Chapter V: Economic Generalisations and Reality

CHAPTER V ECONOMIC GENERALISATIONS AND EEALTTY 1. IT is a characteristic of scientific generalisations that they refer to reality. Whether they are cast in hypothetical or categorical form, they are distinguished from the propositions of pure logic and mathematics by the fact that in some sense their reference is to that which exists, or that which may exist, rather than to purely formal relations. In this respect, it is clear, the propositions of Economics are on all fours with the proposition of all other sciences. As we have seen, these propositions are deductions from simple assumptions reflecting very elementary facts of general experience. If the premises relate to reality the deductions from them must have a similar point of reference. It follows, therefore, that the belief often expressed by the critics of Economics, that it is a mere system of formal inferences having no necessary relation to reality, is based upon misconception. It may be admitted that our knowledge of the facts which are the basis of economic deductions is different in important respects from our knowledge of the facts which are the basis of the deductions of the natural sciences. It may be admitted, too, that for this reason the methods of economic science—although not the 104 v ECONOMIC GENERALISATIONS AND REALITY 105 tests of its logical consistency—are often different from the methods of the natural sciences. But it does not follow in the least that its generalisations have a "merely formal" status—that they are "scholastic"

deductions from arbitrarily established definitions. Indeed, it may be urged that, on the contrary, there is less reason to doubt their real bearing than that of the generalisations of the natural sciences. In Economics, as we have seen, the ultimate constituents of our fundamental generalisations are known to us by immediate acquaintance. In the natural sciences they are known only inferentially. There is much less reason to doubt the counterpart in reality of the assumption of individual preferences than that of the assumption of the electron.1 It is true that we deduce much from definitions. But it is not true that the definitions are arbitrary. It follows, too, that it is a complete mistake to regard the economist, whatever his degree of "purity", as concerned merely with pure deduction. It is quite true that much of his work is in the nature of elaborate processes of inference. But it is quite untrue to suppose that it is only, or indeed mainly, thus. The concern of the economist is the interpretation of reality.

The business of discovery consists not merely in the elucidation of given premises but in the perception of the facts which are the basis of the premises. The process of discovering those elements in common experience which afford the basis of our trains of deductive reasoning is economic discovery just as much as the shaking out of new inferences from old premises. The theory of value as we know it has developed in 1 See the classical discussion of this matter in Cairnes' Character and Logical Method of Political Economy, 2nd edition, pp. 81-99. See also Hayek: Collectivist Economic Planning, pp. 8-12.

106 SIGNIFICANCE OF ECONOMIC SCIENCE OH. recent times by the progressive elaboration of deductions from very simple premises. But the great discovery, the Mengerian revolution, which initiated this period of progress, was the discovery of the premises themselves. Similarly with the other foundations we have discussed. The perception and selection of the basis of economic analysis is as much economics as the analysis itself. Indeed it is this which gives analysis significance. 2. At the same time it must be admitted that the propositions which have hitherto been established are very general in character. If a certain good is scarce, then we know that its disposal must conform to certain laws. If its demand schedule is of a certain order, then we know that with alterations of supply its price must move in a certain way. But, as we have discovered already,1 there is nothing in this conception of scarcity which warrants us in attaching it to any particular commodity. Our deductions do not provide any justification for saying that caviare is an economic good and carrion a disutility. Still less do they inform us concerning the intensity of the demand for caviare or the demand to be rid of carrion.

From the point of view of pure Economics these things are conditioned on the one side by individual valuations, and on the other by the technical facts of the given situation. And both individual valuations and technical facts are outside the sphere of economic uniformity. To use Strigl's expressive phrase, from the point of view of economic analysis, these things constitute the irrational element in our universe of discourse.2 But is it not desirable to transcend such limita1 See above, Chapter II., Sections 1, 2, 3. 2 Strigl, op. cü„ p. 18.

v ECONOMIC GENERALISATIONS AND REALITY 107 tions? Ought we not to wish to be in a position to give numerical values to the scales of valuation, to establish quantitative laws of demand and supply? This raises, in a slightly different form, some of the questions we left unanswered at the conclusion of the last chapter. No doubt such knowledge would be useful. But a moment's reflection should make it plain that we are here entering upon a field of investigation where there is no reason to suppose that uniformities are to be discovered. The "causes" which bring it about that the ultimate valuations prevailing at any moment are what they are, are heterogeneous in nature: there is no ground for supposing that the resultant effects should exhibit significant uniformity over time and space. No doubt there is a sense in which it can be argued that every random sample of the universe is the result of determinate causes. But there is no reason to suppose that the study of a random sample of random samples is likely to yield generalisations of any significance. That is not the procedure of the sciences. Yet that, or something very much like it, is the assumption underlying the expectation that the formal categories of economic analysis can be given substantial content of permanent and constant value.1 A simple illustration should make this quite clear.

Let us take the demand for herrings. Suppose we are confronted with an order fixing the price of herrings at a point below the price hitherto ruling in the market. Suppose we were in a position to say, "According to the researches of Blank (1907-1908) the elasticity of demand for the common herring (Clupea harengus) is l¯3; the present price-fixing order therefore may be 1 Note the qualification "permanent and constant value". Before the above conclusion is dismissed as too drastic, the remarks below on the positive value of investigations of this sort should be examined.

