Real productivity from the point of view of society belongs
to physical equipment and materials which embody technical
know-how; profit belongs to capital as finance embodied in busi-
ness organization, and interest (the reward of waiting) belongs
to rentier wealth. Marshall was evidently conscious of the diffi-
culty of identifying the reward of waiting with the productivity of
physical capital goods; when capital comes into the analysis,
the smoke screen of ambiguity which covers the whole argument
of the Principles grows thicker than ever; but the neo-neoclassi-
cals do not seem to have felt any difficulty about it. (A recent
textbook, The Neoclassical Theory of Production and Distribu-
tion by Professor C. E. Ferguson, is valuable because the author,
as he declares, has faith in the theory and is not afraid to make
it clear and definite. 9 He sets out a number of propositions in
which inputs consist of labor and physical capital goods of vari-
ous kinds, following Walras; then he turns to the economy as a
whole and treats capital as a whole as an input which can be
treated in the same way as the input of, say, the services of a par-
ticular kind of machine. Wages are accounted for by the mar-
ginal product of labor, and profit by the marginal product of
capital. )
Apart from logical incoherence, the flaw in the new orthodoxy
destroys the validity of its message. The deepest layer in neo-

9 C. E. Ferguson, The Neoclassical Theory of Production and Distribu-

tion (Cambridge: Cambridge University Pre!,s, 1969), see p. XVII.
xiv / Introduction

classical thought was the conception of society as a harmonious

whole, without internal conflicts of interest. Society, under the
guidance of the hidden hand, allocates its resources between
particular uses in such a way as to maximize utility; society de-
cides the allocation of resources between present consumption
and accumulation to permit greater consumption in the future.
Accumulation is represented by Robinson Crusoe transferring
some of his activity from gathering nuts to eat to making a
fishing rod; or by the sturdy peasant who cuts timber in the
forest to build himself a durable hut. 10 Here saving means a
sacrifice of present consumption or leisure to increase produc-
tivity for the future; saving and investment are two aspects of
the same behavior. Keynes destroyed this part of the analogy
by showing that, in a private enterprise economy, investments
are made by profit-seeking firms and it is they who decide for
society how much it will save. But he let the rest of the analogy
stand. He was immediately concerned with a situation where
investment, on any criterion, was manifestly too low; he main-
tained that, while beneficial investments were to be preferred,
any investment was better than none. But once Keynes has be-
come orthodox, the case is altered. If we are to be guaranteed
near-full employment the question comes up, what form should
employment take? The neo-neoclassicals have dodged that ques-
tion. Adopting the slogan that the rate of return on investment
to an individual firm measures, corresponds to, or is derived
from, the marginal product of capital to society, they have re-
constructed the case for laisser faire.
The new doctrine is now coming to a crisis. The first part of
the doctrine--that the amount of investment is controlled by
how much society wants to save-was discredited in the great
slump. The second part, that the form of investment is con-
trolled by the principle of maximizing the welfare of society,

10 Cf. Marshall, Principles, p. 351.

 Introduction / xv

is being discredited by the awakening of public opinion to the

persistence of poverty--even hunger-in the wealthiest nations,
the decay of cities, the pollution of environment, the manipula-
tion of demand by salesmanship, the vested interests in war, not
to mention the still more shocking problems of the world outside
the prosperous industrial economies. The complacency of neo-
laisser faire cuts the economists off from discussing the economic
problems of today just as Say's Law cut them off from discuss-
ing unemployment in the world slump.
It seems that this second crisis, like the first, is due to the
uncritical acceptance of the apologetic that seemed plausible
(though it was never logical) in the late nineteenth century. In
these essays I attempt to find the roots of modern orthodoxy in
the neoclassical tradition.
It seems that modern orthodoxy is mainly based upon Walras,
which narrows its scope. The tradition of Marshall, though full
of confusions and sophistries, was much richer. Many of the
problems that we used to discuss in the 1930s have been lost
from the canon. I hope that a re-examination of the old-fashioned
questions will help to clear the way for a more penetrating dis-
cussion of the problems of today.
CONTENTS

INTRODUCTION vii

1 STATIONARY STATES 3
Walras 4
Kingdom Come 9
Marshall 13
The Wicksell Process 14

2 THE SHORT PERIOD 16

Supply Price 18
Expectations 22
Effective Demand 23

3 INTEREST AND PROFITS 25

Walras 26
Marshall 27
Keynes 30
xviii / Contents

The Neo-neoclassics 32
The Productivity of Investment 33
The Pseudo-production Function 34
Ricardo and von Neumann 39
The Rate of Exploitation 42
Savings and Investment 44
Effective Demand 49

4 INCREASING AND DIMINISHING

RETURNS 52
Irreversibili ty 53
"Marginal Products" 55
Economies of Scale 58

5 NON-MONET ARY MODELS 64

Market Prices 65
A One-commodity Economy 67
Money and "Real Forces" 71
Money in a Golden Age 74
"Real" Instability 75

6 PRICES AND MONEY 77

The Theory of Interest and Money 79

Counter-revolution and Restoration 82
The Chicago School 85
 Contents / xi.~

The Theory of Employment 88

Inflation 90·
The Unit of Account 95

7 THE THEORY OF THE FIRM 97

Perfect and Imperfect Competition 98
Monopoly and Oligopoly 102
Choice of Technique 103
Macro and Micro Theory 107

8 GROWTH MODELS 109

Harrod 109
Profits and Saving 117
Innovations 125
The Meaning of Neutral Inventions 125
The Vintage Model 129
Neoclassical Vintages 132
Induced Bias 139

CONCLUSION 141
INDEX 145
Economic Heresies
 1
STATIONARY
STAT E S To find a stationary economy in real
life we should look for some corner of
the world untouched by war and trade where tradition rules and
the cycle of production and distribution repeats itself from year
to year, from generation to generation, without changes in popu-
lation, technical innovations, or concentration of wealth. But, in
such a society, prices, incomes, and property are also ruled by
tradition. Analogies with modern capitalism may be found in it,
but they will be too far-fetched to be convincing. The stationary
state in economic theory was not supposed to describe any ac-
tual society. It was an analytical device intended to throw light
upon relationships in the changing world in which the econo-
mists were living.
For Adam Smith, Ricardo, and Marx the central subject of
discussion was the accumulation of means of production and of
property. In a stationary state there is no accumulation. The
neoclassical school, which came into fashion in the second half
of the nineteenth century, introduced two quite distinct ways of
eliminating accumulation from models which were intended in
other respects to correspond to reality. One was to consider the
situation, so to speak, today, with the physical stocks of com-
modities and means of production that happen to be in existence;

3
4 / Economic Heresies

the other was to consider the situation at Kingdom Come when

the process of accumulation has been completed and no one
finds it worthwhile to acquire anything more. These two opposite
kinds of stationary states are unfortunately often confused in
modern textbooks. 1

WALRAS
The first kind is the basis of Walras' market where the rela-
tive prices of commodities are determined by supply and de-
mand. Walras (and his modern disciples) tells us more about
the commodities than about the people concerned. Each trader
enters the market with something to offer. Is he a specialist?
If so, his command of purchasing power depends very much
upon the price that his particular commodity commands in
terms of other things. He may do very well out of the market
or he may come away with less than will feed his family till the
next meeting. This aspect of the matter is very little discussed.
To ensure that there must be an equilibrium pattern of prices
reconciling the supply of arbitrarily given stocks of various
kinds of commodities with the demand, which is governed by
whatever happens to be the tastes and desires of the traders, it
is necessary to allow for the possibility of a zero price for a
commodity for which supply exceeds demand at any positive
price. If some of the traders have nothing else to offer except
such a commodity, what is to become of them?
There is one case that has been observed in real life which
corresponds pretty well to the Walrasian conception of equilib-
rium between supply and demand arrived at by a process of
"groping" through bids and offers by traders. This is in a pris-
oner-of-war camp.2 The men are kept alive more or less by

1Cf. above, p. xiii.

2See R. A. Radford, "The Economic Organization of a P.o.w.
Camp," Economica (November 1945).
 Stationary States / 5

official rations and they receive parcels from the Red Cross once
a month. The contents of the parcels are not tailored to the
tastes of the individual recipients, so that it is possible for each
to gain by swapping what he wants less for what he wants more.
A market is formed when the parcels are opened and prices,
offered and bid, are quoted in terms of cigarettes. Trading and
re-trading take place until demand is equated to supply for each
commodity (there are not likely to be any zero prices in such a
situation!) and each trader, at the prices ruling, has no further
desire to exchange one thing for another. 3 Each trader has an
initial endowment (his parcel) more or less the same as every
other and each comes away with a roughly equal value of con-
sumables. The problem of the distribution of consumption being
governed by prices is therefore not very important.
Anyone who happened to prefer just what was in his parcel
need not trade. Each swaps only to get something that he likes
better than what he has. Thus trade makes everyone subjectively
better off. (This is a good advertisement for trade which does
not apply to specialist producers, say, of cocoa beans or rubber,
who may find one day that the laws of supply and demand have
reduced them to misery.)
In the prison-camp market, cigarettes are used as a unit of
account and, perhaps, as a medium of exchange in three-cor-
nered transactions, but there is no store of value, no "link be-
tween the present and the future." 4 All commodities are con-
sumed within the month and a fresh set of prices is established
when a fresh lot of parcels arrive. In this sense it is a non-mone-
tary economy. Though prices are quoted in a single unit, the
value of an ounce of each commodity really consists in its po-
tential purchasing power over all other commodities. The overall
price level in terms of cigarettes is no more significant than the

3 No doubt a sense of what is proper behavior rules out the forma-

tion of monopolies; cf. above, p. x.
4 S~e Keynes, General Theory, p. 293.
6 / Economic Heresies

price level in terms of pounds of cheese, or of anything else.

The whole point of this case is that the parcels are simply
given. Each item has its opportunity cost in terms of other things
that it might be exchanged for but there are no costs of produc-
tion and no choice of what to produce. To extend the notion of
a non-monetary stationary state to an economy with production
going on is not so easy. We have to assume that there are given
stocks, not of consumable commodities, but of "factors of pro-
duction." There is a given labor force, an area of land with
particular types of soil in particular locations, a certain amount
of productive equipment, such as buildings, roads, and machines
of various kinds, and stocks of raw materials. The equipment
and stocks were produced in the past but the amount of each
kind in existence "today" is quite arbitrary. The raw materials
are used up and reproduced week by week and the equipment
is kept intact in the process of production, like well-farmed land.
Workers offer their services for wages and owners of land,
machines, and so forth, offer the services of means of production
for a hire-price or rent. (It is misleading, as we shall see in a
moment, to call machines "capital" and their rent "profit.")
Recipients of income buy commodities produced by the factors,
according to their needs and tastes and according to the pur-
chasing power that each commands. There is no separate source
of income from organizing production. (Managers are a type of
worker.) Workers may hire machines or owners of machines
hire workers, or there may be a disembodied spirit, an auc-
tioneer, who registers all bids and offers. Prices are quoted in
terms of some unit of account. At intermediate stages in the
bargaining process there may be an excess or a deficiency of
demand for a particular commodity. Its price is then raised or
lowered, and its output increased or reduced, as the case may
be. The stock of equipment cannot be altered but insofar as par-
ticular machines are versatile they can be directed to one use or
another according to which offers the best rent.
 Stationary States / 7

This argument is very hard to grasp, for a process which

would take a long (perhaps indefinite) time to work out is
conceived to be instantaneous. But the story is not meant to be
taken literally. The only point of it is to argue that there is a
set of prices, wages, and rents that provides an equilibrium
position.
In equilibrium, the supply and demand for each commodity
are equal. This means that, with the ruling prices and his own
income, no individual wants to buy more of one thing or less of
another than he is doing. Similarly, with the ruling prices at
which commodities can be sold and the ruling levels of wages
and rents, no producer would find some other combination of
factors profitable. The price of a pound, or yard, or pint of any
commodity is just sufficient to cover its average share of the cost
of wages for particular types of labor, replacement of raw mate-
rials, and rents for the particular pieces of equipment that are
producing the flow of output in which it forms part.
If the supply of any particular machine is in excess of demand,
its rent is zero. Similarly, if labor were in excess of demand,
wages would be zero. This is clearly incompatible with equilib-
rium, for the labor force could not be kept in being with nothing
to eat. To get out of this difficulty it is assumed that technical
conditions are such that there is substitutability between factors
of production, in the sense that the output of a commodity can
always be increased by using a larger physical amount of one
factor with a fixed amount of the others. When the use of one
factor alone is increased, the proportional increment of product
is less than the proportional increment of the factor; there are
diminishing returns between factors or falling marginal pro-
ductivity of the increasing factor as the physical proportions of
factors change. The operation of the auctioneer ensures that, in
equilibrium, no factors are employed in a combination where
one enjoys increasing returns; if it did, the feturn on employing
it would be greater than its hire-price and more would be used.
8 / Economic Heresies

In equilibrium the factors are used in such proportions, in

the production of each commodity, that the value of the marginal
product of each (in terms of the unit of account) is not less than
its hire-price per physical unit. Thus if, at a certain stage in the
bargaining process, some labor were unemployed, the wage
would be reduced and it would become profitable to employ
more labor with the given physical amount of other factors.
This is plausible enough when the other factor is land. Agricul-
tural technique can be adapted to a wide range of intensity of
cultivation. Machines are not so versatile, but the adaptation
may be supposed to be made, up to a certain limit, by shift
working or there may be better and worse designs among the
arbitrarily given stock of machines so that a lower wage rate
makes machines with a lower output per man-hour worth using.
In that case, in equilibrium (provided that the total stock of
machines is more than enough to provide full employment at a
subsistance wage) the least productive machine in use in some
or all lines may have zero rent, like Ricardo's marginal land.
In this model a low wage rate does not create a problem of
effective demand. The lower the wages, the higher the rents.
Workers consume less and property owners more.
Each piece of land and each machine receives its appropriate
rent, depending on its technical productivity, the availability of
other factors, and the demand for its product. There is no gen-
eral rate of profit on the value of capital or expected rate of
return on new investment. If we introduced into the picture a
rate of interest-a price for purchasing power today to be re-
paid (or re-borrowed) at a future date-the equilibrium of the
system would be upset.
Interest may be regarded as a hire-price for finance but it is
quite unlike the wages and rents of factors of production. It is
expressed as value per unit of value while they are expressed as
value per unit of a physical service-a man-hour of labor of a
particular type or the use for a year of a particular machine.
 Stationary States / 9

With a standard rate of interest in the market, each machine

and each acre would have a capital value such that its rent
divided by that value was equal to the rate of interest. These
values would bear no regular relation to the past cost or present
reproduction cost of machines. It would then be profitable to
produce those for which value exceeded cost. Expectations of
change, investment, and saving have to be brought into the story
and the auctioneer has to be conceived to be capable of register-
ing bids and offers spread over an indefinite future. Moreover,
there has to be some story to account for how the rate of interest
is determined.

KINGDOM COME
A story is provided in the model of the other kind of sta-
tionary state, though not a very plausible one. There, the rate of
interest is determined by the tastes and habits of the owners of
property. They require a certain return on their wealth-the
"reward of waiting"-in order to prevent them from consuming
it in "present gratifications." 5 They get this return by lending
finance to entrepreneurs who use it to acquire and operate
means of production which earn profits. So long as the rate of
interest-the hire-price of finance-is less than the rate of re-
turn to be confidently expected on investment, the stock of
capital goods is accumulating. The stationary state is reached
when the two are equal.
For the sake of a convenient label, we may call this model
Pigovian, for it was Pigou who drew out the concept of a sta-
tionary state from the others that it is mingled with in Marshall's
Principles. In the Pigovian model the stock of equipment is not
just arbitrarily given "today." The rate of interest is a supply

1\ See above, p. xii.

10 / Economic Heresies

price for capital. In stationary equilibrium, the value of the stock

of capital goods in existence is such that the value of the an-
nual net profit covers this supply price.
This model, though not particularly convincing, is much less
difficult to grasp than the first one. We are not confined to a thin
slice of time "today." No change is occurring but time rolls on
from the past into the future. The stock of equipment and the
amount of wealth are constant because the owners and operators
choose that they should be. No one is saving or making net invest-
ment because no one wants to do so. Production is organized,
not by a ghostly auctioneer but in business firms which operate
plants and employ labor. Equipment is being kept intact, not
because it happens to exist, but because the firms decide to keep
it intact. (We can now admit amortization of capital as an ele-
ment in cost of production, which is difficult to accommodate
in the first model.) However, since there are perfectly confident
expectations that the future will be exactly like the past, there
is no more scope for "enterprise" than in the Walrasian model.
The firms must be supposed to pay their managers and earn just
enough gross profit to keep the value of their capital intact and
to pay the standard rate of interest on it.
We can now leave the bewildering calculation of relative
values and introduce a price level in terms of money. The wage
rate, the hire-price of labor, is fixed in terms of a unit of money.
(There may be different rates for different levels of skill, etc., but
the composition of the labor force has become adjusted to the
pattern of demand, just like the stocks of equipment.)
Now, with given money-wage rates, a given corpus of tech-
nical knowledge, and a uniform rate of profit on capital, there
is a determinate set of money prices for all commodities and
means of production. (It is possible to borrow from the other
model an arbitrarily fixed supply of "land" but it is more con-
genial to this model to suppose that all means of production are
reproducible.) Technical conditions specify the input-output
 Stationary States I 11

table for the whole economy in terms of labor and means of

production, each in its own physical unit. The requirement of a
uniform rate of profit settles relative prices, including the wage
rate in terms of any commodity, and the money-wage rate settles
money prices. (If money as a medium of exchange is in use, the
quantity of it in existence is just what is required to pay wages
and carry out transactions at the ruling prices.) 6
The flow of money incomes-wages and interest-is purchas-
ing the flow of output of consumption goods; the composition of
output is such that the consumers are willing to buy what is
offered at the ruling prices. The stock of equipment is appropri-
ate to producing this output while keeping itself intact. There
may be other techniques known but those that have been chosen
are those which (at the ruling prices and wage rates) make it
possible to earn a profit equal to the ruling rate of interest on
finance invested. None can earn more and any that earned less
would not have been installed.
The price of each product is such that it can pay the wage
for all the labor required to produce it directly, and indirectly
through the replacement of stocks of materials and wear and
tear of plant, while paying the rate of profit on the value of all
the capital directly and indirectly required to produce it.
The cost of labor in terms of his own product to each em-
ployer is such that the excess of the value of output over the
wage bill pays all other costs. Thus the cost of labor in terms of
product is less the greater the value of capital per man employed.
The real wage in the cost-of-living sense depends upon the level
of prices of those commodities which workers want to buy.
Given the rate of profit, the level of real wages in both senses
depends upon the technology in use.

6 Since the rate of interest has to be equal to the "reward of waiting"

there is no scope for introducing a rate of interest based on the demand
and supply of money.
12 / Economic Heresies

The micro-equilibrium of the system depends upon the rule

that competition is obliging the firms to produce a given output
at minimum cost. To each individually, the wage rate, the rate
of interest, and all prices are given independently of his own be-
havior; he combines the factors of production in such a way that
the marginal net product of each is not less than its supply price.
That is, in considering how much of each factor to employ, the
cost of other factors and the selling price of the product are
taken into account. (This is a different concept from the mar-
ginal physical product of the Walrasian model; Marshall set the
fashion for confusing them in his famous footnote about the
marginal shepherd, 7 by cooking the example so as to make them
identical. )
But now we come to a serious snag. There is nothing in the
model to show that the available labor force is being employed.
The owners of property have as much as they are willing to
own at the ruling rate of interest and the firms are operating as
much plant as will yield the corresponding rate of profit when
the wage bill and the income from interest is being spent on the
consumption goods that are being produced. They are quite con-
tent. What about the number of workers who need jobs? (This
point was picked out by Harrod in terms of a growing economy.
The rate of investment that absorbs saving makes the employers
quite happy, but the "warranted" rate of growth of the stock of
capital which this produces is not in any way regulated to fit
the "natural" rate of growth of the effective labor force.)
It is here that confusion between the two models very often
occurs. The argument is switched back to tlle first model where
the wage bargain can be made in real terms. When there is re-
dundant labor, the real wage in terms of each commodity falls.
It becomes profitable to employ more labor per unit of "capital"

7 Principles, pp. 516-517.

 Stationary States / 13

up to the point where the marginal product of labor is brought

down to equality with the lower real wage.
This argument falls between two stools. The "quantity of
capital" is neither a list of stocks of fully specified means of
production, as in the first model, nor a sum of value embodied
in forms appropriate to the ruling rate of profit as in the second.
No comprehensible explanation has ever been given of what it
is supposed to be.
The highly unsatisfactory nature of these two models and the
still more unsatisfactory mixture between them are generally con-
cealed by elaborating analysis of their micro-properties-particu-
lar prices and so forth-which leaves their macro-outlines in a
haze.

MARSHALL
The notion of the supply price of capital being the "reward
of waiting" was invented by Marshall, but he never really recon-
ciled himself to the confines of a stationary state. In his vision
of contemporary capitalism, as opposed to his formal analysis,
"progress" is taking place. He can best be understood if we set
his argument in a kind of near-enough golden age with steady
overall accumulation going on and a more or less constant over-
all rate of profit. Profits in particular industries go up and down
around a central "normal" level, and the total stock of capital is
continuously growing. This model, which we may label Marshall-
ian, though it is only one element in Marshall's complex of
doctrines, has something in common with the classics, since it
depicts growth; but it is radically different in its theory of profits.
For the classics, the real-wage rate is given in terms of the com-
modities that the workers consume; the rate of profit then
emerges as a residual. For Marshall, the rate of profit is given
and the real-wage rate in terms of all commodities emerges as
a residual.

B
14 / Economic Heresies

But then another flaw in the argument appears. In all the talk
in the Principles (as opposed to the formal analysis) it is not the
saving of rentiers but the energy of entrepreneurs which governs
accumulation. The individual businessman, with firmness and
·elasticity of character, is striving to expand his own business and
in doing so adds to the national stock of productive capacity.
"The building of an additional floor on the factory or putting an
extra plough on one farm, does not generally take a floor from
another factory or a plough from another farm; the nation adds
a factory floor or a plough to its business as the individual does
to his." 8 In the famous passage 9 which anticipates Keynes, a
slump occurs when confidence fails-investment declines, un-
employment reduces the demand for consumer goods and so
multiplies itself. Clearly it is the confidence of the entrepreneurs
in future profits that has failed, not the desire of rentiers to add
to their wealth. But if the rate of profit dominates the rate of
interest and the entrepreneur dominates the rentier, there is noth-
ing in the story to say what determines the "normal" rate oi
profit. Still less is there anything to provide the moral justifica-
tion for rentier income that Marshall sought to derive from the
need to reward the "sacrifice" to capitalists of owning capital.

THE WICKSELL PROCESS

There is another kind of mixture of the two models which is
associated particularly with the name of Wicksell. In his story
the economy is stationary in the sense that there is no technical
progress, but saving is going on. The given state of knowledge is
embodied in a hierarchy of techniques of production which can
be arranged in order of levels of output per head and of "capital"
per man employed. There is full employment of a constant labor

8 Principles, pp. 535-536. This is in contrast to the supply of "land,"

which is fixed.
9 Ibid., p. 711.
 Stationary States / 15

force and "capital" accumulates by installing successive tech-

niques, moving up the hierarchy. The marginal product of "capi-
tal" is falling as time goes by and consequently the rate of in-
terest falls. Equilibrium in the sense of the first model must
mean that, at each moment of time, the stock of capital goods
is adjusted, not to a single rate of interest but to the spectrum
of interest rates appropriate to various lengths of future time,
while equilibrium in the sense of the second model means that
the rate of saving is appropriate to the expected return on rentier
wealth. We have to imagine correct foresight of a complicated
future development combined with the blind "groping" of Wal-
rasian markets.
It is very difficult to find assumptions that will make this story
self-consistent (Wicksell himself gave it up in despair) and it
hardly seems worthwhile to do so, for the notion of accumula-
tion and technical change without any evolution of technical
knowledge is unnatural. In the progressive capitalist economies,
adaptation takes place along with investment. There is no hier-
archy of techniques already fully blueprinted-the blueprints are
drawn only for the technique that will be used. Moreover, con-
tinuous accumulation is unlikely to be associated with a falling
rate of profit. The problem of choice of technique is important
for developing countries but for them the main point is to reach
full employment in the first place. It is important also for fully
planned socialist economies. For them, there is a "cost of wait-
ing" in the sense that a project which will yield output at a later
date is pro tanto less eligible than one yielding sooner. This con-
cept can be expressed in a notional rate of interest to be taken
into account in planning investment but a rate of profit on the
existing stock of "capital" has no meaning for them.
There does not seem to be any place anywhere where the
"Wicksell process" of accumulation under equilibrium condi-
tions with a falling rate of profit has application. It was an at-
tempt to integrate two incompatible models which are much
better kept separate.
 2
THE SHORT
PER I 0 D Marshall discussed the influence of
demand upon supply in terms of a
succession of three periods or phases. When the supply of a
commodity "is limited to the stores that happen to be at hand,"
demand alone determines price; next, productive capacity being
given, demand may influence the rate of output over a certaID.
range; finally, "in the long run" productive capacity is adjusted
to demand and prices are governed by cost of production, in-
cluding profit at the normal rate on the investment involved. 1
The distinction between the first phase and the second is not
very useful. As Marshall himself pointed out: "Nearly all deal-
ings in commodities that are not very perishable, are affected
by calculations of the future." 2 For manufactured goods of
which retailers hold stocks, the concept of "market clearing
prices" makes no sense. The second phase, however, introduces
an invaluable concept, which sharply distinguishes the Mar-
shallian school of thought from the tradition of Walras-that is,
the "short period" during which the stock of plant is unchanged
while its utilization can be varied.

1 Principle.!, p. 337.
2 Ibid.

16
 The Short Period I 17

This corresponds to the relations of production in capitalist

industry. At any moment capacity is limited by buildings, equip-
ment, and know-how already in existence. An industrial firm has
committed finance to more or less long-lived installations on
which it expects to recover a net profit over some years of oper-
ations. It is committed also to employing staff through contracts
which cannot easily be terminated. On the other hand, ordinary
labor can be employed week by week or even day by day and
running expenses for power, raw materials, and so on vary with
weekly output.
When closely examined, the distinction between a stock of
plant and its degree of utilization, between variable and fixed
costs or sunk costs and escapable costs, cannot be made quite
precise. Some costs are sunk forever, some are necessary per
week or per shift irrespective of the amount of output being pro-
duced; in a time of general scarcity of labor, a manager may be
just as reluctant, when there is a fall in sales in his particular
market, to stand off skilled workers (who may never come
back) as to reduce staff. Similarly, the bottleneck that checks
increasing output may be the availability of labor-especially
of skilled manpower-rather than the existence of plant. But the
general notion of a distinction between changes in utilization and
changes in productive capacity is indispensable for the analysis
of industrial activity.
The essential idea is that a short-period situation is one in
which productive capacity happens to be whatever it is. But a
situation with specific plant in existence today is not to be iden-
tified with the Walrasian concept of a given stock of factors of
production; its role in analysis is quite different. Unlike the
Walrasian concept, Marshall's short period is a moment in a
stream of time in which expectations about the future are influ-
encing present conduct, and it belongs to a monetary economy
in which the division of proceeds between wages and profits
emerges from the relation of money prices to money-wage rates.
18 / Economic Heresies

With the aid of this concept, we can analyze price policy in im-
perfect competition, the effects in the present of uncertainty
about the future, and the meaning of equilibrium in a process of
growth, all of which are ruled out by the assumptions of a Wal-
rasian market.
We can make use of the distinction between the long- and
short-period concepts without being committed to any faith in
equilibrium being established in the long run. Indeed, it is ab-
surd to talk of "being in the long period," or "reaching the long
period," as though it were a date in history. (Marshall himself
thought of the economy as tending toward long-run equilibrium
but never actually being there.) It is better to use the expressions
"short period" and "long period" as adjectives, not as substan-
tives. The "short period" is not a length of time but a state of
affairs. Every event that occurs, occurs in a short-period situa-
tion; it has short-period and long-period consequences. The
short-period consequences consist of reactions on output, em-
ployment, and, perhaps, prices; the long-period consequences
concern changes in productive capacity.

