Any attempt either to forecast the trend of economic development, or to influence it by measures based on an examination of existing conditions, must presuppose certain quite definite conceptions as to the necessary course of economic phenomena. Empirical studies, whether they are undertaken with such practical aims in view, or whether they are confined merely to the amplification with the aid of special statistical devices of our knowledge of the course of particular phases of trade fluctuations, can at best afford merely a verification of existing theories; they cannot in themselves provide new insight into the causes or the necessity of the trade cycle.
This view has been stated very forcibly by Professor A. Lowe.
“Our insight into the theoretical interconnections of economic cycles, and into the structural laws of circulation,” he says, “has not been enriched at all by descriptive work or calculations of correlations.” We can entirely agree with him, moreover, when he goes on to say that “to expect an immediate furtherance of theory from an increase in empirical insight is to misunderstand the logical relationship between theory and empirical research.”
The reason for this is clear. The means of perception employed in statistics are not the same as those employed in economic theory; and it is therefore impossible to fit regularities established by the former into the structure of economic laws prescribed by the latter. We cannot superimpose upon the system of fundamental propositions comprised in the theory of equilibrium, a trade cycle theory resting on unrelated logical foundations. All the phenomena observed in cyclical fluctuations, particularly price formation and its influence on the direction and the volume of production, have already been explained by the theory of equilibrium; they can only be integrated as an explanation of the totality of economic events by means of fundamentally similar constructions. Trade cycle theory itself is only expected to explain how certain prices are determined, and to state their influence on production and consumption; and the determining conditions of these phenomena are already given by elementary theory. Its special task arises from the fact that these phenomena show empirically observed movements for the explanation of which the methods of equilibrium theory are as yet inadequate. One need not go so far as to say that a successful solution could be reached only in conjunction with a positive explanation of elementary phenomena; but no further proof is needed that such a solution can only be achieved in association with, or by means of, a theory that explains how certain prices or certain uses of given goods are determined at all. It is not only that we lack theories that fulfill this condition and that fall outside the category best described as “equilibrium theories”—theories that are characterized by taking the logic of economic action as their starting point; the point is rather that statistical method is fundamentally unsuited to this purpose. Just as no statistical investigation can prove that a given change in demand must necessarily be followed by a certain change in price, so no statistical method can explain why all economic phenomena present that regular wave-like appearance we observe in cyclical fluctuations. This can be explained only by widening the assumptions on which our deductions are based, so that cyclical fluctuations would follow from these as a necessary consequence, just as the general propositions of the theory of price followed from the narrower assumptions of equilibrium theory.
But even these new assumptions cannot be established by statistical investigation. The statistical approach, unlike deductive inference, leaves the conditions under which established economic relations hold good fundamentally undetermined; and similarly, the objects to which they relate cannot be determined as unequivocally as by theory. Empirically established relations between various economic phenomena continue to present a problem to theory until the necessity for their interconnections can be demonstrated independently of any statistical evidence. The concepts on which such an explanation is based will be quite different from those by which statistical interconnections are demonstrated; they can be reached independently. Moreover, the corroboration of statistical evidence provides, in itself, no proof of correctness. A priori we cannot expect from statistics anything more than the stimulus provided by the indication of new problems.
In thus emphasizing the fact that trade cycle theory, while it may serve as a basis for statistical research, can never itself be established by the latter, it is by no means desired to deprecate the value of the empirical method. On the contrary, there can be no doubt that trade cycle theory can only gain full practical importance through exact measurement of the actual course of the phenomena it describes. But before we can examine the question of the true importance of statistics to theory, it must be clearly recognized that the use of statistics can never consist in a deepening of our theoretical insight.
Even as a means of verification, the statistical examination of the cycles has only a very limited value for trade cycle theory. For the latter—as for any other economic theory—there are only two criteria of correctness. First, it must be deduced with unexceptionable logic from the fundamental notions of the theoretical system; and second, it must explain by a purely deductive method those phenomena with all their peculiarities that we observe in the actual cycles. Such a theory could only be “false” either through an inadequacy in its logic or because the phenomena it explains do not correspond with the observed facts. If, however, the theory is logically sound, and if it leads to an explanation of the given phenomena as a necessary consequence of these general conditions of economic activity, then the best that statistical investigation can do is to show that there still remains an unexplained residue of processes. It could never prove that the determining relationships are of a different character from those maintained by the theory.
