Capital and Its Structure

8/7/2019 Capital and Its Structure

1/138

CAPITALAND ITS STRUCTURE


2/138

CAPITALAND ITS STRUCTURE

BYLUDWIG M. LACHMANN

SHEED ANDREW S A ND McMEEL, INC.SUBSIDIARY OF UNIVERSAL PRESS SYNDICATEKANSAS CITY


3/138

ToM.L.This edition is published in cooperation with the programs ofthe Institute for Humane Studies, Inc., Menlo Park, California;and Cato Institute, San Francisco, California.Capital and Its Structure Copyright 1978 by the Institute forHumane Studies.

Jacket and cover design copyright 197ft by Cato Institute, SanFrancisco.The original edition of Capital and Its Structure was publishedby Bell & Sons, Ltd., on behalf of the London School of Economicsand Political Science in 1956.All rights reserved. Printed in the United States of America.No part of this book may be used or reproduced in any mannerwhatsoever without written permission except in the case ofreprints in the context of reviews. For information write SheedAndrews and McMeel, Inc., 6700 Squibb Road, Mission, Kansas66202.

Library of Congress Cataloging in Publication DataLachmann, Ludwig M.Capital and its structure.

(Studies in economic theory)Includes bibliographical references and index.1. Capital. I. Title. II. Series.HB501.L223 1978 332'.O4l 77-82807ISBN 0-8362-0740-8ISBN 0-8362-0741-6 pbk.Lachmann, Ludwig M


4/138

CONT E NT SCHAP. pagePREFACE TO THE SECOND EDITION viiPREFACE xiii

I THE ORDER OF C A P I T A L 1II ON EXPECTATIONS 20

III PROCESS ANALYSIS AND CAPITAL THEORY 35IV THE MEANING OF CAPITAL STRUCTURE 53V CAPITAL STRUCTURE AND ECONOMIC PROGRESS 72

VI CAPITAL STRUCTURE ANDASSET STRUCTURE 86VII C A P I T A L IN THE T R A D E C Y C L E 100

INDEX 129


5/138

PREFACEFor a long time now the theory of capital has been under acloud. Tw enty years ago , wh en Professor Knight launchedhis attack on the capital theories of Boehm-Bawerk andWicksell, there opened a controversy which continued for years

on both sides of the Atlantic. Tod ay very little is heard of allthis. The centre of interest has shifted to other fields.In practice of course problems concerning capital have byno means lost their interest. There can be few economists whodo not use the word 'capital' almost every day of their workinglives. But apart from some notable exceptions, economistshave ceased to ask fundamental questions about capital. It ispertinent to enquire why this has happened. It would seemthat there are three major reasons to account for this curiousneglect.In the first place, many economists have evidently come tobelieve that we do not require the conceptual framework of atheory of capital in order to discuss problems germane tocapital, or at least those problems in which practical interesthas of late been greatest, such as investment. In other words,the view appears to have gained ground that a theory of capitalis not really necessary. Th is, as I shall attempt to show inthis book, is an erroneous view . It is hardly possible to discussthe causes and consequences of a change in a stock without someknowledge of the nature and composition of this stock; or, it isonly possible to do so if we are prepared to abstract from allthose features of the situation which really matter. In thediscussion of capital problems, as of any other problems, wecannot dispense with a coherent frame of reference.A second reason for the present-day neglect of the theoryof capital has probably to be sought in the contemporarypreoccupation with quantitative precision of statement andargum ent. Most contemporary economics is presented in aquantitative garb. This is not the place to enquire into thereasons for this predilection. To some extent of course econ-omists, in spending so much effort on quantifying the terms inxiii


6/138

PREFACEwhich they present their theories, wittingly or unwittinglymerely reflect the spirit of our age.But why should such quantification be more problematicalin the theory of capital than it is in other fields of econ-om ic study? In m ost business transactions capital is treatedas a quantity. In every balance sheet we find a cap italaccount.The fact remains, however, that in spite of protracted effortsit has proved impossible to find a quantitative expression forcapital which would satisfy the rigorous requirements of econ-omic thought. Most economists agree today that, except underequilibrium conditions, a 'quantity of capital' is not a mean-ingful concept. In this book an attempt is made to follow upsome implications of this conclusion. But the fact that theconcept of capital has for so long proved refractory to allattempts at quantification is almost certainly one of the reasonsfor the lack of interest, and hence of progress, in the theory ofcapital.A third reason, closely related to the one just mentioned,appears to lie in the rather peculiar nature of the relationshipbetween capital and knowledge. Th e various uses made ofany durable capital good reflect the accumulated experienceand knowledge gained, in workshop and market, by those whooperate it. But modern economic theory cannot easily copewith change that is not quantitative change; and knowledgeis as refractory to quantification as capital is. Th e acquisitionand diffusion of knowledge certainly take place in time, butneither is, in any meaningful sense of the word, a 'function' oftime . Modern economists, uneasily aware of the problem,have tried to avoid it by assuming a 'given state of knowledge'.But such an assumption, if taken literally, would obviously pre-vent us from considering economic change of any kind. Forinstance, as Mrs. Robinson has pointed out, 'a "change inmethods of production in a given state of knowledge" is, strictlyspeaking, a contradiction in terms'. W ith very durable capitalgoods the assumption becomes quite untenable. Our railwaysafter all are not run by people w ith the technical knowledge of125 years ago.The theory of capital is a dynamic theory, not merely becausemany cap ital goods are durable, but because the changes in use

x iv


7/138

PREFACEwhich these durable capital goods undergo during their life-time reflect the acquisition and transmission of knowledge.

Our own approach in this book follows another trend ofmodern economic thought, not towards the 'objective' andquantifiable, but towards the subjective interpretation ofph eno m ena . Of late ma ny economists have exercised theiringenuity in fashioning their science in accordance with therigid canon s of Logical Em piricism. Even the theo ry of valuehas been m ade to conform to the strict rules of the b ehaviou rists:nowadays we are not supposed to know anything abo ut hu m anpreferences un til these hav e been 'reve aled ' to us. Bu t few ofthese efforts have been successful. T he fact rem ains th at th etwo greatest achievements of our science within the last hun-dred years, subjective value and the introduction of expecta-tions, became possible only when it was realized that the causesof certain phenomena do not lie in the Tacts of the situation'but in the appraisal of such a situation by active minds.The generic concept of capital without which economistscannot do their work has no measurable counterpart amongm ate rial objects; it reflects the e ntrepren euria l ap praisa l of suchobjects. Beer barre ls an d blast furnaces, ha rbou r installationsand hotel-room furniture are capital not by virtue of theirphysical properties but by virtue of their economic functions.Something is cap ital because the m arket, the consensus of entre -preneurial minds, regards it as capable of yielding an income.This does not mean that the phenomena of capital cannot becomprehended by clear an d unam biguous concepts. T hestock of capital used by society does not present a picture ofchaos. Its arrange me nt is not arbitrary . Th ere is some orderin it. Th is book is devoted to the exploration of the problemsof the order of capital.The chief object of this book is thus to rekindle interest inthe fundamental problems of capital rather than to present aclosed system of generalizations about them; to outline a newapproach and to show that i t can be applied, with some pro-mise of success, to a n um be r of such problem s rang ing from th eproductivity of capital to the demise of the 'strong boom'; topoint out the implications of certain economic facts whichhave been long neglected; and, above all, to emphasize the

xv


8/138

PREFACEtransmission of knowledge, the interaction of minds, as theultimate agent of all economic processes.I am painfully aware of the fact that this book leaves manyvital questions unanswered. It could hardly be otherwise.But it is my hope that others will follow and make their con-tributions to the theory of capital. There can be few fieldsof economic enquiry today which promise a richer harvest thanthe systematic study of the modes of use of our materialresources.It is not impossible that at some time in the future the con-cept of capital structure, the order in which the various capitalresources are arranged in the economic system, w ill be given aquantitative expression; after all, any order can be expressedin numbers. For many reasons such a development wou ld bemost welcome. But this book has been written to meet thepresent situation in which we badly need a generic concept ofcapital, but in which all attempts to express it in quantitativeterms have thus far been unsuccessful.

My greatest debt of gratitude is to Professor F. A. Hayekwhose ideas on capital have helped to shape my own thoughtmore than those of any other thinker. T o Professor F. W .Paish who, during his stay at this University in 1952 as a V isit-ing Trust Fund Lecturer, undaunted by a heavily loaded time-table, read several chapters in draft form, I am indebted formuch sagacious comm ent and advice. In writing the finalversion of Chapter VI I have drawn heavily on his unrivalledknowledge of the intricacies of modern business finance. Butneedless to say, the responsibility for what I say is entirely mine.I owe more than I can express in words to m y friends in theUniversity of the Witwatersrand for their steady help andencouragement, in particular to Mr. L. H. Samuels and Mr.T . van Waasdijk, who patiently read draft after draft, and fromwhose helpful comment and suggestions I have derived muchunearned profit.I also wish to express my gratitude to the Research Com-mittee of this University who by their generous financialassistance have considerably eased my task.Lastly, I have to thank the Royal Economic Society, theeditor and publishers of the Manchester School of Economic and

xvi


9/138

PREFACESocial Studies, Messrs. George A llen & Un win, the McGraw-HillBook Com pany, In c., and Messrs. Routledge & Kegan Paul forpermission to quote passages from works published by them.I also wish to acknowledge my gratitude to the authors ofthese passages.