SUPPLY PRICE
A short-period situation mayor may not be in equilibrium
from a long-period point of view. In a situation which is in
equilibrium, no one is kicking himself. Expectations are being
fulfilled. Plant, operated at a normal level of utilization, is pro-
ducing a flow of output which is being sold at prices that prom-
ise to yield a satisfactory rate of profit on the investments con-
cerned. Labor of appropriate skill and training is available to be
employed. When a sudden unforeseen change has recently oc-
curred, long-period equilibrium does not obtain; the stock of
plant and the composition of the labor force are found to be
inappropriate. They cannot be altered overnight but their utiliza-
tion can be changed to make the best of the situation meanwhile.
 The Short Period I 19

An out-of-equilibrium situation may be a seller's or a buyer's

market. In a seller's market, the level of demand is such that it
would be possible to sell more than the capacity rate of output
at prices that cover average total costs (including all overheads
and an allowance for amortization) and yield a net profit. In a
buyer's market, it is impossible to sell capacity output at a re-
munerative price. The distinction is not precise because capacity
output is not a clear-cut conception. There may be an interme-
diate range of rates of output that cannot be classified unambigu-
ously, but a rough-and-ready distinction is sufficient for the main
argument.
The reaction of output and price to unforeseen changes in
demand depends upon the competitive situation among the
producers concerned. In some types of trade (mainly for agri-
cultural produce) commodities are thrown on the market and
sold for what they will fetch; but for manufactures it is the
other way round-the producer declares a price and sells what
the market will take.
Marshall assumed that a higher rate of utilization of plant
would be accompanied by higher prices. In the Pigovian sys-
tem this was systematized in the notion that, in conditions of
perfect competition, the level of output is always such that mar-
ginal cost is not less than price, provided that the price covers
average prime cost. If so, in a seller's market prices would be
pushed up to the point where demand is cut back to equality with
capacity output; in a buyer's market, high-cost capacity would
be shut down and those plants kept in operation for which aver-
age prime cost was not greater than price. Then any plant that
is working at all is working up to capacity.
The experience of an all-round buyer's market in the 1930s
shocked us into realizing (what Marshall always knew) a that

a See Principles, p. 4S8.

20 / Economic Heresies

prices may be held above prime costs and plants worked at less
than full capacity; and the experience of seller's markets in re-
cent times has shown that long delivery dates and rationing of
customers accompany prices held below the level that chokes
off excess demand. In short, imperfect competition is the general
rule in manufacturing industry.
In modern industrial capitalism, market structures and the
policies of sellers are very various. Where a single monopolistic
firm has a strong hold upon a market, or where two or three
powerful oligopolists are maneuvering in it, there is a great deal
of scope for individual policy. In the general run of more or less
competitive industries, the most common behavior seems to be
as follows. Firms make their plans and calculate their average
costs of production on the basis of a normal or standard rate of
utilization of plant. Moderate variations of output above and
below the normal level leave prices unchanged but a strong
swing in demand, or a change in costs due, say, to a change in
wage rates or in the price of a raw material, calls for a reconsider-
ation of prices. When a number of firms are supplying the same
market, no one wants to be the last to cut prices or the first to
raise them, for fear of losing customers to competitors. From
this the institution of price leadership arises. A convention is
established that all await a change made by one firm and all
follow it immediately. The leader pursues a policy that suits its
own convenience, but it is in the position of a reigning monarch
among baronies. Its independence is limited by the need to
avoid offending the interests of the other members of the group.
Moreover, disputes over the succession break out from time to
time.
In a normal situation, the prices set by the leader enable it to
make a comfortable rate of profit, while other firms, smaller in
size, less efficient, or struggling newcomers, have higher costs
and lower margins. In a buyer's market, the institution of price
leadership prevents the competitors from cutting each other's
 The Short Period / 21

throats; when costs rise, it enables them all to defend themselves

from losses. In a strong seller's market, price leadership may
have rather a tendency to keep prices down below the "per-
fectly competitive" level with a view to nursing the market
through a period of shortage.
The best simple generalization seems to be that (so long as
wage rates and the prices of the elements entering into prime
costs are constant) moderate swings of demand have no effect
at all on prices (the short-period supply curve is perfectly
elastic). But this does not mean that they have no effect upon
profits. With constant prices, the excess of receipts over costs
is greater the higher the rate of output, for overhead costs are
independent of utilization and even prime costs per unit may
fall as output increases up to the limit where capacity is being
strained. The ex post realized profit on an investment is higher
the higher the average level of utilization of plant over its life-
time.
Moreover, there is a justification for Marshall's view that an
increase in demand for the products of a particular industry will
lead to an increase in prices, provided that it is believed to be
strong enough to justify increasing investment in productive
capacity; but this occurs not so much because marginal costs are
pushed up as because the firms concerned consider that they
need more profits to finance the investment, and that they are
justified by proper business principles in exacting them, while
they find them easy to earn in the conditions of a seller's market
created by an expansion of demand ahead of the growth of ca-
pacity.
On the other hand, in a buyer's market, maintaining or even
raising prices (as may happen under monopoly or strong price
leadership) is unlikely to prevent profits from falling.

B*
22 / Economic Heresies

EXPECT AnONS
The third stage in Marshall's story of the adaptation of supply
to demand is much less satisfactory. First, as we have seen, he
does not give a comprehensible account of the level of the
normal rate of profit which enters into the determination of
prices "in the long run." Second, he seems to imply that, when
new competition is attracted into a market by exceptional
profits, it will increase capacity gradually until profits are re-
duced to the normal level. He fails to point out that, in such a
case, there is likely to be an overshoot which causes profits to
fall sharply, instead of sliding gently down to the "normal" level.
Third, giving an optimistic account of the operation of the
economy, he concentrates on the effect of a rise in demand lead-
ing to an increase in capacity, not of a fall causing it to shrink.
Once investment has been made and businesses established, the
process of reducing productive capacity is slow and painful. As
Dennis Robertson used to say, the short period is not the same
length at both ends.
The link between a short-period situation with given plant
and the changes in productive capacity which will follow from
it is constituted by the state of expectations generated within it.
When a process of steady growth is going on and expectations
are being realized, the changes taking place at each moment are
harmonious; they will lead to changes in the amount and the
composition of productive capacity that fit with the development
of demand.
Complete equilibrium is never found in reality, but it is ap-
proximated when plans are based upon long-term calculations.
Fluctuations around the normal level of utilization of plant then
have limited consequences. A boom is recognized as a boom.
High profits are enjoyed without investment being speeded up;
a fall of sales is weathered through as a temporary misfortune.
Instability arises from the influence of current experience upon
 The Short Period / 23

expectations. When a seller's market is expected to last, it leads

to rapid investment which may cause an overshoot and kill the
seller's market. But in a buyer's market, productive capacity
is kept in being hoping for a recovery, so that if recovery does
not occur, the buyer's market persists.

EFFECTIVE DEMAND
Marshall was discussing the demand and supply of particular
commodities. The analysis is even more important when applied
to the movements of effective demand as a whole.
In a harmonious situation, expectations are capable of being
fulfilled. Productive capacity is growing at the same rate as
demand in the markets that it supplies; the level of profits ex-
pected for the immediate future is inducing such a level of in-
vestment as will generate such a level of profits as will justify
these expectations.
In a boom, expectations are self-contradictory. Profits are
high because investment is going on, and investment is induced
by expectations of profit which are due to that investment.
Sooner or later, growth in the stock of productive capacity com-
peting in the market will overtake growth in demand; the pros-
pects of profit on a further increase in capacity are dimmed; a
fall in the rate of investment then reduces actual profits.
A depression is a situation of self-fulfilling pessimism. Ex-
pectations of profit are low, therefore investment is discouraged,
therefore sales are below normal capacity operation, therefore
profits are low; therefore gloomy expectations are proving cor-
rect. (In trade-cycle theory, a revival grows out of a depression
when the stock of equipment is reduced relatively to demand,
just as the crises of a boom grow out of an overshoot, but it
may be doubted whether an upturn ever occurs of itself, with-
out some fresh external stimulus to effective demand.)
Keynes' General Theory arose from the attempt to diagnose
24 / Economic Heresies

the situation of a general and prolonged buyer's market. The

hard core of the analysis is concerned with a short-period posi-
tion with given productive capacity and given expectations of
future profits. This accounts for the paradox that what is strictly
speaking a static theory opened the way to a great outburst of
dynamic analysis. Keynes was concerned, above all, to show
that there is no "natural" tendency toward equilibrium with full
employment; therefore government policy is necessary to make
the private-enterprise system work in a tolerable manner. He
was, of course, mainly preoccupied with the question of reme-
dies for unemployment; he merely glanced at the problems of
inflation in a seller's market 4 and his long-period analysis was
very sketchy. It was left to Harrod to transpose The General
Theory into long-period terms, showing that an uncontrolled
capitalist economy cannot be expected either to maintain sta-
bility or to produce growth at a satisfactory rate.
Since the Keynesian revolution became orthodox, the govern-
ments of all the advanced industrial nations have been very
much concerned to preserve near-full employment for workers
and highly profitable markets for capitalists. New influences are
playing upon the movements of effective demand, which are out-
side the purview of Marshall, let alone the general equilibrium
of Walras.

4 He analyzed inflation later, in How to Pay for the War (London:

Macmillan, 1940).
 3
INTEREST AND
PRO FIT In an economy where manufacture
is carried on by artisans, the earn-
ings of labor, capital, and enterprise cannot be distinguished as
separate sources of income. Skill, knowledge, work, business
sense, and ownership of the appropriate means of production,
bound up together, are supplying particular commodities to
particular markets. In a competitive economy (where the regula-
tion of just prices has broken down) the income to be made
from a particular commodity is strongly affected by supply and
demand. Social income is, so to say, divided vertically into re-
ceipts from separate commodities. When employment for wages
becomes the main form of production, the division is horizontal,
between income from work and income from property. Profit as
a distinct category of income is a characteristic of industrial
capitalism.
Marx, following the hard-headed classical economists, at-
tributed profit to the power of capital to exploit labor. The neo-
classicals rejected this point of view but they never succeeded
in producing an alternative theory of profits that was both co-
herent and plausible.

25
26 / Economic Heresies

WALRAS
When Walras introduced a rate of interest into his timeless,
non-monetary market, he gave two completely incompatible ac-
counts of it, which, no doubt, he hoped would come to the same
thing. 1
In one story, there is a certain commodity which yields a re-
turn in the form of a perpetual annuity at some percentage rate
upon its value. The hire-prices of all the physical factors of pro-
duction are still determined by supply and demand in the market,
but now the prospective earnings of each piece of property are
capitalized at the general rate of interest so as to give its present
value. Walras himself is rather vague about the capital value of
a worker regarded as a factor of production. One of his latter-
day disciples has the courage to carry his ideas to their logical
conclusion, that is, to capitalize the future earnings of the labor
force, so that prospective net national income is represented as a
return equal to the rate of interest on the capital value of the
total stock of factors of production. 2 But even he cannot suggest
any way of accounting for what the level of this rate of interest
is.
The second story in Walras is connected with saving. Every
seller in the market immediately spends his receipts upon some-
thing but he is at liberty to buy means of production, say ma-
chines, which are valued for their future earning power. At any
moment there are given conditions of production for each type
of machine and the price of a machine, regarded as a product,
is determined by supply and demand along with all the other
commodities. The ratio of the current hire-price of a machine

1 See Elements of Pure Economics, trans. W. Jaffe (London: Allen &

Unwin, 1954), Lesson 23.
2 See J. R. Hicks, Capital and Growth (Oxford: Clarendon Press,
1965), p. 264.
 Interest and Profit / 27

to its cost represents its current rate of return. The rate of

profit is then established by the machine for which this ratio is
highest. The prospective hire-price of other machines is capi-
talized at this rate. The value of all but the most profitable ma-
chine is then found to be less than its cost of production. Only
the most profitable type of machine is worth having. The savers
are buying machines of this type.
But Walras failed to point out that if savers are guided by
current values they will be misguided. To make correct invest-
ments they need to know the future course of relative prices of
all commodities and types of machine. This model also has been
worked out in neo-neoclassical terms; 3 but it seems impossible
to reconcile the contradiction between the assumption of correct
foresight for each individual over an indefinite future and the
daily higgling of a Walrasian market.

MARSHALL
In the Pigovian stationary state, which formalizes the static
element in Marshall's system, the rate of interest is the return on
the rentiers' wealth (the reward of waiting) which is just suffi-
cient to induce them to keep it in being. Finance is lent and
borrowed, in indefinite amounts, at this rate. Consequently the
prices of commodities and the allocation of resources between
different uses are such that every investment of capital earns a
rate of profit equal to this rate of interest. Such a theory is quite
hollow; it merely repeats the assumption that in a stationary
state, the rate of profit is equal to the reward of waiting.
In Marshall's account of a growing economy there is a great
deal of verbal confusion between various meanings of the rate
of interest. In his terminology, the long-term rate of interest is

3 Cf. M. Morishima, Equilibrium, Stability, and Growth (Oxford:

Clarendon Press, 1964), Section III.
28 / Economic Heresies

identified with the rate of profit on capital and this, in turn, is

sometimes, but not always, identified with rentier income (the
reward of waiting). On the other hand, the short-term rate of
interest or rate of discount is a phenomenon of the money mar-
ket; it can be influenced by the behavior of the banks or by
movements of the supply of gold. 4 (Wicksell similarly distin-
guishes between the "natural rate of interest," which means the
rate of profit, and the "money rate of interest," which is the cost
of borrowing.)
The terminology can be revised as follows. Profit is the net
return to a firm on its invested capital. Interest (a complex of
rates for various types of loan) is the hire-price of finance; the
yield of placements is the rate of return that a rentier receives
on the capital value of his assets./; The last two are connected,
for the rate of interest in the money market influences the sec-
ondhand value of placements, but the range of transactions that
they cover is not identical. An important element in the complex
of interest rates is the charge for bank loans (in Marshall's day,
the rate of discount on bills); from the point of view of a bank,
interest on loans is one part of its gross receipts, not a return on
capital, while rentier wealth may include elements such as real
estate not corresponding to the liabilities of business firms. In all
this, the most important point is to isolate Marshall's conception
of the rate of profit on capital.
At any moment, investment is going on; firms already in exist-
ence are planning to enlarge their productive capacity and new
businesses are being started up. Investors, looking into the fu-
ture, reckon what prices they can expect for additional output
and what wages and other costs they will have to pay, and they

4 Cf. E. Eshag, From Marshall to Keynes (Oxford: Blackwell, 1963),

Chapter III.
5 Of course, in reality the "reward" of owning wealth is owning wealth,
whether or not it yields income; the "reward" of saving is an addition
to wealth.
 Interest and Profit / 29

know what additional equipment a given sum of money can buy

at current prices. They thus calculate the rate of profit to be
expected on investment. (This may be expressed either as the
rate of discount which reduces the value of the expected gross
profits spread over future time to equality with the present cost
of investment, or as the permanent annuity that the investment
could secure by amortization and reinvestment, maintaining the
capital intact over an indefinite future.) Each investor goes in
for the scheme that promises the highest return. In normal times,
for the representative investor, expected prices and costs (here
is the missing link in the argument) will work out so as to give
an expected yield on the investment equal to the "normal rate
of profit." Marshall insisted that the rate of profit (which he
called the long-term rate of interest) can be seen only at the
frontier of investment, looking forward, but if "normal" condi-
tions normally obtained, the actual realized rate of profit would
generally turn out to be equal to the expected rate. There is a
tendency for the rate of profit to be evened out throughout the
economy: or rather there is a pattern of profit rates-lines
which are easy to enter on a small scale may have a lower rate
of profit than that enjoyed by the great firms; or within one in-
dustry, at a moment of time there may be struggling or decaying
firms doing badly compared to the "representative firm" which
at that moment is in its prime. (As an observation of contempo-
rary family businesses, Marshall's story of "trees in the forest"
was apt, although its role in his theory was not convincing.)
The diffusion of profits throughout the economy is maintained
by the short-period mechanism. Where demand is expanding
ahead of supply in some line, prospective profits are seen to be
high. Not only a large proportion of new investment will flow in
that direction, but also amortization funds from less successful
lines. Thus the push and pull of demand are continually mold-
ing the stock of capital into the form which yields the normal
return.
The value of all capital goods in existence today is found by
30 / Economic Heresies

capitalizing their current net earnings at a rate corresponding to

the normal rate of profit. The rate of interest which is paid on
borrowed finance normally accommodates itself to the ex-
pected rate of profit, with an allowance for risk; but it may be
influenced by monetary factors which displace it from its proper
level. When the rate of interest is too low, speculation sets in,
rash investments are made, prices are driven up. Too Iowa rate
of interest thus causes a temporary and unhealthy rise in pros-
pective profits. It was left to Keynes to point out that too high
a rate of interest causes depression and low profits.
In this part of the argument, Marshall has tacitly abandoned
the idea that the rate of profit is equal to the reward of waiting,
for, if it were, investment would not be going on. And the influ-
ence of the monetary rate of interest on the rate of profit is only
an unfortunate aberration. So what does determine the normal
rate of profit? Marshall evidently hoped that his readers would
not notice that he does not say.

KEYNES
Keynes cleared up the verbal confusions of the neoclassics
by drawing a sharp distinction between the rate of profit and the
rate of interest, that is, between the return on real investment
accruing to entrepreneurs and the cost of borrowing which influ-
ences the return on secondhand placements received by rentiers.
But he did not attempt to supply a theory of the rate of profit
in the long run. His argument was concerned purely with a short-
period situation. The expected rate of profit, which he called the
marginal efficiency of capital, is an estimate of future returns to
be obtained on investments in productive capacity; it is neces-
sarily uncertain and it is influenced by subjective psychology-
the state of the animal spirits of the investors.
The actual rate of profit being earned on the capital already
in existence has no meaning in the short-period situation that
 Interest and Profit / 31

Keynes was discussing. The overall total of gross profits per

annum is whatever it is; it cannot be reduced to net profit with-
out knowing the future in order to calculate what depreciation
should properly be deducted from it; furthermore, the rate of
net profit involves a calculation of the value of capital; the his-
toric cost or the current reproduction cost of stocks and equip-
ment are irrelevant; they reflect past conditions, not the future.
The value of the stock of capital, in this situation, can mean only
its expected future earnings discounted at some appropriate rate.
If we knew what the rate of profit was, we could use it as the
rate of discount, calculate the value of capital, and show that it
is yielding the rate of profit. But this in no way helps to find out
what the rate of profit is.
The complex of yields which represents return on capital from
the rentier point of view is determined by the interplay of the
preferences of owners of wealth and the stocks of money and var-
ious other kinds of placements (bonds, shares, etc.) in existence.
The level of prices is established from day to day in the market.
The monetary authorities, through the banking system, can in-
fluence the level of interest rates by operating on the supply of
money.6
It is clear enough that a fall in the rate of interest (in a given
state of expectations) raises the capital value of all income-
bearing placements, of real estate and of house property which
yield rents in cash or in kind; it cannot have any direct effect
upon the value today of equipment being used in industrial pro-
duction. It may have an important influence in stimulating
house building and lowering future rents; its effect on industrial
investment is not so clear (except that small businesses may find
it easier to get bank loans). On this point Keynes rather lost his
grip on the distinction between the rentier and the entrepreneur.

.fl Cf. below, p. 79.

32 / Economic Heresies

His discussion of "the state of long-term expectations" is devoted

to the Stock Exchange rather than to the accumulation of means
of production.
Where he allowed his mind to play upon long-term problems,
his conceptions are still more obscure. In particular, the sug-
gestion that the euthanasia of the rentier could be brought about
merely by establishing a permanently low rate of interest now-
adays seems fantastical.

THE NEO-NEOCLASSICS
The neo-neoclassics, 7 who tried to reconstruct traditional or-
thodoxy after the Keynesian Revolution, slipped back into the
habit of identifying the rate of profit with the rate of interest
and reasserted the doctrine that the rate of return measures the
marginal productivity of capital from the point of view of society
as a whole, without attempting to explain what it meaml.
The neoclassical scheme of ideas was intended to present an
industrial economy as a scene of rationality and social harmony
under the guidance of the "hidden hand" of competitive market
forces. Marshall had some reservations; the clearest statement
came from J. B. Clark. "What a social class gets is, under natural
law, what it contributes to the general output of industry." 8 On
this view, the profit received by the capitalist is due to the con-
tribution to output of his capital. Capital equipment contributes
to output (along with education and training) by raising the
productivity of labor; a command of finance permits a capi-
talist to provide equipment, employ labor, and receive profits.
The neo-neoclassical revival of pre-Keynesian theory took over
J. B. Clark's identification of capital as profit-earning finance

7 See, in particular, R. M. Solow, Capital Theory and the Rate of

Return (Amsterdam: North Holland Publishing Company, 1963).
8 "Distribution as Determined by the Law of Rent," Quarterly Journal
of Economics (April 1891).
 Interest and Profit / 33

with capital as a stock of means of production. Leaving land on

one side, "capital" and labor are the "factors of production"
and their "rewards" correspond to their "marginal productiv-
ities." 9 The basis of this doctrine seems to be a confusion be-
tween the idea of the productivity of investment and the pro-
ductivity of "capital."

The Productivity of Investment The productivity of invest-

ment to society is not a very precise idea but it has an important
meaning. We can imagine an independent peasant family, or a
cooperative society like a kibbutz, deciding how much of their
labor to devote to improving the land or how much of the pro-
ceeds of sales to a surrounding market economy to devote to
buying productive equipment. The cost of an investment is more
work or less consumption in the present and the benefit is an in-
crease in the productivity of work in the future. Neither the cost
nor the proceeds are homogeneous and both contain psycho-
logical elements; the relation between them can be represented
as a rate of return only by adopting some more or less arbitrary
convention of measurement. However, the general idea of a
present sacrifice yielding future advantages is clear enough.10
What has it got to do with the rate of profit on capital? In such
a community, the current output of consumption goods, and the
future benefit of higher consumption or more leisure, will be
distributed among its members on some principle or other; the
means of production belong to the community as a whole and
the distinction between income from work and income from
property has no meaning for them.
Under laisser-faire capitalism, the division of net output be-
tween consumption and investment is decided for society by
profit-seeking entrepreneurs. There is no mechanism in the sys-

9 This proposition is categorically reaffirmed by C. E. Ferguson in

The Neoclassical Theory of Production and Distribution, p. 215.
10 Cf. Solow, op. cit., p. 154.
34 / Economic Heresies

tem even to ensure that all available labor is employed for one
or the other. When there is unemployment, the cost to society of
some additional investment is not much more than zero, indeed
it is negative if we bring the misery of unemployed workers into
the account, but capitalists would have to pay wages to get it
done.
In a progressive near-full employment economy, maintaining
a growing national income with a more or less constant rate of
profit, the effective labor force is evidently increasing through
growth of numbers and rising output per head (otherwise growth
with a constant rate of profit would not be possible). The prod-
uct from the point of view of society of the investment which is
going on includes the growth of the real-wage bill as well as the
additional profit. If this is the "marginal product of capital" it
much exceeds the rate of profit. In order to know how the bene-
fit to society will be divided between wages and profits in the
future we need to know the rate of profit; there is nothing here
to tell us how it comes to be what it is.
In one sense, modern capitalism has something in common
with a cooperative where the benefit of investment is set against
its cost. A government may consider that near-full employment
has been achieved with too large an amount of consumption and
too little investment for the future good of the economy. The
government then wants to make the market for consumption
goods less profitable and investment more attractive. It finds it
by no means easy to do so, for to reduce the profitability of the
market discourages investment, but one way or another it may
succeed. It is being guided by some general view of national in-
terest, which might perhaps be expressed in terms of the pro-
ductivity of investment but has nothing to do with the rate of
profit on capital.

The Pseudo-production Function Another deep-seated con-

fusion arises from failing to distinguish between comparisons of
 Interest and Profit / 35

stocks of capital in imagined equilibrium positions and ac-

cumulation going on through time, such as the Wicksell process
·of "capital deepening." 11 (Both are very unreal concepts but
it is necessary to set them up in order to see what they are in-
tended to mean.)
In a Pigovian stationary state, with a given rate of interest,
we are to suppose that there are a number of different possible
methods of producing a given rate of output; competition be-
tween profit-maximizing firms has led to capital being embodied
in forms which yield a rate of profit equal to the rate of interest.
On this basis we can construct a pseudo-production function
showing all the possible points of equilibrium in an imagined
"given state of technical knowledge" which is intended to illus-
trate the supposed effect of relative "factor prices" (the real-
wage rate and the rate of profit) on the choice of technique. The
techniques are set out in order of net output per man employed.
At each rate of profit, the eligible technique is the one which
permits the highest real-wage rate to be paid when that rate of
profit obtains. Any method of production which is not eligible
at some rate of profit is inferior and is not included in the sched-
ule of techniques. For each pair of techniques there is a rate
of profit at which they are equally eligible. One technique re-
quires a higher value of capital per man than the other (at the
prices corresponding to that rate of profit) and produces a net
value of output per man just sufficiently higher to pay the addi-
tional profit required; with a small difference in the rate of
profit, one or the other ceases to be eligible.
It was in this context that the "res witching" controversy arose.
At labor-value prices, the cost of equipment required for each
technique is proportional to the labor-time required to produce

11 This confusion is very clearly seen in Professor Samuelson's "Sum-

ming up" of the "reswitching" controversy. "Paradoxes of Capital
Theory," Quarterly Journal of Economics (November 1966).
36 / Economic Heresies

it. A higher real-wage rate then entails a higher cost of capital

per man employed. A technique with a higher cost of capital
cannot be eligible (at any rate of profit) unless it has a higher
output per man. Thus, in such a case, the order of techniques
in terms of capital per man is the same as the order in terms of
output per man. The pseudo-production function then looks like
the "well-behaved production function" of the neo-neoclassics,
on which an addition to "capital" per man produces an addition
to output per man. The switches of techniques on the pseudo-
production function are then always forward, a lower rate of
interest causing a technique with a higher output per man to
become eligible. 12
But, as Ricardo realized, labor-value prices are a very special
case. They rule only when the capital to labor ratio and the time-
pattern in which costs are incurred are the same in all lines of
production. In the general case, the rate of profit as well as the
real-wage rate enters into relative costs; since the two move in
opposite directions (a lower rate of profit entails a higher real-
wage rate) the cost of the equipment required for anyone tech-
nique (in terms of a unit of net output) may rise or fall with
the rate of profit. Over a range where the cost of the equipment
for the technique with the lower output per man rises with the
rate of profit by sufficiently more than for the adjacent technique
with a higher output, there will be a backward switch, so that
a higher wage rate is associated with a more "labor intensive"
technique, that is, with a lower output per man. For the neo-
neoclassicals this was a paradox. It upsets the notion of a pro-
duction function exhibiting substitution between labor and "cap-
ital."