108 SIGNIFICANCE OF ECONOMIC SCIENCE CH. expected to leave an excess of demand over supply of two million barrels". How pleasant ifc would be to be able to say things like this! How flattering to our usually somewhat damaged self-esteem vis-a-vis the natural scientists! How impressive to big business! How persuasive to the general public! But can we hope to attain such an enviable position? Let us assume that in 1907-1908 Blank had succeeded in ascertaining that, with a given price change in that year, the elasticity of demand was l·3. Rough computations of this sort are not really very difficult and may have considerable utility for certain purposes. But what reason is there to suppose that he w¾s unearthing a constant law? No doubt the herring meets certain physiological needs which are capable of fairly accurate description, although it is by no means the only food capable of meeting these needs. The demand for herrings, however, is not a simple derivative of needs. It is, as it were, a function of a great many apparently independent variables. It is a function of fashion; and by fashion is meant something more than the ephemeral results of an Eat British Herrings campaign; the demand for herrings might be substantially changed by a change in the theological views of the economic subjects entering the market.

It is a function of the availability of other foods. It is a function of the quantity and quality of the population. It is a function of the distribution of income within the community and of changes in the volume of money. Transport changes will alter the area of demand for herrings. Discoveries in the art of cooking may change their relative desirability. Is it possible reasonably to suppose that coefficients derived from the observation of a particular herring market at a v ECONOMIC GENERALISATIONS AND REALITY 109 particular time and place have any permanent significance—save as Economic History"? Now, of course, by the aid of various devices it is possible to extend the area of observation over periods of time. Instead of observing the market for herrings for a few days, statistics of price changes and changes in supply and demand may be collected over a period of years and by judicious "doctoring" for seasonal movements, population change, and so on, be used to deduce a figure representing average elasticity over the period. And within limits such computations have their uses. They are a convenient way of describing certain forces operative during that period of history. As we shall see later on, they may provide some guidance concerning what may happen in the immediate future. Eough ideas relating to the elasticity of demand in particular markets are indeed essential if we are to make full use of the more refined tools of economic analysis. But they have no claim to be regarded as immutable laws. However accurately they describe the past, there is no presumption that they must continue to describe the future.

Things have just happened to be so in the past. They may continue to be so for a short time in the future. But there is no reason to suppose that their having been so in the past is the result of the operation of homogeneous causes, nor that their changes in the future will be due to the causes which have operated in the past. If we wanted to be helpful about herrings we should never dream of relying on the researches of the wretched Blank who was working in 1907-8. We should work the whole thing out afresh on the basis of more recent data. Important as such investigations may be—and nothing that is here said on their method110 SIGNIFICANCE OF ECONOMIC SCIENCE CH. ological status should be regarded as derogating from their very considerable practical value—there is no justification for claiming for their results the status of the socalled "statistical" laws of the natural sciences.1 But, it might be said, is not the difference between the results of such investigation and the postulates on which, as was shown in the last chapter, the main generalisation of Economics depends, a matter only of degree rather than kind? It has been shown that if there were not a hierarchy of ends, but if the different ends were all of equal importance, the results of conduct would be quite indeterminate, and even the most elementary generalisations of the theory of value would not be applicable. There is no guarantee that this will not happen. It is only a matter of probability that the conditions making such propositions applicable will persist. In exactly the same way it can be shown analytically that circumstances are conceivable in which the demand curve may have a positive inclination. Yet if this were frequent many of the best accepted generalisations of deductive theory would not be applicable. Again, it is only a matter of probability that this is not the case. Wherein is the difference of kind between this assumption and the assumption that the elasticity of demand for herrings is l·3?

The argument is weighty. And it may be freely conceded that in this sense the difference is a difference of degree rather than a difference of kind. But it is surely open to the reply that the difference of degree is so great as to justify our acting as if it were a 1 On the problems discussed above very interesting remarks are to be found in Halberstaédter, Die ProbUmatik des wirtschafüichen Prinzips.

v ECONOMIC GENERALISATIONS AND KEALITY 111 difference of kind. It might be the case that valuations were of such a peculiar nature that conduct was indeterminate. But it is so overwhelmingly unlikely that we are warranted in neglecting the possibility. It is not so unlikely that the demand function may be positive, but there is still a very strong probability that this is not the rule, but the exception. On the other hand, when we are dealing with the valuation of particular products and the elasticity of demand derived therefrom, for the reasons already set forth, there is surely an overwhelming probability that constancy is not to be expected. Here, indeed, we have the historicorelative in excelsis. The fact that we can arrange our preferences in an order is a fact of so much greater a degree of generality than the actual momentary order of preference of any individual that we are surely justified in regarding them as possessing, in our universe of discourse at least, a difference of status.

And while it is arguable that in the future much valuable work will be done in attempting to ascertain these momentary values, it seems more important, if a sense of proportion is to be maintained, that their limitations should be realised than that stress should be laid on the formal similarity with the broad qualitative foundations which constitute the basis of the science as we know it. Perhaps, indeed, this is another of the methodological differences between the natural and the social sciences. In the natural sciences the transition from the qualitative to the quantitative is easy and inevitable. In the social sciences, for reasons which have already been set forth, it is in some connection almost impossible, and it is always associated with peril and difficulty. It seems clear, from what has happened already, that less harm is likely 112 SIGNIFICANCE OF ECONOMIC SCIENCE CH. to be done by emphasising the differences between the social and the natural sciences than by emphasising their similarities.1 3. If this is true of attempts to provide definite quantitative values for such elementary concepts as demand and supply functions, how much more does it apply to attempts to provide "concrete" laws of the movement of more complex phenomena, price fluctuations, cost dispersions, business cycles, and the like.