L. M. LACHMANNUniversity of the WitwatersrandJohannesburgSeptember,

xvii


10/138

CHAPTER ITHE ORDER OF CAPITAL

The realm of economics consists of many provinces betweenwhich, in the course of time, a fairly high degree of inter-regional division of lab ou r has evolved. N atu rally , develop-ment in some of these regions has been faster than in others.There are some 'backward areas', and a few of them actuallyap pe ar to m erit description as 'distressed are as '. No ne seemsto have a better claim to this unenv iable status than th e T heo ryof C apital. In fact it would hardly be an exaggeration to saythat at the present time a systematic theory of capital scarcelyexists.Considering the degree of division of labour just mentionedthis surely is an astonishing state of affairs. T he re can h ard lybe a field of economic thought, pure or applied, in which thewo rd 'cap ital ' is no t m ore or less constantly em ployed. W ehe ar of a world-w ide cap ital sho rtage. I n discussions on theconvertibility of currencies we are asked to distinguish between'cu rren t' an d capital transactions. A nd it is clear th at the'economic integration of Western Europe' requires that someat least of the industrial resources of these countries be re-grouped and change their form; in other words, that it entailsa modification of Europe's capital structure.Yet, in the Theory of Capital the present state of affairs is aswe have described it. T he produ ct imp orted and used by theother economic disciplines is not a standardize d produ ct. T h eword 'capital ' , as used by economists, has no clear and unam-biguous m ean ing. Sometimes the wo rd denotes the mate rialresources of produ ction, sometimes their money value. Some-times it means money sums available for loan or the purchaseof assets. W hile to some economists 'capit al' has come tomean nothing but the present value of future income streams.The conclusion suggests itself that no progress made in thetheory of capital could fail to pay handsome dividends in theform of 'external economies' to be reaped by all those whohave to work with the notion of capital.1


11/138

2 CAPITAL AND ITS STRUCTUREThe root of the trouble is well known: capital resources areheterogeneous. Capital, as distinct from labour and land, lacksa 'natural' unit of measurement. W hile we may add headto head (even woman's head to man's head) and acre to acre(possibly weighted by an index of fertility) we cannot add beerbarrels to b last furnaces nor trucks to yards of telephone wire.Yet, the economist cannot do his work properly without ageneric concept of capital. Where he has to deal with quanti-tative change he needs a comm on denominator. Almostinevitably he follows the business man in adopting moneyvalue as his standard of measurement of capital change. Th ismeans that whenever relative money values change, we loseour common denominator.In equilibrium, where, by definition, all values are consistentwith each other, the use of money value as a unit of measure-ment is not necessarily an illegitimate procedure.1 But indisequilibrium where no such consistency exists, it cannot beapplied. Th e dilemm a has been known ever since Wickselldrew attention to it.2 But in most current discussions oncapital the whole problem with its manifold implications, wh ichgo far beyond the confines of the theory of capital, is, as a rule,allowed to be ignored.In confronting this dilemma it seems best to start by settingforth a few fundamental facts about capital.All capital resources are heterogeneous. The heterogeneitywhich matters is here, of course, not physical heterogeneity,but heterogeneity in use. Even if, at some future date, somemiraculous substance were invented, a very light metal perhaps,which it was found profitable to substitute for all steel, wood,copper, etc., so that a ll capital equipm ent were to be made fromit , this would in no way affect our problem. Th e real economicsignificance of the heterogeneity of capital lies in the fact thateach capital good can only be used for a limited number o f pur-poses. We shall speak of the multiple specificity of capital goods.

1 'But although, even in the analysis of a stationary equilibrium, the inclusionof the "quantity of capital" among the determinants of that equilibrium meansthat something which is the result of the equilibrating process, is treated asif it were a d atum , this confusion was m ade relatively innocuous by the essentiallimitations of the static method, which while it describes the conditions of astate of equilibrium, does not explain how such a state is brought ab out.' F . A.von Hayek : Profits, Interest and Investment, p p . 8 3 - 4 .2 Lectures on Political Economy, Vol . I , p . 202.


12/138

THE ORDER OF CAPITAL 3Each capital good is, at every moment, devoted to what inthe circumstances appears to its owner to be its 'best', i.e. itsmost profitable use. T h e wo rd 'best5 indicates a position ona scale of alternative possibilities. Ch ang ing circum stanceswill change th at position. Un expected change m ay open upnew possibilities of use, and make possible a switch from yester-day 's 'best ' to an even better use. O r, it m ay compel a switchfrom 'present best' to 'second best' use. H enc e, we can no t besurprised to find that at each moment some durable capitalgoods are not being used for the purposes for which they wereoriginally designed. These new uses m ay, from the point ofview of the owners of the capital goods, be 'better' or 'worse',m ore or less profitable th an th e original ones. In e ach casethe change in use means that the original plan in which thecapital good was m ean t to play its pa rt has gone astray. Inmost of the arguments about capital encountered today thesefacts and their implications, many of them crucial to a clearunderstanding of the nature of economic progress, are almostcompletely ignored.It is hard to imagine any capital resource which by itself,operated by human labour but without the use of other capitalresources, could tu rn ou t any ou tpu t at all. Fo r most purposescapital goods have to be used jointly. Complementarity is of theessence of cap ital use. But the heterogeneous cap ital resourcesdo not lend themselves to combination in any arbitrary fashion.For any given number of them only certain modes of comple-mentarity are technically possible, and only a few of these areeconom ically significant. It is am ong the latter tha t the en tre-pren eur has to find the 'optimu m com bination '. Th e 'best*m ode of com plemen tarity is thus not a 'da tu m '. It is in noway 'given' to the entrepreneur who, on the contrary, as a rulehas to spend a good deal of time and effort in finding out whatit is. Ev en whe re he succeeds quickly he will no t enjoy hisachievement for long, as sooner or later circumstances willbegin to change again.

Unexpected change, whenever it occurs, will make possible,or comp el, chang es in the use of cap ital goods. I t will thu scause the disintegration of existing capital combinations.Even w here it opens up new an d prom ising possibilities for someresources it will open them u p for some, no t for all. T he rest


13/138


14/138

THE ORDER OF CAPITAL 5capital' , conceived of it as homogeneous and measurable, likeany other stock. Fro m Ricard o onw ards, of course, their ma ininterest was in the distribution of incomes. Ca pital was ofimportance mainly as the 'source' of profit in the same way aslabour was the source of wages and land the source of rent.And as the rate of profit was regarded as tending towardsequality in all uses of capital, the problem was really posedonly un de r equilibrium conditions. Fo r only here is therea determinate and homogeneous quantity which we may call'output' , and as profit is conceived as part of homogeneousou tpu t, so its 'source' would a pp ear to be equally h omog eneous.The notion of a homogeneous capital stock no doubt wasborrowed, as are so many of our concepts, from accountingpractice which makes a (homogeneous) money sum appear inthe capita l account. Now, the classical economists were cer-tainly not unaware of some of the dangers of describing econ-omic mag nitudes in money terms. But in this case the da nge rseemed to be avoided, and the notion of a 'real stock' gained acertain plausibility from the employment of two devices whichplayed a prom inen t pa rt in the classical doc trine. O n the onehand, the labour theory of value made it possible and neces-sary to reduce capital values to labour values, i.e. to homo-geneous labour units. O n the other han d, there was theconcept of the Wages Fund which not merely served to renderthe stock of capital homogeneous but also reduced all capitalgoods to consumption goods measurable in labour units.In the neo-classical schools which rose in the last threedecades of the last century the focus of interest is still dis-tribution, explained now by the marginal productivity prin-ciple. H ere factors of production of various classes betw eenwhich there is no longer over-all homogeneity, though themembers of each class are still regarded as homogeneous,produce a homogeneous product. The Wages Fund isabandoned, but the new problem which emerges now thatcapital goods can no longer be regarded as consumption goodsin statu nascendi, viz. the effect of capital change on capitalvalues in terms of consumable o utp ut, is ignored.T h e analysis is still couched in equilibrium terms. Butcapital is no longer measured in labour, or any other cost,units, and no one ever makes it clear how capital is to be