12 This was the case of Professor Samuelson's famous "Surrogate Pro-

ductionFunction." It was actually a special form of the pseudo-production
function. See "Parable and Realism in Capital Theory," Review 0/ Eco-
nomic Studies (June 1962).
 Interest and Profit / 37

This analysis provides a very striking illustration of the fact

that the old neoclassicals had failed to give a definition of a
"quantity of capital" (except for the case where it can be mea-
sured as "labor embodied," which was not to their fancy) and
it shows that the concept of the "marginal product of capital" is
unseizable; certainly, at a switch point, comparing one technique
with another, profit per man is proportional to the value of
capital per man, so that, in a certain sense, the return on invest-
ment is equal to the rate of profit. But this is true only because
all prices of inputs and outputs are such that the rate of profit is
the rate corresponding to that switch point. 13
But the whole argument is only negative. Such a thing as a
pseudo-production function does not exist in nature. There is no
sense in arguing about whether it is "likely" to be well-behaved
or not, and however perfectly well-behaved it might be, it could
not tell us anything about how the rate of profit comes to be
what it is.
The pseudo-production function appears to be important only
when it is confused with an actual production function which
shows how investment made today will affect output in the
future.
Wicksell (though he abandoned the attempt) at one time tried
to make use of a simplified pseUdo-production function (in
which techniques are specified only by the length of the "period
of production") to find the relation between the rate of profit
(which he called the natural rate of interest) and the "marginal
product of capital." According to this line of thought, the stock
of means of production in existence today operating techniques
now known came into existence by embodying savings made in
the past. The Wicksell process of accumulation in a "given state
of technical knowledge" requires a rise in "capital" per man,

13 Cf. L. L. Pasinetti, "Again on Capital Theory and Solow's 'Rate of

Return,''' Economic lournal (June 1970).
38 / Economic Heresies

going on through time, to be associated with a falling rate of

profit. Thus, accumulation is seen as creeping along a production
function which was always known and does not alter as the proc-
ess goes on. Suppose that ever since Adam left paradise a single
state of technical knowledge has obtained and investment has
been slowly increasing the stock of capital. Then as the rate of
profit falls, technology must gradually pass each switch point at
which two techniques are equally profitable and at each point the
return on investment is equal to the rate of profit.
This is evidently absurd. If we were to take the story seriously,
we should have to suppose that the stock of capital at any mo-
ment has been chosen in the light, not of one rate of profit, but
of a complex of rates corresponding to different periods of fu-
ture time. And we would have to suppose that the stock of
capital equipment at any moment was not embodying a single
technique appropriate to a particular set of prices and rate of
profit, but was composed of fossils of past investments made in
the light of expectations of higher rates of return than those now
ruling.
No doubt it is perfectly possible to work all this out on stated
assumptions, but such an analysis does not even pretend to apply
to either the past or the future of the economy that we are living
in.
 Interest and Profit / 39

RICARDO AND VON NEUMANN

While the neoclassical tradition was running into the sand,
there was a revival of interest in the classics. 14 In the classical
theory of the rate of profit, the real wage is treated as part of
the necessary costs of production.
In Ricardo's corn economy, the output of corn produced by
a man-year of work on marginal land is a technical datum. The
corn-wage is also a technical datum, given by the needs of sub-
sistence. Output minus wage is the annual profit per man em-
ployed. The wage rate and the length of time from harvest to
harvest determine the corn-capital required to employ a man.
Profit per man over capital per man, each as a quantity of corn,
is the rate of profit. The rate of profit emerges from the technical
data of the system because the necessary wage is part of the
specification.
The rate of profit being determined in the production of the
wage good, competition sets the corn-prices of all other products
so that they yield the same rate of profit. The corn-value of the
output of a man in any industry, minus the wage, provides a
gross profit per man employed which is sufficient to keep capital
intact and to yield the standard rate of profit on the corn-value
of the capital goods associated with employing him. Ricardo
himself was mainly interested in the prospective fall in profit
per man employed as increasing total employment extended the
margin of cultivation. Moreover, he was distracted by the objec-

14 This was being carried out mainly under the influence of Piero
Sraffa. His article of 1926, which set off the theory of imperfect compe-
tition, his preface to Ricardo's Principles (1951), and finally the Produc-
tion of Commodities by Means of Commodities (1960) constitute a
criticism of the theory of distribution in terms of marginal productivity
which the neo-neoclassics have not been able to answer, though they have
attempted to dodge it by arguments such as those described above.
40 / Economic Heresies

tion that the real wage cannot be treated as a single homo-

geneous product into giving up the corn model and pursuing the
will-o'-the-wisp of an "invariable measure of value." 15
After being lost to sight for a century, the pure classical
theory of profits was worked out by von Neumann.1 6 In his
model, the necessary wage consists of a specified basket of com-
modities. These commodities are produced by labor with the aid
of a stock of commodities--equipment, raw materials, and the
like-all of which are produced within the system by labor and
themselves. The commodities are combined in the proportions
which produce the fastest maintainable growth rate of the output
of baskets of wage goods. As the flow of output of wage goods
increases, employment of labor grows (either the population is
growing at just the right rate or there is an indefinite reserve
of potential labor, living on nuts in the jungles, ready to take
employment when the standard real wage is offered). In any
period, the surplus of production over the wage of the labor
employed and the replacement of the means of production used
up constitute net profit. The physical elements in the net profit
are in the same proportion as the stock of means of production
and the wage fund. Thus the ratio of net profit to the stock of
capital is unambiguous. The technical conditions of production
and the real wage determine the rate of profit.
There is an element in the von Neumann model which might
be taken to suggest a resemblance to Walrasian marginal pro-
ductivity. It is physically possible to produce some or all of the
commodities with various proportions of others. The optimum
proportions are characterized by marginal productivities equal
to prices. But at any point on a von Neumann path, the optimum

15 See Piero Sraffa, Preface to Ricardo's Principles. Works and Corre-

spondence of David Ricardo (Cambridge: 1951), Vol. I.
16 See "A Model of General Economic Equilibrium," Review of Eco-
nomic Studies, XIII (1) no. 30 (1945-1946).
 Interest and Profit / 41

proportions for the whole output have already been chosen; the
stock of means of production already exists in the correct pro-
portions. All· relative prices are appropriate to costs of produc-
tion including profit at the standard rate. This is totally different
from the Wall' asian situation) where stocks of commodities,
means of production, and labor are given, at any moment, in
arbitrary proportions and relative prices and wage rates are set-
tled by supply and demand. Von Neumann's equations describe
the equilibrium conditions of an optimum path; they cannot, in
the nature of the case, say anything about what happens to an
economy when it is outof equilibriumP Nor does von Neumann
say anything about the pseudo-production function. His economy
is bound to the one technique dictated by the real-wage rate.
(If the wage rate were specified in calories instead of quantities
of particular commodities, we might compare economies where
the workers were fed, say, with wheat or with potatoes. In the
latter, the real cost of the wage would be lower and the rate of

17 When a planning authority is provided with a job lot of means of

production and wishes to maximize employment at some future date, it
has a wide choice of possible policies even if it is confined to a single
technology. At one extreme, it might find the bottleneck commodity,
collect a set of inputs in the optimum proportions from the stocks avail-
able, discarding the surplus amounts of those in more than the optimum
ratio to the bottleneck commodity, and set output growing at the maxi-
mum rate. (This is the "turnpike" policy.) At the other extreme, it might
begin by producing only the bottleneck commodity until there was
enough of it to make some other the bottleneck, and so on, until the
stock had been built up to the optimum proportion with the commodity
of which there was the largest supply (relative to requirements) in the
original job lot. Which policy within this possible range would be best
must depend upon the detailed specification of the original stocks of
commodities and the technical equations. There cannot be an a priori
assumption that the turnpike policy will be eligible. This way of looking
at things, of course, is leaving out the main problem that arises in reality.
In von Neumann's world, labor comes into existence only when there
is a real wage to feed it. Actual planners are worried about workers who
already exist.
42 / Economic Heresies

profit higher. The relative prices of all commodities would be

different in the two cases, and different techniques of production
might be eligible.)
Von Neumann assumed that the whole surplus was continu-
ously being invested in enlarging the stock of commodities and
increasing employment. Then total net profit and total net in-
vestment are identically the same thing. The rate of profit is
equal to the rate of growth.
Let us vary his assumptions by supposing that owners of
property consume part of the output of wage goods. The rate of
profit and the level of wages cannot be affected, for they are
fixed by technical conditions. The rate of growth would be re-
duced. Here is an important clue which will be picked up later.

THE RATE OF EXPLOITATION

In Ricardo's corn model, profit is pure exploitation. The
workers have to seek employment because they have no access
to land and no means to live from harvest to harvest. The func-
tion of the capitalist is to engage to pay rent to the landlord and
to advance corn-wages to the workers. He is taking advantage of
their necessity to make them produce a profit for him.
But this exploitation is not to be deplored. It is the only way
that wealth can be increased. The landlords consume their share
of the corn in supporting feudal retainers. The capitalist con-
sumes very little of his share; he invests it in employing more
workers and producing more profit.
Marx enlarged the conception of accumulation through ex-
ploitation. Competition between capitalists drives them to re-
duce costs by increasing output per head so that they "ripen
the productive power of social labor as though in a hot-house."
In Volume I of Capital, Marx seems to predict that, as cap-
italism develops, real wages will fluctuate around the level which
was established when industrial employment first began to take
 Interest and Profit I 43

over from an economy of artisans and peasants. As output per

head increases, with constant wages, the rate of exploitation is
rising. The tension between rising production and constant or
falling consumption for the mass of the population will bring an
explosion. But in Volume III there are hints of a different prog-
nosis, according to which the rate of exploitation will tend to be
constant. If so, the real-wage rate must be rising in step with
output per head. 18 The diagnosis of Volume I seems to fit with
modern experience in the so-called developing countries where
the level of wages at which capitalist investors can recruit labor
is kept low by the supply of would-be workers with no means
to live; the share of profit in proceeds in the enclaves of modern
industry is extremely high.19
On the other hand, in the successful industrial economies,
where near-full employment prevails, where trade unions are
strong, and social legislation aims to eliminate desperate misery,
it seems that a fairly constant rate of exploitation tends to be
established so that a rising overall level of real wages becomes
normal. This (up till now) has saved capitalism from the con-
tradictions that Marx expected would destroy it, both by fending
off the indignation of the workers and by keeping a market ex-
panding for goods and services that can be profitably supplied
to them.
Once we remove the postulate that the real wage is technically
determined, the classical theory of profits loses precision; but it

18 Marx believed that the development of technology must be such

as to raise the ratio of capital to output. Then, if the share of profit were
constant, the rate of profit must be falling. In modern conditions, it
seems, there is a tendency rather to keep the value of output per man
employed and the value of capital per man rising more or less in step,
so that a constant share of profit in net output and a constant rate of
profit in capital are not incompatible.
19 Cf. P. J. Loftus, "Labour's Share in Manufacturing," Lloyds Bank
Review (April 1969).
44 / Economic Heresies

still provides the basis for an account of how the system oper-
ates.

SAVINGS AND INVESTMENT

Let us return to von Neumann's model and alter his assump-
tions in two respects; there may be a certain range of real-wage
rates in terms of larger or smaller baskets of commodities and
the whole net profit need not be invested in expanding the sys-
tem-part may be consumed by rentiers.
Now, if we compare two paths with the same real-wage rate
and the same techniques in use but different proportions of profit
consumed, the rate of profit (as we saw above) is the same in
both; the growth rate in each is equal to the rate of profit multi-
plied by one minus the proportion of profits consumed. When
the whole net profit is consumed, the growth rate is zero. (This
is a kind of stationary state that has some features in common
with the Pigovian model.)
Then (at a point where total employment is the same on
both) compare two paths (having the same technique) with
equal proportions of profits consumed, but one with a higher
growth rate than the other. The former has a higher rate of
profit. 20 Its real-wage rate is lower for two reasons: first, the
proportion of investment to consumption is higher; second, the

20 If we release the assumption of a rigid technique, the higher rate of

profit and lower real wage may be supposed to have led to the selection
of different techniques from the spectrum of possibilities which is in
common for the economies being compared. The technique which is
eligible at a higher rate of profit may have either a higher or a lower
output of a given basket of commodities per unit of labor. (This propo-
sition was established in the double-switching controversy.) When the
pseudo-production function is well-behaved, a higher rate of profit is
associated with a lower output per head, which lays a further burden
upon the workers. (This proposition is derived from the golden rule or
neo-neoclassical theorem. See below, p. 136.)
 Interest and Profit / 45

amount of profit being greater, the amount of consumption by

rentiers is greater. Or, if we compare two paths with the same
growth rate, that which has the lower proportion of profits con-
sumed has the higher real-wage rate.
These propositions are summed up in the formula, when

7r is the rate of profit on capital, g is the rate of expansion of the

economy, s'" and sp are the proportion of saving in wages and

profits respectively. In the classical theory, the real wage is fixed
and profit emerges as a residual, depending upon technical con-
ditions; in this theory, the rate of profit is determined by the
combined effect of saving and investment, and the real wage
emerges as a residual.
All this is merely a set of formulae. The question at issue is
how the allocation of labor and means of production between
investment and consumption is carried out under laisser-faire
capitalism. Is it, as the classics thought was obvious, the indus-
trialist who invests and ploughs back profits to expand his busi-
ness, or is it (as the neoclassics seemed to maintain) the house-
holder who decides how much of his income to consume and
hands over the rest to be invested?
In writing the General Theory it took Keynes a "long struggle
to escape" from the neoclassical view but much earlier he had
described the system flourishing before 1914 in classical terms:

Thus this remarkable system depended for its growth on a double

bluff or deception. On the one hand the labouring classes accepted
from ignorance or powerlessness, or were compelled, persuaded,
or cajoled by custom, convention, authority and the well-estab-
blished order of Society into accepting, a situation in which they
could call their own very little of the cake, that they and Nature
and the capitalists were co-operating to produce. And on the
other hand the capitalist classes were allowed to call the best part

c
46 / Economic Heresies

of the cake theirs and were theoretically free to consume it, on

the tacit underlying condition that they consumed very little of it
in practice. 21

Under capitalism, from the first, the function of profits was

to be saved and saving, in the main, took the form of investing
the profits accruing to a business in its own expansion. Some part
of profits was handed over to the households of the capitalists
or paid as interest to those who provided finance, but if the main
purpose of profit had been to support rentier consumption "the
world would long ago have found such a regime intolerable." 22
Adopting the classical view, we can supply the missing link
in the Marshallian model. Firms are carrying out schemes of
investment with a view to increasing their operations. Earned
and unearned incomes (in the language of the British Inland
Revenue) are being paid out to households, and money is flow-
ing back from households to firms for the purchase of goods and
services.
If we postulate that the budget is balanced and that any net
saving out of earned incomes is offset by private house building,
it follows that the overall sales value of the goods and services
being bought by households from firms, over any period, ex-
ceeds the wage bill directly and indirectly incurred in producing
them by the amount of the wage bill for new investment plus
expenditure out of profits (in which should be included the
greater part of the salaries that the captains of industry allow to
themselves). Here is the source of net profit. The equivalent of
the prime costs of production of goods sold to the public is
recovered from expenditure of their own wage bill; gross profits
are recovered from the wage bill of the investment sector and

21 Keynes, Economic Consequences of the Peace, pp. 16-17 (London:

Macmillan, 1919).
22 Ibid.
 Interest and Profit / 47

expenditure on consumption out of profits. Setting off amortiza-

tion against the cost of replacements, net profit is equal to the
value of net investment plus the value of rentier consumption.
As Kalecki puts it, the workers spend what they get and the
capitalists get what they spend. 23
Technical conditions and the level of profits determine the
level of money prices relatively to money-wage rates and so
determine the level of real wages in terms of any basket of com-
modities.
A surplus or deficit in the foreign balance on income account
is added to or set off against home investment. A deficit in the
budget and an excess of house building over saving from earned
income tell in the same direction as net investment (a surplus in
the budget or in saving from earned income reduce net profit
correspondingly) .
All this is concerned with actual flows of payments. The rate
of profit which governs investment plans is not an actual pay-
ment. It is an expression of expectations of future prices and
costs. Only when an economy is growing smoothly in the condi-
tions of a golden age, with expectations being continuously ful-
filled and therefore renewed, does the realized rate of profit have
a definite meaning. The conditions of our formula are then ful-
filled. Profit is generated in the sale of consumption goods (just
as, in the corn economy, it is generated in the production of
the wage good); the rate of profit obtainable in industry in gen-
eral enters into the prices of investment goods that firms sell to
each other or into the book value of those that they produce for
themselves. The rate of profit (when there is no net saving out
of earned income) is equal to the rate of accumulation divided
by the proportion of profits saved.
Reality is never a golden age. There are disturbances due to

23 Cf. below, note 8, p. 119.

48 / Economic Heresies

markets in which supply and demand rule, mistaken expecta-

tions, and unforeseen events. The rate of profit on capital is
neither uniform throughout an economy nor steady through
time. Nevertheless, the concept of the normal rate of profit de-
termined by investment and the propensities to save provides the
framework of a general theory within which detailed analysis
can be built up.
The normal rate of profit must be sharply distinguished from
the rate of interest. The reward of waiting-the rate of return
on rentier wealth-is determined in the money market. With
the facilities that modern institutions provide, marketable place-
ments are much less risky than productive assets; the level of
their yields is normally much below the prospective rate of profit
that attracts real investment.
The function of legal and financial institutions, including the
Stock Exchange, is to reduce lenders' risk and so facilitate the
supply of finance to industry. Nowadays the major part of in-
dustrial investment is financed from retained profits, and no
duubt this was just as much the case in the era of Marshall's
family businesses as it is under the regime of managerial capital-
ism. At the same time new businesses are always being started;
family concerns are being sold to the public and issues of securi-
ties may be made by firms who prefer to finance expansion that
way. There is therefore a need for outside saving as well as re-
tention of profits. On our assumptions, the overall rate of outside
saving is equal to the excess of investment over retentions, but
new savings are not necessarily directly available to be bor-
rowed. A small trickle of new demand for placements and
money is coming at any moment into the large pool of the
capital market from savers and a trickle of new issues is drain-
ing it off. The level of prices of the whole pool of securities is
constantly changing with the "state of the news." To see the in-
fluence of the supply and demand of finance, we can imagine
that we are examining the market in a state of tranquillity when
 Interest and Profit / 49

a stable rate of profit and rate of growth are confidently expected

to continue to be maintained for an indefinite future. For the
level of interest rates to remain constant then requires that the
pattern of growing demand for placements is matched by the
pattern of supply; in particular, it requires the banking system to
allow the quantity of money to expand to satisfy the liquidity
preference of the owners of the growing total of wealth (as well
as the needs of trade) and, one way and another, to make the
loans to industry which the rentiers do not provide. Thus, when
the demand for securities is growing more slowly than the sup-
ply, the level of interest rates will be tending to drift upward
unless the movement is offset by the banking system, and con-
trariwise.
Self-finance of firms increases the rentiers' wealth as well as
their own savings. Provided the investments financed out of
retentions are successful, they increase the earning capital of the
firm. In a family business, the family may claim a right to enjoy
the benefit; in a public company, the value of shares rises. In-
sofar as this increase in their wealth stimulates rentier consump-
tion, it tends to raise the overall level of the rate of profit on in-
vested capital.
The above provides Marshall with the basis of a theory of the
rate of profit and the rate of interest, but it does not provide
what he was looking for-a justification for rentier income.

EFFECTIVE DEMAND
Keynes reproached the classical economists (whom he did
not distinguish from the neoclassicals) with neglecting the prob-
lem of effective demand. Ricardo conceived that the workers are
obliged to consume their wages in order to live; the landlords
consume their rents and the capitalists either consume or invest
their profits. There is no possibility of a breach between supply
and demand because the product is distributed in real terms.
50 / Economic Heresies

In the corn economy there is no specialization and exchange so

there is no problem of finding a market for whatever is pro-
duced.
Marx paid some attention to the problem of "realizing surplus
value"; the product of his business accrues to a particular cap-
italist in the form of some particular commodities; they must
be sold before the proceeds will pay his wage bill and provide
his profit. Marx repudiated Say's Law and in some passages he
suggests that under-consumption will be the doom of capitalism.
In the main line of his argument, however, the capitalists are
always investing the surplus that comes into their hands so that
the problem of realization solves itself.
Rosa Luxemburg maintained that the capitalist system can
keep up its rate of investment (and therefore its profits) only so
long as it is expanding geographically. Marshall allowed for the
possibility of a collapse of confidence 24 but he did not lay much
emphasis on it and his pupils were propounding the truth of
Say's Law and the Treasury View at the time when they were
struck by the great slump.
Keynes diagnosed the flaw in the laisser-faire system that al-
lowed such a disaster to occur. Since the war, governments in
all capitalist countries have been playing a large part in their
economies and they have succeeded for some time, mainly by
high levels of expenditure largely financed by budget deficits, in
maintaining near-full employment and creating a situation favor-
able to a high overall rate of profit.
Avoiding slumps is all to the good as far as it goes, but now
there is growing up, especially in the United States, a protest
against the wasteful or pernicious lines of production into which
government and industry direct resources, and their failure to
provide for the basic human needs of the population. The neo-

24 See Principles, pp. 710-711.

 Interest and Profit / 51

neoclassical economists cannot take any part in this great debate

as long as they have nothing to contribute to it except the tat~
tered remnants of the laisser-faire doctrine that what is profitable
is right.
 4
INCREASING
AND DIMINISHING
RET URN S The expression "increasing and
diminishing returns to scale" im-
plies some kind of symmetry between these phenomena but in
origin they have nothing in common. The notion of diminishing
returns was developed from Ricardo's theory of rent; increasing
returns, from Adam Smith's principle that the division of labor
depends upon the extent of the market.
The classical economists were concerned with a process of
historical development. A number of confusions and contradic-
tions have arisen from the neoclassical attempt to squeeze their
concepts into the mechanical equilibrium of a stationary state.
The concept of constant returns to scale, in the technical
sense, means that each physical input required for a given output
-man-hours of labor of specific skill and energy, machines of
specific types, materials, sites, and so forth---can be regarded
as homogeneous within itself, and that a given proportionate in-
crease in each input will bring about an equal proportionate in-
crease in output. Diminishing returns arise from the fact that
some inputs, in particular those that are given by nature, can-
not be increased at will. To produce a certain proportionate
increase in output then requires a more than proportionate in-

52
 Increasing and Diminishing Returns / 53

crease in other factors. There are still conditions of constant

returns in a technical sense; if all factors were increased, output
would increase in the same proportion. On the other hand, the
economies of large-scale production which give rise to increasing
returns operate by changing the nature of the inputs. Output per
man-hour grows as work becomes more specialized; equipment
can be designed to produce a larger output at lower cost, larger
supplies of materials can be more finely graded, and so forth.
It is not a question of the proportions in which given physical
inputs are used but rather a question of the specification of the
inputs themselves. When some inputs have to be provided on a
large minimum scale-say a railway network-strictly constant
returns can be realized only for increases in output which are a
multiple of the capacity of the indivisible inputs. For ranges of
increases in between there are increasing returns due to sub-
optimal utilization of the input. (To make the Walrasian sys-
tem work we have to assume divisibility of all factors-other-
wise the services of some items would fall to a zero price just
after they had been built at great expense.) This concept is logi-
cally distinguishable from the economies of specialization but
the two are likely to be mixed up together in any actual case.
The main difficulty about these conceptions is connected with
time. A change (in output, in prices, or in costs) is an event,
taking place at a particular moment, that alters the situation in
which the change took place.

IRREVERSIBILITY
The notion of a functional relationship between output and
costs can make sense only in strictly short-period analysis. When
the specification of inputs and of methods of work remains un-
changed from year to year, output may rise and fall, as more of
variable inputs are applied to one that is fixed, up and down a
supply curve which remains independent of the direction of
change. Such conditions may be approximately fulfilled when

c·
54 / Economic Heresies

the amount of output of a particular crop depends upon the

application of man-hours of work, over a yearly cycle, to a
particular area of agricultural land. There must have been ir-
reversible investments made in the past in clearing the land, in
drainage, irrigation, and so forth; but once the investment has
been made, productive capacity is kept intact in the course of
operating it, so that henceforth investment is indistinguishable
from the "natural resources" in which it is embedded; the short-
period situation is quasi-permanent. Similarly, in industry with
given equipment, output per head may fall or rise as older plant
is bought into or put out of use.
But a long-period supply curve is a very treacherous concept.
To increase productive capacity requires investment. The larger
capacity will exist at a later date than the smaller capacity which
preceded it. In general it will be different in its technical nature,
for three reasons. First, technical change is continually going
on in the industrial economies. New plants will embody tech-
niques formerly untried. Second, the mere fact of expanding
capacity involves technical adaptations even when they are
applications of general principles already known. (The notion
of a "book of blueprints" exhibiting "the state of technical
knowledge" has played a part in doctrinal controversy, not in
realistic analysis. In reality techniques are blueprinted only when
they are about to be used.) Third, large installations often re-
quire investments of a quasi-permanent type which alter the
whole situation forever after.
Marshall was uneasily aware of the problem of irreversibility.
He thought of an increase in output as taking place through
time. A lower point on his falling supply curve is at a later date.
When output has once expanded from B to A, a retraction of
output back to B would take place at lower costs than obtained
when B was the rate of output in the first place. 1 This was a way

1 See Principles, Appendix H.

 Increasing and Diminishing Returns / 55

of smuggling technical progress, learning by doing, and irreversi-

ble investment into the static theory.
The most important example of this way of thinking was
the "infant industry case" as an exception to the presumption in
favor of free trade. It is sufficiently obvious that when one coun-
try is trying to catch up upon the advanced technology of an-
other, it must protect its industry from lower-cost competition
until it has cut its teeth. In the process of development the scale
of industry may grow but the main point is not the scale but
the time that it takes for workers and managers to learn the
business and for accumulation to provide the installations that
it requires. Since there was no room for time in the neoclassical
model, the argument had to be framed in terms of economies of
scale. This, like Marshall's irreversible supply curve, is an ex-
ample of common sense breaking in and thereby wrecking the
logical structure of the equilibrium model.
Economists have not much emphasized the opposite kind of
irreversibility-the destruction of resources, the devastation of
amenities, and the accumulation of poison in air and water.
Pigou made a great point of "external diseconomies" such as
the smoke nuisance but, within the confines of his stationary
state, he could not emphasize permanent losses. It has been left
rather to the natural scientists to sound the alarm, while ortho-
dox economists, unperturbed, continue to elaborate the pre-
sumption in favor of laisser faire.

"MARGINAL PRODUCTS"
A second problem presented by the concepts of diminishing
and increasing returns, was concerned with the relation of
"marginal products" to factor prices. In the Walrasian sta-
tionary state, the higgling of the market and recombination of
factors are supposed to settle all marginal productivities and
all hire-prices by a simultaneous process. There is, in a certain
56 / Economic Heresies

sense, a rising marginal cost for each commodity; if the output

of anyone commodity were to be increased, it would have to
attract factors of production from other uses so that their price
in terms of the commodity in question would have to be raised.
But such an increase in output is only notional. When the
supplies of all factors of production are given, output of one
commodity can increase only if other outputs are reduced. A
change in the pattern of demand means that some factors of
production are released, where demand has fallen, to be
transferred to the production of the commodities for which
demand has risen. Before we can say what happens to the price
of a commodity of which output rises, we must know what
specific factors of production are released by the fall in output
of other commodities.
To find the marginal product of a specific factor, say a certain
type of machine, we have to consider what output would be lost
if a unit of this factor were withdrawn. This loss is the reduction
in output of the commodity that this machine was used to make
minus the increase in output of other things due to deploying the
labor and other factors cooperating with the machine in other
uses. The physical marginal product is thus a very complex en-
tity, while the value of the marginal product has no unambigu-
ous meaning, since the pattern of prices, of factors and com-
modities, is altered by the change in productive capacity. Thus
it is hard to understand what is meant by saying that a factor
(say a park of machines of a particular type) receives a reward
(say, the hire-price per machine year) 2 equal to the value of
its marginal product. 3
Marginal product in Ricardo's scheme has a quite different

2 We have to assume that machines do not require amortization for it

is impossible to distinguish gross and net product in Walrasian terms.
Cf. above, p. 10.
3 Cf. below, pp. 68-69, for the definition of marginal productivity in
a one-commodity economy.
 Increasing and Diminishing Returns / 57

meaning. In the simplest form of Ricardo's model, the only

output is "corn," which stands for all agricultural produce, and
it is the only wage good. Capitalist farmers are accumulating
corn in order to expand future output. To employ a man from
harvest to harvest requires a specific investment of corn-the
wage fund-which is equal to the wage bill for a year. Labor
and capital are inseparable-the unit of input isa man-year of
work together with the investment of corn in advancing a year's
wage. Capitalist farmers maximize profits by deploying labor
in such a way as to equalize the intensive and extensive margins
of cultivation, that is, so that the additional output of corn from
adding a man-year of work on the best land is not less than can
be obtained by increasing the area of cultivation (neglecting the
cost of breaking in new ground). Thus marginal product per
man-year falls as employment and output expand over suc-
cessively less fertile land. (Rent absorbs the difference in the
productivity of better and worse land, so that the farmer receives
the same average return for each man he employs.) Now, the
marginal product of an additional man employed provides the
wage per man-year and the profit on the capital required to
employ him. It is far from being the case that each "factor"
separately receives its marginal product. Man-plus-capital earns
the marginal product (which is equal to the product of a man-
year of work on marginal, no-rent land). The wage-bill for the
ulan-year is deducted from this product and what remains is
the profit on the capital required to employ a man. The princi-
ple remains the same when capital includes equipment and
stocks of materials, though the problem of valuation is then
not so simple as when output and capital are made of the same
stuff.
In Ricardo's scheme, the corn-wage was fixed by the needs
of subsistence, so that as output per man (net of rent) was fall-
ing, the rate of profit on capital was being eroded. If we like to
postulate a constant rate of profit on capital, then, in such a
58 / Economic Heresies

case, the real wage would be falling with the marginal produc-
tivity of a man-plus-capital as total employment increases.
Marshall understood the difference between marginal pro-
ductivity in Ricardo and in Walras but he made it very difficult
for his readers to see the point. 4

ECONOMIES OF SCALE
The application of the idea of marginal productivity to the
case of increasing returns caused even more trouble. Marshall
thought of the economies of scale as mainly internal to an in-
dividual firm operating a single plant. There were also external
economies due to the development of an industry as a whole.
He did not think of any limit to economies of scale. "As the
industry grows, the firm grows." Thus (at constant money-wage
rates) cost per unit of output was a decreasing function of out-
put. But he maintained that prices are equal to average cost
including an allowance for normal profit. Thus prices must fall
with costs. However, for each firm, marginal cost is less than
average cost; therefore less than price.
This was Marshall's famous dilemma. 5 How can competitive
conditions be reconciled with increasing returns?
Pigou tried to rescue Marshall by postulating an optimum
size of firm at which long-period costs are at a minimum. Then,
to enjoy normal profits, the firm must be working beyond the
point of minimum cost to just such an extent that the excess of
marginal cost (which is equated to price) over long-period aver-
age cost is sufficient to yield the required profit. 6 When price is