In the last ten years there has been a great multiplication of this sort of thing under the name of Institutionalism, "Quantitative Economics", "Dynamic Economics ", and what not ;2 yet most of the investigations involved have been doomed to futility from the outset and might just as well never have been undertaken. The theory of probability on which modern mathematical statistics is based affords no justification for averaging where conditions are obviously not such as to warrant the belief that homogeneous causes of different kinds are operating. Yet this is the normal procedure of much of the work of this kind. The correlation of trends subject to influences of the most diverse character is scrutinised for "quantitative laws". Averages are taken of phenomena occurring under the most heterogeneous circumstances of time and space, and the result is expected to have significance. In Professor Wesley Mitchell's Business Cycles,3 1 On the matters discussed in this section I am much indebted to conversations with Dr. Machlup.

2 On the aspect of Institutionalism discussed below, Professor Wesley Mitchell's essay on '1 he Prospects of Economics in the Trend of Economics (edited Tugwell) should be consulted. On the general position of the school, Bee Morgenstem, Bemerkungen úber die Problematik der Amerikanischen Inslitutionalisten in the Saggi di Storia e Teoria Economica in onore e recordo diO¡useppe Praia, Turin, 1931; Fetter, art. America, Wirtschaflstheorie der Oegenwart, Bd. 1, pp. 31-60. See also the review of the Trend of Economics by the late Professor Allyn Young, reprinted in his Economic Problems New and Old, pp. 232-260. 3 Business Cycles, 2nd edition, p. 419.

v ECONOMIC GENERALISATIONS AND REALITY 113 for instance, a work for whose magnificent collection of data economists are rightly grateful, after a prolonged and valuable account of the course of business fluctuations in different countries since the end of the eighteenth century, an average is struck of the duration of all cycles and a logarithmic normal curve is fitted by Davies' Method to the frequency distribution of the 166 observations involved. What possible meaning can inhere in such an operation? Here are observations of conditions widely differing in time, space, and the institutional framework of business activity. If there is any significance at all in bringing them together, it must be by way of contrast. Yet Professor Mitchell, who never tires of belittling the methods and results of orthodox analysis, apparently thinks that, by taking them all together and fitting a highly complicated curve to their frequency distribution, he is constructing something significant—something which is more than a series of straight lines and curves on half a page of his celebrated treatise.1 Certainly he has provided the most mordant comment on the methodology of "Quantitative Economics"

that any of its critics could possibly wish. There is no need to linger on the futility of these grandiose projects. After all, in spite of their recent popularity, they are not new, and a movement which has continually invoked a pragmatic logic may well be judged by a pragmatic test. It is just about a 1 On this see Morgenstern, International vergleichende Konjunktur}orschung(Zeitschrift fur die Gesammte Staatswisse7ischaft, vol.lxxxiii.,p.261). In the second edition of his book. Professor Mitchell attempts to meet Dr. Morgenstern's strictures in an extensive footnote, but so far as I can see, beyond urging that his observations for China relate to coast towns (!), he does not go beyond a reiteration that "the distribution of the observations around their central tendency is a matter of much theoretical interest" (Bus¡msa Cycles, 2nd edition, p. 420). 8 114 SIGNIFICANCE OF ECONOMIC SCIENCE OH.

hundred years ago since Richard Jones, in his Inaugural Lecture at King's College, London,1 sounded the note of revolt against the "formal abstraction" of Eicardian Economics, with arguments which, if more vividly expressed, are more or less exactly similar to those which have been expressed by the advocates of "inductive methods" ever since that day. And time has gone on, and the "rebels" have become a highly respectable band of expert authorities, the pontifical occupants of chairs, the honoured recipients of letters from the Kaiser, the directing functionaries of expensive research institutes. . . . We have had the Historical School. And now we have the Institutionalists. Save in one or two privileged places, it is safe to say that, until the close of the War, views of this sort were dominant in German University circles; and in recent years, if they have not secured the upper hand altogether, they have certainly had a wide area of power in America. Yet not one single "law" deserving of the name, not one quantitative generalisation of permanent validity has emerged from their efforts.

A certain amount of interesting statistical material. Many useful monographs on particular historical situations. But of "concrete laws", substantial uniformities of "economic behaviour", not one—all the really interesting applications of modern statistical technique to economic enquiry have been carried through, not by the Institutionalists, but by men who have been themselves adept in the intricacies 1 Richard Jones, Collected Works, pp. 21 and 22. The comparison is not altogether fair to Jones, who may have been very well justified in some of his criticisms of the Ricardian system. The true precursor of modern "Quantitative Economics" was Sir Josiah Child, who attempted to prove that the concomitance of low interest rates and great riches was an indication that the latter was the result of the former.

v ECONOMIC GENERALISATIONS AND REALITY 115 of the "orthodox" theoretical analysis. And, at the end of the hundred years, the greatest slump in history finds them sterile and incapable of helpful comment—their trends gone awry and their dispersions distorted.1 Meanwhile, a few isolated thinkers, using the despised apparatus of deductive theory, have brought our knowledge of the theory of fluctuations to a point from which the fateful events of the last few years can be explained in general terms, and a complete solution of the riddle of depressions within the next few years does not seem outside the bounds of probability. 4. But what, then, are we to say of the more detailed kind of realistic studies? Having ascertained the persistence of the fact of scarcity, the multiplicity of factors of production, ignorance of the future, and the other qualitative postulates of his theory, is the economist then excused from the obligation of maintaining further contact with reality?

The answer is most decidedly in the negative. And the negative answer is implicit in the practice of all those economists who, since Adam Smith and Cantillon, have contributed most to the development of Economic Science. It has never been the case that the exponents of the socalled orthodox tradition have frowned upon realistic studies. As Menger pointed out years ago, at the height of the MetL·äenstreit,z the analytical school have never been the assailants in these controversies. Economics is not one of those 1 The discredit of the Historical School in Germany is very largely due to the failure of its members to understand the currency disturbances of the War and the postWar period. It is not improbable that the utter failure of " Quantitative Economics" to understand or predict the great depression may be followed by a similar revulsion. It would certainly be difficult to imagine a more complete or more conspicuous exposure.