15/138

6 CAPITAL AND ITS STRUCTUREmeasured'.8 Reasoning based on the marginal productivityprinciple can no longer be applied to capital even wherethe change is, within the framework of 'comparative statics',from one equilibrium to another. A fortiori it is impossibleto speak of capital in quantitative terms in conditions ofdisequilibrium.More recently the focus of interest in discussions on capitalhas shifted to Investment, defined as the 'net addition to thecapital stock'. Now , it is possible to define the economicforces engendering investment in terms which avoid thequantification of capital, or even the very concept of capital.Professor Lerner, for instance, has defined his 'marginalefficiency of investment' exclusively in terms of present outputforgone and future output obtained.4 But in any realisticdiscussion of the 'inducement to invest' it is clearly impossibleto ignore existing capital resources, on the use and profitabilityof which the new capital cannot but have some effect.A theory of investment based on the assumption of a homo-geneous and quantifiable capital stock is bound to ignoreimportant features of reality. Owing to its very character itcan on ly deal with quantitative capital change, investment anddisinvestment. It cannot deal with changes in the composition ofthe stock. Yet there can be little doubt that such changes inthe composition of the stock are of fundamental importance inmany respects, but in particular with regard to the causes andeffects of investment. As long as we cling to the view that allcapital is homogeneous, we shall only see, as Keynes did, theunfavourable effects of investment on the earning capacity andvalue of existing capital goods, since all the elements of ahomogeneous aggregate are necessarily perfect substitutes foreach other. Th e new capital competes with the old andreduces the profitability o f the latter. Once we allow forheterogeneity we must also allow for complementarity betweenold and new capital. The effect of investment on the profit-ability of old capital is now seen to depend on which of thevarious forms of old capital are complementary to, or sub-stitutes for, the new capital. The effect on the complements

Joa n Robinson: The Rate of Interest, and other Essays, p. 54 .4 A. P. Lerner: *On the Marginal Product of Capital and the Marginal Effi-ciency of Investment', Journal of Political Economy, February 1953, see especiallypp. 6-9.


16/138

THE ORDER OF CAPITAL 7will be favourable, on the substitutes unfavourable. The'inducement to invest' will therefore often depend on the effectthe new capital is expected to have on the earning capacity ofold capital complementary to it. In other words, investmentdecisions, as to their magnitude, and even more as to the con-crete form they are likely to take, depend at each moment onthe prevailing composition of the existing capital stock. Ingeneral, investments will tend to take such concrete forms asare complementary to the capital already in existence. A realunderstanding of the investment pattern is therefore impossibleas long as we cling to the homogeneity hypothesis. Investmentis thus seen to be ultimately a problem of the capital order.At each moment it reflects, both as regards its quantitativevolume and its concrete form, the possibilities left open by theexisting capital order.We must now return to our main task in this chapter. Weshall attempt to present very briefly a preliminary view of thechief problems with which we shall be concerned in this book.The problem of capital as a 'source' of profit is not amongthem. In Chapter V , to be sure, we shall have occasion toexam ine certain aspects of the theory of interest, but in the con-text of this book this will be to us a mere side-line. The mainsubject-matter of this book is the Capital Structure. When weturn our attention to the relationship between capital andinterest we do it for the light that interest sheds on capital, notvice versa. Those economists to whom the concept of capitalwas in the main an instrument in their search for the explana-tion of the interest phenomenon naturally conceived of it asa hom ogeneous aggregate consisting of value units. By con-trast, our conception of capital is that of a complex structurewhich is functionally differentiated in that the various capitalresources of which it is composed have different functions.The allocation of these functions, and the changes which itsmode undergoes in a world of change, is one of our mainproblems.

In this book we shall endeavour to outline an approach tocapital problems which is both realistic and directly based onthe definition of economic action: realistic in that we deal withthe world of unexpected change; directly based on the definitionof economic action in that we start from the fact that capital


17/138

8 CAPITAL AND ITS STRUCTUREresources are scarce resources with alternative uses.5 To us thechief problem of the theory of capital is to explain why capitalresources are used in the way they are; why in a given situationsome alternatives are rejected, others selected; what governsthe choice or rejection of alternative uses when unexpectedchange compels a revision of plan . There are two broadanswers to these questions, the first of which is perhaps rathertrite, but the second of which has not as yet been given theprominent place it deserves in economic thought. Th e firstanswer is that, of course, capital goods must be used in such away as to produce, directly or indirectly, the goods and servicesconsumers want at prices they are prepared to pay. This is afamiliar them e. Th e urge to maximize profits warrants thebelief that after some trial and error capital goods will in realitybe used in such a fashion. But there is a second answer.Capital uses must 'fit into each other'. Each capital good hasa function which forms part of a plan. Capital goods with nosuch function will not be maintained. Th e fact that capitalgoods which do not 'earn their keep' will be discarded warrantsthe belief that a tendency towards the integration of the capitalstructure really exists. But wh ile, on the one hand, the scopeof this phenomenon is wider than is comm only recognized, thetendency does not operate unimpeded.It is obvious that capital equipment for which no labour canbe found to work it, is useless and w ill be scrapped. Once weabandon the assumption that all capital is homogeneous thescope of this phenomenon is seen to extend beyond that of thecomplementarity between labour and capital. Capital equip-ment may have to be scrapped because no capital combinationcan be found into which it would fit.On the other hand, the scrapping of surplus capital, andconsequently the integration of the capital structure, may inreality be delayed for a number of reasons. For one thing ,expectations of a future different from the present may spoilthe simplicity of our theorem. If capital owners think that

8 We do not wish to imply, of course, that other capital theories are not basedon, or do not conform to, the definition of economic action and the praxeologicalaxioms it entails. But all too often the link is rather tenuous.There is a 'missing link' in most equilibrium theories; they all have to assumethat, once the data are given, the problem of how equilibrium is reached hasbeen solved. By contrast we shall concern ourselves with the 'pat h ' which menhave to follow in building up capital combinations and using them.


18/138

THE ORDER OF CAPITAL 9complementary factors will be available in the future, they willprefer to wa it. Moreover, for heavy durable equipment theannual cost of maintenance is probably relatively small. Th ismeans that even a small profit may suffice to keep capital goodsin existence. Ow ners of displaced capital goods will then tryto find complementary resources by offering their ownersco-operation on favourable terms. This, of course, is what w emeant by 'switching capital goods to second-best use'. Beforedisplaced capital goods are scrapped attempts will thus bemade to lure other capital goods, potentially complements tothem, out of the combinations of which they happen to formpart. T o some extent these attempts will be successful. Th isis important as it gives rise to certain dynamic processes weshall study in Chapter III . As we shall see, equilibriumanalysis cannot be applied to them.For the present we m ay conclude that a tendency towards theintegration of the capital structure exists, but that it mayencounter resistance from optimistic expectations and thepossibility of multiple use. W hile the former will lead to'surplus stocks' and the maintenance of other forms of visibleexcess capacity which have of late attracted the attention ofeconomists, the latter will give rise to a kind of invisible excesscapacity, a counterpart to 'disguised unem ploym ent'. Capitalresources will be used in ways for which they were not p lanned,but these uses will be discontinued the moment complementaryresources make their appearance.It is evident that only a morphological theory can beexpected to cope with such problems. W hether in reality anintegrated capital structure, in the sense that every capitalgood has a function, can exist in a world of unexpected changeremains to be seen. But we m ay say that the desire to maxi-mize profits on existing capital goods and the obvious futilityof maintaining those that cannot, either now or in the fore-seeable future, be fitted into the existing structure, warrantthe belief that economic action will at each moment tend inthe direction of such an integrated structure, even though thismay never be completed.All this has implications for the theory of investment. Wecannot exp lain how either existing resources are being replaced,whether by their replicas or otherwise, or what kind of new


19/138

10 CAPITAL AND ITS STRUCTUREcapital goods is being created, without having first of all learnthow existing capital is being used. Th e shape in wh ich newcapital goods make their appearance is determined largely bythe existing pattern, in the sense that 'investment opportunities'really mean 'holes in the pattern'.In the traditional view, maintained by Keynes and hisfollowers, new investment follows the success of similar existingcapital combinations. If shipping lines have been profitablemore ships will be built. W hile, to be sure, the marginalefficiency of capital is defined as an expectational magnitude,we are always given to understand that high profits on existingcapital offer a strong incentive to invest while low profits donot. As we shall see, this is another illegitimate generalizationbased on the homogeneity hypothesis. A number of invest-ment opportunities actually owe their existence to the failureof past capital combinations to achieve the purposes for whichthey were designed. This problem will be discussed also inChapter III.