4 Cf. above, p. 12.

5 Cf. G. L. S. Shackle, The Years of High Theory, p. 11 et seq. (Cam-
bridge: 1967).
6 The equilibrium price can be presented as equal to minimum average
cost by including a lump-sum normal profit per annum in total cost. In
 Increasing and Diminishing Returns / 59

higher and output is greater than this, supernormal profits are

attracting in new competition and forcing the firm back. Con-
trariwise when output is less. To make room for increasing re-
turns, Pigou then had to rely upon purely external economies,
or "economies of large scale to the industry." Each firm was
always working under conditions of rising marginal cost, but an
increase in the number of firms would lower average cost at the
minimum for all of them. (This is a simplified account of an
intricate argument which was broken off before it was re-
solved.) 7 This fanciful construction, although it was demolished
by Piero Sraffa more than forty years ago, is still used as the
basis of the "theory of the firm" in modern textbooks.
The next problem was to introduce the "laws of returns" into
a theory of the relative prices of commodities.
In the Walrasian stationary state all supplies of factors are
physically specified and fixed in amount. Each pattern of de-
mand then produces a particular pattern of relative prices. (In
the P.O.W. camp, if there were a larger proportion of Sikhs, who
do not smoke, the cigarette prices of other items in the parcels
would generally be higher.) 8 Pigou did not think of physical
factors (except "land") as being specified and fixed; nor did
he go to the other extreme (which came into fashion after his
time) of thinking of the stock of capital equipment as a large
lump of putty. He did think of the total of "resources" as being
somehow given.
Formalizing Marshall's vague suggestions, he identified in-

Wicksell's version of this argument, profits are zero in equilibrium so

that price is equated to minimum cost to the firm. See Lectures (London:
Routledge, 1934, vol. 1, p. 26). But for Wicksell the rate of interest
enters into costs of production. Cf. below, p. 97.
7 See "Increasing Returns and the Representative Firm: A Sympo-
sium," Economic Journal (March 1930).
8 See above, p. 4.
60 / Economic Heresies

dustries with commodities, and he divided the industries into

those where diminishing returns predominate, so that the supply
price of the commodity is rising with output, and those where
economies of scale to the industry predominate, so that supply
price is falling.
A change in the pattern of demand would release "resources"
from some commodities to be embodied in means of production
for other commodities. This put into his head the idea that to
reduce the output of commodities "subject to diminishing re-
turns" and transfer resources to commodities "subject to in-
creasing returns," by a system of taxes and subsidies, would
bring about an increase in total real output and in welfare. How-
ever, he soon recognized that this was based upon a false sym-
metry between increasing and diminishing returns. 9 A reduction
in demand for a commodity produced with the aid of a scarce
factor reduces the rent of the factor. This is a transfer of wealth,
not a saving of cost to society.
Abstracting from scarce factors, what remained of the argu-
ment seems to be as follows. Each commodity is produced by
a competitive industry which sells it at a price corresponding to
its cost of production including normal profits. Some commodi-
ties are more susceptible to increasing returns than others, so
that if "resources" were moved between industries to take ad-
vantage of the difference, the loss of economies of scale in those
where output was reduced would be less than the gain where
output was increased. The pattern of demand is strongly affected
by relative prices (in general, commodities are substitutes for
each other) so that demand would be shifted by taxing some

9 Pigou revised the argument originally put forward in Wealth and

Welfare after the error was pointed out by Allyn Young in his review of
the book (Quarterly Journal of Economics, August 1913). In successive
editions of The Economics of Welfare, it became more incomprehensi-
ble at each attempt. Wealth and Welfare was published in 1912, the first
edition of Economics of Welfare in 1920, and the fourth in 1932.
 Increasing and Diminishing Returns / 61

commodities and subsidizing others (the net revenue being

zero). Total money income is given (there is full employment of
workers at constant wage rates and a fixed total of "waiting"
receiving a given rate of interest). The price to the consumer of
taxed commodities would be raised by little, if anything, more
than the tax (because they have little economies of scale to
lose) while the price to the consumer of the subsidized com-
modities would be reduced by more than the subsidy, because of
the gain of economies. Thus the real income of consumers would
be increased.1O
This argument was never treated seriously as a recommenda-
tion for policy and nowadays it seems to have dropped out of
the canon of orthodox teaching. Pigou put it forward as an ex-
ample of the theoretical exceptions to the rule that perfect com-
petition, in conditions of laisser faire, produces the optimum
distribution of given resources between alternative uses. Here
again common sense was breaking in, but he managed to catch
it and wrap it up in the assumptions of static equilibrium before
it could do much harm.
All these difficulties and confusions connected with the con-
cepts of diminishing and increasing returns arise from the neo-
classical attempt to escape from time. When we set the argu-
ment in what I have called the .Marshallian model-a growing
economy with a constant normal rate of profit on capital-they
appear much less intractable. Irreversibility is no problem. Time
marches in; there is no need to pretend that the past is the same
thing as the future. In the Marshallian model, the dilemma be-
tween competition and falling costs disappears. If the normal
rate of profit is constant, as the economy expands, it follows that
as output per head rises, money prices fall relatively to money-
wage rates. By assuming a constant rate of profit Marshall has

10 Cf. R. F. Kahn, "Notes on Ideal Output," Economic Journal

(March 1935).
62 / Economic Heresies

assumed that prices are kept in line with costs; competition may
be highly imperfect, in the sense that each firm has considerable
freedom in setting prices; the number of independent firms in
anyone market may be falling; but still the economy is competi-
tive in the broad sense; all he needs is to assume that firms gen-
erally prefer to take advantage of falling costs to expand sales
rather than to try to hog a monopolistic profit by restricting the
growth of output. l1 In this model there is no great importance
to be attached to the distinction between economies of scale and
technical progress, nor between economies internal to a firm,
economies accruing to an industry producing a particular com-
modity or economies resulting from the general development of
industry, transport, distribution, and finance. 12 As time rolls on,
output of all kinds is increasing; productivity rises more for
some commodities than others and relative prices alter accord-
ingly. So long as the rate of profit on capital is constant through
time, long-run normal prices are governed by costs. The forces
of demand-the distribution of purchasing power, needs and
tastes of the consumers, and persuasive skill of salesmanship-
influence the composition of output. The only effects of de-
mand upon prices arise where there are bottlenecks created by
specialized factors of production in limited supply which cannot
be broken by technical innovations or where economies of scale
are concentrated upon a particular commodity. Only thus does
the composition of output react upon costs of production and
so on relative prices. Supply-price rising and falling with the
sale of output of particular commodities then appears as a quite
minor complication. (Marshall, it seems, puffed it up out of all
proportion in order to bring supply and demand into the fore-
front of his doctrines.)

11Cf. below, p. 102.

12The last type of economies was the subject of the famous address
of Allyn Young, "Increasling Returns and Economic Progress," Economic
Journal (December 1928).
 Increasing and Diminishing Returns / 63

But once we bring historical time into the argument, it is not

so easy to present the free play of the market as an ideal mecha-
nism for maximizing welfare and securing social justice. Mar-
shall himself admitted that accumulation and employment de-
pend upon expectations of an uncertain future. His short-period
theory is a theory of instability and in historical terms his theory
of distribution based on "rewards" of "factors of production"
becomes meaningless. Economic history is notoriously a scene
of conflicting interests, which is just what the neoclassical econo-
mists did not want to discuss.
 5
NON-MONETARY
MOD E L S Much of traditional doctrine is set
out in terms of an economy which
operates without money, implying that the "real" system is
operating behind "the veil of money" which the economist must
tear aside. Money, however, in this view is not only a veil, it
can somehow interfere and distort the real relations which would
obtain without it. The main point of this doctrine is that, in a
market in which all transactions are conducted in kind, supply
creates demand, goods are the demand for goods, so that with-
out money there could never be under-consumption or over-
production; there could never be involuntary unemployment or
under-utilization of productive capacity.
But what does the absence of money entail? The Walrasian
model is not only without money, it is without time. Goods are
exchanged against goods "today." Prices are set at the level
that clears the market so that there is no carry-over to tomorrow.
The essential characteristic of the model is not the absence of
money but the absence of any effect of expectations about the
future on present behavior.

64
 Non-monetary Models / 65

MARKET PRICES
In an actual market, even the simplest, a trader sells for the
sake of acquiring purchasing power that he can use later-hours
or years later as he pleases. Any durable good that is expected
to be in fairly general demand in the future provides a vehicle
for purchasing power. Clearly it is a great convenience to all
concerned when some particular commodity is recognized as a
general medium of exchange; 1 when convention has endowed a
commodity with general acceptability the demand for it in its
capacity as money overshadows its direct use. Trade has ceased
to be swapping of goods for goods; demand is divorced from
supply; equilibrium is not guaranteed. Money then gets the
blame for the fact that the future is uncertain.
As we know very well from experience, an equilibrium price
will be established when dealers know what the equilibrium
price is. (The leading case of this phenomenon was the opera-
tion of the gold standard before 1914.) A chance fall in price
today below the equilibrium level is quickly corrected by buying
for stock; a chance rise, by running stocks down. Production
is carried on in the knowledge of what costs it is worthwhile to
incur. When dealers have to guess the future course of prices, a
fall today often leads to selling, causing a further fall, and con-
trariwise. Producers take time to adjust supplies; an increase in
demand leading to a high price is followed by an excessive in-
crease in output that cracks the market. Uncertainty, not money,
is the cause of the trouble.
There is a quite different sense in which the Walrasian model
is non-monetary, that is, there is no general price level. Each
seller is interested in the purchasing power of his own particular

1 Cf. R. W. Clower, Monetary Theory, Penguin Modern Economic

Readings (London: Penguin Books, 1969), pp. 9-14.
66 / Economic Heresies

product over particular things that he wants to buy. But this,

which should be the strongest point in Walrasian analysis, gen-
erally seems to be smoothed over in expounding it. The model
is used to show that a competitive equilibrium has the char-
acteristic that (with given supplies of physical factors of produc-
tion and a given list of products) no more could be produced
of anyone thing without producing less of another. This is de-
scribed as an optimum position. An increase in the supply of
anyone product moves the system to a superior position. But
superior from whose point of view? This must be looked at.
Let us suppose that, after a position of equilibrium has been
established, there is an increase in the supply of one commodity,
say peanuts, others remaining unchanged. The peanut price of
other commodities is raised. A new equilibrium is established in
which the distribution of income among the traders is affected
by the change in prices that has occurred. When the elasticity
of demand for peanuts in terms of things in general is unity, the
sellers of peanuts purchase the same total of goods as before
while the rest of the traders purchase more peanuts. (There may
be changes in the relative prices of other commodities and
changes of income among the sellers of peanuts.) When the
demand is elastic, the sellers of peanuts purchase more of other
commodities. (Provided that there are no inferior goods for
which demand falls when real income rises, the direction of a
change in the quantity of commodities in general is unambigu-
ous.) Now the rest are getting more peanuts and giving more
of their own products, that is, they are consuming less of each
others' products. The sellers of peanuts (taken together) are
better off. Some of the rest, for whose commodities peanuts are
a substitute, are likely to be worse off. When the demand for
peanuts is inelastic, the sellers of peanuts are worse off. They
are giving more peanuts for less of other commodities. The rest
(taken together) are better off.
Where supply and demand rule in the modern world, that
is, in trade in primary products, demands are generally highly
 Non-monetary Models / 67

inelastic. A good harvest may be a disaster-the farmer hanged

himself in the expectation of plenty. Moreover, of course, the
instantaneous re-establishment of equilibrium after a change in
supply is a myth. The rest of the trading community does not
necessarily gain from a sharp fall in the income of one group of
producers. When a large part of the market for British textiles
was in the colonies, a fall in the price of tea or cocoa caused
unemployment in Lancashire.
All this has nothing to do with the existence of money. Con-
flicts of interest are a necessary characteristic of a system in
which value depends upon scarcity.
The Walrasian system claims to provide a theory of general
equilibrium while it is often said that Marshall with his one-at-a-
time method provides only partial equilibrium. In fact, Walras
provides only half a system for he discusses the prices of com-
modities without discussing the incomes of the people who trade
them. The Pigovian model (for what it is worth) is a general
equilibrium system though, like the Walrasian one, it is con-
cerned only with comparisons of equilibrium positions, while the
Marshallian system is more general than either, for it permits
the discussion of processes going on through time.

A ONE-COMMODITY ECONOMY
A quite different kind of non-monetary model is set up in
terms of an economy with only one commodity. There are no
relative prices-supply and demand have nothing to bite on-but
net output per annum of the community is divided into income
from work and income from property. In Ricardo's corn econ-
omy, the wage rate is fixed as a quantity of corn. (If wages were
actually paid in bags of corn, corn would fulfill one of the most
important functions of money.) The wage fund which reappears
every year at harvest time as a quantity of corn is both physical
capital and value of capital. The excess of the year's output of
corn over replacement of capital and consumption is the year's
68 / Economic Heresies

saving which is added to the stock of capital for investment in

expanding employment over the coming year. There can be no
deficiency of demand and there is no problem of finding an outlet
for investment; more workers, in Ricardo's world, are always
available to take employment when offered the standard corn-
wage. The model eliminates instability and uncertainty in order
to concentrate on one problem-the distribution of the product
of industry between the classes of society. The elimination of
money is incidental to the elimination of uncertainty.
The neo-neoclassicals also make use of a one-commodity
world. In their model, capital as a wage fund is neglected. (They
want to say that the marginal physical product of labor is equal
to the wage, not to the wage plus interest on working capital.)
Capital is a stock of physical means of production. We can
adapt the corn model to these requirements by supposing that
the wage bill is paid in arrears out of the year's harvest and by
introducing seed corn which is owned by capitalist employers.
Land is a free good. There is a well-behaved production func-
tion in labor-time and seed corn. At the beginning of any cycle
of production there is a stock of seed corn available to be in-
vested and a labor force available to be employed. Competition,
not only between workers for jobs, but also between employers
for hands, ensures that the wage rate is set at the highest level
that is compatible with full employment. The wage, that is to
say, is equal to the marginal physical product of labor-the out-
put that would be lost if one less man were employed, with the
same total amount of seed corn. The return per ton on the
stock of corn is then equal to its marginal product, that is to
the output that would be lost if a small amount, say one ton less
of seed, were used with the same amount of labor-time, minus
the replacement of the ton of corn. 2

2 It -is usual to appeal to Euler's theorem to demonstrate that the

marginal products, multiplied by the respective amounts of the factors,
add up to the total net product. (See, for instance, WickseU, Lectures
 Non-monetary Models / 69

The special features of this construction are, first, that any

given quantity of physical capital can provide employment for
any amount of labor (the elasticity of substitution between them
is positive over an indefinite range-the production function
never cuts the axes of the diagram in which it is drawn) and a
change in the ratio of capital to labor can be made without any
cost of adaptation. 3 Second, the wage bargain between a worker
and an employer is made in terms of his own product. Third,
investment consists in adding something to a heap of means of
production that already exists without requiring any change
in it.
This model was constructed to provide a bridge between the
conception of the stock of capital as a set of physical inputs
which assist labor to produce output and as a fund of finance
which enables the employers of labor to make profits. The
bridge breaks when the peculiar assumptions of the one-com-
modity world are withdrawn from under it. 4
The one-commodity assumption makes it possible to define

vol. I, pp. 127-128.) In the com model, the point can be simply shown.
When one less man is employed, a quantity of seed com equal to the
average used per man is released and re-allocated to the remaining labor
force. The marginal product of labor is equal to the average net output
per man minus the net addition to output due to the seed corn released,
that is, minus the marginal product of corn multiplied by corn per man.
Total net output is equal to the marginal product of labor multiplied by
the quantity of labor plus the marginal product of seed corn multiplied
by the quantity of com. We take, as the unit of the quantity of each
factor, the unit in which marginal productivity is expressed. Here we take
a man-year of work as the unit of labor and a ton as the unit of seed
corn.
3 In many neo-neoclassical models this conception is applied also to
technical progress. For instance: "Capital is made up of a large number
of identical meccano sets which never wear out and can be put together
. . . to incorporate the latest technical innovations in successive editions
of the Instruction Book." T. W. Swan, "Economic Growth and Capital
Accumulation," Economic Record (November 1956).
4 Cf. L. L. Pasinetti, "Switches of Technique and the Rate of Return
in Capital Theory," Economic Journal (September 1969).
70 / Economic Heresies

the marginal product of labor; the other assumptions, in par-

ticular indefinite substitutability of factors and a chronic scarcity
of labor relative to demand for it, are necessary to support the
proposition that wages tend to equal marginal products.
Even with all these provisos it is still not true to say that
the return per unit of capital corresponds to the marginal prod-
uct of investment from the point of view of society as a whole.
Consider a Wicksell process of accumulation of capital with
a constant labor force, in the setting of the corn economy.
Every year some part of the net output of corn is saved and
added to the revolving stock of seed to be used next year. Ac-
cording to the rules of the model, seed corn per man employed
is higher next year, net output per man is higher, though in a
smaller proportion, the corn-wage is higher, and the return per
ton of corn is reduced. Now compare the situations at two dates
between which an appreciable addition has been made to the
stock of corn. At the later date the permanently maintainable
net output of corn (with a constant amount of work) is higher
than at the earlier date by some number of tons per annum.
This increment of output may be described as the product of the
investment. In the corn economy the productivity of investment
(which is in general both complex and vague) I) can be neatly
expressed as the ratio of the increment of net output to the incre-
ment of the stock of seed corn brought about by investing the
savings made over the period.
Each year the return per ton of seed corn is equal to the mar-
ginal productivity of the stock in existence. This rate of return
is lower at the second date, and the level of wages is correspond-
ingly higher. Total net income at the second date is equal to the
income at the earlier date plus the increment which represents
the product of investment. In respect to the level of income of

5 Cf. above, p. 33.

 Non-monetary Models / 71

the earlier date, the workers get exactly what the owners of seed
corn lose; the increment of income also is divided between
wages and profits in shares appropriate to the new position. Thus
the increment of income due to the investment that has been
made consists partly of wages. 6 At each point in the process of
accumulation, the rate of profit (equal to the marginal product
of corn-capital) is less than the product of the investment made
over the past period.
In discussion of this subject, starting with Wicksell himself,
and continuing to the present day,7 there is much confusion
between comparisons of positions of equilibrium corresponding
to different rates of profit (a pseudo-production function) and
the effects of accumulation as a process going on through time;
and between the return enjoyed by owners of wealth (the reward
of waiting) and the productivity of investment from the point
of view of society as a whole. The assumptions of the one-
commodity economy enable us to sort out these distinctions,
though it is more often used to confuse them.

MONEY AND "REAL FORCES"

These models are non-monetary in the strict sense that they
purport to operate without a medium of exchange. It is not
possible to set up anon-monetarymodel, in this sense, for an in-
dustrial economy. Where workers do not own the means of
production that they operate and where there is specialization
there must be money in the sense of some vehicle for general

6 This was the original meaning of the 'Wicksell Effect' (see C. O.

Uhr, "Knut Wicksell, a Centennial Evaluation," American Economic Re-
view, December 1951). I borrowed this term, perhaps illegitimately,
for the difference in the value of a given physical stock of capital goods
at different rates of profit.
7 Cf. above, p. 15.
72 / Economic Heresies

purchasing power in which wages are paid. 8 The models of

Marshall, Wicksell, and Pigou are monetary in this sense but,
in the orthodox teaching which Keynes had to attack, "money"
was used in a wider and vaguer sense. There was a dichotomy
between "real" forces which determine the relative prices of
commodities and factors of production and "monetary" forces
which are responsible for the general price level and for move-
ments of the national income as a whole. Thus "money" had
both a wide metaphysical sense, as something opposed to what
is "real," and a precise narrow sense as the actual monetary
arrangements operating through the supply of gold and the
institutions and policies of the banking system. Thus the ortho-
dox doctrine (though rarely precisely stated) implied that the
"real" forces establish equilibrium while aberrations such as in-
flation and unemployment are "the fault of money" and could
be avoided by a correct policy of the monetary authorities.
The relation between monetary and real forces was expressed
by Marshall in terms of the discount rate and the long-term rate
of interest and by Wicksell in terms of the money rate and the
real rate of interest. (As we have seen, the long-term or real rate
of interest is identified with the rate of profit on capital.) The
argument can be set out in terms of what I have called the
Pigovian model-a stationary state in which the rate of interest
is governed by the reward of waiting-that is, in which it is set
at such a level that the rentiers are willing to own just the
amount of wealth that is in existence. The system is monetary
in the sense that a medium of exchange is in use but the mone-

8 A model such as that set up by von Neumann or Sraffa does not

need to mention money. All relative prices are determined by technical
conditions and the rate of profit. Wages can be expressed in physical
terms (though they could equally well be expressed by a money-
wage rate and a level of money prices). But these models are systems
of equations expressing equilibrium relationships. They cannot be used
to discuss the behavior of the human beings who inhabit them.
 Non-monetary Models / 73

tary system cannot have any permanent effect upon any of the
real relationships within it. 9
In the Pigovian stationary state, the rate of profit is equal to
the rate of interest governed by the reward of waiting. The
money-wage rate and the rate of profit determine all prices
(given technical conditions) and the money value of national
income and of total wealth. The supply of the medium of ex-
change is equal to the demand for it. Let us call this the supply
of cash. The theory of real and money rate of interest is then
set out as follows. Suppose that, in a position of equilibrium,
cash is arbitrarily increased without any change in income hav-
ing occurred. The redundant cash cannot find holders at the
current level of prices. Consequently the price of interest-bear-
ing bonds is bid up. Now the rate of interest has been reduced
below the equilibrium level. Rentiers are getting less than the
reward of waiting and they are holding a larger nominal total
of wealth. They start to draw on capital to increase consumption;
at the same time the entrepreneurs, finding the rate of interest
lower and the rate of profit higher because of the increased de-
mand for goods, want to increase investment. The two sectors,
competing against each other for labor, drive up money wages.
Wages and prices rise until the demand for cash has risen to
absorb the extra supply at the equilibrium rate of interest. (Both
Marshall and Wicksell told the story this way round. They did

9 Before Keynes, there was a great deal of confusion in the neoclassi-

cal scheme as to the effect of the wage bargain on the distribution of
income. Pigou himself was converted only after a struggle to the view
that a change in money-wage rates primarily affects money prices, not
real wages. (See "Money Wages in Relation to Unemployment," Eco-
nomic Journal, March 1938.) Yet this was in fact essential to his own
scheme. If prices are equal to marginal costs, an all-round rise in money-
wage rates, which raises all marginal costs proportionately, must cause
a corresponding rise in money prices. It was a mistake in terms of his
own model to suppose that the wage bargain is made in real terms.
74 / Economic Heresies

not remark that a rise in the rate of interest does not bring·
money-wage rates down in the rapid and painless manner that
the argument requires.)
This way of arguing, however, is not legitimate. The station-
ary state is essentially timeless. It is not equipped to deal with
unexpected events. It can be used only for comparisons of equi-
librium positions. All that we can say on the basis of this analy-
sis is that, in Pigovian stationary states, there is one stock of
cash appropriate to each level of money-wage rates. When the
wage rate and the stock of cash are not in harmony, the sys-
tem is not in equilibrium.

MONEY IN A GOLDEN AGE

The argument can be extended to the model of steady growth
with a normal rate of profit constant through time. The effective
labor force is growing at a steady rate and entrepreneurs, one
with another, are causing accumulation to take place at the same
rate. The wage bill in terms of money is rising at the same rate.
When the growth of effective labor force is only a growth of
numbers, money-wage rates are constant; when output per head
is growing, money wages rise in step so as to keep the price level
constant. (Technical progress is neutral in Harrod's sense, so
that the ratio of the value of capital to the value of output is
constant.) The supply of cash must be increasing at the same
rate as the wage bill. This comes about automatically if firms
are financing the difference between this month's investment and
last month's saving by borrowing from the banks and the bank-
ing system is allowing the supply of cash to expand appropri-
ately. The proportion of household income saved is constant and
so is the proportion of profits retained by firms to finance in-
vestment. In each period, household savings being placed by the
purchase of securities, together with retained profits, are suffi-
cient to finance the investment of the last period. The increase
 Non-monetary Models / 75

from one period to the next in the value of investment being met
by borrowing from the banks, the quantity of money is growing
at the same rate as the money value of the stock of capital. The
rate of interest is constant and the supply of finance grows with
the demand for it.
This is not a system in equilibrium; there is no mechanism to
keep it on its path. The only point of setting it up is to see where
it is liable to go wrong.

"REAL" INSTABILITY
The sources of disturbance in the golden age are not confined
to the operations of the monetary system. Keynes broke down
the old dichotomy; he showed that the "real" forces can by no
means be relied upon to establish equilibrium, however well the
monetary system behaves. The monetary system may, indeed,
contribute to disturbances, and monetary management may do
something, though not much, to dampen disturbances due to
other causes, but it has a minor influence, either for good or ill.
Supply and demand in the sphere of commodity trade; the in-
stability of investment in uncontrolled private enterprise; the
interplay of money-wage rates and money prices; these are the
sources of disturbances. If the "real" forces were behaving prop-
erly, it would not be difficult to get the monetary system to work
properly too.
The old dichotomy still haunts modern theory. It has been
revived in a curious form by the Chicago school.1° The argu-
ment is that when the national income, in real and money terms,
is growing smoothly at a steady rate, it is found that the stock
of money is growing at the same rate (at least a definition of

10 See Milton Friedman, "The Role of Monetary Policy," American

Economic Review, March 1968, reprinted in The Optimum Quantity of
Money (Aldine Press, Chicago, 1969).
76 / Economic Heresies

money can be found that makes this true). Therefore, to cause

the stock of money to grow steadily is all that is necessary to
ensure steady growth in the national income. The non-monetary
theory, that the real forces tend to establish equilibrium, thus
reaches its apotheosis in the doctrine that money is the only
thing that matters.
 6
PRICES AND
M 0 N EY The archetypal quantity theory for-
=
mula, MV PT, like any identity.
has to have its terms defined in such a way as to make it hold.
Keynes' identities, Y == + == +
C I ==
C S; S I, have the great
advantage that they correspond to columns in the national ac-
counts, income, consumption, investment, and saving. (The
formula comes right because Y, I, and S are all net of deprecia-
tion, and the budget and the foreign balance are either boiled
in appropriately with S and I or set out separately.) Keynes, in
fact, embraced the modern system of national income accounts 1
in order to be able to convince his critics that I = S, hot
+
S aM. The elements in the quantity equation are not so trans-
parent. M must be defined as the quantity in existence at a mo-
ment of time of a specific list of items, say, coins, notes, and
bank deposits (whether inclusive or current accounts only).
T is an index of transactions-is it to include all transactions
made during a year, or only those connected with the production
and distribution of the real national income? Similarly, P, an
index of prices, must be appropriate to the list of transactions.

1 See General Theory, pp. 102-103 and How to Pay for the War, ap-
pendix by E. Rothbarth.

D
77
78 / Economic Heresies

Then V, the velocity of circulation of the money included in M,

can only mean PTI M. Alternatively, if M includes all kinds of
money used in connection with the transactions listed in T, and
V is the average of the number of times that each item is used in
a year, then M is PT IV. The so-called Cambridge equation,
MIP = kR, where R is "resources" (presumably, an index of real
income) and k is the ratio of money balance to "resources," was
even more vapid-it only says that the real value of the stock
of money is its nominal value divided by an appropriate index
of prices.
The truisms are intended to be used, in a looser way, to ex-
hibit causal relationships, like ~ Y =.J.PJ),
s
where s is the mar-
ginal propensity to save. If the quantity equation had been read
in the usual way, with the dependent variable on the left and
the independent variables on the right, though rather vague, it
would not have been silly. Suppose that, between one year and
the next, PT rises; either activity has increased-employment
and output are higher this year than last, or the general price
level has risen because of a rise in costs in money terms; then
if the quantity of money has not increased, the velocity of cir-
culation must have risen. But it was not taken so. It was used
as the basis of the argument that a change in the quantity of
money will produce a more or less proportionate change in the
price level.
I do not think that its supporters ever really believed it, for,
if they had, they would have joined the money cranks in the
great slump and proclaimed: "It can all be done with a foun-
tain pen."
The reason why the equation was read left-handed was that it
grew up side by side with a body of doctrine couched in "real"
terms which consisted mainly of an exposition of the conditions
of equilibrium. Employment, accumulation, real wages, and the
production and consumption of commodities were looked after
 Prices and Money / 79

in Volume I of the Principles of Economics, and there was noth-

ing left to discuss in Volume II except the supply of money and
the general price level.