2 Die Irrthümer des IHstorismv,s, Preface, pp. iii. and iv.

116 SIGNIFICANCE OF ECONOMIC SCIENCE OH. social sciences which are always discussing method before proceeding to deliver the goods; if it had not been for the Historical School there would have been no methodological controversy save such as related to the status of particular propositions. The procedure of "orthodoxy" has always been essentially catholic. The attacks, the attempts to exclude, have always come from the other side. The analytics have always acknowledged the importance of "realistic" studies, and have themselves contributed much to the development of the technique of investigation. Indeed, it is notorious that the most important work of this kind has come, not from this or that "rebel" group who were calling in question the application in Economics of the elementary laws of thought, but rather from just those men who were the object of their onslaught. In the history of applied Economics, the work of a Jevons, a Menger. a Bowley, has much more claim on our attention than the work of, say, a Schmoller, a Veblen, or a Hamilton. And this is no accident. The fruitful conduct of realistic investigations can only be undertaken by those who have a firm grasp of analytical principle and some notion of what can and what cannot legitimately be expected from activities of this sort.

But what, then, are legitimate expectations in this respect? We may group them under three headings. The first and the most obvious is the provision of a check on the applicability to given situations of different types of theoretical constructions. As we have seen already, the validity of a particular theory is a matter of its logical derivation from the general assumptions which it makes. But its applicability to a v ECONOMIC GENEEALISATIONS AND EEALITY 117 given situation depends upon the extent to which its concepts actually reflect the forces operating in that situation. Now the concrete manifestations of scarcity are various and changing; and, unless there is continuous check on the words which are used to describe them, there is always a danger that the area of application of a particular principle may be misconceived. The terminology of theory and the terminology of practice, although apparently identical, may, in fact, cover different areas.

A simple illustration will make this clear. According to pure monetary theory, if the quantity of money in circulation is increased and other things remain the same, the value of money must fall. This proposition is deducible from the most elementary facts of experience of the science, and its truth is independent of further inductive test. But its applicability to a given situation depends upon a correct understanding of what things are to be regarded as money; and this is a matter which can only be discovered by reference back to the facts. It may well be that over a period of time the concrete significance of the term "money" has altered. If then, while retaining the original term, we proceed to interpret a new situation in terms of the original content, we may be led into serious misapprehension. We may even conclude that the theory is fallacious. It is indeed well known that this has happened again and again in the course of the history of theory. The failure of the Currency School to secure permanent acceptance for their theory of Banking and the Exchanges, in other respects so greatly superior to that of their opponents, was notoriously due to their failure to perceive the importance of including Bank Credit in their conception of money. Only by continuous 118 SIGNIFICANCE OF ECONOMIC SCIENCE OH.

sifting and scrutiny of the changing body of facts1 can such misapprehensions be avoided. Secondly, and closely connected with this first function of realistic studies, we may expect the suggestion of those auxiliary postulates whose part in the structure of analysis was discussed in the last chapter By inspection of different fields of economic activity we may expect to discover types of the configuration of the data suitable for further analytical study. Again, we may take an example from the theory of money. It will be clear from an inspection of the actual procedure of banks of issue that the efEect upon the supply of money in the widest sense of given additions to the reserve of precious metals will depend upon the exact nature of the law and practice concerning reserve requirements. It follows, therefore, that in the full elaboration of the theory of money we must introduce alternative assumptions, taking account of the various possibilities in this respect. It is clear that these are not possibilities which are necessarily easily exhausted by general reflections on the nature of banks of issue. Only close study of the facts is likely to reveal which assumptions are most likely to have a counterpart in reality, which assumptions, therefore, it is most convenient to make.

And, thirdly, we may expect of realistic studies, not merely a knowledge of the application of particular theories, and the assumptions which make them appropriate to particular situations, but also the exposure of areas where pure theory needs to be reformulated and extended. They bring to light new problems. 1 Professor Jacob Viner's Canadian Balance of International Indebtedness and Professor Taussig's International Trade provide classical examples of this kind of investigation.

v ECONOMIC GENERALISATIONS AND REALITY 119 The best example of the unexplained residue is provided by those fluctuations of trade which have come to be known as the trade cycle. Elementary equilibrium theory, as is well known, does not provide any explanation of the phenomena of booms and slumps. It explains the relationships in an economic system on a state of rest. As we have seen, with a certain extension of its assumptions it can describe differences between the relationships resulting from different configurations of the data. But it does not explain without further elaboration the existence within the economic system of tendencies conducive to disproportionate development. It does not explain discrepancies between total supply and total demand in the sense in which these terms are used in the celebrated Law of Markets.1 Yet unquestionably such discrepancies exist, and any attempt to interpret reality solely in terms of such a theory must necessarily leave a residue of phenomena not capable of being subsumed under its generalisations.

Here is a clear case where empirical studies bring us face to face with the insufficiencies of certain generalisations. And it is perhaps in the revelation of deficiencies of this kind that the main function of realistic studies in relation to theory consists.2 The theoretical economist who wishes to safeguard the implications of his theory must be continually "trying out", in the explanation of particular situations, the generalisations he has already achieved. It is in the examination of particular instances that lacunæ in the structure of existing theory tend to be revealed. 1 On all this see Hayek, Monetary Theory and the Trade Cycle, chaps i. and ii., passim. 2 Another important function, this time in relation to practice, will be discussed in the next section.