All capital goods have to fit into a pattern or structure.What determines the structure? In the first place, there arethe various production plans, which determine the use to wh icheach capital good will be put. But if we are to speak of astructure, these plans must be consistent with each other.What makes them so? The market compels the readjustmentof those production plans which are inconsistent with eitherconsumers' plans or other production plans. From the pushand pull of market forces there emerges finally a network ofplans which determines the pattern of capital use.But the economic forces which integrate the capital structurevia the network of plans should not be confused with the force,discussed above, which causes the re-integration of the structureby discarding surplus equipment. The former serve tominimize, though they cannot altogether eliminate, the impactof unexpected change by removing inconsistency of plans.The latter only comes into operation after unexpected changehas happened and caused some capital to be displaced. Theformer will be discussed at some length in Chapter IV.We hope to have persuaded the reader that once we abandonthe homogeneity hypothesis we are compelled to adopt amorphological approach to the problems of capital, which


20/138

THE ORDER OF CAPITAL 11must supersede purely quantitative reasoning. For thequantitative concept of a homogeneous stock we have to sub-stitute the concept of a functionally differentiated CapitalStructure.6As yet we have left the concept of Capital undefined. Wenow define it as the (heterogeneous) stock of material resources.In thus defining it we follow Walras in stressing heterogeneity('les capitaux proprement dits') and Irving Fisher in refusingto draw the traditional distinction between land and capital.In fact, all the three possible criteria of distinction betweenland and other material resources are readily seen to beirrelevant to our purpose.When capital is defined, with Boehm-Bawerk, as the 'pro-duced means of production' land is, of course, excluded. Butto us the question which matters is not which resources areman-made but which are man-used. Historical origin is noconcern of ours. Our interest lies in the uses to wh ich aresource is put. In this respect land is no different from otherresources. Every capital com bination is in fact a combinationof land and other resources. Changes in the composition ofsuch combinations are of just as much interest to us whereland is, as where it is not affected.A second criterion of distinction between land and othercapital resources is based on the contrast between the 'fixed'nature of the supply of land and the variable character ofthe supply of other material resources. The criterion isquantitative in a purely physical sense which is not necessarilyeconom ically relevant. There are even today large tracts ofundeveloped land in the world which could be brought intoproductive use by combining them with capital resources.In other words, there is 'physical' land which is not a source ofeconomic services. Th e conditions in which such economictransformation will take place are precisely a problem in thetheory of capital. T o ignore them is to ignore one of the mostsignificant dynamic aspects of capital.

Professor S. Herbert Frankel has for some time expounded a similar view.'Capital . . . is, apar t from the symbolism of accounting, always "con crete"in the sense tha t it is embedded in, and a ttuned to, the particular purposes andstate of knowledge which led to its "creation* \ It is but temporarily incorporatedin ever changing forms and patterns suited to the evanescent ends for which itis designed' (The Economic Impact on Under-developed Societies, 1953, p. 69).


21/138

12 CAPITAL AND ITS STRUCTUREProfessor Hayek has defined capital as 'those non-permanentresources which can be used only . . . to contribute to thepermanent maintenance of the income at a particular level'.7Permanent resources thus fall outside the scope of this definitionand outside the scope of the theory of capital. But we cannotadopt this definition as we cannot ignore the uses to whichpermanent resources are put. There is no reason to believethat in their case the pattern of resource use is fundamentallydifferent from that of non-perm anent resources. In fact, aswe said above, the two are almost invariably used together.This does not mean that the replacement, the recurrent needfor which is the distinguishing characteristic of non-permanentresources, is of no interest to us. In so far as replacement doesnot take the form of the production of mere replicas it is ofinterest to us as it changes the composition of the capital stock.But what distinguishes our approach from Professor Hayek'sis that we are not concerned with the maintenance of income ata particular level. We are not interested in that long periodwhich must elapse before the income-stream from non-permanent resources dries up, but in the series of short periodsduring which resources are shifted from one use to another, andin the repercussions of such shifts. The causes and reper-cussions of these shifts are more or less the same, whether theresources shifted are permanent or not (except for very short-lived resources). As long as the period during which incom e-streams from non-permanent resources might be exhaustedremains beyond our horizon, there is no reason why we shoulddistinguish between permanent and non-permanent resources.The growing predominance of very durable capital equipmentin modern industrial society appears to underline the point-lessness of the distinction and justify our neglect of it.We shall now briefly set out the logical structure of the argu-ment thus far presented in this chapter.Heterogeneity of Capital means heterogeneity in use;Heterogeneity in use implies Multiple Specificity;

Multiple Specificity implies Complementarity;Complementarity implies Capital Combinations;Capital Combinations form the elements of the CapitalStructure.7 The Pure Theory of Capital, p. 54 .


22/138

THE ORDER OF CAPITAL 13We are living in a world of unexpected change; hencecapital combinations, and with them the capital structure, willbe ever changin g, will be dissolved and re-formed. I n th isactivity we find the real function of the entrepreneur.We must now give some consideration to the method ofanalysis wh ich we shall emp loy in this book. M ultiple Speci-ficity, as we saw, is a characteristic of the capital resourceswh ich form the subject-matter of this book. Th eir mode ofuse chang es as circum stances cha ng e. Th eir life-story thusfalls naturally into a sequence of periods during each of whichwe note their use for a specific purpose in conjunction withlab ou r services an d other cap ital goods. Th is fact alreadypoints to the need for Period Analysis in studying the patternof capital use. It is not, however, a mere m atte r of time, bu tof human action in t ime.The employment of a number of capital goods in a capitalcom bination d uring a given period is em bodied in a prod uctionplan ma de at the beginning of the period. Th e plan thusprovides a scheme of orientation, a frame of reference forsubsequent action. Th is pa tter n of resource use will be con-tinued as long as the plan succeeds in the sense that a 'target'envisaged in it is attain ed. T h e m ethod to be employed indescribing such events might thus be strictly called Tlan-Period-Analysis' . A nd inasmuch as we have to go beyond th e'unit period', and consider what happens in the next period asa result of what happened in this, we shall speak of ProcessAnalysis. In th e context wh ich alone interests us this m ean sthat if the plan fails the capital combination will be dissolvedand its constituent elements turned to other uses, each withinthe range permitted by its multiple specificity.The method of Process Analysis has been described byProfessor Lindahl in a famous passage:

Starting from the plans and the external conditions valid atthe initial point of time, we have first to deduce the developmentthat will be the result of these data for a certain period forwardduring which no relevant changes in the plans are assumed tooccur. Next we have to investigate how far the developmentduring this first periodinvolving as it must various surprisesfor the economic subjectswill force them to revise their plansof action for the future, the principles for such a revision being


23/138

14 CAPITAL AND ITS STRUCTUREassumed to be included in the data of the problem. And sinceon this basis the development during the second period isdetermined in the same manner as before, fresh deductionsmust be made concerning the plans for the third period, andso on .8In this book we shall thus employ the method of processanalysis based on plans and those entrepreneurial decisionswhich accompany their success and failure. But we shall notindulge in building 'dynamic models' based on 'behaviourfunctions' expressed in terms of 'difference equations'. Ourreason for this refusal is that to assume that entrepreneurialconduct in revising plans at the end of successive periods is, inany objective sense, determined by past experience and thuspredictable, would mean falling into a rigid determinism whichis quite contrary to everyday experience.Men in society come to learn about each other's needs andresources and modify their conduct in accordance with suchknowledge. But the acquisition of this knowledge follows no

definite pattern, certainly no time-pattern. Knowledge is notacquired merely as time goes by.All human conduct is, of course, moulded by experience,but there is a subjective element in the interpretation of experienceto ignore wh ich would be a retrograde step. Different menreact to the same experience in different ways. Were we con-concerned with 'Macro-dynamics' all this might not mattervery much . Probability might provide a convenient way out.But we are concerned with the conduct of the individual entre-preneur. A rigid determinism in these matters wou ld appearto reflect an outmoded, pre-Knightian approach. Th e econo-metricians have thus far failed to explain why in an uncertainworld the meaning of past events should be the only certainthing, and why its 'correct' interpretation by entrepreneurscan always be taken for granted.To assume 'given behaviour functions' or 'entrepreneurialreaction equations' is simply to deny that entrepreneurs arecapable of interpreting historical experience, i.e. experiencewhich does not repeat itself. In other words, to make theseassumptions is to say that entrepreneurs are automata and have

8 Erik L indahl: Studies in the Theory of Money and Capital (George Allen & UnwinLtd.), pp. 38-9.


24/138

THE ORDER OF CAPITAL 15no minds. Observation, however, bears out the contentionthat entrepreneurial minds exist and function.The entrepreneurial interpretation of past experience findsits most interesting manifestation in the formation of expecta-tions. Expectations, i.e. those acts of the entrepreneurialmind which constitute his 'world', diagnose 'the situation' inwhich action has to be taken, and logically precede the makingof plans, are of cruc ial im po rtan ce for process analysis. Amethod of dynamic analysis which fails to allow for variableexpectations due to subjective interpretation seems bound todegenerate into a series of economically irrelevant mathe-matical exercises.It thus seems clear tha t a study of capital problems in a w orldof unexpected change has to be conducted by means of processanalysis, and that the application of this method presupposesa study of entrepreneurial expectations.