THE THEORY OF INTEREST AND MONEY

Keynes sloughed off the quantity theory in several stages. In
the General Theory he emerged in a shining new skin. To find
the determinants of the general level of prices, he now main-
tained, we must look first to the level of money-wage rates; the
level of effective demand is a minor influence which he was in-
clined to believe in, though it was not essential to his argument. 2
The volume of transactions varies with the level of effective de-
mand, and the principal determinant of changes in effective
demand is changes in the level of investment.
Money plays a secondary role. The quantity of money is con-
trolled by the banking system. When effective demand is begin-
ning to rise, it induces an increase in average overall velocity of
circulation as money balances are moved from the inactive to
the active circulation. If the banking system fails to allow the
quantity of money to increase, the demand for active balances
will tend to raise the level of interest rates, which causes V to
rise to the required extent-for instance, rentiers are induced to
exchange holdings of money for bonds when the yield of the
latter rises. In a period of unemployment, an increase in the
quantity of money can do some good. It tends to lower interest
rates and to permit the "fringe of unsatisfied borrowers" to get
finance.
Relatively to given expectations oj profit, a fall in interest
rates will stimulate investment somewhat, and, by putting up
the Stock Exchange value of placements, it may encourage ex-

2 See "Relative Movements of Real Wages and Output," Economic

lournal (March 1939).
80 / Economic Heresies

penditure for consumption. These influences will increase effec-

tive demand and so increase employment.
The main determinant of the level of interest rates is the state
of expectations. When bond-holders have a clear view of what
is the normal yield which they expect to be restored soon after
any temporary change, the banking system cannot move interest
rates from what tbe-y are expected to be. It is the existence of
uncertainty or "two views" that makes it possible for the banks
to manipulate the money market. But even when the rate of
interest can be moved in the required direction, it may not have
much effect. The dominant influence on the swings of effective
demand is swings in the expectation of profits.
This is the burden of the main argument. But Keynes' ideas
were not always definite, precise, and consistent. Emphasis on
monetary factors varies from one part of his analysis to another.
In his utopian vision of a future without wars, population growth,
or major inventions, he foresaw a world in which the need for
accumulation will have come to an end; both the social return
and the private rate of profit on investment will have fallen very
low. Provided that the rate of interest is brought down corre-
spondingly, the euthanasia of the rentier will have removed the
worst vices of capitalism (though there could still be fun in
speculating on the Stock Exchange). But there may be some
ultimate bottom stop to the rate of interest so that it might get
stuck at too high a level and bar the entry to paradise regained.
This argument certainly flatters the monetary system, not so much
because of the notion that liquidity preference might ultimately
check accumulation, as because of the tacit assumption that suc-
cessive reductions in the rate of interest could keep accumula-
tion going in face of a falling rate of profit. (I remember that,
when I came to Chapter 17, reading the proofs of The General
Theory, I wrote that for the first time I was finding the argu-
ment difficult to follow; Keynes replied, in effect, that he was
not surprised for he found it difficult himself.)
 Prices and Money I 81

In the main argument, which concerns immediate short-

period situations, while Keynes was dethroning "monetary"
theory, he yet gave money great importance, even in the title
of his book, The General Theory of Employment, Interest and
Money. This was for three reasons. First, he came to it from
the Treatise and the Tract. The tradition in which he was work-
ing connected the general price level with "money" (as opposed
to the "real forces" at work in Volume I of the Principles of
Economics); he came at the problem of effective demand from
that side.
Second, he was influenced by a particular historical episode
in which he himself played a part. For some years after the de-
parture of sterling from the gold standard in 1931, the exchanges
were strong so that the level of domestic interest rates was in-
sulated from competition for international reserves; in 1934
the gilt-edged rate of interest was still relatively high (by the
standards of those days), the economy was still in a slump, and
there was a large conversion operation falling due. Keynes, in a
speech delivered as Chairman of the National Mutual, argued
that interest rates were too high and ought to be reduced. Gilt-
edged rose sharply; this assisted the Bank of England (if it did
not persuade them) to institute a period of relatively cheap
money which helped to promote a boom in house-building that
was in any case under way and brought some relief to the
country while the world slump had scarcely begun to lift out-
side.
This episode confirmed Keynes' conviction that the rate of
interest is a monetary phenomenon, not bound by "real forces,"
and at the same time gave him, perhaps, an exaggerated im-
pression of how much good it could do.
Third, in the orthodox system that he had to attack, the
rate of interest, confused with the rate of return on investment,
was the regulating mechanism which caused savings to be in-
vested and secured eqUilibrium with full employment. He had
82 / Economic Heresies

to make every possible concession to this point of view in order

to get a hearing. It would have been much simpler to start by
assuming a constant rate of interest and a perfectly elastic supply
of money. But then his whole case would have been dismissed
as a misunderstanding of the orthodox position. He was
obliged to accept the presumptions of his critics in order to
explode them from within.

COUNTER-REVOLUTION AND RESTORATION

The doctrines of the General Theory (though, as Keynes said,
"moderately conservative") were felt to be shocking. The con-
cessions which he made to orthodoxy about the rate of interest
were used to provide a mollifying version of his system of ideas
which turned it back onCe more into a variant of the quantity
theory.
Professor (now Sir John) Hicks first stepped into the field
with his IS and LM curves. Construct a diagram with the rate of
interest on the vertical axis and national income on the horizon-
tal axis. The vertical axis represents some kind of index of the
level of the complex of interest rates on loans and placements
of all kinds. This is a convention which Keynes himself used; he
conducted a large part of his argument in terms of the rate of
interest, though reminding the reader from time to time that
this is a severe simplification of a very complex concept. The
horizontal axis is labeled income. This presumably means net
national income (in a closed economy) though the argument
seems to require gross national product at constant prices or in
wage units, or perhaps output in units of employment.
The IS curve slopes down to the right, indicating that output
is a decreasing function of the rate of interest. A level of output
corresponding to a level of IS implies a propensity to consume;
to each level of I, the rate of investment per annum (net or
gross?), corresponds a certain level of consumption. This is
 Prices and Money / 83

intended to be a representation of the multiplier (the relation

of an increase in consumption brought about, presumably with-
out time lag, by an increase in the rate of investment) though in
the diagram it seems to mean the average relation of consump-
tion to income. No doubt these points 'Jf definition could be
cleared up, but there is a more serious difficulty. What is the
meaning of making the rate of investment a function of the rate
of interest?
Keynes' contention was that a fall in the rate of interest rela-
tively to given expectations of profit would, in favorable cir-
cumstances, increase the rate of investment. This was rather a
hazy part of his argument. Kalecki amended it to show that,
with given prospects of profit, a cheapening and increased avail-
ability of finance may increase investment plans being made to
be carried out over the immediate future. When it does so, then
as plans are realized in increased expenditure on investment
and the multiplier gets to work on increasing consumption, cur-
rent receipts of firms rise. Assuming that their plans for the fu-
ture are influenced by present experience, it follows that a fur-
ther rise of investment will take place. This generates a boom
which will not last because after some time the growth in the
stock of productive capacity competing in the market will over-
take the increase in total expenditure and so bring a fall in cur-
rent profits per unit of capacity, with a consequent worsening of
the expected rate of profit on further investment.
Keynes was sometimes apt to collapse the future into the
present in a confusing way. His account of a boom is to say that
a high rate of investment causes a fall in expected profits as the
supply of productive capacity increases. 3 But one thing he could
never have said is that a permanently lower level of the rate
of interest would cause a permanently higher rate of investment.

3 General Theory. p. 136.

84 / Economic Heresies

Now consider the other curve in the diagram-LM slopes up

to the right. Here the causation is reversed-a higher level of
output causes the rate of interest to be higher. Evidently there
is a hard and fast fixed quantity of money (gold or cowrie
shells?) without which transactions cannot take place. Thus a
higher output, requiring more money in active circulation, leaves
less available to satisfy liquidity preference and so is associated
with a higher rate of interest.
Here the simplification of allowing the interest rate to stand
for the terms on which finance can be obtained proves treacher-
ous. Are we to suppose that loans are harder to get in a boom
than in a slump? One of the best known lessons of monetary
history (which Keynes often repeated) is that a fall in activity
leads to a collapse of confidence and a rise in interest rates,
whereas, at a time of high activity, high expectations of profit
affect the confidence of lenders as well as borrowers.
However, in this scheme, LM is an increasing function of the
level of income and IS is a decreasing function of the rate of
interest. There is one level of the rate of interest and level of
income at which the curves cut. This is the equilibrium position
corresponding to the given fixed amount of money. Here we
have the quantity theory in its purest form. If the equilibrium
level of income is below that corresponding to full employment,
let the authorities increase the supply of money so as to shift LM
to the right until it cuts IS at the full employment level. But now
a piece of Keynes' long-run speculations is introduced. There
may be a minimum level below which the rate of interest will
not fall however much the supply of money is increased. If this
level is above the rate of interest shown by IS at full employ-
ment, monetary policy alone cannot do the trick. This repre-
sents the doctrine of the liquidity trap. (If Keynes' own ideas
were to be put into this diagram, it would show IS as the volatile
element, since it depends upon expectations of profit; the case
where full employment cannot be reached by monetary means
 Prices and Money / 8S

would be shown by IS falling steeply and cutting the income

axis to the left of full employment.)
The concept of the liquidity trap arises from approaching the
problem of unemployment from the quantity theory. According
to that theory in its simplest form, the elasticity of demand for
money in terms of goods is equal to unity so that an increase in
the quantity of money leads to a proportional rise in the prices
of goods; but, after Keynes, the qualification was introduced
that the demand for money in terms of bonds may be highly
elastic, so that an increase in the quantity of money runs into
hoarding and fails to raise prices. On this basis was erected a
number of fantastical notions, such as the view that falling prices
are good for trade, because, by raising the amount of real wealth
represented by a fixed stock of money, they encourage consump-
tion. The whole complex of ideas was somehow spliced onto a
Walrasian version of neoclassical theory and used to bind up
the wounds which the great slump had inflicted on laisser-faire
orthodoxy.
Axel Leijonhufvud's On Keynesian Economics and the Eco-
nomics of Keynes (Oxford University Press, 1968) is valuable
because he destroys this construction by its own internal con-
tradictions and clears away a great deal of rubbish, while re-
maining strictly within the framework of monetary theory.

THE CHICAGO SCHOOL

The quantity theory that was being expounded at the Univer-
sity of Chicago, while Keynes was wrestling with the General
Theory, was much more robust and self-confident than the
wishy-washy "Cambridge" version. In 1934, Simons was main-
taining that the slump was due to two main causes. First, trade
unions exercising monopoly power in the labor market would
not let money-wage rates fall; second, "It is no exaggeration to
say that the major proximate factor in the present crisis is com-
86 / Economic Heresies

mercial banking" 4-the reason being that any movement in

business earnings leads to an expansion or contraction of
credit which drives prices up or down. The instability of PT is
due to the instability of MV. The remedy that he proposed was
100 per cent bank reserves, so that the government would have
complete control of the supply of money. The best might be to
keep M constant but there is a difficulty: "The obvious weakness
of fixed quantity, as a sole rule of monetary policy, lies in the
danger of sharp changes on the velocity side, for no monetary
system can function effectively or survive politically in the face
of extreme alternations of hoarding and dishoarding." {) The cor-
rect rule is to maintain "the constancy of some price index,
preferably an index of prices of competitively produced com-
modities." 6
This noble simplicity has been a good deal sophisticated by
the modern Chicagoans, led by Milton Friedman. A great part
of their work consists in historical investigations of the relation-
ship between changes in the supply of money and national in-
come in the United States. The correlations to be explained
could be set out in quantity theory terms if the equation were
read right-handed. Thus we might suggest that a marked rise in
the level of activity is likely to be preceded by an increase in the
supply of money (if M is widely defined) or in the velocity of
circulation (if M is narrowly defined) because a rise in the
wage bill and in borrowing for working capital is likely to pre-
cede an increase in thtf value of output appearing in the statistics.
Or that a fall in activity sharp enough to cause losses deprives
the banks of credit-worthy borrowers and brings a contraction in

4 Henry C. Simons, "A Positive Program for Laissez Faire," reprinted

in Economic Policy for a Free Society (Chicago: Chicago University
Press, 1948), p. 54.
{) "Rules Versus Authorities in Monetary Policy," Economic Policy
for a Free Society, p. 164.
i Ibid., p. 183.
 Prices and Money / 87

their position. But the tradition of Chicago consists in reading

the equation from left to right. Then the observed relations are
interpreted without any hypothesis at all except post hoc ergo
propter hoc.
There is an unearthly, mystical element in Friedman's thought. 7
The mere existence of a stock of money somehow promotes
expenditure. But insofar as he offers an intelligible theory, it
is made up of elements borrowed from Keynes. 8 An increase in
the basis of credit, say by open-market operations, permits the
banks to satisfy part of the "fringe of unsatisfied borrowers" or
to offer loans on easier terms; part of additional bank lending
goes to various financial intermediaries and part goes into the
market for bonds. A general easing of interest rates puts up the
Stock Exchange. In various ways this permits investment plans
to be carried out that otherwise would have been frustrated for
lack of finance, as well as encouraging purchases, especially of
consumer durables, both because loans are easier to get and
because, with a rise in the capital value of placements, rentiers
reduce their rate of saving. Thus, other things equal, an increase
in the quantity of money promotes an increase in activity.
The difference between Friedman and Keynes is not in the
analysis (insofar as it is intelligible) but in emphasis. The
general implication of Friedman's doctrines is that money is very
important, not as a symptom but as a cause of instability. At one
time he seemed to suggest that correct control of monetary pol-
icy could stabilize the economy, but in a later pronouncement
he maintained that it is too difficult for the authorities to hit off
the right policy and that wrong policy exaggerates instability;
therefore the best policy is just to keep the quantity of money
expanding at the rate (say, 4 per cent per annum) which would

7 See The Optimum Quantity of Money, Chapter 1-

8 Cf. Don Patinkin, "The Chicago Tradition, the Quantity Theory and
Friedman," Journal of Money, Credit and Banking (February 1969).
88 / Economic Heresies

be appropriate if the economy was expanding at that rate. 9 This

is a return to Simons' ideal of a constant M, adapted to modern
notions of growth, without any of the reservations which made
him hesitate to recommend it. 10
In both these schools, Keynes' theory of the rate of interest
related to liquidity preference has been twisted, one way or the
other, into a vendon of the quantity theory; the essence of the
quantity theory is that there is a definable and recognizable
quantity, M, the movements of which have a powerful influence
upon the movements of PT. In short, the whole argument of
both schools consists in reading the quantity equation from left
to right instead of from right to left.

THE THEORY OF EMPLOYMENT

Keynes himself, of course, was not contending with the bastard
progeny of his own ideas. He had to combat the old orthodoxy,
which lay much deeper. Hicks confused the issue by presenting
his purely monetary construction as the theory of the "classics,"
just as Dennis Robertson confused an analysis of "the supply
and demand of loanable funds" with an argument about "the
real forces of productivity and thrift." The old orthodoxy was
rooted in Say's Law. "What constitutes the means of payment
for commodities is simply commodities. Each person's means of
paying for the productions of other people consist in those which
he himself possesses." 11 Or as Marshall put it:

9 Cf. pp. 75-76. Friedman, The Optimum Quantity of Money, p. 48.

10 Friedman himself makes Simons out to have been a Keynesian and
considers that he underestimated the importance of the quantity of
money. See "The Monetary Theory and Policy of Henry Simons,"
Journal of Law and Economics (October 1967): reprinted in The Opti-
mum Quantity of Money.
11 J. S. Mill, quoted by Marshall, Principles, p. 710.
 Prices and Money / 89

The whole of man's income is expended in the purchase of serv-

ices and of commodities. It is indeed commonly said that a man
spends some portion of his income and saves another. But it is a
familiar economic axiom that a man purchases labour and com-
modities with that portion of his income which he saves just as
much as he does with that which he is said to spend. He is said to
spend when he seeks to obtain present enjoyment from the serv-
ices and the commodities which he purchases. He is said to save
when he causes the labour and the commodities which he pur-
chases to be devoted to the production of wealth from which he
expects to derive the means of enjoyment in the future.l 2

Saving makes available real resources-labor and means of

production-which will be used for investment. (Marshall ad-
mitted that the mechanism breaks down when confidence fails,
but his disciples in the Treasury did not follow up that line of
thought.) The latter-day neoclassicals have made the basis of
the old orthodoxy much clearer than it was at the time when
Keynes was trying to diagnose it. In their models it is explicitly
assumed that there is and has always been correct foresight, or
else "capital" is malleable so that the past can be undone (with-
out cost) and brought into equilibrium with the future; in short,
they abolish time. But this is not enough to ensure full employ-
ment. They have also to assume that the wage bargain is made
in terms of product; the real-wage rate finds the level at which
the stock of "capital" is squeezed up or spread out to employ
the available labor force. Keynes took it for granted that in an
industrial economy wage rates are set in terms of general pur-
chasing power; and he brought the argument down from the
cloudy realms of timeless equilibrium to here and now, with an
irrevocable past, facing an uncertain future. Money then comes
into the argument as "the link between the present and the fu-

12 Pure Theory of Domestic Values (1879), reprinted in Scarce Tract

series (London: London School of Economies, 1930), p. 34.
90 / Economic Heresies

ture." The General Theory is a "monetary theory" only in the

sense that relationships and institutions concerned with money,
credit, and finance are necessary elements in the "real" economy
with which it is concerned.

INFLATION
Of all the conclusions of the Keynesian Revolution, the most
disruptive of orthodoxy was the proposition that there is no
such thing as an equilibrium of the general price level. The price
level in an industrial economy is a historical accident. The main
influence upon the level of prices, at any moment, is the level
of money-wage rates, and the level of money-wage rates, at any
moment, is the result of movements that have taken place over
the distant or recent past. Prices, certainly, may move relatively
to money-wage rates, over the long run with changes in produc-
tivity and over the short run with changes in the level of profit
margins (the degree of monopoly), but these movements, which
affect the level of real wages, are confined within narrow limits
by technical and market relationships, while the level of money-
wages and prices is not tethered to anything and may change (at
least upward) without any limit at all.
The orthodox theory that Keynes was attacking maintained
that a cut in money-wage rates means a cut in real wages and
that a cut in real wages induces an increase in employment.13
Keynes' argument was not the one which has been foisted on
him by the bastard Keynesians-that money-wage rates are
rigid for institutional reasons. It was that if wages could be cut,
in a slump, it would make the situation worse, because it would
lead to falling prices and expectations of further falls, so that
investment was discouraged, while the fall in the money value of

13 See The Report of the Macmillan Committee on Finance and Indus-

try, 1931, Addendum III by T. E. Gregory.
 Prices and Money / 91

assets would reduce the availability of credit and is liable to

break the banks.14
There was one grain of truth in the orthodox doctrine: if one
country can succeed in lowering money-wage rates relatively to
those of its trade rivals, it gains a competitive advantage. This
is one of the "beggar-my-neighbor remedies for unemployment."
Keynes pointed out that the same advantage could be gained
in much less painful manner by depreciating the exchange rate.
While Keynes was immediately concerned with the causes and
consequences of a deficiency of effective demand, he also pro-
vided the basis for the analysis of inflation. In spite of his own
optimism, the argument of the General Theory suggested that
it would be by no means a simple matter to cure capitalism of
its major defect and leave the rest of its mechanisms intact. It
seemed obvious that the continuous full employment would be
accompanied by a continuous fall in the value of money which
might disrupt the basis of the whole system. A rise of prices,
however it is caused, cannot in itself produce continuing in-
flation. A rise of prices of goods sold to the public reduces the
expenditure in real terms of household incomes and increases
the share of profit in value added. The immediate impact of the
inflation exhausts itself in a change in the distribution of real
income between firms and households. But this sets the stage for
a change in money incomes. There are two channels for further
inflation-via profits and via wages. If firms expect the favor-
able situation to last, they may step up plans for investment;
dividends may be increased; share prices are likely to rise,
causing capital gains; thus rentier incomes are increased in
money terms. More immediately, the balance of power in wage-
bargaining shifts to the workers' side. Firms, seeing good pros-

14 See Keynes, "The Consequences for the Banks of the Collapse of

Money Values" (August 1931). reprinted in Essay in Persuasion (Lon-
don: Macmillan, 1951), p. 168 ff.
92 / Economic Heresies

pects of profits, are reluctant to provoke strikes; in a situation of

general high employment there are acute scarcities of some types
of labor; the cost of living has recently risen. Money-wage rates
for some groups catch up on or overtake the rise of prices, and
other groups have a strong claim to match their increases. Thus
rises in money incomes and expenditure increase demand and
rises in wage rates increase costs. The effect of the original rise
in prices is frustrated, and prices rise again.
The distinction between "demand-pull" and "cost-push" is
not very useful when applied to the market for goods but it has
an important meaning when applied to the market for labor. An
excess demand for labor creates a situation in which firms,
competing for hands, raise effective wages by various devices
above the rates agreed with the trade unions. This also, of
course, creates a situation favorable to increasing agreed rates.
Thus demand-pull encourages cost-push. Pure cost-push is seen
when there is unemployment and slack demand for labor but
the trade unions are able to enforce rises in wage rates all the
same.
The direction in which the vicious spiral of wages and prices
is spinning may have an influence on the level of real wages.
When the movement starts from prices (say, as a result of a
sharp upswing in effective demand or of an increase in indirect
taxes) it may be impossible for money-wages to catch up; when
the movement starts from the side of wages there may be a delay
in the adjustment of prices, so that there is, at least for some
time, an improvement in real wages, at least for the best or-
ganized groups of workers. (One of the main difficulties of so-
called incomes policy is to persuade the trade unions that they
cannot benefit from raising money-wage rates, because to some
extent they can.)
Keynes' argument about the relation of wages to exchange
rates also takes on a new meaning in an inflationary situation.
A country where money-wage rates, relative to output per head,
 Prices and Money / 93

rise faster than in the rest of the capitalist world is liable

to develop a deficit in the balance of trade; this undermines con-
fidence and produces a deficit in the balance of payments. A
depreciation of the currency, in so far as it is effective in redress-
ing the balance of trade, increases inflationary pressure and is
liable to wipe out the benefit by causing wages to rise all the
faster. Thus there is a second vicious spiral, of wages and ex-
change rates as well as of wages and prices.
The prediction that continuous inflation must sooner or later
undermine confidence in the currency and lead to hyper-inflation
turns out to have been exaggerated. The system has proved
capable of adapting itself, with surprising success, to a continu-
ously falling value of money. All the same, its consequences are
extremely demoralizing. The distribution of income thrown up
by the market economy can be tolerated as long as every in-
dividual feels that his position in it is due to fate or to his own
merits. When it becomes clear that the relative incomes of in-
dividuals are mainly determined by the bargaining position of
the group to which they belong, the ethics of the system-a fair
day's work for a fair day's wage-disintegrates, industrial dis-
cipline is -undermined, and the tradition of public service gives
way to a general scramble for advantage-even doctors and
school teachers are exasperated at the erosion of their position
to the point of striking for more pay.
An incomes policy which would check inflation by preventing
overall money incomes from rising faster than overall real out-
put would require a general acceptance of some pattern of re-
wards for various kinds of work. Once traditions have been
questioned, there is no acceptable criterion for deciding what it
should be. Still less is there any acceptable criterion for decid-
ing the general distribution between work and property, espe-
cially since the old argument that the rich are necessary to so-
ciety because they provide savings has been discredited. More-
over, even if it were possible to find an acceptable incomes
94 / Economic Heresies

policy, to apply it would require a fundamental change in the

traditional powers of both workers and employers which neither
side is willing to accept. Perhaps the modern revival of a doc-
trine so unconvincing as the quantity theory of money can be
explained as a refuge from the uncomfortable thought that the
general level of prices has become a political problem.
Kalecki's interpretation of the General Theory was less opti-
mistic than Keynes'. He foresaw a political trade cycle; govern-
ments would vacillate between fear of inflation and fear of un-
employment; the stop-go cycle would overlay a general trend
of accumulation. "The regime of the 'political business cycle'
would be an artificial restoration of the position as it existed
in nineteenth-century capitalism. Full employment would be
reached only at the top of the boom, but slumps would be rela-
tively mild and short lived." 15 It is now necessary to add that
the system appears to grow progressively more difficult to con-
trol as time goes by. During periods of high activity prices and
money incomes rise. During the slack periods they do not fall.
Indeed they may continue to rise. The doctrine that a small
percentage of statistical unemployment is sufficient to keep
prices constant cannot hold, for as soon as it is discussed in
public the trade unions hear about it and become determined to
prove it false. The notion that a fall in demand lowers prices
also becomes dubious. The profit margins set by the powerful
firms do not correspond to the pure monopoly prices of static
theory. They are set at a level calculated to yield a satisfactory
return on some normal or standard average level of utilization
of capacity. A fall in sales raises unit costs. The seller has no
motive for lowering prices and will feel it appropriate to raise
them. Thus, the political trade cycle is overlaid by a chronic
tendency to rising prices. Inflation is favorable to profits, for

15 "Political Aspects of Full Employment," The Political Quarterly

(October/December 1943).
 Prices and Money / 95

over and above the return due to "value added" by incurring

costs of production there is an extra element due to the passage
of time. Stocks bought at one date can be sold at a higher price
merely because they are sold later. The product of labor paid at
today's wage rates will live to compete with products paid for
at higher wages. When investment is planned and debts incurred
on the basis of expectations that inflation will continue, a check
to rising prices would cause acute financial embarrassment and
might precipitate a sharp slump. An inflationary economy is in
the situation of a man holding a tiger by the tail.
The soothing doctrines of the bastard Keynesians have been
a very poor preparation for the actual problems of modern
capitalism.

THE UNIT OF ACCOUNT

Continuous inflation is a great nuisance for the governments
and for a majority of the citizens of the countries where it oc-
curs. It is also a nuisance for accountants and economists. Infla-
tion destroys the convention that "a shilling is a shilling." The
purchasing power of money has to be related to the time at
which it is to be spent; the rate of profit on investment and the
rate of interest are not the same in terms of money as in terms
of purchasing power. The actual realized profit over a past
period can be deflated by whatever seems to be the most appro-
priate index number, though it is never simple and obvious what
index number is appropriate. This concerns the man of words
who is recording past history. The man of deeds who plans in-
vestment or places his wealth on the Stock Exchange has to con-
sider a future which is yet unknown. The expectations which
guide his conduct mayor may not turn out to have been well
founded. All this adds great complications to an analysis which
is sufficiently complicated when the purchasing power of money
has an agreed objective meaning. In order to discuss old-
96 / Economic Heresies

fashioned questions we have adopted the old-fashioned conven-

tion of assuming a constant actual and expected purchasing
power of money over consumption goods, so that the rate of
profit on capital has the same meaning in real and in money
terms. When the questions have been dealt with on this basis,
it will be necessary to take them to pieces again to discuss the
complications that have to be incorporated into the analysis in
a world in which there is no unit of value that has an agreed and
unambiguous meaning.
 7
THE THEORY OF
THE FIR M The so-called theory of the
firm that was being debated
before imperfect competition came into fashion 1 (and which
survives in many modern textbooks) arose from the attempt to
find an answer to "Marshall's dilemma." 2 If competition means
that each producer can sell as much as he pleases at the going
market price, then to maximize profits he goes on expanding
output so long as marginal cost is less than price. But if long-
period average costs fall as output expands, because of econ-
omies of scale, marginal cost is less than average cost. There is
no position of long-period equilibrium until one firm has estab-
lished a monopoly. To resolve this contradiction, Pigou intro-
duced the idea of an optimum size of firm. A firm, on this view,
consists of a unit of the factor of production, "management";
there are diminishing returns, after a certain point, from the ap-
plication of the other factors. labor and capital, to this unit. Dis-
economies of large-scale mauagement set in, offsetting the econ-
omies of specialization. The long-period average-cost curve for

1 See "Increasing Returns and the Representative Firm: A Sympo-

sium," Economic Journal (March 1930).
:I Cf. above, p. 58.

97
98 / Economic Heresies

the firm has a U shape; at the minimum point, in equilibrium,

long-period marginal cost, long-period average cost, and the
price of the commodity being produced are all equal. (The argu-
ment is simple only when the firms in one industry are produc-
ing a single homogeneous output.) The optimum size of firm
relative to the market in which it is operating must be small
enough to establish a sufficiently large number of firms to keep
competition going.
At each moment the firm is maximizing its current profits by
seIling the output at which marginal short-period cost is equal
to price. When price exceeds long-run average cost a super-
normal profit is attracting new competition; when it is below,
investment is being siphoned off into other industries. Costs in-
clude the rate of interest on finance. In equilibrium, price (and
short-period marginal cost) exceeds average prime cost by a
sufficient margin to permit quasi-rents to accrue at the level
which will provide for replacement and normal profit on the in-
vestment involved at a rate equal to the ruling rate of interest.
A variant of the scheme was set out in Hicks' Value and
Capital (a book which had an important influence in the re-
vival of orthodoxy after Keynes). There, it is tacitly assumed
that each industry consists of a fixed number of firms so that,
for every commodity, the price (equal to marginal cost) is an
increasing function of the level of output. In this scheme, Wal-
rasian prices governed by supply and demand take the place of
Pigovian costs of production including normal profits.