It might be shown, for instance, by statistical investigation that a general rise in prices is followed by an expansion of production, and a general fall in prices by a diminution of production; but this would not necessarily mean that theory should regard the movement of price as an independent cause of movements of production. So long as a theory could explain the regular occurrence of this parallelism in any other way, it could not be disproved by statistics, even if it maintained that the connection between the two phenomena was of a precisely opposite nature. It is therefore only in a negative sense that it is possible to verify theory by statistics. Either statistics can demonstrate that there are phenomena the theory does not sufficiently explain, or it is unable to discover such phenomena. It cannot be expected to confirm the theory in a positive sense. The possibility is completely ruled out by what has been said above, since it would presuppose an assertion of necessary interconnections, such as statistics cannot make. There is no reason to be surprised, therefore, that although nearly all modern trade cycle theories use statistical material as corroboration, it is only where a given theory fails to explain all the observed phenomena that this statistical evidence can be used to judge its merits.
Thus it is not by enriching or by checking theoretical analysis that economic statistics gain their real importance. This lies elsewhere. The proper task of statistics is to give us accurate information about the events that fall within the province of theory, and so to enable us not only to connect two consecutive events as cause and effect, a posteriori, but to grasp existing conditions completely enough for forecasts of the future and, eventually, appropriate action, to become possible. It is only through this possibility of forecasts of systematic action that theory gains practical importance. A theory might, for instance, enable us to infer from the comparative movements of certain prices and quantities an imminent change in the direction of those movements: but we should have little use for such a theory if we were unable to ascertain the actual movements of the phenomena in question. With regard to certain phenomena having an important bearing on the trade cycle, our position is a peculiar one. We can deduce from general insight how the majority of people will behave under certain conditions; but the actual behavior of these masses at a given moment, and therefore the conditions to which our theoretical conclusions must be applied, can only be ascertained by the use of complicated statistical methods. This is especially true when a phenomenon is influenced by a number of partly known circumstances, such as, e.g., seasonal changes. Here very complicated statistical investigations are needed to ascertain whether these circumstances whose presence indicates the applicability of theoretical conclusions were in fact operative. Often statistical analysis may detect phenomena that have, as yet, no theoretical explanation, and which therefore necessitate either an extension of theoretical speculation or a search for new determining conditions. But the explanation of the phenomena thus detected, if it is to serve as a basis for forecasts of the future, must in every case utilize other methods than statistically observed regularities; and the observed phenomena will have to be deduced from the theoretical system, independently of empirical detection.
The dependence of statistical research on preexisting theoretical explanation hardly needs further emphasis. This holds good not only as regards the practical utilization of its results, but also in the course of its working, in which it must look to theory for guidance in selecting and delimiting the phenomena to be investigated. The oft-repeated assertion that statistical examination of the trade cycle should be undertaken without any theoretical prejudice is therefore always based on selfdeception.
On the whole, one can say without exaggeration that the practical value of statistical research depends primarily upon the soundness of the theoretical conceptions on which it is based. To decide upon the most important problems of the trade cycle remains the task of theory; and whether the money and labor so freely expended on statistical research in late years will be repaid by the expected success depends primarily on whether the development of theoretical understanding keeps pace with the exploration of the facts. For we must not deceive ourselves: not only do we now lack a theory that is generally accepted by economists, but we do not even possess one that could be formulated in such an unexceptionable way, and worked out in such detail, as eventually to command such acceptance. A series of important interconnections have been established and some principles of the greatest significance expounded; but no one has yet undertaken the decisive step that creates a complete theory by using one of these principles to incorporate all the known phenomena into the existing system in a satisfactory way. To realize this, of course, does not hinder us from pursuing either economic research or economic policy; but then we must always remember that we are acting on certain theoretical assumptions whose correctness has not yet been satisfactorily established. The “practical man” habitually acts on theories that he does not consciously realize; and in most cases this means that his theories are fallacious. Using a theory consciously, on the other hand, always results in some new attempt to clear up the interrelations that it assumes, and to bring it into harmony with which theoretical assumptions; that is, it results in the pursuit of theory for its own sake.