120 SIGNIFICANCE OF ECONOMIC SCIENCE OH. But this is not in the least to say that the solutions of the problems thus presented are themselves to be discovered by the mere multiplication of observations of divergences of this sort. That is not the function of observation, and the whole history of the various "inductive revolts" shows that all studies based on this expectation have proved utterly fruitless. This is particularly true of trade cycle theory. So long as the investigators of this problem were content with the multiplication of time series and the accumulation of coefficients of correlation, no significant advance was discernible. It was not until there arose men who were prepared to undertake the entirely different task of starting where elementary theoretical analysis leaves off and deriving from the introduction of further assumptions of the elementary qualitative nature we have already examined, an explanation of fluctuation which is compatible with the assumptions of that analysis, that progress began to be made.

There can be no better example of the correct relationship between the two branches of study. Realistic studies may suggest the problem to be solved. They may test the range of applicability of the answer when it is forthcoming. They may suggest assumptions for further theoretical elaboration. But it is theory and theory alone which is capable of supplying the solution. Any attempt to reverse the relationship must lead inevitably to the nirvana of purposeless observation and record. Moreover—and this brings us back to the point from which we started—there is no reason to believe that the generalisations which may be elaborated to explain the residues thus discovered will be anything but general in character. For reasons which we have v ECONOMIC GENERALISATIONS AND REALITY 121 already examined, the hope of giving permanent and particular content to the categories of pure analysis is vain. By "trying out" pure theory on concrete situations and referring back to pure theory residual difficulties, we may hope continually to improve and extend our analytical apparatus. But that such studies should enable us to say what goods must be economic goods and what precise values will be attached to them in different situations, is not to be expected. To say this is not to abandon the hope of solving any genuine problem of Economics. It is merely to recognise what does and what does not lie within the necessary boundaries of our subjectmatter.

To pretend that this is not so is just pseudo-scientific bravado. 5. But to recognise that Economic laws are general in nature is not to deny the reality of the necessities they describe or to derogate from their value as a means of interpretation and prediction. On the contrary, having carefully delimited the nature and the scope of such generalisations, we may proceed with all the greater confidence to claim for them a complete necessity within this field. Economic laws describe inevitable implications. If the data they postulate are given, then the consequences they predict necessarily follow. In this sense they are on the same footing as other scientific laws, and as little capable of "suspension". If, in a given situation, the facts are of a certain order, we are warranted in deducing with complete certainty that other facts which it enables us to describe are also present. To those who have grasped the implications of the propositions set forth in the last chapter the reason is not far to seek. If the "given situation"

122 SIGNIFICANCE OF ECONOMIC SCIENCE CH. conforms to a certain pattern, certain other features must also be present, for their presence is "deducible" from the pattern originally postulated. The analytic method is simply a way of discovering the necessary consequences of complex collocations of facts—consequences whose counterpart in reality is not so immediately discernible as the counterpart of the original postulates. It is an instrument for "shaking out" all the implications of given suppositions. Granted the correspondence of its original assumptions and the facts, its conclusions are inevitable and inescapable. All this becomes particularly clear if we consider the procedure of diagrammatic analysis. Suppose, for example, we wish to exhibit the effects on price of the imposition of a small tax. We make certain suppositions as regards the elasticity of demand, certain suppositions as regards the cost functions, embody these in the usual diagram, and we can at once read off, as it were, the effects on the price.1 They are implied in the original suppositions. The diagram has simply made explicit the concealed implications.

It is this inevitability of economic analysis which gives it its very considerable prognostic value. It has been emphasised sufficiently already that Economic Science knows no way of predicting out of the blue the configuration of the data at any particular point of time. It cannot predict changes of valuations. But, given the data in a particular situation, it can draw inevitable conclusions as to their implications. And if the data remain unchanged, these implications will certainly be realised. They must be, for they are implied in the presence of the original data. 1 See, e.g., Dalton, Public Finance, 2nd edition, p. 73.

v ECONOMIC GENERALISATIONS AND REALITY 123 It is just here that we can perceive yet a further function for empirical investigation. It can bring to light the changing facts which make prediction in any given situation possible. As we have seen, it is most improbable that it can ever discover the law of their change, for the data are not subject to homogeneous causal influences. But it can put us in possession of information which is relevant at the particular moment concerned. It can give us some idea of the relative magnitude of the different forces operative. It can afford a basis for enlightened conjectures with regard to potential directions of change. And this unquestionably is one of the main uses of applied studies—not to unearth "empirical" laws in an area where such laws are not to be expected, but to pro` vide from moment to moment some knowledge of the varying data on which, in the given situation, prediction can be based. It cannot supersede formal ' analysis. But it can suggest in different situations what formal analysis is appropriate, and it can provide at that moment some content for the formal categories.