PERFECT AND IMPERFECT COMPETITION

The short-period analysis of prices in both schemes depends
upon competitive conditions, not in a vague MarshaIlian sense,
but on the strict assumptions of an indefinitely large number of
independent sellers in a perfect market, which entails a perfectly
elastic demand, at the ruling market price, for the product of
 The Theory of the Firm / 99

each seller. Each firm is producing its short-period capacity out-

put (unless it has temporarily gone out of business because the
ruling price is below its average prime cost). The limit on output
is set by rising marginal cost; for any greater output, marginal
cost would exceed the selling price.
In the slump it was sufficiently obvious that plants were not
being operated at full capacity with rising marginal costs; the
upshot of the debate which broke out in the 1930s was that
firms set prices by adding a gross margin to prime costs; below
designed capacity, prime costs per unit of output is a constant
or decreasing function of the level of output; if prime cost is
identified with marginal cost, clearly it is much less than price.
To reconcile this with the assumption of profit-maximizing pol-
icy, the idea was introduced that marginal revenue is related to
price by the formula e/e-I, where e is the elasticity of demand
from the point of view of the individual seller; but since this e,
if it exists at all, can only be a calculation in the minds of in-
dividuals concerned with price policy, it does not add much to
the argument. 3
Even in prosperous times it is unusual for most plants to
be working to capacity-if capacity were the limit to output
there would be no need for advertisement. In a normal situation,
it seems, there are many firms which would produce a larger
output if it could be sold at the going price. Customers distribute
themselves among rival sellers according to inertia, proximity,
genuine differences in needs or tastes, and response to the blan-
dishments of salesmanship. Moreover, when an acute seller's
market is being enjoyed with full-capacity operation, it is pru-
dent to allow delivery dates to lengthen rather than to choke off
excess demand by high prices. Thus the system of analysis ac-

3 For an account of my own contribution to this debate, see the

Preface to the second edition of The Economics of Imperfect Competi-
tion (London: Macmillan, 1969).
100 / Economic Heresies

cording to which price equals marginal cost, so that the level of

gross profits is governed by the excess of marginal cost over
average prime cost, is seen to be without application.
Wi~h this, the notion of wages equal to value of marginal
product also collapses. When a plant is being worked below
designed capacity with constant average prime costs, a loss of
one man-shift of work entails a loss of the average value of out-
put of one man.
In general in modern industry, it seems that the wage bill is
about half of value added. In the typical case, then, the value
of marginal product of labor is twice the wage. Hicks was quite
correct in saying that to abandon the assumption of perfect
competition "must have very destructive consequences for eco-
nomic theory" if economic theory means nothing more than
Walrasian general equilibrium. 4
The long-period aspect of the Pigovian scheme is even less
convincing. The individual firm is not supposed to be aiming
at the optimum size. It is aiming at maximizing the flow of net
profit to be got in any situation. Then whenever a firm finds
itself with a rate of profit in excess of the rate of interest, it
surely must be carrying out investment in order to get more
profit in the future. The argument is concerned with a stationary
state, with fixed "resources"; it is intended to show how given
resources are allocated between different uses; a constant total
of "capital" is washing about between different industries finding
the level at which the rate of profit is equalized. But once profit-
maximizing firms are allowed into the story, how can accumula-
tion be kept out?
The essence of the competitive process is that some firms take
business away from others. Those which are successful grow
faster than industry as a whole, those which are least successful

4 See Value and Capital, p. 83 (Oxford: Clarendon Press, 1939).

 The Theory of the Firm / 101

cease to exist. Pigou's concept of managerial diseconomies of

scale, perhaps, can be applied to the kind of business where "the
entrepreneur" is a particular individual. As a business grows
beyond the scope of one-man management it runs into difficul-
ties. 5 But this is an exceptional case. At any moment there may
be a number of individuals who have found a satisfactory niche
and manage to maintain independence, but the majority of busi-
nesses are either growing, being forced out of existence by the
growth of others, or being absorbed into some larger organiza-
tion.
Why do firms grow? Some contemporary writers are inclined
to treat growth as a specially modern phenomenon arising from
the divorce between control and property in the modern cor-
poration, legally owned by a floating population of shareholders
and operated by a cadre of salaried managers; they seem to sug-
gest that there was a past period to which the textbook scheme ap-
plied. Yet obviously the successful family businesses of the early
nineteenth century must have been just as keen on growth as
any modern corporation. Anyone who is in business naturally
wants the business to survive (particularly if his own heirs and
successors are involved) and to survive it is necessary to grow.
When a business is prosperous it is making profits; for that very
reason it is threatened with competition; it would be feckless to
distribute the whole net profit to the family for consumption;
part must be ploughed back in increasing capacity so as to sup-
ply a growing market, to prevent others coming in, or to diver-
sify production if the original market is not expanding. Anyone,
by growing, is threatening the position of others, who retaliate
by expanding their own capacity, reducing production costs,
changing the design of commodities, or introducing new devices
of salesmanship. Thus each has to run to keep up with the rest.

5 See E. A. G. Robinson, The Structure of Competitive Industry, Cam-

bridge Handbook (London: Nisbet, 1931).
102 / Economic Heresies

As we have seen, the very fact that investment is going on is

generating opportunities for profitable sales, 6 so that as long as
growth goes on, it can go on. The determination of firms to grow
by reinvesting profits was characteristic of capitalism from the
start; indeed, if it were not the case, capitalism would never
have happened.

MONOPOLY AND OLIGOPOLY

The way out of Marshall's dilemma is in the opposite direc-
tion. Where competition is vigorous, there must be a tendency
toward monopoly, which is often held up at the stage of oligop-
oly when a few powerful firms prefer armed neutrality to the
final battle for supremacy.
Marshall accounted for growth by economies of scale 7 which
give a firm a competitive advantage by reducing costs of pro-
duction. This is of importance where technology demands large
indivisible investments but in general the advantage to a firm of
size is mainly in size itself-that is, in financial power. In Mar-
shall's day, a particular business operated in a particular in-
dustry in which it had the technical know-how and the market
connections required. Now the large corporation can jump from
one industry to another, employing its own experts or buying
up a smaller concern already established there. The modern de-
velopment of conglomerates provides clear evidence that it is
financial power, rather than technical economies of scale, that
permits firms to continue to grow when they are already large.
While the reduction in the number of independent firms gen-

6 See above, p. 46.

7 In Marshall both internal and external economies accrue to the
individual firm. As usual with him, the concepts are not clear-cut. Pigou
distinguished between economies of scale to the firm and economies of
scale to the "industry" producing a particular commodity. This is a
logical set of concepts which it is not easy to apply in reality.
 The Theory of the Firm I 103

erates monopoly in particular industries in particular countries,

the breakdown in the barriers between industries and between
national economies increases competitiveness. In the textbook
theory of the firm, a monopolist, faced by a known and station-
ary demand-curve for the commodity that he controls, restricts
output to the level at which marginal revenue is equal to mar-
ginal cost and so extracts the maximum possible profit from the
market. There are, certainly, examples of monopolies which
conform more or less to the textbook pattern, but in general the
great firms are far from restricting output-they are continu-
ously expanding capacity, conquering new markets, producing
new commodities, and exploiting new techniques. The level of
profit margins and the rate of profit on investment that they en-
joy are in general higher than those in stagnant markets where
competition still prevails, because in expanding markets they
can catch the profits that they need to finance expansion. Mod-
ern industry is a system not so much of monopolistic competi-
tion as of competitive monopolies.
The command of finance by the great firms gives them free-
dom to follow their own devices, manipulating not only the
market economy but also national and international policy.
("What's good for General Motors is good for the United
States.") The breach which this makes in the textbook scheme
is much more serious than the abandonment of the doctrine that
prices are governed by marginal costs which followed from the
recognition of imperfect competition. It destroys the basis of
the doctrine that the pursuit of profit allocates resources between
alternative uses to the benefit of society as a whole.

CHOICE OF TECHNIQUE
It is an absurd, though unfortunately common, error to sup-
pose that substitution between labor and capital is exhibited
by a movement from one point to another along a pseudo-pro·
104 / Economic Heresies

duction function. 8 Each point represents a situation in which

prices and wages have been expected, over a long past, to be
what they are today, so that all investments have been made in
the form that promises to yield the maximum net return to
the investor. The effect of a change in factor prices cannot be
discussed in these terms. Time, so to say, runs at right angles
to the page at each point on the curve. To move from one point
to another we would have either to rewrite past history or to
embark upon a long future. In dynamic conditions, changes in
the composition of demand, changes in technique, and changes
in costs of specific factors of production are continuously going
on. Investments are always made in less than perfect knowledge
of present possibilities and less than perfect confidence in ex-
pectations about the future. The stock of capital in existence
today is not that which would have been chosen if the future,
that is now today, had been correctly foreseen in the past. It
is not composed of units of the most appropriate technique; it
contains numerous fossils from earlier periods of techniques
which were chosen in conditions different from those obtaining
today. Nor is it ever being maintained in a constant form. It
is continually being done over as gross investment replaces one
set of capital goods by another set appropriate to a new complex
of expectations. To discuss the choice of technique, we must
look, not at the total stock of capital as at a point on a pseudo-
production function, but at the investment plans which are
being made at each moment.
In the Pigovian scheme any firm can borrow as much or as
little as it pleases at the ruling rate of interest. In an equilibrium
position, no firm is planning to make any net investment, for if
it expanded its productive capacity, managerial diseconomies
would cause average costs to rise and the additional returns from

8 E.g., R. M. Solow, "On the Rate of Return: Reply to Pasinetti,"

Economic Journal (June 1970).
 The Theory of the Firm / 105

an increment of output would not be enough to cover the incre-

ment on the interest bill for the additional finance. The tech-
nique of production that it has chosen is controlled by the rate of
interest and the level of wages, according to the rule that a given
output is produced at minimum cost. In a dynamic economy the
rate of interest may, perhaps, be supposed to have some influ-
ence on the amount of investment which is being planned at any
moment, 9 but there is no reason why it should influence the
choice of technique. With the finance that it is planning to in-
vest, the firm must be supposed to prefer a plan promising a
greater increment of profit to one promising less, irrespective of
what it had to pay for the finance. 10 But the problem of choos-
ing between plans is indefinitely complicated; decisions may
actually be made on hunch or on some conventional rule such
as a pay-off period. 11 When sophisticated estimates are made of
discounted cash flow, it is the expected rate of profit that comes
into the calculation, not the rate of interest. There is an im-
portant way, however, that the distribution of available finance
between firms affects the techniques that are adopted-that is,
when the minimum size of an efficient installation is very large.
Then only a powerful firm can attempt it. Smaller firms have
to be content with less ambitious projects. The powerful firm
undertakes such an investment only when it has sufficient control
over the market to be confident of a satisfactory return,12 while
the small-scale competitive producers have to be satisfied with a
lower rate.

9 Cf. above, p. 31.

10 Cf. M. Kalecki, Essays in the Theory of Economic Fluctuations
(London: Allen & Unwin, 1939).
11 N. Kaldor and J. A. Mirrlees, "A New Model of Economic Growth,"
Review of Economic Studies (June 1962).
12 This point was forcefully made by Schumpeter. See Capitalism,
Socialism and Democracy (New York: Harper, 1942), Chapter 8. See
also J. K. Galbraith, The New Industrial State (New York: Houghton
Mifflin, 1967), Chapter 19.
106 / Economic Heresies

The most important influence upon the choice of technique is

not the cost of finance or "factor prices" but the rate of invest-
ment relative to the availability of labor. When, as may happen
in the early stages of industrialization, an individual firm can
employ as much labor as it likes at a constant wage rate, it may
be supposed to find the technique that promises the highest re-
turn per unit of investment and carry on its expansion by
gradually increasing employment with the same type of equip-
ment. If a new technique is offered which is superior to that in
use, in the sense that at current prices it both reduces the cost
of investment per man and reduces the wage bill per unit of
output, then a keen profit-maximizer will install it, but there
is no great compulsion to do so.
The situation is very different in an environment of near-full
employment. A large firm whose plants provide an appreciable
proportion of the jobs in particular regions has to consider,
when planning investment, how much more labor it will be
able to recruit. It will generally find it necessary to carry out
expansion, at least partly, by raising investment per man em-
ployed. It is not provided with a predigested "book of blue-
prints" of techniques; it must find out what the possibilities are
and assess them as best it may. Nor is there any reason to sup-
pose that the process necessarily involves "capital deepening"
and a fall in the rate of profit. In the course of exploring ways of
raising output per head it will often succeed in developing su-
perior techniques. The successful firms have no great objection
to allowing money-wage rates to rise; they may even be bidding
for labor by offering various inducements to attract men from
other employers. Small firms using labor-intensive techniques
must then mechanize or go out of business. Those which survive
may well find themselves more prosperous in the end. Since, as
output per head rises, prices are likely to rise less than in pro-
portion to wage rates, it is possible to see long spells of accumu-
lation in which real-wage rates are rising but the rate of profit
 The Theory of the Firm / 107

is not falling. In this sense, "substitution of capital for labor"

is the essence of industrial development, but it has nothing what-
ever to do with the factor prices shown on a pseudo-production
function.

MACRO AND MICRO THEORY

There have been many accounts of the behavior of particular
firms (investigations connected with anti-monopoly legislation in
various countries are a rich source) and statistical inquiries into
the behavior of gross margins, the profitability of different types
of organization, and so forth. This has mainly been pure de-
scription without benefit of theory or it has befuddled itself with
attempts to fit into an inappropriate analytical scheme. A theory
of the firm appropriate to a dynamic economy is in its infancy.13
Meanwhile, it is necessary to develop a general theory of ac-
cumulation within which a micro theory can be elaborated. At
the first stage, a firm can be simply identified with the capital
that it controls; the size and number of firms making up the
whole industrial structure are not important in themselves. The
interaction between firms, however, is important as a determinant

13 A "new wave" was started twelve years ago by Edith Penrose with
The Theory of the Growth of the Firm (Oxford: Blackwell, 1959), which
has been followed up by W. Baumol, Business Behavior, Value and
Growth (New York: Harcourt, Brace & Jovanovich), R. Marris, Mana-
gerial Capitalism (London: Macmillan, 1964), M. Gordon, The Invest-
ment, Financing and Valuation of the Corporation (Homewood, Ill.:
Richard D. Irwin, 1964), and many others. In each of their models the
policy of the firm is to aim at growth, restrained by a diversity of limi-
tations. Any simple formula to describe the motivation of firms is un-
likely to be satisfactory because their behavior is highly complex and
various. The neo-neoclassical hypothesis that the aim of a firm is to
maximize the present value of its shares does not seem to say anything
very precise, for the main influence on the present value of shares is
the expectations which the market holds about the future growth of
their value.
108 / Economic Heresies

of accumulation and technical progress in industry as a whole.

The behavior of a particular firm may be discussed in terms of
its reaction to prospective profits, but accumulation cannot be
explained in terms of prospects of profit which have an objective
basis apart from the investment that is induced by them. When
firms are cautious and reluctant to invest except for a high rate
of return, the return that they actually get will be low, because
sluggish investment and high profit margins restrict effective de-
mand. The prospect of profit for each depends on what the rest
are doing.
In any case, accumulation cannot be accounted for only by
the prospect of profits. If investors were solely concerned to
find the best return on the finance that they command, the less
successful firms would stop investing and place their funds by
buying shares of the more successful. As Keynes remarked, "En-
terprise only pretends to itself to be mainly actuated by its own
prospectus, however candid and sincere." 14 The state of the
"animal spirits," which is largely a function of the energy and
competitiveness of groups of firms, is the most important factor
in capitalist development, though it by no means follows that
the most energetic enterprise necessarily produces the most
beneficial results for society as a whole.

14 General Theory, p. 160.

 8
GROWTH
MOD E L S For the classical economists, eco-
nomic growth brought about by
capital accumulation and technical progress was the central
problem; in the neoclassical era it was little discussed, except
vaguely by Marshall, who retained something of the tradition of
Ricardo; after the Keynesian Revolution it came back into fash-
ion.
The treatment of growth in von Neumann's ultra-classical
model is brutally simple. A technically specified wage is the cost
of labor and bodies are becoming available to carry it out at the
rate at which the output of wage goods is growing. The first
long-run Keynesian model was proposed by Harrod. For him,
the "natural" rate of growth of the effective supply of labor is
given exogenously and the rate of growth of the economy may
or may not keep up with it. In the neo-neoclassical models that
have since proliferated, the natural rate of growth is automati-
cally realized by some kind of equilibrating mechanism.

HARROD
The great strength of Harrod's model is that it is not an equilib-
rium scheme. It is a projection into the long-period of the con-
cepts of the General Theory. Accumulation comes about through

E 109
110 / Economic Heresies

decisions taken by profit-seeking finns and there is no guarantee

that the rate of investment in uncontrolled private enterprise
will be either steady or at a desirable level. Unfortunately, his
own exposition of his model 1 is almost as confusing as the inter-
pretations that neo-neoclassicals have put upon it.
The share of net saving in net income, s, is determined by the
propensity to save of the public; the capital to output ratio, v, is
given by technical conditions. Therefore there is only one pos-
sible maintainable growth rate, g = s/v. This is the "warranted"
rate of growth. It is important to realize that this does not mean
the rate of growth that firms will actually undertake or the rate
that they decide or desire to carry out in the given conditions.
It is the rate that they would have to carry out in order to be
satisfied, after the event, with what they have done, so as to be
willing to continue. The warranted rate of growth is an expres-
sion of the thriftiness conditions of the economy. A high war-
ranted rate of growth (relatively to the desire of firms to ac-
cumulate) generates under-consumption and so reduces actual
growth. A low warranted rate generates inflationary conditions
and stimulates growth.
The concept of s and v being exogenously determined gives
rise to the problem that has become known as Harrod's knife-
edge, though Sir Roy himself repudiates it. In long-period terms:
the formula, g = s/v, is equivalent to I/K = I/Y . Y/K, the
ratio of net investment to the stock of capital is equal to the
share of net investment in net income multiplied by the income
to capital ratio, all in value terms. (The money prices of con-
sumer goods are assumed constant, so that values can be ex-
pressed in money.) The formula grew out of a trade-cycle theory

1 See "An Essay in Dynamic Theory," Economic Journal (March

1939) Towards a Dynamic Economy (London: Macmillan, 1949), and
"A Comment on Joan Robinson's 'Harrod After Twenty One Years,'''
Economic Journal (September 1970).
 Growth Models / 111

and Sir Roy seems reluctant to admit that net income and in-
vestment have a precise value only in conditions of steady
growth or that the meaning of v that satisfies the formula
v = slg is the value of capital over the value of net income. For
him v seems to mean the incremental capital to output ratio
somehow expressed in physical terms.
The main mechanism in the trade-cycle theory was of the
"capital-stock-adjustment" type but it can equally well be inter-
preted in terms of expectations of profit. In any given situation,
with given physical productive capacity, an increase in the rate
of gross investment raises the level of current gross profits above
what it was in the immediate past. If the improvement in pros-
pects is expected to last, investment will increase further and so
profits will rise further; in short, a boom develops. Contrariwise
when the rate of investment falls. This part of the argument is
concerned with assumptions about the actual behavior of an
actual economy. The "warranted rate of growth" is a meta-
physical concept. It concerns the existence of a possible equilib-
rium path, not the stability of any path that an economy may be
following. The problem of the knife-edge is the problem of the
one and only possible value of g compatible with exogenously
given values of s and v.
Taking the Harrod formula out of Sir Roy's hands, we can
attempt to find out the meaning of the assumptions which it
requires.
The neo-neoclassicals seized upon Harrod's model and thrust
it into a pre-Keynesian mould. 2 The rate of saving governs the
rate of investment. The "warranted rate of growth" is realized,
whatever it may be. When v < Sin, there is more saving than is,

2 See, in particular, Trevor W. Swan, "Growth Models: of Golden

Ages and Production Functions" in Economic Development, ed. K. Ber-
rill (London: Macmillan, 1964). International Economic Association
Conference at Camagori.
112 / Economic Heresies

needed to look after n, the natural rate of growth (presumably

the argument always starts from a position of full employment);
g then exceeds n. There is a well-behaved production function
in output, labor, and "capital." Excess saving is raising the "cap-
ital" to labor ratio, the rate of profit is falling, and the real-wage
rate rising. As v rises, the amount of saving required to look
after growth at the natural rate is increasing; the rate at which
«deepening" is going on is decelerating, until equilibrium is
reached with g = n = slv. This is nothing more than the Wick-
sell process, in a one-commodity world, 3 superimposed upon
long-run steady growth. (The story is also told backwards. When
v > sin, decumulation sets in, with g < n and v falling, until
the equilibrium value of v is established.)
To retain the Keynesian character of the growth model, we
must interpret it in a different way. We must introduce another
term into the argument: the rate of accumulation that firms,
taken together, are willing to bring about. Harrod's central prop-
osition is that when the firms cannot carry out accumulation at
the warranted rate (for instance, because slv > n) or when they
are too slack to do so, under-consumptionand slumpy conditions
prevail, so that there is stagnation or decline over the long run
(though occasional short-lived booms may occur). Thus a high
value of s, for Harrod, plays just the opposite of its neoclassical
role. For him, far from promoting a high rate of growth, it is an
impediment to any growth at all.
To understand this paradox, we must examine the meanings
of g, v, and s in more detail.
One of the most useful and important innovations in the Har-
rod scheme is the treatment of technical progress, but since the
analysis is complicated we shall attempt to deal with one layer
at a time. At this stage we assume that the "natural" rate of

3 Cf. above, p. 70.

 Growth Models / 113

growth, n, is given only by the rate of increase of the labor force.

To strip the model to its essentials we postulate:

that is, net income per annum is exhaustively divided into con-
sumption and net industrial investment per annum and it is ex-
haustively divided into wages and profits. (Wages are taken to
include all "earned income" except the high salaries of business
executives, which should be included in the profits of their
firms.) K is the value of the capital stock. The growth rate, g, is
11K; this entails that the capital to income ratio, v or KIY, is
constant through time. The rate of profit, 7r, is PIK. At any
point on a path of steady growth that is actually being realized~
the initial conditions, including the physical composition of the
stock of equipment, must be compatible with the growth rate
that is going on. The rate of profit on capital and the share of
wages and profit in net income are constant through time; the
question which we have to consider is how they are related to
g, v, and s.
The value of capital per man employed depends primarily
upon technical conditions and upon the rate of profit. There
may be a pseudo-production function showing other possibilities.
but it would not come into the story, for the technique appro-
priate to the growth rate and the rate of profit has already been
installed at any given point on the growth path, and is being ex-
panded with each item in proportion.
The assumption of a uniform rate of profit implies that the
model is competitive in the long-period sense-no firm can
make a monopolistic super-normal profit by restricting entry into
its market. There is no need to assume perfect competition in
the short-period sense that all plant is always operated at capac-
ity with sharply rising marginal costs, so that gross profit margins
are determined by marginal cost minus average prime cost~
114 / Economic Heresies

fluctuating with every seasonal or random change in demand.

We may assume that firms set prices by a mark-up on prime cost
in such a way that, if normal capacity operation is realized on
the average, receipts will cover total cost including amortization
and yield a net profit per annum that corresponds to the rate of
profit that they hope to enjoy. (Over a period when average
utilization exceeds the normal level, actual net profit exceeds ex-
pectations, and contrariwise. On our tranquil path, we may
suppose that normal utilization is realized on the average, though
not necessarily without variations week by week.)
The degree of monopoly, or ratio of gross margins to prime
cost, now comes into the determination of v, the capital to in-
come ratio. The length of the standard working day, the preva-
lence of multiple shifts and so forth are subsumed under the
given long-period technical conditions, but the degree of utiliza-
tion of plant is connected with the short-period price-policy of
firms. (We must suppose that working hours may vary with
overtime or that the normal level of employment is less than 100
per cent of available labor.) With given plant, a higher degree of
monopoly means a lower ratio of normal to full-capacity opera-
tion and therefore a higher cost of investment per unit of output.
So much for v; we must now consider s. On what assumptions
could we find s, the share of net saving in net income, to be given
independently of the rate of profit on capital and the share of
profit in net income? (Sir Roy discusses the influence of the rate
of interest received by rentiers on the subjective desire to save
but he neglects the effect of the distribution of income between
wages and profits.) Weare not obliged to assume that every
family saves the same proportion, s, of its income. There may be
different proportions of saving by rich and poor, provided that
the distribution of income between families remains constant
through time, but poverty and wealth must not be correlated
with earned and unearned income. Rentier property-bonds,
shares, and cash-must be randomly distributed through the
 Growth Models / 115

population so that the representative family is drawing income

from wages and from interest and is saving a certain proportion
of its total income from both sources. (There must be a bank-
ing system which keeps the quantity of money growing at the
right rate to provide for the growing wage bill and for any hold-
ings of cash due to liquidity preference, so as to keep the rate
of interest constant through time at a level compatible with the
rate of accumulation that is going on.) The firms retain enough
gross profit to keep capital intact (in physical and in value
terms). We may suppose either that net profit is fully distributed
to shareholders and finance for net investment raised by means
of new issues of shares and bonds; or we may suppose that
rentier income includes capital gains due to investment of re-
tained profits and that the saving ratio, s, covers this part of in-
come as well as the rest. In such a world, the firms (taken to-
gether) are free to make the level of prices and the rate of profit
what they please by the level at which they set gross margins.
Let us suppose money-wage rates are fixed once and for all.
Comparing a higher with a lower degree of monopoly, the prices
of consumer goods are higher and the prices of investment goods
are adjusted accordingly. With a given level of employment, the
wage bill is the same in the two positions; where prices are
higher gross profits are higher. (When Ap is the excess of prices
and G is gross profits, AG = Ap (G
p
+ W).) The extra gross
profit is being paid out for replacements in the investment sec-
tor and as income to households. The higher prices reduce the
purchasing power of wages but the increment of rentier income
exactly compensates. The volume of outlay for consumption is
higher by the same amount as the value of goods sold at the
prices fixed by the firms; sY, net saving in money terms, is
higher by the same amount as I, net investment; A, amortization
allowances in money terms, are higher by the same amount as
D, outlay for maintenance of stock of physical capital.
116 / Economic Heresies

We now see the force of the expression "the degree of monop-

oly." Provided that, when the price leader in each market sets
a certain level of gross margins on his own prime costs all sellers
abide by the corresponding prices, each getting a margin which
depends on his own costs, then the price leader can set prices
as he pleases. But there is always a danger that some cad, not
content with his share, will try to increase his sales by under-
cutting, and margins will come tumbling down. The freedom of
a price leader is limited by the dispersion of costs and the ag-
gressiveness of potential competitors. Thus, in a general broad
sense, the less competitive is the general situation in an economy,
the higher the "degree of monopoly," measured by the ratio of
gross margins to prime cost, is likely to be.
Now consider the formula, g = s/v. The rate of profit may be
supposed to have an influence upon v, the capital to income
ratio, but this cannot be relied upon to get us off the knife-edge.
Given sand g, there is only one value of v compatible with
equilibrium. With a single technique, v may vary over a certain
range with the rate of profit (the value of a given stock of equip-
ment in terms of output rises or falls with the rate of profit ac-
cording to the time-pattern of production); on the pseudo-
production function there may be no rate of profit that yields
the required value of v (allowing for utilization) or there may
be several. (This was established in the "reswitching" debate.)
Even if there is a convenient value of v, corresponding to one
rate of profit, there is no mechanism in the system to bring it
into being. There is no way in which a rate of profit determined
by short-period price-policy can be supposed to find the value
that, if it obtained in long-period conditions, would be com-
patible with the right value of v.
Evidently the knife-edge is a chimera. The problem is created
by the unnatural assumption that s, the ratio of net saving to net
income, is determined by the psychology of households rather
than by the requirements of firms. All the same, in exploring it
we have learned something of value.
 Growth Models / 117