The plan of this book is conceived in the following manner:In Chapter II we start by establishing a few systematic gener-alizations abo ut expectations. I n doing so we shall have todelve much more deeply into the subject than has been thecase in recent discussions. Argu me nts ab ou t how people bringthe various possible outcomes of actions they envisage into con-sistency with each o ther will be seen to be qu ite inad equ ate forou r purpo se. T h e formation of expectations is a m om ent inthe process of the acquisition of knowledge and has to bestudied as such. I n C ha pte r II I we shall show how processanalysis can be applied to capital problems; how entrepreneursreact to unexpected change by forming and dissolving capitalcombinations in the light of experience gained from workingwith them . In Ch apter IV we ask how a capital structurecan come to exist in a world of unexp ected chan ge. We shallsee tha t, thoug h in the real world a capital structure in tegratin gall existing cap ital resources, even if it ever cam e in to ex istence,could not exist for any length of time, integrating as well asdisintegrating forces are always in operation, and much canbe learned from a study of their modus operandi.The Chapters II to IV constitute the theoretical nucleusof the book. In the following three chap ters the ideas setforth in the earlier part will be applied to a variety of capital


25/138

16 CAPITAL AND ITS STRUCTUREproblems. In Chapter V w e attempt to show that, in so far asthe accumulation of capital can be said to prompt economicprogress, it does so by prompting certain typical changes in thecomposition of the capital stock. In Chapter V I we raise thequestion whether structural relationships exist in the sphere ofproperty rights and claims as well as in that o f physical capitalresources, and if so, how the two spheres are interrelated. InChapter VII we hope to show that if capital accumulationentails changes in the composition of the capital stock, it alsoentails certain consequences for the course of the Trade Cycle;that industrial fluctuations are frequently due to intersectionalmaladjustment, and that in the circumstances which usuallyaccompany economic progress this is often almost inevitable.Finally, some attention will be given to the question of howsuch maladjustments can be, and are in reality, overcome.

The argument so far set forth in this chapter, derived as it isfrom heterogeneity and multiple specificity of capital, has anumber of implications which will come up for fuller dis-cussion at various points in the book. But a few of them which,in a sense, underlie all that follows should be noted alreadynow.What we have so far said in this chapter serves to sharpenour understanding of the function of the entrepreneur. Weusually say that the entrepreneur 'combines factor services'.So he does, but the statement is too wide and not preciseenough since it suggests that the relationship between the entre-preneur and the owners of resources, human and material, issymm etrical in all cases. Labour, of course, is hired and dis-missed. But the entrepreneur's function as regards capital isnot exhausted by the hire of services. Here his function is tospecify and make decisions on the concrete form the capitalresources shall have. H e specifies and modifies the shape andlayout of his plant, which is something he cannot do to histypists, desirable though that may seem to him . As long as wedisregard the heterogeneity of capital, the true function of theentrepreneur must also remain hidden . In a homogeneousworld there is no scope for the activity of specifying.This fact is of some practical importance. For it followsthat the entrepreneur carries, and must carry, a much heavier


26/138

THE ORDER OF CAPITAL 17responsibility towards the owners of his capital than towardshis workers, since as regards the resources of the former heenjoys a much wider range of discretion than as regards thoseof the latter. It m eans no t merely th at 'wo rkers' control ofindu stry' is impossible. It also m eans tha t cap ital own ers,havin g delegated the p ower of specification to the entre pren eur,are 'uncertainty-bearers' in a sense in which workers are not.They are in fact themselves entrepreneurs of a special kind.The whole relationship between manager-entrepreneurs andcapitalist-entrepreneurs will be taken up for discussion towardsthe end of Chapter VI.From time to time, in particular in the last three chapters ofthis book, we shall employ the notion of 'economic progress' .By this we m ean a n increase in real income per head . W emust note that to assume progress is not necessarily the samething as to assume the 'dy na m ic' conditions of a wo rld of unex-pected chang e. O n the one han d, of course, dynam ic con-ditions m ay lead not to progress bu t to disaster. O n the otherhand, most recent discussions of progress have been couchedin terms not of a dynamic world, but of the model of an'expanding economy', of Cassel's 'uniformly progressiveeconomy' slightly modified by Messrs. Harrod and Hicks.This model embodies the notion of 'growth', of progress at aknown and expected ra te . Its significance for the real w orld,however, is doubtful. Already the m etap ho r 'grow th' issingularly inappropriate to the real world as it suggests aprocess during which the harmony of proportions remainsund isturbed. No r can we, after wh at has been said, anylonger believe that progress will manifest itself in the capitalsphere merely in the form of capital accumulation, i.e. purelyqu antitativ e growth. In w hat follows we shall always assumethat progress takes place in conditions of unexpected change.Inevitably unexpected change entails some capital gains andlosses. H enc e we cann ot say th at progress is eithe r accom -pan ied or caused by the accum ulation of cap ital. But themalinvestment of capital may, in some cases, by providingexternal economies, become the starting-point of a process ofdev elopm ent. A railway line bu ilt for the exploitation of somemineral resource may be a failure, but may nevertheless giverise to more intensive forms of agriculture on land adjacent to


27/138

18 CAPITAL AND ITS STRUCTUREit by providing dairy farmers with transport for their produce.Such instances play a more important part in economic pro-gress than is commonly realized. Th e ability to turn failureinto success and to benefit from the discomfiture of others isthe crucial test of true entrepreneurship. A progressiveeconomy is not an economy in which no capital is ever lost,but an economy which can afford to lose capital because theproductive opportunities revealed by the loss are vigorouslyexploited. Each investment is planned for a given environ-ment, but as a cumulative result of sustained investmentactivity the environment changes. These changes in environ-ment did not appear on the horizon of any of the entrepre-neurial planners at the time when the plan was conceived.All that matters is that new plans which take account of thechange in environment should be made forthwith and oldplans adjusted accordingly. If this is done as fast as the newknowledge becomes available there will be no hitch in theconcatenation of processes, of plan and action, which we callprogress.What we have just said has some relevance to the problemsof the 'economic development of underdeveloped areas' whichhave in recent years been so extensively discussed by economistsand others.9 Economic progress, we saw, is a process whichinvolves trial and error. In its course new knowledge isacquired gradually, often painfully, and always at some costto somebody.10 In other words, some capital gains and lossesare inevitable as durable capital goods, in the course of theirlong lives, have to be used for purposes other than those forwhich they were originally designed. Such capital losses havebeen frequent concomitants of economic progress in the historyof almost all industrial countries, and have on the whole donemuch good and little harm.But a question arises in this connection which has, to ourknowledge, rarely been adequately discussed: Who will bearthe risk? Where economic development is financed by risk

In these discussions Professor Frankel has shown himself an undaunted criticof the quantitative notion of capital which most other writers on the subjectaccepted withou t question. In particu lar, he frequently w arned his fellow-economists 'not to regard the calculations of the private entrepreneur in terms ofestablished accounting symbolisms as in any sense an automatic or mechanicalprocess' (S. H. Frankel, op. cit., p. 66).10 For examples see S. H. Frankel, op. cit., pp. 101-2.


28/138

THE ORDER OF CAPITAL 19capita l, cap ital own ers, of course, be ar the risk of loss. Bu trisk capital is now rapidly becoming very scarce everywherein the world, even in those countries in which it is not actuallybeing taxed out of existence. O n the other ha nd , the politicalmyths of the twentieth century being what they are, politiciansin most underdeveloped countries abhor the very thought offoreign risk-bearing capitalists, or, if they do not, at least haveto pretend before the electorate that they harbour such feel-ings. The reason is not far to seek: He who bears the riskmu st also app oint the ma nag ers. In the prevailing climate ofopinion the economic development of such countries willtherefore largely have to be promoted with the help of loancap ital. Th is mea ns tha t the capital losses will have to beborne by those who are probably least able to bear them, viz.the inhabitants of these areas themselves, whose rescue frompoverty has been, after all, the ostensible purpose of the wholeoperation!


29/138

CHAPTER IION EXPECTATIONS

The explicit introduction of expectations into the economictheories of the last thirty years has given rise to a host of newproblems. Of these the most fundamental is the questionwhether expectations have to be regarded as independent'data* or as the results of economic processes. As yet, econ-omists appear to disagree on the answer; hence the unsatis-factory state of the theory of expectations.Evidently expectations are not economic results in the sensein which prices and output quantities are. N o econom icprocess determines them . A 10 per cent rise in the price ofapples may just as well give rise to an expectation of a furtherrise as to that of a future fall. It all depends on the circum-stances accompanying the rise, and different people may givethese circumstances a different interpretation. It follows thatall those dynamic theories which are based on 'differenceequations', 'accelerators', etc., simply by-pass our problem.At best we may regard them as provisional hypotheses to beemployed as long as the wider questions remain unanswered.It is, however, equally impossible to treat expectations asdata in the same way as we treat consumers' tastes., Fromwhatever angle we look at them, expectations reflect economicexperience and are affected by changes in it. In this fact liesan important difference between a change in tastes and achange in expectations. A change in taste of course may alsobe due to experience, but it need not be. I may give upsmoking a certain brand of cigarettes because I have found thatit affects my throat, but I may also give it up because I nolonger like it and 'have lost the taste for it'. There^ lies beh indtastes an irreducible substrate which rational analysis cannotgrasp, which may be of interest to the psychologist, but whichdefies the analytical tools of the economist.Expectations, on the other hand , always embody problematicalexperience, i.e. an experience which requires interpretation. Itis the task of the theory of expectations to elucidate the problems20