Of course, if other things do not remain unchanged, the consequences predicted do not necessarily follow. This elementary platitude, necessarily implicit in any scientific prediction, needs especially to be kept in the foreground of attention when discussing this kind of prognosis. The statesman who said "Ceteris paribus be damned!" has a large and enthusiastic following among the critics of Economics! Nobody in his senses would hold that the laws of mechanics were invalidated if an experiment designed to illustrate them were interrupted by an earthquake. Yet a substantial majority of the lay public, and a good many soi-disant economists as well, are continually criticising well124 SIGNIFICANCE OF ECONOMIC SCIENCE OH. established propositions on grounds hardly less slender.1 A protective tariff is imposed on the importation of commodities, the conditions of whose domestic production make it certain that, if other things remain unchanged, the effect of such protection will be a rise in price. For quite adventitious reasons, the progress of technique, the lowering of the price of raw materials, wage reductions, or what not, costs are reduced and the price does not rise. In the eyes of the lay public and "Institutionalise economists the generalisations of Economics are invalidated. The laws of supply and demand are suspended. The bogus claims of a science which does not regard the facts are laid bare. And so on and so forth. Yet, whoever asked of the practitioners of any other 1 See, e.g., the various statistical "refutations" of the quantity theory of money which have appeared in recent years. On all these the comment of Torrens on Tooke is all that need be said. "The History of Prices may be regarded as a psychological study. Mr. Tooke commenced his labours as a follower of Homer and Ricardo, and derived reflected lustre from an alliance with those celebrated names; but his capacity for collecting contemporaneous facts preponderating over his perceptive and logical faculties, his accumulation of facts involved him in a labyrinth of error. Failing to perceive that a theoretical principle, although it may irresistibly command assent under all circumstances coinciding with the premises from which it is deduced, must be applied with due limitation and correction in all cases not coinciding with the premises, he fell into a total misconception of the proposition advanced by Adam Smith, and imputed to that high authority the absurdity of maintaining that variations in the quantity of money cause the money values of all commodities to vary in equal proportions, while the values of commodities, in relation to each other, are varying in unequal proportions. Reasonings derived from this extraordinary misconception necessarily led to extraordinary conclusions. Having satisfied himself that Adam Smith had correctly established as a principle universally true that variations in the purchasing power of money cause the prices of all commodities to vary in equal proportions, and finding, as he pursued his investigations into the phenomena of the market at different periods, no instances in which an expansion or contraction of the circulation caused the prices of commodities to rise or fall in an equal ratio, he arrived by a strictly logical inference from the premises thus illogically assumed, at his grand discovery—that no increase of the circulating medium can have the effect of increasing prices'' (The Principles and Operation of Sir Robert Peel's Act of 1844 Explained and Defended. 1st edition, p. 75).

v ECONOMIC GENERALISATIONS AND REALITY 125 science that they should predict the complete course of an uncontrolled history? Now, no doubt, the very fact that events in the large are uncontrolled,1 that the fringe of given data is so extensive and so exposed to influence from unexpected quarters, must make the task of prediction, however carefully safeguarded, extremely hazardous. In many situations, small changes in particular groups of data are so liable to be counterbalanced by other changes which may be occurring independently and simultaneously, that the prognostic value of the knowledge of operative tendencies is small. But there are certain broad changes, usually involving many lines of expenditure or production at once, where a knowledge of implications is a very firm basis for conjectures of strong probability. This is particularly the case in the sphere of monetary phenomena. There can be no question that a quite elementary knowledge of the Quantity Theory was of immense prognostic value during the War and the disturbances which followed. If the speculators who bought German marks, after the War, in the confident expectation that the mark would automatically resume its old value, had been aware of as much of the theory of money as was known, say, to Sir William Petty, they would have known that what they were doing was 1 The alleged advantage of economic "planning"—namely, that it enables greater certainty with regard to the future—depends upon the assumption that under "planning" the present controlling forces, the choices of individual spenders and savers, are themselves brought under the control of the planners. The paradox therefore arises that either the planner is destitute of the instrument of calculating the ends of the community he intends to serve, or, if he restores the instrument, he removes the raison d'etre of the "plan". Of course, the dilemma does not arise if he thinks himself capable of interpreting these ends or—what is much more probable— if he has no intention of serving any other ends but those lie thinks appropriate. Strange to say this not infrequently happens. Scratch a would-be planner and you usually find a woulb-be dictator.

126 SIGNIFICANCE OF ECONOMIC SCIENCE OH. ridiculous. Similarly, it becomes more and more clear, for purely analytical reasons, that, once the signs of a major boom in trade have made their appearance, the coming of slump and depression is almost certain; though when it will come and how long it will last are not matters which are predictable, since they depend upon human volitions occurring after the indications in question have appeared. So, too, in the sphere of the labour market, it is quite certain that some types of wage policy must result in unemployment if other things remain equal: and knowledge of how the "other things" must change in order that this consequence may be avoided makes it very often possible to predict with considerable confidence the actual results of given policies. These things have been verified again and again in practice. Today it is only he who is blind because he does not want to see who is prepared to deny them. If certain conditions are present, then, in the absence of new complications, certain consequences are inevitable.

6. None the less, economic laws have their limits, and, if we are to use them wisely, it is important that we should realise exactly wherein these limitations consist. In the light of what has been said already, this should not be difficult. The irrational element in the economist's universe of discourse lies behind the individual valuation. As we have seen already, there is no means available for determining the probable movement of the relative scales of valuation.1 Hence in all our analysis we take 1 It should be observed that this is not the same as saying that there is no means available for defining the probable movement of the demand curve. It is important to realise that the demand curve is to be conceived as derived from the more fundamental indifference system, and it is to this latter that our proposition relates.

v ECONOMIC GENERALISATIONS AND REALITY 127 the scales of valuation as given. It is only what follows from these given assumptions that has the character of inevitability. It is only in this area that we find the regime of law. It follows, therefore, that economic laws cannot be held to relate to movements of the relative scales, and that economic causation only extends through the range of their original implication. This is not to say that changes in values may not be contemplated. Of course, changes in values are the main preoccupation of theoretical Economics. It is only to say that, as economists, we cannot go behind changes in individual valuations. We may explain, in terms of economic law, relationships which follow from given technical conditions and relative valuations. We may explain changes due to changes in these data. But we cannot explain changes in the data themselves. To demarcate these types of change the Austrians1 distinguish between endogenous and exogenous changes. The ones occur within a given structure of assumptions. The others come from outside.