PROFITS AND SAVING

The problem of the knife-edge disappears when we recognize
that profits provide the main source of saving and that invest-
ment generates the profits that it needs. The principal source of
finance for gross investment is retention of gross profits by
firms. (Expenditure on investment precedes the receipt of the
profits to which it gives rise; in a growing economy firms must
be borrowing from banks the difference between this month's
outlay and the receipts of, say, six months ago.) 4 When invest-
ment exceeds retentions, there is saving by households. (In our
simplified model there is no excess of saving over the investment
carried out by firms-no private house-building, budget deficit,
or balance of trade-so that household income exceeds con-
sumption only by the excess of investment over retention of
profits by firms.) (; Household savings are borrowed by the firms
to finance the excess of investment over their own retentions,
directly by sales of securities, or indirectly through the banking
system, which is providing deposits to satisfy the liquidity pref-
erence of rentiers, over and above their short-term lending to
firms.
In our model, the issue of equities is treated simply as a form
of borrowing and rentiers are regarded as treating shares simply
as income-yielding placements, not as a controlling interest in
firms. The relation between borrowing by issuing shares and on
bonds of various kinds is a very intricate subject (complicated
by the legal fiction that interest is a cost but dividends are not).
Here we leave all this on one side and postulate that the only
form of long-term borrowing is the issue of shares. 6

4 Cf. p. 74 above.
15 Income of households is here exclusive of capital gains. See below,
p. 120, note 12.
6 We must therefore suppose that when banks provide deposits to
satisfy the liquidity preference of rentiers, they hold shares of firms.
118 / Economic Heresies

In the simple case where there is no saving out of wages, the

rate of profit on a steady growth path is given by the formula
7r = g/sp-the level of net profits is such as to provide net sav-
ing per annum equal to net investment. Then, whatever gv may
be, S is equal to it, for the share of profit adjusts in such a way
as to make it SO. 7 Provided that the firms are willing to carry out
investment at such a level as to make g = n, that is, to realize
the natural rate of growth; and provided that it is physically
possible for them to do so-the initial conditions at any moment
are appropriate; and politically possible-the real-wage rate,
governed by technical conditions and the rate of profit given by
7r = g/sp, is not below the tolerable level, then growth at the

natural rate takes place. But even if all the other conditions are
fulfilled, growth at the natural rate will not be realized if firms
lack the energy to carry it out. There is no law of nature that
the "natural" rate of growth should prevail. This marks the dis-
tinction between a Keynesian and a neo-neoclassical growth
model.
We must now consider the relation between the degree of
monopoly and the rate of profit. When there is no saving out of
earned income, the rate of profit is independent of the degree
of monopoly, but the real-wage rate is not. By setting higher
prices (given money-wage rates) the firms can increase the
profit that would be obtained from a given volume of sales to

This is forced upon us by our simplifying assumptions which exclude

bonds and government debt. But equally in a more realistic case, when
the quantity of money is increased to provide a placement for household
saving, the banks must be acquiring assets which the savers do not
fancy.
7 Keynes simplified his model the other way. In the main part of the
argument of the General Theory liquidity preference is presented as a
choice between bonds and money. There are allusions to the yield of
shares as a rate of interest, but they are not fully worked out. However,
in the Treatise the main argument is conducted in terms of equities.
 Growth Models / 119

the public but they cannot ensure that the volume of expenditure
by the public will increase accordingly. Even if additional profits
were paid out to rentiers instantaneously, the firms would get
back only (1 - sp) II P of expenditure from every increase in P.
Thus the rate of profit, 7r = g/ Sp, is independent of the degree
of monopoly. A higher level of prices, however, reduces the real
wage; the wage rate corresponding to a given value of 7r is not
independent of the degree of monopoly. (With a given level of
employment, a higher degree of monopoly entails a lower level
of utilization of plant and a higher value of K/Y. Therefore,
with a given rate of profit it entails a higher share of profit in
income and a lower real wage per man employed.) 8
We must now introduce saving out of earned income. Luigi
Pasinetti proposed a neat model in which the rate of profit is
equal to the rate of growth divided by capitalists' saving even
when there is some saving out of wages. 9 He divided the economy
into two classes, capitalists and worker-rentiers who earn wages
and receive profits on their accumulated savings; and he assumed
that the rate of profit is the same on capital owned by both
classes. 1o We can elaborate on the argument by making some

8 This is the long-period version of Kalecki's famous theory: the

workers spend what they get and the capitalists get what they spend.
When there is no saving out of wages, gross profit, over any period, is
equal to gross investment plus capitalists' consumption; and the share
of wages in proceeds is the inverse of the degree of monopoly. See
Essays in the Theory of Economic Fluctuations (London: Allen & Unwin,
1938), p. 76.
9 See "The Rate of Profit and Income Distribution in Relation to the
Rate of Economic Growth," Review of Economic Studies (October
1962).
10 His critics pointed out that if the share of wages in net income is
sufficiently high it is possible for the proportion of total net saving pro-
vided by the worker-rentiers to exceed that provided by capitalists even
though the propensity to save of the latter is higher. In such a case, the
worker-rentiers are acquiring capital faster than the capitalists so that no
120 / Economic Heresies

further distinctions. Firms are obliged to retain at least enough

of gross profit to keep capital intact and they normally retain a
large proportion of net profit as well. Households may be divided
into three classes: rentiers who derive their whole income from
placements; worker-rentiers with mixed incomes, and workers
whose whole income is derived from wages. (Wages include all
earned income.) 11
We may assume that equities are held mainly by wealthy
rentiers. It is now convenient to include capital gains due to
investment of retained profits in the income of shareholders.12
Taken together, the rentiers have a higher than average pro-
pensity to save, though there may be individuals among them
who are dissipating wealth inherited from the past.
The worker-rentiers are mainly concerned with saving-up to
spend later (through pension contributions, etc.) so that ratio
of net saving to total income for this class is relatively small.
They may be supposed to have a higher liquidity preference than

equilibrium is possible until the capitalists have ceased to own an ap-

preciable proportion of the stock of capital; the whole net income then
accrues to worker-rentiers and their propensity to save governs the share
of saving in income, the s of Harrod's formula. Having got themselves
back on to Harrod's knife-edge, the critics claim that the marginal pro-
ductivity theory of distribution then becomes true. See J. E. Meade and
F. H. Hahn, "The Rate of Profit in a Growing Economy," Economic
Journal (June 1965), J. E. Meade, "The Outcome of a Pasinetti Proc-
ess" (A Note), Economic Journal (March 1966), and P. A. Samuelson
and F. Modigliani, "The Pasinetti Paradox· in Neo-classical and More
General Models," Review of Economic Studies (October 1966).
11 See above, p. 113.
12 This is a matter of accounting conventions. When retentions are
treated as saving, rentier income is treated as consisting only of what
is paid out to them by firms. A sale of securities to finance consumption
is then treated as dis-saving. When capital gains are included in rentier
income, retentions are excluded from saving, and unrealized capital gains
are included in it. The share of saving in profits, Sp, is not affected by
the way the accounts are set up, but it is affected by the actual behavior
of rentiers.
 Growth Models / 121

the wealthy rentiers so that the income on their placements is

much less than the rate of profit. The workers who own no
property provide no net saving. Each class (taken as a whole)
may be supposed to have its own propensity to save; the mixed-
income class applies their propensity to save to income without
distinguishing its source. The share of saving out of wages as a
whole depends upon how much of the total wage bill goes into
these incomes and how much to workers whose propensity to
save is zero. Similarly for profits. On a steady growth path the
distribution of income between classes remains constant. We can
therefore regroup incomes according to their origin and postulate
that Sp, the share of saving in net profit, is considerably greater
than sw, the share of saving in wages.13
Saving out of wages tends to reduce the rate of profit. The for-
g-sw (~)
mula now becomes 71" = . On the other hand, saving
sp
out of wages gives some leverage to allow the degree of monop-
oly to affect the rate of profit. A higher level of prices relatively
to money wages (with a given rate of growth being maintained)
entails lower real wages; consequently less saving out of wages.
Profits must therefore be higher by a sufficient amount to allow
saving out of profits to make up the deficiency. When real wages
are less by - A W, saving out of profits is higher by an amount
equal to Sw A W. Therefore, A P = s'c A W.
sp
As we have already seen, when Sw = sp (Harrod's s), A P =
- A W. When Sw = 0, (71" =
g/sp) , A P =
O. So long as Sw < sp,
A P is less than - A W. The effect of the degree of monopoly

13 In the "anti-Pasinetti" case, where the workers are acquiring a grow-

ing share of total capital, steady growth is not possible; the overall
propensity to save (Harrod's s) is falling as time goes by and the rate of
profit rising. It is hard to understand how this is supposed to provide
support for neoclassical theory. See above, p. 119, note 10.
122 / Economic Heresies

upon the rate of profit is greater the smaller the difference be~
tween Sw and Sp.14
There is one more aspect of household savings that must be
considered. The excess of investment over retentions of profits
and borrowing from banks is borrowed from households. But
firms are under no obligation to borrow just because households
have savings that they want to place in securities. If all invest~
ment were financed by retentions there could be no net saving
out of the incomes paid out to households as wages and divi~
dends. Yet every household is free to save as it pleases.
Kaldor has suggested a mechanism which reconciles this ap~
parent contradiction. 15 In a simplified form his argument is as
follows. Divide all rentiers into old shareholders and new savers.
The income of shareholders consists of dividends and capital
gains. Net investment financed by retentions causes the value of
shares to rise. In tranquil conditions, with a constant rate of
interest and rate of profit, there is a constant valuation ratio, v,
the ratio of the stock~exchange value of the equity of a company,
V, to the value of its earning assets, K. (This must not be con~
fused with Harrod's v, the capital to income ratio). For the
typical company, which is growing at the growth rate per annum
of the economy, g K= 6.K
= 6.V
V .
but a K, the annual net mvest~
ment of the firm is equal to a V, the annual increment of value
of its outstanding shares, only when the valuation ratio is equal
to one. If there are no new issues, 6. V accrues to the existing
shareholders as capital gains. When part of a K is financed by
new issues, capital gains are less than 6. V. The shareholders of
the typical company (whose rate of profit is 71' and rate of growth

14 I am indebted to Dr. Amit Bhaduri for some discussions of this

point.
15 "A Neo-Pasinetti Theorem," Review of Economic Studies (October
1966).
 Growth Models / 123

g) get the benefit of profits some time after they have accrued
to the firm. When P is the net profit of, say, last year, and r is the
retention ratio of that firm, the shareholders receive (l - r) P as
dividends, this year, and rPv as capital gains. As a continuous
income per unit of capital of the firm, they receive (I - r) 7T +
r 7T v.
If there were an excess of positive new household saving com-
ing on!to the market for placements over the supply of new
securities generated by the borrowing of firms, the valuation
ratio would be driven up. A higher valuation ratio means a
higher annual income for shareholders corresponding to a given
rate of profit. Equilibrium is established when the expenditure
for consumption of old shareholders exceeds their receipts of
dividends sufficiently to require a sale of securities (realization
of capital gains) that offsets the excess demand for securities
coming from the new savers. The banking system is assumed to
be generating a sufficient increase in the quantity of money to
offset liquidity preference at the rate of interest at which net
saving out of incomes paid to households, taken as a whole, is
equal to net borrowing by firms. On this view, new issues by
firms tend to keep up the rate of interest and to keep down the
rate of profit.1 6 (This very recent argument cannot strictly be
regarded as an old-fashioned question, but it is necessary to
complete the construction of a Keynesian growth model.)
The purpose of a growth model of this type is not to predict
equilibrium but to map out the possible causes of disturbances.
The assumptions of our model are too simple for testable hy-
potheses to be drawn from it at this stage, but it suggests some
interesting lines of thought.
For instance, we have seen that in some circumstances a

16 Kaldor offers the formula 1"( = g(1 - u)/r where u is the proportion of
investment financed by new issues.
124 / Economic Heresies

higher degree of monopoly may generate a higher rate of profit

with a constant rate of growth. This does not mean that a rise
in profit margins (with a constant rate of investment) necessarily
increases profits. In any given week, the volume of expenditure
for consumption goods depends upon the level of money incomes
of the recent past. A rise of prices this week reduces the volume
of sales and may cause unemployment. But in a modern cap-
italist economy where the government is concerned to maintain
effective demand, the reduction of employment is offset, one way
or another, by additional expenditure. Then the firms, by mutual
consent, can make the rate of profit whatever they like.
Second, the legal system of property which obtains in the capi-
talist world is out of line with economic reality. The capital gains
which accrue to rentiers are the "reward" for no service. 17 More-
over, they introduce an inherently inflationary element into the
economy. Finance which is spent upon investment creates in-
comes which can be spent over again for consumption.
Third, as the mixed-income class grows accustomed to placing
their savings in equities (directly or through institutions such as
investment trusts set up to accommodate them), an increasing
amount of capital gains enter incomes with a relatively high
propensity to consume, so that the overall ratio of saving to in-
come is drifting down; the rate of profit corresponding to a given
rate of growth therefore tends to rise as time goes by.
Of course, in the untranquil world there are many influences
upon the Stock Exchange as a whole and on the valuation of
particular firms as well as those that operate in the calm atmos-
phere of imagined steady growth. They are likely on the whole
to contribute to instability, since expected profits cast a shadow

17 Cf. J. K. Galbraith: "No grant of feudal privilege has ever equalled,

for effortless return, that of the grandparent who bought and endowed
his descendants with a thousand shares of General Motors or General
Electric." The New Industrial State, p. 394.
 Growth Models / 125

before and tend to increase consumption out of profits just when

investment is increasing.
In general, thriftiness plays a different role in this model from
its role in Harrod's system. A high warranted rate of growth, due
to a high value of s, tends to cause slumpy conditions and to
inhibit growth. In our model, a higher propensity to save permits
a higher level of real wages at a given rate of growth. It pushes
back the "inflation barrier" at which real wages reach the toler-
able minimum and so makes it possible for the firms to grow
faster (if they are willing to do so) ; but firms serving the market
for consumer goods do not like saving, which reduces their
profits; they do all they can to keep it at bay with advertize-
ment and innovations that generate psychological obsolescence,
as well as by generating demand by turning unrealized needs
into conscious desires.

INNOVA nONS
Technical progress was not easy to fit into the neoclassical
concept of stationary equilibrium. Marshall's treatment of a
growing economy with a constant normal rate of profit implies
technical progress, and he evidently mixed acquired knowledge
in with the conception of economies of scale (which take place
with a growth of output but are not lost with a decline) 18 but
the whole question was left very vague.

The Meaning 0/ Neutral Inventions Pigou discussed the ef-

fects of "inventions" in terms of a comparison of stationary equi-
librium positions. He divided inventions into those which save
labor, those which save "capital," and a wide neutral band in
between which saves both. Hicks reduced neutrality to a point,

18 Principles-Appendix H.

F
126 / Economic Heresies

putting all improvements on one side or the other, into the labor-
saving and the capital-saving categories. (This terminology,
which is still widely used, is very confusing. All technical im-
provements, except those that merely save time,19 increase out-
put per head at some point in the process of production; in cur-
rent discussions "labor-saving" is sometimes to be taken in the
straightforward sense of reducing labor required for a given out-
put and sometimes in the Hicksian sense of more labor-saving
than neutral.)
The concept of a neutral invention is one that leaves the rela-
tive shares of wages and profits in proceeds the same in the new
equilibrium position as in the old. Here there is an ambiguity;
the technical nature of the change cannot determine the relative
shares by itself, without reference to the capital to labor ratio
in the new position. Hicks proposed, as the definition of neu-
trality, that the invention raises the marginal product of each
"factor" equally when the "capital" per unit of labor is the same
in the new position as the old. In what sense is "capital" to be
taken? Let us suppose that we are comparing two equilibrium
positions with the same labor employed, while in Alpha output
per head is higher than in Beta as a result of superior technical
knowledge, and let us suppose that the share of wages in the
larger net output of Alpha is the same as in the smaller net out-
put of Beta, that is, the relation between the two equilibria is
neutral. Now suppose that the real-wage rate in Alpha is higher
than in Beta in the same proportion as net output per head.
Then, if the rate of profit is the same in the two positions, the
value of capital per man is higher in Alpha in the same pro-

19 For instance, a change of methods might be supposed to produce

two crops a year, where there was formerly one, from the same land
with the same amount of work. If each six-months crop is exactly
half the former yearly crop, there is a saving of finance without any
saving of factors of production.
 Growth Models / 127

portion as net output. If Hicks' constant "capital" per man is

measured in labor time-that is, the value of capital divided by
the real-wage rate-then the return per unit of capital in this
sense has been raised in the same proportion as the wage rate;
this satisfies the Hicks criterion of neutrality. If the argument is
put into terms of the "one-commodity world" and capital is
measured by a physical quantity of the commodity, then, if the
real wage is higher in Alpha in the same proportion as net out-
put per head, and commodity-capital per man is constant, the
rate of profit is higher in the same proportion as the wage rate.
But to have the same "capital" per man at a higher rate of profit
implies a drastic fall in the propensity to save. On the other
hand, if the propensity to save in Alpha is the same as in Beta,
presumably the rate of profit is lower in Alpha (in a Pigovian
stationary state). Then commodity-capital in Alpha must be
higher than in Beta in a greater proportion than output. For rela-
tive shares to be constant, there must be a production function in
Alpha (in terms of labor and the commodity as inputs with the
commodity as output) of unit elasticity of substitution.
Harrod cut through these conundrums by proposing as the
definition of neutrality a situation in which both the rate of profit
and relative shares are unchanged. What was more important,
he departed from the artificial concept of an invention as a
shock moving equilibrium from one position to another; he con-
ceived of technical progress going on continuously by a succes-
sion of innovations. Neutrality implies that innovations are scat-
tered evenly throughout the economy so that output per head
is raised at the same rate at all stages of the process of produc-
tion; if we simplify the economy to two sectors, one producing
commodities and one producing plant, then, when technical
progress is neutral, and the rate of profit is constant, plant per
man, measured in labor time, is constant and the value of capital
per man is rising at the same rate as the real-wage rate.
(We can now describe a bias in technical progress, on either
128 / Economic Heresies

side of neutrality, as capital-using when, if the rate of profit were

constant, the value of capital per man required by new tech-
nology would be rising faster than output per man; and as cap-
ital-saving in the contrary case.)
Harrod postulates a continuous, steady, and neutral rate of
technical progress, given by God and the engineers, which raises
output per head at a steady rate when the accumulation of cap-
ital is going on at a steady rate. The "natural" rate of growth is
compounded of the rate of growth of the labor force and the
rate of growth of output per head. 20 When the natural rate of
growth is being realized, the capital to income ratio, the relative
shares of wages and profits and the rate of profit are all constant
through time.
There is something contradictory in postulating a uniform rate
of profit throughout an economy in which technical progress is
going on. Some firms are always taking advantage of new ideas
faster than others and enjoying a higher rate of profit on their
investments. Moreover, technical progress alters the nature of
commodities and the requirements of skill and training of
workers. However, there does not seem to be much hope of
dealing with such problems until the main lines of a simplified
analysis have been established. We therefore make the drastic
assumptions that commodities and workers retain their physical
characteristics and all technical change is concentrated in the
design of equipment. Then output per head of consumer goods
has an unambiguous meaning; on a steady growth path, the
value of capital per man in terms of consumption goods is rising
at the same rate as output per man and the capital to income
ratio is constant. And we assume that the proportions of high-

20 Sir Roy later introduced the idea of an optimum rate of growth,

which is a much more complicated concept. See "Optimum Investment
for Growth" in Problems of Economic Dynamics: Essays in Honour 0/
Michal Kalecki (Oxford : Pergamon Press, 1966).
 Growth Models / 129

and low-profit investments remain constant through time, so

that there is a constant overall rate of profit on capital. On this
basis (though admittedly it is not very solid) we can apply the
preceding argument in terms of g, n, and iT to a steady growth
path with neutral technical progress.

The Vintage Model When technical progress takes the form

of designing improved equipment which reduces the labor and
raw materials required per unit of output, we must suppose that
each round of gross investment, say per year, goes into the
newest and best equipment, while inferior equipment installed in
the recent past is still in use. 21 The physical stock of capital
equipment in existence at any moment depends upon the length
of service life of plant, for this determines how many types, or
vintages, of equipment, dating from earlier years, are in use at
any moment alongside the latest and best which has just been
installed. The longer the life of plant, the lower the output of
labor equipped with the oldest plant and the lower the average
output per head for the labor force as a whole. To avoid compli-
cations we confine the argument to the case of constant employ-
ment.
What determines the length of service life? To avoid some
intricate points which add nothing of interest to the analysis, we
may postulate that the potential physical lifetime of plant is
longer than the service life, and that each plant has to be worked
in exactly the same way over its life, neither gaining nor losing
efficiency as time goes by. Its output therefore remains the same
over its service life and the quasi-rent that it yields falls as the
real-wage rate rises. (We shall continue to use the convention
of constant prices of consumer goods, so that the rise in the real-

21 See W. E. G. Salter, Productivity and Technical Change (Cam-

bridge, 1960). Salter confines the theoretical part of his argument to
perfect competition, which is an unnecessary· restriction.
130 / Economic Heresies

wage rate comes about through a rise in the money-wage rate.)

On the steady growth path, the wage rate rises at the same rate
as output per head. When a plant is first installed it has a
higher quasi-rent than any older plant. Next year, its output is
the same and its running costs have risen; the cost of materials,
power, etc., bought from other firms is constant (like the prices
of consumption goods) for rising wages offset increases in aver-
age output per head in producing them; the rise in prime costs
for the plant is the rise in its own wage bill. (The interest cost
of working capital has gone up correspondingly.) We may sup-
pose that each plant is used until its quasi-rent has fallen to zero.
It is then scrapped and the labor that was working it finds em-
ployment on the latest, most superior plant that is newly put into
production.
In the simplest form of this argument, it is assumed that there
are no prime costs except wages; then the plant is scrapped
when its whole output is just less than sufficient to pay the wage
bill. 22 In any case, prime costs (on our assumptions) rise each
year with the rise in the wage bill. The time which it takes (at
a given growth rate) to wipe out the quasi-rent of plant of any
one vintage depends upon the ratio of quasi-rent to wages of
the latest and best plant. In short, the length of life of plant is
a decreasing function of the share of wages in the value of net
output. The share of wages depends upon the value of capital
per man of a plant when new and on the rate of profit. To
avoid overburdening the argument we may now return to the
simple formula, 7T = g/sll; and to postpone discussion of a point

22 It has been argued that this concept is incompatible with imperfect

competition (see D. Mario Nuti, "The Degree of Monopoly in the
Kaldor-Mirrlees Growth Model," Review of Economic Studies, April
1969). However, part of wages go to quasi-overhead labor which would
be required to keep the plant going at a minimum level of utilization so
that prices still exceed short-period marginal cost even when quasi-rent
has fallen to zero.
 Growth Models / 131

which we will take up later, we assume for the time being that
the cost of capital per man embodied in a plant when new is
independent of the rate of profit. 23 Then the share of wages
varies only with the rate of profit. The higher the rate of profit,
the greater the difference between the wage rate (at any mo-
ment) and the value of output per man with the latest equip-
ment and therefore the longer the time which it will take to re-
duce its quasi-rent to zero (the growth. rate being given) and the
larger the number of vintages that will be in use at the date when
it is scrapped. The greater the length of life of equipment, the
lower the average output per man employed. At the same time,
the proportion of the labor force that has to be re-equipped at
each round is less when the life of plant is longer, so that the
real resources required for gross investment are less.
By comparing paths exactly alike in all respects except for
differences in the rate of profit, we can trace out another dimen-
sion of the pseudo-production function. In terms of comparisons
of Pigovian stationary states with different rates of profit, we
traced out the vertical dimension of the pseudo-production func-
tion, showing the choice of technique in a "given state of knowl-
edge" and we saw that it may be so badly behaved as to contain
backward-switch points, below which a technique with a higher
output per head is associated with a higher rate of profit.24
Along the horizontal dimension, showing the length of life of
plant (at a given growth rate) the pseudo-production function is

23 There is only one superior technique invented at each round and

labor-value prices of individual commodities and means of production
rule as on Professor Samuelson's "surrogate"pseudo-production function.
(Cr. above, p. 36, note 12.)
24 See above, p. 36. For the time being we are assuming that the cost
of plant when new is independent of the rate of profit. When this dimen-
sion of the pseudo-production function is drawn with the share of wages
in net output and the rate of profit as coordinates, a section of it con-
forming to our assumption is a straight line of which the slope repre-
sents the value of capital per man employed.-
132 / Economic Heresies

quite well behaved. A higher rate of output per head (with

shorter life of equipment) is associated with a lower rate of
profit. In a certain sense, it is associated with a higher ratio of
capital to labor. There is no need to compare the stocks of cap-
ital in terms of value (which is a treacherous concept when the
rate of profit is a variable); the point is that a shorter length
of life entails a larger stock of capital in the sense that a higher
rate of gross investment is required to maintain it. On anyone
path there is a constant proportion of the labor force occupied
with investment and a constant stock of capital in terms of labor
embodied, producing an output of commodities rising at the
growth rate. As we move along the pseudo-production function
from a higher to a lower rate of profit, the proportion of the
labor force in investment rises (for a larger amount of plant is
being replaced every year). Thus the pseudo-production function
has the well-behaved characteristic of showing a higher capital
to labor ratio (in the relevant sense) associated with a lower
rate of profit.
Moreover, we can find here an analogy with the Wicks ell
process of "deepening" the stock of capital. A rise of output per
head would be brought about by an increase in the share of
gross investment in output, leading to a shorter length of life
of plant. We must consider the significance of this for neoclas-
sical theory.

Neoclassical Vintages The model cannot work well in a pre-

Keynesian setting, in which saving governs accumulation. In
some versions,25 it is taken for granted that the ratio of saving
to net income (Harrod's s) is determined by the propensity to
save out of household income. Then, at any moment, there is a

25 See Ferguson, op. cit., Chapters 13 and 14, for a comprehensive ac-
count of the vintage models that have been proposed by neo-neoclassical
writers.
 Growth Models / 133

certain volume of saving per annum and there is a rate of in-

terest at which firms are induced to carry out sufficient gross in-
vestment to produce net investment equal to this amount of
saving. That rate of interest then determines the rate of profit,
which influences the length of life of plant. 26 But this is not pre-
Keynesian, it is bastard Keynesian. For Keynes, the influence
of the rate of interest on the rate of investment (which in any
case is rather a weak point in his system) 27 is produced by its
relation to the expected rate of profit (or "marginal efficiency of ·
capital"). To stimulate investment, the rate of interest (the cost
of borrowing) has to be less than the expected rate of profit by
a sufficient premium for risk, which, on a tranquil, steady growth
path, must be assumed to be negligible. When the rate of profit
is established, the banking system must be supposed to see to it
that the rate of interest does not get out of line (otherwise equi-
librium would be upset), but if there is nothing else in the story
to determine the rate of profit, the rate of interest cannot do it.
Another argument consists in showing that, in a certain very
special sense, the rate of return on investment to society as a
whole is equal to the rate of profit. This is an ingenious use of
the vintage model, which is worth repeating.
Taking all the assumptions of steady growth with perfect
competition, let us postulate that there is a clear-cut division
between the production of plant (which is the only capital good)

26 In Professor Arrow's model, the propensity to save governs the

share of gross investment in value of output. (This is finding an answer
by changing the question.) The real-wage rate at any moment satisfies
the condition of full employment. Then he is home. He set out to ex-
hibit "learning by doing" but in fact he offers an ordinary vintage model
in which technical progress is embodied in the design of equipment. The
only difference is that the rate of progress is an increasing function of
the amount of gross investment. See "The Economic Implications of
Learning by Doing," Review of Economic Studies (June 1962).
27 Cf. above, p. 83.