30/138

ON EXPECTATIONS 21ou r experience (an d th a t of others in so far as it is accessibleto us) sets us in judging the uncertain future, as well as toclarify the modus interpretandi. It is a task with which econ-omists thus far do not seem to have come to grapple.Experience is the raw material out of which all expectationsare formed. But not all m ateria l is equa lly useful, no t allexperience is equally relevant to a given situation. Th ere isa subjective element in the acts of the mind by which we selectthose portions of our experience we allow to affect our judg-m en t of the future. But this subjectivism ofinterpretationis some-thing altogether different from the subjectivism of want whichunderlies ou r utility theo ry. T he former yields provisionaljudgments to be confirmed by later experience, imperfectknowledge capab le of being perfected. T h e latter can provideus with no new knowledge: we either have a want or do nothave it.In a society based on division of labour men cannot actwitho ut knowing each other's needs an d resources. Suchknowledge need not be, as some have thought, 'perfect' , butit must be relevant knowledge, knowledge of the demand andsupply conditions in the m arkets in which one hap pen s to dea l.There is no difficulty in conceiving of such knowledge in a'stationary state' , a world without change, since here we neednot ask how people came by their knowledge any more thanwe need ask how this improbable state of affairs came about:both belong to the category of, now irrelevant, 'bygones'.But we can also conceive of a quasi-stationary state in whichchanges are few and far between, and each change has had itsfull repercussions before the nex t chan ge takes place. Th isquasi-stationary state is the background of most neo-classicaleconomics. Against it the method of comparative staticsshows itself to full adv an tag e. I n this state knowledge isguid ed by prices functioning as signposts to action . H er e it isby observing price changes that consumers learn which goodsto substitute for which, arid producers learn which line ofproduction to aban don and which to turn to. He re we maysay that the price system integrates all economic activity.We m ay reg ard the price system as a vast network of com m uni-cations through which knowledge is at once transmitted fromeach marke t to the remotest corners of the economy. Every


31/138

22 CAPITAL AND ITS STRUCTUREsignificant change in needs or resources expresses itself in aprice change, and every price change is a signal to consumersand producers to modify their conduct. Thus people gainknowledge about each other by closely following market prices.But in the world in which we are living change does notfollow such a convenient pattern. Here knowledge derivedfrom price messages becomes problematical. It does not ceaseto be knowledge, but 'does not tell the whole story'. Manychanges may happen simultaneously. Parts of our comm uni-cations network may be 'jammed' and messages delayed.When a number of messages is received it is no longer clearin wh ich order they were 'sent'. Moreover, even if there isno delay in transmission, today's knowledge m ay be out of datetomorrow, hence no longer a safe guide to action. Worst ofall, in a world of continuous change much may be gained bythose 'speculators' who prefer to anticipate tomorrow's changestoday rather than adjust themselves to those recorded in thelatest message received. Their action will affect prices wh ichothers take as their point of orientation, and which, if thesespeculators turn out to be wrong, may mislead others intoactions they would not have taken had they known the realcause of the price change.It is here neither necessary nor possible to follow up all theramifications of the problem what constitutes relevant know-ledge in a world of continuous change. This theme will betaken up again in Chapter IV . Action based on pricemessages conveying misleading information is, as we shall seein Chapter VII, often an important factor in the Trade Cycle.For our present purpose it is sufficient to realize:First, that in a world of continuous change prices are nolonger in all circumstances a safe guide to action;Second, that nevertheless even here price changes do transmitinformation, though now incomplete information;Third, that such information therefore requires interpretation(the messages have to be 'decoded') in order to be transformedinto knowledge, and all such knowledge is bound to be imper-fect knowledge.In a market economy success depends largely on the degreeof refinement of one's instruments of interpretation. O n theother hand, every act is a source of knowledge to others.


32/138

ON EXPECTATIONS 23The formation of expectations is nothing but a phase in thiscontinuous process of exchange and transmission of knowledgewhich effectively integrates a market society. A theory ofexpectations which can explain anything at all has thereforeto start by studying this phase within the context of the processas a whole. If it fails to do this, it can accomplish nothing.Its first task is to describe the structure of the mental actswhich constitute the formation of expectations; its second task,to describe the process of interaction of a number of individualswhose conduct is orientated towards each other.For anybody who has to make a decision in the face of anuncertain future the formation of an expectation is incidentalto the endeavour to diagnose the situation in which he has toact, an endeavour always undertaken with imperfect know-ledge. Th e business man who forms an expectation is doingprecisely what a scientist does when he formulates a workinghypothesis. Both, business expectation and scientific hypothesisserve the same purpose; both reflect an attempt at cognitionand orientation in an imperfectly known world, both embodyimperfect knowledge to be tested and improved by later experi-ence . Each expectation does not stand by itself but is thecumulative result of a series of former expectations which havebeen revised in the light of later experience, and these pastrevisions are the source of whatever present knowledge wehave. O n the other hand , our present expectation, to berevised later on as experience accrues, is not only the basis ofthe action plan but also a source of more perfect future know-ledge. Th e formation of expectations is thus a continuousprocess, an element of the larger process of the transmission ofknowledge. Th e rationale of the method of process analysis, aswe shall learn in Chapter III, is that it enables us to treat theex ante 'data' of action as provisional hypotheses to be revisedin the light of later experience.We have said that the formation of expectations is inci-dental to the diagnosis of the situation as a whole in which onehas to act. How is this done? We analyse the situation, aswe see it, in terms of forces to which w e attribute various degreesof strength. We disregard what we believe to be minor forcesand state our expectations in terms of the results we expect theoperation of the major forces to have. W hich forces we regard


33/138

24 CAPITAL AND ITS STRUCTUREas major and minor is of course a matter of judgment. Herethe subjective element of interpretation is seen at work. Ingeneral, we shall be inclined to treat forces working at randomas minor forces, since we know nothing about their origin anddirection, and are therefore anyhow unable to predict the resultof their operation. We treat as major forces those about whoseorigin and direction we think we know something. This meansthat in assessing the significance of price changes observed inthe past for future changes we shall tend to neglect those webelieve to have been due to random causes, and to confine ourattention to those we believe due to more 'permanent' causes.Hence, in a market economy, there are some price changeswhich transmit knowledge and are acted upon, and there arealways others wh ich are disregarded, often wrongly, and there-fore becom e econom ically *functionless\ Th is is an importantdistinction to which we shall return at the end of the chapter.Having stated our expectations at the start of a 'period', wetest them at its end by comparing actual with expected results,attempting to infer therefrom whether our initial diagnosis offorces and their strength was correct. Th is process, like allverification of hypotheses, is indirect and therefore often in-conclusive. Again, it requires interpretation and yields imper-fect knowledge. We may have been right for the wrongreason. Or, though we now may know that our originalhypothesis was wrong, we do not know how we could have beenright. Th e same result, say, a price change, may have beenbrought about by a number of different constellations of majorforces, hence the need for further hypotheses and further testing.Expectations are thus phases of a never-ending process, theprocess by which men acquire knowledge about each other'sneeds and resources. For our present purpose we shall drawthree conclusions from this fact:First, when at any point of time we look backwards at ourpast course of action we find that all our past expectations forma continuous sequence, whether they turned out to be rightor wrong. For we learnt from all of them .Second, there are problems of consistency, both interpersonaland intertemporal. Different people may hold differentexpectations at the same time; the same person may holddifferent expectations at different times. These are quite


34/138

ON EXPECTATIONS 25insoluble problems as long as we regard expectations as inde-pendent of each other. Why should they be consistent witheach other?But if we look at the process as a whole, the prospect becomesmore hopeful; successful expectations, which stand the test, are,on the whole, more likely to reflect 'real forces' than un-successful expectations. And those whose expectations arenever successful are likely to be eliminated by the marketprocess. Moreover, as we shall see in Chapter IV , the marketalso tends to evolve institutions which mitigate interpersonaland intertemporal inconsistency.Third, the results of past mistakes are there not merely toprovide lessons, but to provide resources. In revising our ex-pectations we not only have the know ledge, often dearly bought,of past mistakes (our own and others') to learn from, but alsotheir physical counterpart, malinvested capital. Malinvestedcapital is still capital that can be adapted to other uses. Thisis the main problem of the theory of capital in a worldof unexpected change. We shall come to deal with it inChapter III.Thus far we have endeavoured, all too briefly, to indicate atleast some of the foundations on w hich, in our view, a theory ofexpectations which truly reflects economic processes whichintegrate a society founded on specialization and exchange,must be based. In the light of what we have thus learnt weshall now turn to a critical examination of some other attemptsto present the problem of expectations in a systematic manner.