We can see the relevance of these distinctions to the problem of prognosis if we consider once more the implications of the theory of money. Given certain assumptions with regard to the demand for money, we are justified in asserting that an increase in the volume of any currency will be followed by a fall in its external value. This is an endogenous change. It follows from the original assumptions, and, so long as they hold, it is clearly inevitable. We are not justified in asserting, however, as has been so often asserted 1 See especially Strigl, Aenderungen in den Daten der Wirtschaft (Jahrbücherfür Nationak>kcmomie und Statistiìc, vol. oxxviii., pp. 641-662).

ÌŽ8 SIGNIFICANCE Oï ECONOMIC SCIENCE <* in recent years, that if the exchanges fall, inflation must necessarily follow. We know that very often this happens. We know that governments are often foolish and craven and that false views of the functions of money are widely prevalent. But there is no inevitable connection between a fall in the exchanges and a decision to set the printing presses working. A new human volition interrupts the chain of "causation". But between the issue of paper money and the fall in its external value, no change in the assumed disposition to action on the part of the various economic subjects concerned is contemplated. All that happens is, as it were, that the exchange index moves to a lower level. A more complicated example of the same distinction is provided by the Reparations controversy. Suppose that it could be shown that the external demand for German products was very inelastic, so that in the short period, at any rate, the degree of necessary transfer burden over and above the burden of paying the domestic taxes was very great. In such circumstances it might be argued that the present crisis was directly due to purely economic factors.

That is to say that, up to the point at which panic supervened, the various complications were entirely due to obstacles implicit in the given conditions of world supply and demand.1 But suppose it can be shown that the prime cause of the present difficulty was financial panic, induced by the fear of political revolt at the magnitude of the original tax burden, then it cannot be argued that the train of causation was wholly economic. The political reaction to the 1 This is the limiting case discussed in Dr. Machlup's Transfer und Prtübewegung (Zeitschrift far Nationcdolamomie, vol.i., pp. ðöõ-5(!l).

v ECONOMIC GENERALISATIONS AND REALITY 129 tax burden intervenes. The "transfer crisis" arises from exogenous causes.1 Now there can be no doubt that this distinction is not always easy to draw. In some cases there may be a functional connection between rates of remuneration and increments of the quantity and the quality of the working population. How is this to be regarded? So far as the response is concerned, it is endogenous. But so far as the configuration of the market demand is concerned, it is exogenous. New people with new scales of relative valuation appear. Again, as Professor Knight has often pointed out, the situation is further complicated by the fact that in some societies there exist definite financial incentives to certain individuals to produce changes in the data. Resources are devoted to changing technical knowledge by research, and the tastes of economic subjects by persuasion. In respect of such changes the distinction is difficult to apply. We must admit that the system is "open". Nevertheless, over a large part of the field the classification is intelligible enough and a positive aid to clear thinking. Until matters have been clarified very much further its retention seems essential.

1 Professor Souter says that words fail him to describe the type of mind that takes any pleasure in drawing such distinctions (op. cü., p. 139). But surely, methodological considerations apart, there are very solid reasons of convenience for observing them. I venture to suggest that if Professor Souter had been asked to advise any Government on such questions there would have come a point at which, having diagnosed the "economic" factors, he would have turned and said, "But then, of course, there is the political problem; will people stand it?" And he might have added with Cantillon, "But that ia not my business". Or, as true blue Hegelian, taking all knowledge for his province, he might have then launched forth on a disquisition of what is and what is not politically possible. But he would have made the distinction. Exactly how he labelled it we might argue about in a friendly way afterwards.

9 130 SIGNIFICANCE OF ECONOMIC SCIENCE OH. In the same way it should be recognised that in the discussion of practical problems, certain kinds of exogenous changes, apparently closely connected with changes within the chain of economic causation, are not infrequently involved. In the sphere of monetary problems the danger that falling exchanges may induce the monetary authorities of the area involved to embark on inflation, will certainly be considered germane to the discussion. In the sphere of tarifl policy, the tendency of the granting of a protective tariff to create monopolistic communities of interest among domestic producers is certainly a probability which should not be overlooked by the practical administrator. Here and in many other connections there is a penumbra of psychological probabilities which, for purely practical reasons, it is often very convenient to take into account.1 No doubt the kind of insight required into these problems is often of a very elementary order—although it is surprising how many people lack it. No doubt most of the probabilities involved are virtual certainties. Men in possession of their senses are not likely to question them as working maxims of political practice. Still, not all 1 I venture, as in the first edition, to draw attention to the actual words used in this prescription. It is more accuracy in mode of statement, not over-austerity in speculative range, for which I am pleading. I am very far from suggesting that, when discussing practical problems, economists should refrain from contemplating the probability of those changes in the data whose causation falls outside the strict limits of Economic Science. Indeed, I am inclined to believe that there is here a field of sociological speculation in which economists may have a definite advantage over others. Certainly it is a field in which hitherto they have done very much more than others—one has only to think of the various discussions of the possible forms of a Tariff Commission in a democratic community or the necessary conditions of bureaucratic administration of productive enterprise to see the sort of thing I have in mind. All that I am contendng is the desirability of recognising the distinction between the kind of generalisation which belongs to this field and the kind whicb belongs to Economics proper.