F*
134 / Economic Heresies

and the production, with the aid of plant, of commodities for

consumption. Workers in the investment sector, with the aid of
the equipment that they operate, maintain the stock of equip-
ment that they need, at the same time as they produce plant for
the consumption-good sector.
Now suppose that ten vintages of plant coexist in the con-
sumption-good sector, each manned by a cohort of 100 teams
of men. One plant employs one team throughout its life. There
are no prime costs except the wage bill. Taking a year as the
gestation period for plant, each vintage is used for ten years. At
the end of that time the real wage has risen to absorb its whole
output and it is scrapped. Now, when plant of vintage VlO is
being constructed, the households, by consuming less than usual,
release resources to have 101, instead of the usual 100, plants
built. Thereafter investment returns to 100 plants a year. To
man the extra plant, a team must be taken from vintage VI
which is entering its last year of life. Next year only 99 teams
are released when the remaining VI plants are scrapped. A team
has to be taken from V 2 to man the hundredth Vn plant, one
from V3 to man the hundredth V 12 plant, and so on until VlO
enters the last year of its life. One team is then transferred to
V 19- At the end of the year the remaining 100 teams are released
and go to V 20 • The normal position is then restored.
Now, the additional output, over and above what would have
been available without the extra V 10 plant, in the first year con-
sists of the output of one V 10 team minus the output of one VI
team. The VI output was scarcely more than the real wage of a
team at the rate then ruling. Thus the additional output this
year is approximately equal to the quasi-rent on a V10 plant in
the first year of its life. Next year the additional output is the
output of a VlO team minus the output of a V 2 team, which is
approximately this year's wage. Over the ten years, it is thus
equal to the series of quasi-rents of a plant, which yields the
normal rate of profit on its initial cost. Thus (assuming that the
 Growth Models / 135

economy was flexible enough to permit one extra plant to be

built without additional cost) the extra consumption is equal to
the rate of profit on the extra investment. 28
This shows that when the economy is growing in equilibrium
with any given rate of profit, the rate of return, in this peculiar
sense, is equal to that rate of profit.
Another favorite argument is to point out that the wage rate
is equal to the marginal product of labor. At any moment, a
small reduction in the labor force would mean that some of the
oldest plant (just about to be scrapped) would cease to be used.
When the wage bill is the only element in prime cost and there
is perfect competition, the reduction in output (the marginal
product) is equal to the reduction in the wage bill. Similarly, a
small increase in the labor force could be accommodated by
postponing the scrapping of some of the oldest plant (but it
would need a rise in the level of effective demand to make it
worthwhile to do so). There is of course no sense in which the
marginal product determines the wage; the age of the oldest
plant, therefore its output, and the wage rate are determined
together by technical conditions and the rate of profit.
Lacking a theory of distribution, the only resort for a neo-
classical vintage model is to treat the economy as though it were
managed by the committee of a kibbutz. 29 (We continue to as-
sume that a constant amount of work is always being done,
though this is not here very reasonable.) If the committee has
decided upon the proportio~ of the labor force to be occupied
in producing plant (taking the conditions of our last example)
and the stock of equipment (in both sectors) has been built up
accordingly, the cooperative will be following exactly the same

28 The above passage is taken (with minor alterations) from a review

of Professor Solow's Capital Theory and the Rate of Return. See Joan
Robinson, Collected Economic Papers (Oxford: Blackwell, 1965), vol. 3.
29 Cf. above, p. 33.
136 / Economic Heresies

path as would be followed by an economy in which the rate of

profit was such as to be associated with the same allocation of
resources between the sectors. As each new plant in the con-
sumption-good sector becomes available, labor to man it is
taken from the least productive plant still in use, thus establish-
ing the same length of life as would occur where incomes con-
sisted of profits and wages, though the distribution of consump-
tion goods among the members of the cooperative may be on
any principles that they find acceptable. (They might use a no-
tional wage rate for accounting purposes, though in the simpli-
fied conditions of our example it would not be necessary--cal-
culations in real terms would be adequate to their require-
ments.)
But what determines the proportion of labor allocated to in-
vestment? Here the neo-neoclassical theorem or golden rule
comes into the argument. 30 Up to a certain point, an increase in
the permanent allocation of labor to the investment sector leads
on to a path (when equilibrium in the length of life of plant has
been established) on which consumption is higher at each phase
of technical progress than it would have been on the former
path. The limit to this process of "deepening" the stock of cap-
ital is reached when the life of plant has been reduced to such
a length that to move one more team of workers over to the in-
vestment sector would add to future output (by shortening the
length of life of plant) no more than would be lost by taking
them away from the consumption-good sector. At this point
there is nothing to be gained from a further rise in the ratio of
the stock of capital measured in labor time to labor currently
employed.
The committee of the cooperative might work this out directly

30 See Joan Robinson, Essays in the Theory of Economic Growth, p.

136 (London: Macmillan, 1963).
 Growth Models / 137

or they might find the notional rate of profit, at any point, cor-
responding to a notional wage rate equal to the output per head of
the oldest plant and compare it to the rate of growth. The rate
of growth is the technological rate of discount or objective "cost
of waiting" for the economy; 31 so long as the notional rate of
profit is greater than this, there is a possibility of gaining con-
sumption in the future (above what it would otherwise be) by
raising the share of investment in output. The committee might
consider it right to aim for the maximum or they might bring a
subjective rate of discount of the future into their calculations
and aim to stop short, with a somewhat longer length of life
of plant (and smaller share of investment) .
When they have decided upon the objective that they propose
to aim at, they have to consider the pace at which they should
go toward it. The most heroic course would be to put all the
labor becoming available into the investment sector as soon as a
certain tolerable minimum level of consumption has been
achieved. This would reach the objective in the shortest possible
time, but it would be extravagant. Investment is accelerating as
men are released (by the rise in output per head) from produc-
ing the constant supply of consumption goods. Part of the labor
in the investment sector has to be occupied with increasing its
own stock of equipment. It might be technically impossible to
tailor the stock of investment-sector equipment so that none of
it became redundant at the moment when investment settled
down to its permanent level. A less drastic course would be to

31 The rate of growth (which is exogenously given by technical prog-

ress) is the rate of return from the point of view of society on a unit of
gross investment embodied in improved technique. The notional profit
includes the surplus of consumption over the notional wage bill. When
the notional rate of profit has been reduced to equality with the rate of
growth and all consumption is included in the notional wage bill there is
no further gain in future consumption to be had by reducing present con-
sumption.
138 / Economic Heresies

work the proportion of labor in the investment sector up to

the level that will be permanently required and then to accumu-
late plant, at the pace which this provides, until a stock with the
required age-composition has been built up, consumption being
allowed to increase meanwhile. Or a longer period of adjustment
might be allowed with a higher level of consumption in the
earlier stages and a smaller acceleration. (The neoclassicals
claim to be able to advise the committee on the ideal program,
taking into account the rate of fall of the marginal utility of con-
sumption goods as consumption per head rises. Or, examining
the program that has been decided upon, they deduce the im-
plicit time-pattern required to justify it.)
Our story of the kibbutz would not throw much light on the
problems facing the authorities in charge of a national invest-
ment plan. We have to set it out only to illustrate the applica-
tion to the problem of accumulation of the neoclassical philoso-
phy of social harmony, separating it from the problems of the
distribution of income between wages and profits.
The maximum stock of capital that the cooperative could aim
to achieve, given by the golden rule, is that which is appropri-
ate to a rate of profit equal to the rate of growth. This rate of
profit obtains when all wages are consumed and all profits saved.
Here is a meeting point between the neo-neoclassics and their
critics, though the former do not usually emphasize the inference
that consumption of unearned income is deleterious to society.32

= = =
32 Here Sp 1, s,. O. Suppose that s. 1 and there is also some
saving out of wages (s .. > 0). Then 'TT' is less than g. Capitalists have
not enough profits to finance investment and they borrow from workers.
The length of life of plant appropriate to this rate of profit is so short
as to reduce consumption at any point on the path below the maximum
because of the excessive amount of labor in the investment sector.
 Growth Models / 139

INDUCED BIAS
We must now open the gate impounding the assumption that
the cost of plant when new is independent of the rate of profit.
When only one superior technique is invented at each round, the
physical specification of plant and the labor embodied in it are
independent of prices; comparing two paths with different rates
of profit, the value of capital per man may be higher or lower,
where the rate of profit is lower. When the two rates of profit
are at separate points on the same pseudo-production function,
the design of plant is also different. When the two points lie on
either side of a forward-switch point, the lower rate of profit is
associated with a higher value of capital. This means that the
share of wages tends to be lower, which tends to make the
length of life of plant longer. A longer life of plant entails a
lower overall value of the stock of capital corresponding to a
given value of plant when new. Thus the more well behaved the
pseudo-production function is in the vertical dimension, the less
so it is in the horizontal dimension. When the two techniques
are divided by a backward-switch point, the lower rate of profit
is associated with a lower value of capital, so that (for a given
rate of growth) the life of plant is all the shorter.
But what is all this about? How can a single pseUdo-production
function be continually re-created as technical progress goes on?
On each path, a succession of new superior techniques are being
invented year by year. How can there be any systematic rela-
tion between separate series, each appropriate to a different
path? One pseudo-production function for a number of Pigovian
stationary states is already an artificial construction; to postulate
a succession of them, all of identical shape, moving at a steady
pace through time, is merely absurd.
Indeed, as soon as we introduce technical progress into the
story, a "natural" rate of growth, exogenously given, is an un-
natural concept. The engineers who design plant are employed,
140 / Economic Heresies

directly or indirectly, by the firms who are going to install it;

prospects of profit influence design. Moreover, actual accumula-
tion does not proceed smoothly. The kind of innovations called
for when there is a high rate of investment going on and labor
is hard to recruit are not the same as those induced by pressure
to cut costs in a depression. And over the long run, history and
geography shape the path that each economy follows. Where
prospects of profit are high, finance is easy to come by; where
labor is scarce, capital-using innovations are favored; where
there is a plentiful reserve of labor in an overpopulated country-
side, capital-saving innovations make rapid development pos-
sible. However capital is measured, the capital to labor ratio
is higher in the United States than in Japan, but they are not two
points on the same production function. 33
There is another aspect of technology which is of far greater
importance than its profitability. After fifty years, Pigou's em-
phasis on the difference between the real cost to society of pro-
ducing saleable goods and the money cost to profit-seeking firms
is beginning to be appreciated. The nature of technology de-
pends very much upon what the public can be induced to put
up with.

33 Cf. above, p. 106.

CONCLUSION

It is easy enough to make models on stated assumptions. The

difficulty is to find the assumptions that are relevant to reality.
The art is to set up a scheme that simplifies the problem so as
to make it manageable without eliminating the essential charac-
ter of the actual situation on which it is intended to throw light.
Keynes found out that the doctrines still orthodox in the inter-
war period were drawn from models which require the assump-
tion that the wage bargain is made in terms of the employer's
product and that the decisions of households to save govern the
rate of investment that firms undertake. These assumptions have
been smuggled back into neo-neoclassical models. All the
pother about the meaning of "capital" has been subsidiary to
this. The special assumptions of a "one-commodity world" are
required for a model in which the real-wage rate tends to the
level that assures full employment. 1 The further assumptions of

1 Cf. above, p. 69. In the one-commodity model, both the pre-

Keynesian assumptions are fulfilled together. In a two-sector model of
the neo-neoclassical type, "capital" has all the same characteristics as in
the one-commodity case, while the consumption good is made of a dif-
ferent physical substance. There is then a price of one in terms of the
other, which varies with the division of net output between the two. The
second pre-Keynesian assumption is brought in to determine this divi-
sion. The assumptions are conscientiously explained by Professor Meade
in A Neoclassical Theory of Growth (London: Allen & Unwin, 1961).

141
142 / Economic Heresies

perfect competition and the instantaneous establishment of equi-

librium are then added in order to be able to demonstrate that
the real wage is equal to the marginal product of labor.
It has often been suggested that this scheme of thought is
the result of ideological bias, but it is unnecessary to raise that
question unless the models can pass the test of consistency and
relevance. In the foregoing essays, a serious objection has been
raised to their logical structure on account of their treatment
of time: they seem to be unable to distinguish between coexisting
differences and sequential change. Even if they could pass the
test of consistency, they would fail on relevance. They are, as
Professor Solow says, "cheap vehicles"; 2 in fact they are too
rickety to stir from the spot where they stand-as soon as any
one of their peculiar assumptions is relaxed the model col-
lapses and we have to start all over again on our own feet.
We can surely agree to start again where Keynes left off.
Who wants to deny that the future is uncertain; that investment
decisions, in a private-enterprise economy, are made by firms
rather than by households; that wage rates are offered in terms
of money, or that prices of manufactures are not formed by the
higgling of a perfectly competitive market?
A model that is intended to be relevant to some actual prob-
lem must take account of the mode of operation of the economy
to which it refers. "Pure theorists" sometimes take a supercilious
attitude to "structuralists" or "institutionalists." They prefer a
theory that is so pure as to be uncontaminated with any material
content. Was Keynes an institutionalist? He took into account
the institutions of a nation-state, of the organization of industry,
the banking system and the Stock Exchange as he saw them.
Since his day, there have been important changes in the setting

2 See "On the Rate of Return: Reply to Pasinetti," Economic Journal

(June 1970).
 Conclusion / 143

in which theory has to operate. Partly as the result of the change

in ideology associated with his name, the governments of cap-
italist industrialized nations play an enormously greater part in
the management of their economies than formerly. Each hopes
to adopt policies that will maintain near-full employment (which
implies a high level of profits) and continuous growth for its
own economy, while avoiding excessive inflation, maintaining a
positive balance of trade on income account and equilibrium in
its balance of payments. The policies that each adopts react
upon the others. The greater internal coherence of national
policies makes international anarchy all the worse. Meanwhile,
the growth of the huge national and international corporations
is establishing independent seats of power which cut across or
manipulate the policies of national governments.
There are signs that the 1970s may prove to be the testing
time for modern capitalism. Is it possible to maintain near-full
employment without undue inflation? Can an international mon-
etary system be devised that will stand up to strains that national
policies put upon it? Even if the crises that are looming up are
overcome and a new run of prosperity lies ahead, deeper prob-
lems will s_till remain. Modern capitalism has no purpose except
to keep the show going. To prevent severe unemployment and
to keep real wages rising secures the adherence of the workers,
growing consumption keeps the public in general complacent,
and opportunities for profit encourage industry to expand.
National economic success is identified with statistical GNP.
No questions are asked about the content of production. The
success of modern capitalism for the last twenty-five years has
been closely bound up with the armaments race and the trade in
weapons (not to mention wars when they are used); it has not
succeeded in overcoming poverty in its own countries, and has
not succeeded in helping (to say the least) to promote develop-
ment in the Third World. Now we are told that it is in the course
of making the planet uninhabitable even in peacetime.
144 / Economic Heresies

It should be the duty of economists to do their best to en-

lighten the public about the economic aspects of these menacing
problems. They are impeded by a theoretical scheme which
(with whatever reservations and exceptions) represents the
capitalist world as a kibbutz operated in a perfectly enlightened
manner to maximize the welfare of all its members.
INDEX

account, unit of, 95-96 profit on, 28, 96; supply price
accounts, national income, 77 of, 13; value of, 31
accumulation, 3-4, 14, 37-38; rate "capital deepening," 35
of, 112; through exploitation, 42 Capitalism, 102, 143; laisser-faire,
amortization, 19; of capital, 10 33,45,50,51; modern, 34
Arrow, K., 133n Capitalism, Socialism and Democ-
auctioneer, 6,7 racy (Schumpeter), 105n
Capital (Marx), 42
bank loans, interest on, 28 Capital Theory and the Rate of
Bank of England, 81 Return (Solow), 32n, 135n
Baumol, W., 107n capital to output ratio, 110
Berrill, K., 111n capital value, 9
Bhaduri, Amit, 122n cash,73
bias, induced, 139-140 change, 53
boom, 22, 23,81, 83,94 Chicago school, 75, 85-88
borrowing, cost of, 28 Clark, J. B., 32
Business Behavior, Value and Clower, R. W., 65n
Growth (Baumol), 107n Collected Economic Papers (Rob-
buyer's market, 19-23 inson), 135n
competition, 97, 98, 100, 102, 103;
imperfect, 18, 20, 62, 97, 98-
Cambridge equation, 78, 85 102, 103; perfect, 98-102, 142
capacity, productive, 14, 16, 17; conglomerates, 102
changes in, 17, 22, 23, 56; in- constant returns to scale, 52, 53
crease in, 54; reducing, 22 cooperatives, 34, 135-138
capital, 89; accumulation of, 15; corn economy, Ricardo's, 39, 42,
amortization of, 10; marginal 50, 57, 67,69n, 70-71
efficiency of, 30; marginal pro- "cost of waiting," 15
ductivity of, 15, 32, 34, 37; "cost-push," 92
quantity of, 13, 37; rate of costs: escapable, 17; fixed, 17;

145
146 / Index

costs (cont'd) equilibrium price, 65

functional relationship between escapable costs, 17
output and, 53; labor, 11; mar- Eshag, E., 28n
ginal, 58, 97, 99, 100; oppor- Essays in the Theory of Economic
tunity, 6; overhead, 21; prime, Fluctuations (Kalecki), 105n,
20, 21, 99; prime average, 19; 119n
sunk, 17; total, 19; variable, 17 Essays in the Theory of Economic
Growth (Robinson), 136n
demand: deficiency of effective, Euler's theorem, 68n
91; effective, 23-24, 49-51; exchange rate, depreciating the, 91
equilibrium between supply and, expectations, 22-23, 95, 104; long-
4, 7-8; influence upon supply, term, 32
16; level of, 19 exploitation, rate of, 42-44
"demand-pull," 92 "external diseconomies," 55
depression, 23,31; cause of, 30 external economies, 58, 59
developing countries, 15
diminishing returns, 52-63 factor prices, 35; relation of "mar-
discount, rate of, 28 ginal products" to, 55
factors of production, 6, 7-8; sub-
earned income, 46 stitutability between, 7
Economic Consequences of the Ferguson, C. E., 33n, 132n
Peace (Keynes), 46n financial institutions, function of,
economic growth, 109-140 48
economic theory, 100 financial power, 102
Economics of Imperfect Competi- firm: optimum size of, 58, 97-98;
tion,99n theory of the, 97-108
Economics of Welfare (Pigou), firms, growth of, 101
60n fixed costs, 17
economies of scale, 58-63, 102, Friedman, Milton, 75n, 86
125 From Marshall to Keynes (Eshag),
economy: corn (Ricardo's), 39, 28n
42, 50, 57, 67, 69n, 70-71; non- full employment, 14, 15, 24, 34,
monetary, 5-6; one-commodity, 61,81,89,91,94, 112, 141
67-71; "real," 90; stationary, 3-
15, 73-74 Galbraith, J. K., 105n, 124n
effective demand, 23-24, 49-51; General Motors, 103
deficiency of, 91 General Theory (Keynes), 5n, 23-
Elements of Pure Economics 24, 45, 77n, 79, 80-81, 82, 83n,
(Walras), 26n 85, 90, 91, 94, 108n, 109, 118n
employment: full, 14, 15, 24, 34, golden rule, 136, 138
61, 81, 89, 91, 94, 112, 141; gold standard, 65, 81
theory of, 88-90 Gordon, M., 107n
equilibrium: between supply and government policy, 24
demand, 4, 7-8; long-run, 18 Gregory, T. E., 90n
 Index / 147

gross national product, 82 Introduction to Monetary Theory

gross profits, 31; retention of, 117, (Clower), 65n
120 inventions, neutral, 125-129
growth: natural rate of, 109, 112- investment, 23, 28-29, 90, 104;
113, 118, 128, 139; of firms, net, 42, 47; productivity of, 33-
101; rate of, 42, 44; "warranted" 34; savings and, 44-49
rate of, 110, 111, 125 Investment, Financing and Valua-
growth models, 109-140 tion of the Corporation, The
(Gordon), 107n
Hahn, F. H., 120n investment plans, 83, 87, 104-106
Harrod, R. F., 12, 24, 74, 109-116, irreversibility, 53-58
120n, 121n, 122, 125, 127-128 15 curve, 82-85
Harrod's knife-edge, 110, 111, 116,
117, 120n Jaffe, W., 26n
Hicks, J. R., 26n, 82, 88, 98, 100,
125-127
hire-price, 6 Kahn, R. F., 61n
household savings, 117-123 Kaldor, N., 105n, 122, 123n
HolV to Pay for the War (Keynes), Kalecki, M., 47, 83, 94, 105n,
24n, 77n 119n
Keynes, John Maynard, 5n, 14,
hyper-inflation. 93
23-24, 30-32, 45, 49, 50, 72,
imperfect competition, 18, 20, 62, 73n, 75, 77, 79-85, 87, 88, 89,
97,98-102, 103 90-92, 98, 108, 118n, 133, 141,
income: distribution of, 93; earned, 142
46; net national, 82; rentier, 14, Kingdom Come, 4, 9-13
49; social, 25; unearned, 46 kibbutz, 33, 135-138, 144
incomes policy, 92-94
increasing returns, 52-63 labor: availability of, 17, 18; cost
induced bias, 139-140 of, 11; excess demand for, 92;
"infant industry case," 55 marginal product of, 70
inflation, 90-95; analysis of, 91; labor force, effective, 34,74
continuous, 93, 95 "labor-saving," 126
input-output table, 10-11 labor-value prices, 36
innovations, 125-140 laisser-faire capitalism, 33, 45, 50,
instability, 124 51
interest, 8-9; definition of, 28; fall "laws of returns," 59
in rates, 31; level of rates, 31; Lectures (Wicksell), 59, 68n
long-term rates, 27-28, 29; Leijonhufvud, Axel, 85
money rate, 28; natural rate of, liquidity preference, 118n
28; profit and, 25-51; rates of, liquidity trap, 85
9, lIn, 27, 30, 81, 83; short- LM curve, 82-84
term rates, 28; theory of money Loftus, P. J., 43n
and, 79-82 Luxemburg, Rosa, 50
148 / Index

machines, hire-price of, 26-27 money-wage rates: level of, 90;

macro-theory, 107-108 lowering of, 91
"Managerial" Capitalism, The Eco- monopoly, 97, 102-103; degree of,
nomic Theory of (Marris), 107n 90, 114, 116, 124
marginal cost, 58, 97, 99, 100 Morishima, M., 27n
marginal net product, 12
marginal physical product, 12 national income accounts, 77
marginal productivity, 58 natural rate of growth, 128
marginal product of capital, 15, 32, neoclassical school, 3, 25
34,37 Neoclassical Theory of Growth, A
marginal product of labor, 70 (Meade), 141n
marginal products, 55-58 Neoclassical Theory of Production
market: buyer's, 19-23; prison- and Distribution, The (Fergu-
camp, 4-6; seller's 19-23; Wal- son),33n
ras', 4 neoclassical vintages, 132-138
market clearing prices, 16 neo-neoclassical theorem, 44n, 136
market prices, 65-67 neo-neoclassics, 32-38
Marris, R., 107n net investment, 42, 47
Marshall, A., 9, 12, 13-14, 16, 17, net national income, 82
18, 19, 21, 22, 23, 24, 27-30, net profit, 17, · 19, 40, 42; source
32, 48, 49, 50, 54, 55, 58, 59, of, 46
61, 62, 63, 67, 72, 73, 88-89, Neumann,. von, 40~2, 44, 72n,
102, 109, 125 109
Marshallian model, 13-14 neutral inventions, 125-129
"Marshall's dilemma," 97, 102 New Industrial State, The (Gal-
Marx, Karl, 3, 25, . 42-43, 43n, braith), 105n
50 non-monetary economy, 5-6
Meade, J. E., 120n, 141n non-monetary models, 64-76
medium of exchange, 65 Nuti, Mario, DOn
micro-equilibrium, 12
micro-theory, 107-108 oligopoly, 102
Mill, J. S., 88n one-commodity economy, 67-71
Mirrlees, J. A., 105n On Keynesian Economics and
modeI(s): growth, 109-140; Mar- the Economics of Keynes (Lei-
shallian, 13-14; non-monetary, jonhufvud), 85
64-76; Pigovian, 9-13; Walras- opportunity cost, 6
ian, 4-9 Optimum Quantity of Money, The
Mojigliani, F., 120n (Friedman),87n
monetary policy, control of, 87 optimum rate of growth, 128n
money, 64, 65, 76, 81, 87, 89; in out-of-equilibrium situation, 19
a golden age, 74-75; prices and, output: capacity, 19; functional re-
77-'Y6; purchasing power of, 95- lationship between costs and, 53;
96; quantity of, 79, 87; real increase in, 54; level of, 19; per
forces and, 71-74; theory of in- man-hour, 53
terest and, 79-82 overhead costs, 21
 Index / 149

Pasinetti, L. L., 37n, 69n, 119, tion of, 46; gross, 31; interest
120n and, 25-51; low, 30; net, 17, 19,
Patinkin, Don, 87n 40, 42; normal rate of, 30, 48;
Penrose, Edith, 107n rate of, 13, 14, 27, 29, 31, 47-
perfect competition, 98-102, 142 48, 50, 124; rate of net, 31;
Pigou, Arthur, 9, 55, 58-59, 60n, reinvestment of, 101-102; re-
61, 72, 73n, 97, 101, 102n, tention of gross, 117, 120; sav-
125, 140 ing and, 117-125; source of net,
Pigovian model, 9-13 31; uniform rate of, 113
placements, yield of, 28 progress, 13
power, financial, 102 property: legal system of, 124;
price(s): equilibrium, 65; factor, rentier, 114
35; labor-value, 36; level of, 31; pseudo-production function, 34-38,
market, 65-67; market clearing, 44n
16; money and, 77-96; relation purchasing power, 4, 5, 6, 65, 95-
of "marginal products" to, 55; 96
supply, 18-21, 62 Pure Theory of Domestic Values
price determination, 4 (Marshall),89n
price leadership, 20-21, 116
price level, 10, 90, 94 quantity of capital, 13, 37
prime costs, 20, 21, 99; average, 19 quantity theory, 77, 79, 85, 88, 94
Principles (Marshall), 9, 12n, 14,
16n, 19n, 50n, 55n, 125n Radford, R. A., 4n, 59n
Principles (Ricardo), 39n, 40n "real forces," money and, 71-74
prison-camp market, 4-6 real-wage rate, 13
Problems of Economic Dynamics: real wages, 11-13
Essays in Honour of Michal rent, 6
Kalecki, 128n "reswitching" controversy, 35, 116
production: factors of, 6, 7-8; returns: constant, 52, 53; diminish-
factors of substitutability be- ing, 52-63; increasing, 52-63
tween, 7; means of, 3; organized, "reward of waiting," 9, 13
10; surrogate function, 36n Ricardo, David, 3, 8, 36, 39, 42,
Production of Commodities by 49, 52, 57, 58, 67-68, 109
Means of Commodities (Srafi'a), Robertson, Dennis, 22, 88
39n Robinson, E. A. G., lOin
productive capacity, 14, 16, 17; Robinson, Joan, 1l0n, 135n, 136n
changes in, 17, 22, 23, 56; in-
crease in, 54; reducing, 22 Salter, W. E. G., 129n
Productivity and Technical Change: Samuelson, P. A., 35n, 36n, 12On,
(Salter), 129n 131n
Productivity of investment, 33-34 savings, 14, 15, 26, 46, 48, 89,
profit margins, level of, 90 112-115; household, 117-123;
profit(s), 13, 25, definition of, 28; investment and, 44-49; profits
diffusion of, 29; expectation of, and, 117-125
80; ex-post realized, 21; func- Say's Law, 50, 88
150 / Index

Schumpeter, Joseph, 105n trade unions, 85, 92, 94

seller's market, 19-23 tradition, 3
service life, determination of, 129 "turnpike" policy, 41n
Shackle, Go L. So, 58n
share of net saving in net income, Uhr, Co Go, 71n
110, 114 underconsumption, 50
short-period situation, 16-24, 62 unearned income, 46
Simons, Henry Co, 85, 86n, 88 unemployment, 34, 85, 91, 92, 94,
slumps, 14, 50, 81, 85, 90, 94, 95, 124
99 unit of account, 95-96
Smith, Adam, 3, 52 utilization, changes in, 17
Solow, Ro Mo, 32n, 33n, 104n,
136n, 142 Valuation of the Corporation (Gor-
speculation, 30 don), 107n
Sraffa, Piero, 39n, 40n, 59, 72n valuation ratio, 122-123
stationary states, 3-15,73-74 "value added," 95
Stock Exchange, 48, 124 Value and Capital (Hicks), 98,
stop-go cycle, 94 lOOn
strikes, 93 value, capital, 9
Structure of Competitive Industry, variable costs, 17
The (Robinson), lOIn vintage model, t"29-132
substitutability: between factors of vintages, neoclassical, 132-138
production, 7
sunk costs, 17 wages, real, 11-13
supply: equilibrium between de- Walras, 4, 24-25, 58, 65
mand and, 4, 7-8; influence of warranted rate of growth, 110,
demand upon, 16 111,125
supply price, 18-21 Wealth and Welfare (Pigou), 60n
"surrogate production function," wealth, transfer of, 60
36n "well-behaved production func-
Swan, To Wo, 69n, 111n tion," 36
'Wicksell effect,' 71 n
Wicksell, Knut, 14-15, 28, 37, 59n,
technical progress, 125-140 68n, 71,71n, 72,73
technique, choice of, 103-107 Wicks ell process, 14-15, 35, 37,
technology, 38, 55 70
"theory of the firm," 59 worker-rentiers, 119-120
Theory of the Growth of the Firm,
The (Penrose), 107n Years a/High Theory, The (Shackle),
thriftiness, 125 58n
Towards a Dynamic Economy yields, 31
(Harrod), UOn Young, Allyn, 60n, 62n
trade, 4-5
trade-cycle theory, 23, 94, 110-111 zero price, 4, 5