Until recently most studies of the problems of expectationswere informed by the view that this is a proper field for theapplication of probability theory. An entrepreneur wh o hasto make a decision the outcome of which is uncertain, is con-ceived of as setting up a probability distribution of possiblequantitative outcomes for each course of action open to him.The next step is usually to 'substitute for the most probableprices actually expected with uncertainty equivalent pricesexpected with certainty'.1 In this way the range of theprobability distribution is compressed to a point, a 'certainty-equivalent'.2 In 1945 we objected to this procedure on the

1 Oscar Lange: Price Flexibility and Employment, p. 31.8 G. L. Shackle: Expectations, Investment and Income, 1938, p. 64.


35/138

26 CAPITAL AND ITS STRUCTUREground that 'by subttituting single-value expectations for theuncertainty range of expected prices we stand to lose morethan we gain, because reaction to price change will largelydepend on the location of the prices affected within the scaleof expected prices5.3 We shall explain later on why wemaintain this view.But meanwhile the whole probability approach to the studyof expectations has come under heavy fire. In the final chapterof his book Expectations in Economics 4 Professor Shackle hassubjected what he calls the 'orthodox view' of the formationof expectations to strong and extensive criticism. His argu-ments are not new,6 but they are none the less effective. Hismain point is 'the irrelevance of estimates of probability (in thesense of relative frequency) to unique or quasi-uniquedecisions'.18 'Few individual enterprises, for example, even inthe course of their whole lives, launch a number of venturesof even broadly similar kinds which is "large" in any senserequired by the theory, or even the practical application, ofprobability principles.' 7 The point is reinforced by theabsence of a 'homogeneous universe of outcomes'. 'For manyimportant kinds of decisions which must be taken in humanaffairs it will be impossible to find a sufficient number of pastinstances which occur under appropriately similar conditions;no well-founded figures of probability for different kinds ofoutcome can be established on the basis of experience.' 8Professor Shackle's criticism of the probability approach tothe problem of expectations is sound enough, though theemphasis, in our view, should be on heterogeneity of situationsrather than on uniqueness of decisions. It seems to us thatProfessor Shackle's argument might lose much of its applic-ability to the real world if the 'uniqueness of decisions' is takentoo literally. Few business decisions are unique in the sensethat they are made only once in a lifetime, and ProfessorShackle only weakens his case by confining it to investment

8 L. M. Lachmann: *A Note on the Elasticity of Expectations', Economica,November 1945, p. 249. G. L. Shackle: Expectations in Economics, 1949.8 Cf. F. H. Knight: Risk, Uncertainty and Profit, Chapter VII, especially pp.224-32, and L. v. Mises: Human Action, Chapter VI, especially pp. 106-15. G. L. Shackle: Expectations in Economics, 1949, p. 127.'Ib id ., p. 110.8 Ibid., pp. 109-10.


36/138

ON EXPECTATIONS 27decisions involving very large sums.9 We have to distinguishbetween uniqueness of decisions and uniqueness of the situa-tions the decisions are taken to meet an d to crea te. T henu m be r of possible business decisions is almost c ertainly smallerthan the number of such possible situations, precisely becausein an uncertain world each decision may have one of severalresults. And the 'outcomes' are here not, as in nature,'external events given to us', but the result of a complexprocess of interaction always accompanied by transmission ofknow ledge. I t therefore seems be tter to base ou r rejection ofthe probability theory of expectations on the inherent hetero-geneity of the situations rather than on the uniqueness ofdecisions.Whether Professor Shackle's positive contribution can be ofmuch help to us in grappling with the problems set out earlierin this cha pter mu st rem ain doubtful. T he object of his studyis the mental processes of an individual who has to takea decision in the face of an un ce rtain future. His theo ry ismodelled on the equilibrium of the isolated individual (Robin-son Crusoe) and stops ther e. It tells us noth ing abou t ma rketprocesses and nothing about the exchange and transmission ofknowledge. T o be sure a plan if it is to make sense must bebased on one self-consistent scheme of expectations',10 but thecreation of such a scheme marks only the beginning of ourproblem s. We have to ask how these expectations are formed,revised if disappointed, and projected into the future whensuccessful. T h e chan ges of knowledge which ProfessorShack le studies in his C ha pter I I I im ply a 'clarifying of ex-pec tations' in a purely formal sense. T he events leadingto such clarification are 'external events', not market trans-actions . I n othe r words, Professor Shackle's is a static theory ,an d chan ge is here co mp rehend ed as once-for-all change withinthe framework of comparative statics.It is noteworthy that the only time that Professor Shacklementions an actual market, this is a market which, for the timebeing, has ceased to function; a market not in operation but insuspense. In trying to account for a certain price phenomenon

9 'No business executive has to decide a hundred times in ten years whether ornot to spend 1,000,000 on a new factory'p. 115.10 Ibid., p. 111.


37/138

28 CAPITAL AND ITS STRU CTU REon the peri-urban land market he finds that 'Evidence takenby the Uthwatt Committee shows that, where the belt of landencircling a town is parcelled up amongst a large number ofseparate ownerships, the market value of each piece is suchthat when the separate values are aggregated, the total isseveral times as great as would b e wa rran ted by any reasonableestimate of aggregate future building development round thetown as a wh ole. I t is as thou gh each actual an d p otentialowner of a plot of land near the town were convinced that,out of a far more than adequate total supply of similarly situ-ated land, the particular plot in question was almost certainto be selected as part of the site for such new houses as will berequired during, say, the next twenty years.9 1XProfessor Shackle regards this as a 'curious phenomenon'and attempts an explanation in terms of his 'potential surprisefunctions'. But a m uch simpler explan ation can be given interms of the market process, or rather, its conspicuous absencein this case.T he function of the capital m arket is to allocate scarce capitalresources am ongst a nu m be r of altern ative uses. Th is is simplewhere these uses are known, not so simple where they are notknow n. Fo r them to be know n, however, it is not eno ughmerely tha t the total qua ntity required is know n. W herethese uses are specific, unless individual uses and their specificrequirements are known, no allocation can take place.In the case un der discussion this is jus t wh at ha s h ap pe ne d:while total requirements can be estimated, individual futurerequirements are unknow n. O n the other ha nd , the need ofland for urb an development is the most imp ortant need. Inthis situation the market safeguards the future provision forthe most important need by suspending the allocation to othersan d creating a reserve stock of lan d. Th is it does by m akingthe price of each plot equal to its value in satisfying the mostimportant need, thus making it impossible for anybody whow ants lan d for oth er purposes, to get it. T h e need for a reservestock will continue until the individual and specific needs areknow n. Th en , bu t only then , the m arket process of allocationc a n be g in . `Market\ in the true economic sense, means a process ofexchange and allocation reflecting the transmission of know ledge. I t

1 1 G. L. Shackle : Expctations in Economics, 1949, p . 9 1 . (His ita l ics.)


38/138

ON EXPECTATIONS 29does not simply m ean tha t prices are quote d. T he pricesquoted may be what they are in order to prevent, and not tofacilitate, dealings. W here this is the case we hav e a market insuspense, not a market in operation. Professor Shackle, like theexperts of the Uthwatt Committee, has become the victim ofverbal confusion.Fro m this exam ple it m ay be seen th at the theory of expecta-tions neglects the market process at its peril.After this critical digression we must turn to a more con-structive task. T he read er, before wh om we set certain ideasabout expectations in the early part of this chapter, will nodoubt expect us to give concrete shape to these ideas byembodying them in an analytical framework within whichconcrete problems can be solved and actual market processesrendered intelligible.12 But there is another reason, intrinsicto our argument, why we should make an attempt in thisdirection.We have described a m arket economy in motion as an im per-fect comm unications network. Th ere are, however, im po rtanteconomic changes which do not find their expression in pricechanges. Th ey constitute the pheno men on of price inflexibilityabout which we shall have to say something in Chapter IV.There are also price changes which do not reflect majoreconomic changes. We said above tha t in a m arket economythere are some price changes which transmit knowledge andare acted upon, and there are always others which are dis-regarded, often wrongly, and therefore become economically ic t ion l e s s \1 3 Evidently it is of great importance to us tofind a generalization on which an adequate criterion of dis-tinction between 'significant' (meaningful) and 'functionless*(meaningless) price movem ents can be based. If such ageneralization could not be found it would become impossibleto hold tha t prices integ rate the m arke t econom y. All wecould say would be : some do and some do not. Th ere aremany difficulties of course in finding such a generalization,foremost among which is one which directly reflects what we12 It is true that for the main purpose of this book, the elucidation of a morpho-logical conception of cap ital, this may not seem strictly necessary. But, as willbe seen in C hapter IV , it may be of help in making us understand the distinctionbetween consistent an d inconsistent capital change.18 See above, p. 24.