* ECONOMIC GENERALISATIONS AND REALITY 131 participants in discussions of this sort are in possession of their senses, and while it is highly desirable that the economist who wishes that the applications of his science should be fruitful should be fully qualified in cognate disciplines, and should be prepared to invoke their assistance, it is also highly desirable that the distinction should be recognised between those generalisations which are economic in the sense in which the word has here been used, and those generalisations of the "sociological penumbra" which do not have the same degree of probability. Economists have nothing to lose by understating rather than overstating the extent of their certainty. Indeed, it is only when this is done that the overwhelming power to convince of what remains can be expected to have free play. 7. All this has a very intimate bearing on the question which we left unanswered at the end of the last chapter. Is it not possible for us to extend our generalisations so as to cover changes of the data? We have seen in what sense it is possible to conceive of economic dynamics—the analysis of the path through time of a system making adjustments consequential upon the existence of given conditions? Can we not extend our technique so as to enable us to predict changes of these given conditions? In short, can we not frame a complete theory of economic development?

If the preceding analysis is correct the prospects are very doubtful. If we were able to ascertain once and for all the elasticities of demand for all possible commodities and the elasticities of supply of all factors, and if we could assume that these coefficients were constant, then we might indeed conceive of a grand calculation which would enable an economic Laplace 132 SIGNIFICANCE OF ECONOMIC SCIENCE OH. to foretell the economic appearance of our universe at any moment in the future. But, as we have seen, useful as such calculations are for judging the immediate potentialities of particular situations, there is no reason for attributing to them permanent validity. Our economic Laplace must fail in that there are no constants of this sort in his system. We have, as it were, to rediscover our various laws of gravitation from moment to moment. But is it not possible in a more formal sense to predict broad changes of the data? We may not be able to foretell particular tastes and the relationships between particular commodities, but by including in our conception of endogenous change, changes such as those indicated above, responses of the population to changes in income, induced invention, and so on, can we not still provide a formal outline of probable developments which shall be useful?

Now there is no doubt that so far as population change is concerned it is possible to conceive of movements as responsive to money incentives. We can conceive, as did the classical economists, of a final equilibrium in which the value of the discounted future remuneration of labour is equal to the discounted costs of bearing, rearing, and training labourers. It is doubtful whether it is very profitable to assume this particular functional connection in dealing with societies other than communities of slave owners. For, save in this case, it must be remembered that we are not entitled to assume, as did the classical economists at one stage, that the costs which are equated to the gains are objective in character: the equilibrium rate outside the slave society is that which will induce the constant supply of labourers, not merely that which v ECONOMIC GENERALISATIONS AND REALITY 133 makes it physiologically possible to support them.

Still, for what it is worth, such an assumption can be made. But even so we have only described in formal terms a condition of final equilibrium. We have done nothing which enables us to predict changes in the ultimate conditions of supply of labourers. The broad vicissitudes of opinion on the optional size of family or the most desirable entourage of slaves—these lie outside the scope of our technique of prediction. Who is to say whether the present influences on the size of the family, which bid fair, if they continue for a few millennia, to reduce the population of Europe to a few hundred thousands of people, will persist, or whether they will give way before the onset of new faiths, new conceptions of duty, new conceptions of the essentials of a good life? We may all venture our guess. But surely economic analysis can have very little to do with it. Nor are the prospects improved when we turn to the sphere of technical change and invention. As Professor Schumpeter has emphasised, it is very difficult to conceive even of equilibrium adjustments here.

Perhaps with some ingenuity it could be done. But how would that help us to predict—what would be necessary for a theory of development in the sense in which we are now using the word—the nature of the changes forthcoming? What technique of analysis could predict the trends of inventions leading on the one hand to the coming of the railway, on the other to the internal combustion engine. Even if we think that, if we know the technique, we can predict the type of economic relationship associated with it, which of course is highly disputable, how can we predict the 134 SIGNIFICANCE OF ECONOMIC SCIENCE OH. technique? As the examples just quoted amply illustrate, it is not at all true that the trend is all in one direction. We need postulate no ultimate indeterminism, if we assume that, from the point of view of our system, such changes are unpredictable. So, too, when we turn to the question of changes in the legal framework within which we conceive the adjustments we study to operate. There is an important sense in which the subjectmatter of political science can be conceived to come within the scope of our definition of the economic. Systems of government, property relationships, and the like, can be conceived as the result of choice. It is desirable that this conception should be further explored on lines analogous to better known analysis. But how can we tell in advance what choice will be made? How can we predict the substance of the political indifíerence systems?

It is well known that the claim has been made to interpret the evolution of political forms in terms of the distribution of "economic" power and the play of "economic" interest. And it would be foolish to deny that, within limits, elucidations of this sort can be provided which are at least intelligible. But on closer examination the limits within which this sort of thing is possible are seen to be much narrower than is often believed to be the case. We may perhaps explain particular political changes in terms of the "interest" of particular groups of producers; the machinery of the market affords at least a loose and superficial index of short period interest which is capable of objective definition. But the plausibility of the more grandiose explanations of this kind rest upon the assumption that the interests of larger groups are equally v ECONOMIC GENERALISATIONS AND REALITY 135 capable of objective definition. And this is not true.

So far from providing a justification for this kind of economic explanation, economic analysis suggests that it is definitely false. The concept of interest involved in all these explanations is not objective, but subjective. It is a function of what people believe and feel. And there is no technique in economics which enables us to forecast these permutations of the spirit. We may forecast their effects when they have occurred. We may speculate with regard to the effects of hypothetical changes. We may consider alternative forms and enquire concerning their stability and tendency to change. But as regards our actual capacity to foretell a process of change, with its manifest dependence on the heterogeneous elements of contingency, persuasion, and blind force, if we are humble, we shall be modest in our pretensions. Thus in the last analysis the study of Economics, while it shows us a region of economic laws, of necessities to which human action is subject, shows us, too, a region in which no such necessities operate. This is not to say that within that region there is no law, no necessity. Into that question we make no enquiry. It is only to say that from its point of view at least there are certain things which must be taken as ultimate data.