The value of business forecasting depends upon correct theoretical concepts; hence there can, at the present time, be no more important task in this field than the bridging of the gulf that divides monetary from non-monetary theories. This gulf leads to differences of opinion in the front rank of economists; and is also the characteristic line of division between trade cycle theory in Germany and in America—where business forecasting originated. Such an analysis of the relation between these two main trends seems to me especially important because of the peculiar position of the monetary theories. Largely through the fault of some of their best-known advocates in Germany, monetary explanations became discredited, and their essentials have, moreover, been much misunderstood; while, on the other hand, the reaction against them forms the main reason for the prevailing skepticism as to the possibility of any economic theory of the trade cycle—a skepticism which may seriously retard the development of theoretical research.14
There is a fundamental difficulty inherent in all trade cycle theories that take as their starting point an empirically ascertained disturbance of the equilibrium of the various branches of production. This difficulty arises because, in stating the effects of that disturbance, they have to make use of the logic of equilibrium theory.15 Yet this logic, properly followed through, can do no more than demonstrate that such disturbances of equilibrium can come only from outside—i.e., that they represent a change in the economic data—and that the economic system always reacts to such changes by its well-known methods of adaptation, i.e., by the formation of a new equilibrium. No tendency toward the special expansion of certain branches of production, however plausibly adduced, no chance shift in demand, in distribution or in productivity, could adequately explain, within the framework of this theoretical system, why a general “disproportionality” between supply and demand should arise. For the essential means of explanation in static theory—which is, at the same time, the indispensable assumption for the explanation of particular price variations—is the assumption that prices supply an automatic mechanism for equilibrating supply and demand.
The next section will deal with these difficulties in more detail: a mere hint should therefore be sufficient at this point. At the moment we have only to draw attention to the fact that the problem before us cannot be solved by examining the effect of a certain cause within the framework, and by the methods, of equilibrium theory. Any theory that limits itself to the explanation of empirically observed interconnections by the methods of elementary theory necessarily contains a self-contradiction. For trade cycle theory cannot aim at the adaptation of the adjusting mechanism of static theory to a special case; this scheme of explanation must itself be extended so as to explain how such discrepancies between supply and demand can ever arise. The obvious, and (to my mind) the only possible way out of this dilemma, is to explain the difference between the course of events described by static theory (which only permits movements toward an equilibrium, and which is deduced by directly contrasting the supply of and the demand for goods) and the actual course of events, by the fact that, with the introduction of money (or strictly speaking with the introduction of indirect exchange), a new determining cause is introduced. Money being a commodity that, unlike all others, is incapable of finally satisfying demand, its introduction does away with the rigid interdependence and self-sufficiency of the “closed” system of equilibrium, and makes possible movements that would be excluded from the latter. Here we have a starting point that fulfils the essential conditions for any satisfactory theory of the trade cycle. It shows, in a purely deductive way, the possibility and the necessity of movements that do not at any given moment tend toward a situation which, in the absence of changes in the economic “data,” could continue indefinitely. It shows that, on the contrary, these movements lead to such a “disproportionality” between certain parts of the system that the given situation cannot continue.
But while it seems that it was a sound instinct that led economists to begin by looking on the monetary side for an explanation of cyclical fluctuations, it also seems probable that the onesided development of the theory of money has, as yet, prevented any satisfactory solution to the problem being found. Monetary theories of the trade cycle succeeded in giving prominence to the right questions and, in many cases, made important contributions toward their solution; but the reason why an unassailable solution has not yet been put forward seems to reside in the fact that all the adherents of the monetary theory of the trade cycle have sought an explanation either exclusively or predominantly in the superficial phenomena of changes in the value of money, while failing to pursue the far more profound and fundamental effects of the process by which money is introduced into the economic system, as distinct from its effect on prices in general. Nor did they follow up the consequences of the fundamental diversity between a money economy and the pure barter economy that is assumed in static theory.
Naturally it cannot be the business of this essay to remove all defects and deficiencies from the monetary theories of the trade cycle, or to develop a complete and unassailable theory. In these pages I shall only attempt to show the general significance for this theory of the monetary starting point, and to refute the most important objections raised against the monetary explanation by proving that certain rightly exposed deficiencies of some monetary theories do not necessarily follow from the monetary approach. All that is wanted, therefore, is, first, a proof, using as our examples some of the best-known non-monetary theories, that the “real” explanations adduced by them do not, in themselves, suffice to build up a complete and consistent theory; second, a demonstration that the existing monetary theories contain the germ of a true explanation, although all suffer, more or less, from that oversimplification of the problem which results from reducing all cyclical fluctuations to fluctuations in the value of money; finally that the monetary starting point makes it possible, in fact, to show deductively the inevitability of fluctuation under the existing monetary system and, indeed, under almost any other that can be imagined. It will be shown, in particular, that the Wicksell-Mises theory of the effects of a divergence between the “natural” and the money rate of interest already contains the most important elements of an explanation, and has only to be freed from any direct reference to a purely imaginary “general money value” (as has already been partly done by Professor Mises) in order to form the basis of a trade cycle theory sufficing for a deductive explanation of all the elements in the trade cycle.