39/138

30 CAPITAL AND ITS STRUCTUREhave called the subjectivism of interpretation: the same pricemovement may be meaningful to one, and meaningless toan oth er perso n. Nevertheless it seems to us possible toconstruct an analytical framework within which:

a. the distinction between meaningful and meaninglessprice movements can be given a clear meaning, andb. the distinction can be seen to be practically significant:meaningful and meaningless price movements do in facthave different results.It may even be possible to link the distinction with that betweenmin or ( 'rando m ') an d major ( 'perm ane nt') forces. W e m ightsay, for instance, that the market will tend to disregard pricechanges believed to be due to rand om causes while paying closeattention to those it believes prompted by a change in theconstellation of fundamental forces.Such a n ana lytical framework we find in wha t, following D r.Lange, we cal l The Pract ical Range' . 1 4

Let us suppose that on a market a 'set of self-consistentexpectations' has had time to crystallize and to create a con-ception of a 'nor m al price ran ge '. Suppose th at any pricebetween 95 and 110 would be regarded as more or less'normal' , while a wider range of prices, say from 80 to 125, would be regarded as possible. We thus have two ranges,an 'inner range' from 95 to 110 reflecting the prevailing con-ception of 'normality', and an 'outer range' associated withw ha t is regarde d as possible price cha nge . M any economistshave started their study of expectations with the notion of a'range', usually in the form of a probability distribution, butonly to discard it at the next moment in favour of a point, a'certainty-equivalent' , ' to substitute for the most probableprices actually expected with uncertainty equivalent pricesexpected with certainty' .1 5 By contrast, we shall endeavourto show that the location and motion of actual prices withinour ranges are of crucial importance for the formation ofexpectations, and that by compressing the range to a point weshould lose the very frame of reference w ithin wh ich ac tua l p ricechanges can alone be meaningfully interpreted and shown tobe relevant to the formation of expectations.

14 Lange, op. cit., p. 30. Ibid., p. 31 .


40/138

ON EXPECTATIONS 31What is the significance of our two ranges for the formationand revision of expectations?As long as actual prices move well within the inner range,betwee n, say, 96 an d 109, such price movements will prob ablybe rega rded as insignificant and due to ran do m causes. Infact, where the 'normality' conception is strongly entrenched,it will be very difficult for the pric e to pass the limits. For assoon as the price approaches the upper or lower limit of theinner range, people will think that the movement 'cannotgo much farther ' and, anticipating a movement in reverse,will sell (near the upper limit) or buy (near the lower limit).In such a situation 'inelastic expectations' will tend to'stabilize' prices within the inner range.But suppose that in spite of sales pressure near the upper,and buying pressure near the lower limit, price either risesabov e 110 or falls below 95 . Suc h an event will sooner orlate r give rise to second though ts. As long as ac tua l pricesmove within the outer range, between 110 and 125, or 80 and95, i t is true, nothing has hap pened which was not regarded aspossible. But the mo re thoughtful mark et operators will takehe ed . T h e m ere fact tha t in spite of the heav y 'speculative*pressure encountered near the limits of the inner range, andengendered by inelastic expectations and the sense of the'normality' of the inner range, price could pass these limits atall is a pointer to the strength of the forces which must havecarried it past such formidable obstacles. Such a move men tcan hardly be due to random causes.But the force that carried the price past the limits, whilestrong, need not be a pe rm an en t force. It ma y spend itselfsooner or later. T h e ma rke t will therefore jud ge the sig-nificance of price movements within the outer range by thesup plem enta ry criterion of the time factor. If prices relapsesoon and return to the inner range this will of course confirmthe prevailing notion of norm ality. But if they stay withinthe ou ter rang e, grad ually opinion will swing rou nd . Firstsome, and then others, will come to revise their notion of'no rm al pr ice '. Such revision will express itself in a newwillingness to buy at a price, say 118, at which formerly onewould have sold, or to sell at a price, say 88, at which formerlyone would have bou ght. This means tha t a price movem ent,


41/138

32 CAPITAL AND ITS STRUCTUREonce i t has been strong enough to overcome resistant pressureat the l imits of the inner range, and reached the outer range,will probably sooner or later be carried further by the veryspecu lative forces w hic h formerly resisted it. Th is is read ilyseen if we reflect that the sales and purchases near the limitsmust have been at the expense of normal stocks, so that a priceof 115 would probably now find the market with low stocks,and a price of 90 with an accumulation of excess stocks whichare now a mere relic of the unsuccessful speculation of the'norm al ists \ A fast m ovem ent within the outer range m aytherefore be just as much due to re-stocking (positive ornega tive) as to the oper ation of m ore per m ane nt forces. Th isis why in such a situation the market keeps a close eye onstock variations .1 6 In fact , in dynamic conditions price move-ments have always to be interpreted with an eye on the'statistical position' of the market which thus becomes a secondsupplementary criterion for diagnosis .Once the price passes the limits of the outer range, risesabove 125 or falls below 80, an entirely new situation faces us.The market, shocked out of i ts sense of normality, wil l have torevise its diagnosis of the permanent forces governing a 'normalsituation' . It m ust no w be co m e clear to every bod y that thehypothesis about the constel lation of fundamental forces whichformed the basis of our range structure has been tested andhas fai led. Bu t wh ile the failure of an expe rime nt ma y invali-da te a hyp othesis, it doe s no t by itself suggest a ne w one . Itfol lows by no means that the real ly operative forces wil lbe recognized at onc e. Th at must depe nd on the ins ight ,vigi lance , and intel l igence of the ma rket. Exp erience shows,for instance, that an inflation is hardly ever recognized assuch in its initial stages, at least in a society which has noprior experience of i t . Alm ost invariably, at on e poin t oranother in this early phase, people wil l think that prices arealready 'too high' , wil l defer purchases and postpone invest-m en t plans. In this w ay, they wil l , by their very failure tounderstand the modus operandi of the fundamental force, miti-gate i ts im pac t for a t ime by redu cing 'effective dem an d'. A n dif, as is not impossible, the inflation stops early enough, they

16 Cf. L. M. Lachmann: 'Commodity Stocks and Equilibrium', Review ofEconomic Studies, June 1936.


42/138

ON EXPECTATIONS 33m ay be right after all! Bu t it is m ore likely th at they will bewro ng. An d in so far as their action entails the un de rm ain -tenance of capital, the ultimate results for society may well bedisastrous.We may conclude that price movements within the innerrang e will be disregarded an d thus be 'functionless\ Pricechange beyond the limits of the inner range may or may not bemeaningful, b ut jud gm en t will here have to depend on supple-mentary criteria such as the time factor and concomitant varia-tions in th e size of stocks. It is only wh en prices move bey ondthe outer range altogether that they become unquestionably'meaningful' , can no longer be disregarded, and convey a'message '. But while the negative con tent of the message isclear enough, viz. the invalidation of the hypothesis whichformed the basis of the former range structure, its positiveconte nt is less so. T he message still requ ires inte rpr eta tion ,and this will depend upon the insight and intelligence of themen in the market .O ur concept of the Ran ge Structure, composed of inner an douter range, seems thus vindicated as a useful tool of analysis,an d o ur refusal to exchange it for a 'certainty-equ ivalen t' w ouldap pe ar to be justified. Fo r ou r concept perm its us to dra w adistinction between price phenomena which are consistentwith the existing structure of expectations, fall 'within theranges' , and thus cause no disappointment, and, on the otherhand, phenomena inconsistent with the existing structure ofexpectations, which fall 'outside the ranges' a revision of whichthey necessitate. W e noticed tha t as long as the price mov e-ment is confined to within the inner range it does not providerelevant new information but merely confirms the soundnessof the diagnosis which found its expression in the existence ofthis range, while movements within the outer range provideinformation of problematical value which, to be useful, has tobe supplemented by observation of other phen om ena. Assoon, however, as the price moves beyon d the limits of the ou terrange, the inadequacy of the diagnosis on which the rangeswere based becomes pa ten t. A new situation has arisen wh ichrequires a new diagnosis and thus a new mental effort.It remains true that, by and large, price changes integratea marke t economy by spreading new knowledge. But not


43/138

34 CAPITAL AND ITS STRUCTUREall price changes are equally significant in this respect. T he irsignificance has to be assessed with respect to a 'given' struc-ture of expectations which finds its expression in a systemof rang es. Th eir practica l effect will dep end on how quicklythe men in the market come to understand what has happenedan d revise their expectations. T o imp ede price change istherefore to withhold knowledge from the m ark et. O n theother hand, it is possible to have 'misleading' price move-m en ts. Ab out them more will be said in later chap ters.


44/138

CHAPTER IIIPROCESS ANALYSIS AND CAPITAL THEORY

The theory of capital has to start from the fact that thecapital goods with which entrepreneurs operate are hetero-geneous. These heterogeneous capital goods have to be usedtogether. Heterogeneity here implies complementarity inuse. The mode of this complementarity, the proportions inwhich the various heterogeneous factors of production arebeing used for a given purpose, m ust find its expression in theProduction Plan. Each such plan is characterized by thecoefficients of production of its input and the output result itenvisages. But wh ile the output result is at first merelyplanned, the decision about coefficients of production has tob

Capital and Its Structure

Documents