II. FISHER’S ANALYSIS OF MONETARY INSTABILITY: SOME CONSTANT FEATURES

Before dealing with the evolution over time of Fisher’s analysis of monetary instability, several constant features of the analysis must be recalled, regarding the largely monetary nature of the business cycle, the real effects of short-run monetary instability, and the global explanatory pattern that he used to account for such instability.

A Constant Explanatory Pattern: The “Debt–Money–Prices” Triangle

In what follows, we argue that, although Fisher’s specific explanations of short-run monetary instability would evolve over time, they had one element in common. They all rested on the cumulative interplay between three key variables: the nominal volume of loans or debts (which we will designate as D),Footnote ¹⁰ the general price level (P), and the volume of deposit currency (M′). This interplay, which Fisher (Reference Fisher1912, p. 363) referred to as “the circle of inflation, loans, deposits, and inflation again,” would provide the common basis for all his successive theories of monetary instability—a fact to which he never seems to have drawn attention. We propose to call this the “debt–money–prices triangle,” as illustrated in Figure 1.

Figure 1. The Debt–Money–Prices Triangle

This interplay can be deconstructed into three specific relations of causality:

• • the P-to-D causality, by which an increase (a decrease) in the general price level brings about an increase (a decrease) in the nominal volume of loans, or debts—including, crucially, those having their source in commercial bank loans;

• • the D-to-M′ causality, by which an increase (a decrease) in bank loans brings about an increase (a decrease) in the volume of deposit currency; and

• • the M′-to-P causality, by which an increase (a decrease) in the volume of deposit currency brings about, other things equal, an increase (a decrease) in the general price level.

Fisher’s investigations into the causes of cyclical monetary disturbances may be interpreted as a search for the main factor responsible for this vicious interplay. As we will see, he would first focus his attention on the P-to-D causality (section III), before turning to the broader D-to-M′-to-P causality (section IV), and, finally, to the specific D-to-M′ causality (section V).

III. FISHER’S EARLY ANALYSIS OF CREDIT CYCLES (1896 to 1911): FOCUSING ON THE P-TO-D CAUSALITY

As mentioned, in his 1911 book, The Purchasing Power of Money, Fisher offered a restatement of the quantity theory, according to which changes in the money stock would tend to produce proportional changes in the price level in the long run. He then restricted the use of the term “money” to designate generally accepted circulating media (i.e., gold coins and banknotes), and held that the volume of deposit currency (M′) would “normally” hold a definite relation to money so defined (Fisher [1911] Reference Fisher and Brown1913, p. 50). Only during “transition periods” or “credit cycles,” covered in Chapter 4 of his book (see below), would this ratio be disturbed. Although Fisher did specify that “periods of transition are the rule and those of equilibrium the exception” ([1911] Reference Fisher and Brown1913, p. 71), he gave the impression, throughout the rest of the book, of considering that deposit currency would generally keep a proportional relation to the monetary base, so that controlling the latter was all that was needed to stabilize the price level. His proposed remedy to monetary fluctuations—namely, his famous “compensated dollar” plan, which he first presented in 1911 (Fisher [1911] Reference Fisher and Brown1913, pp. 337–347)—actually consisted in periodically varying the gold content of the dollar, so as to keep its purchasing power constant; that is, “to change from a gold dollar of constant weight and varying purchasing power to a gold dollar of constant purchasing power and varying weight” (Fisher [1911] Reference Fisher1913, p. 233).Footnote ¹¹ Such a plan, however, would have directly affected only the monetary base, which led several critics to regard it as inadequate to deal with fluctuations in deposit currency.Footnote ¹² Many reviewers of the Purchasing Power of Money also criticized Fisher for overemphasizing the importance of “normal” periods, in which the deposit–reserve ratio could be considered as stable, and downplaying the fact that transition periods (or credit cycles) were actually the normal case.Footnote ¹³ These limits of Fisher’s early analysis having been stressed, the present section focuses on those of his writings of the time—especially Chapter 4 of his 1911 book—in which he did emphasize the cyclical behavior of deposit currency.

From his earlier writings on the subject up until the Great Depression, Fisher assigned a crucial role to the P-to-D causality channel when explaining short-run monetary fluctuations.Footnote ¹⁴ He located the starting point of credit cycles in any change in the price level (P), arguing that this would prompt a variation in the volume of loans (D). This would itself, via its effect on the volume of deposit currency (M′), bring a further change in P, and so on. This variation in D following from an initial variation in P was, in his view, the key relationship behind credit cycles, which he set himself the task of explaining.

Fisher’s first explanation of this P-to-D causality rested upon the particular behavior of the rate of interest. As early as 1896, he argued that an “inequality” of foresight existed between debtors and creditors, the former being more prompt to notice a change in P, which led to a lag in the adjustment of the nominal rate of interest (Fisher Reference Fisher1896, pp. 76–77). As a result, while, under a rising price level, the nominal rate of interest would be slow to rise, the real rate would actually decline (other things equal), which would unduly stimulate the demand for loans until the rate adjustment was complete—and conversely under a falling price level. This provided, in Fisher’s view, an explanation for credit cycles:

At that time, Fisher chose not to detail the reverse causality channel running from D to P, via M′, possibly because he was so convinced that he had found the main anomaly underlying credit fluctuations.Footnote ¹⁵ The thrust of this “inequality-of-foresight” theory was again expressed in Fisher (Reference Fisher1907, pp. 284–287), and would lie at the heart of his explanation of credit cycles in Chapter 4 of The Purchasing Power of Money (Fisher [1911] Reference Fisher and Brown1913). In the latter work, this time, he fully described the triangular interplay between P, D and M′, summarizing the upswing phase as follows:

But, as the above passage illustrates, Fisher kept identifying the “lag in interest” occurring within the P-to-D sequence as the central factor underlying the whole cycle. He had stated, after all, that “the chief object of this chapter is to show that the peculiar behavior of the rate of interest during transition periods is largely responsible for the crises and depressions in which price movements end” ([1911] Reference Fisher and Brown1913, p. 56). On this very point, however, his theory would meet with strong criticism, leading him to adjust his views somewhat.

Indeed, many critics of The Purchasing Power of Money, such as Oliver M. W. Sprague (Reference Sprague1911, pp. 143–144), Mitchell (Reference Mitchell1912, p. 164), and Minnie Throop England (Reference England1912), saw no reason for isolating the rate of interest from other production costs (such as wages, salaries, and raw materials), the adjustment of which was just as (if not more) likely to be lagging behind changes in P—thus providing as good an explanation for the increase in profits and borrowings under a rising price level. England (Reference England1912, pp. 98–101), basing her criticism on empirical studies of her own, thus invited Fisher to modify his theory accordingly. The latter, who may have received similar criticisms beforehand,Footnote ¹⁹ did make such changes in his books Elementary Principles of Economics ([1910] Reference Fisher1911, Reference England1912) and Why Is the Dollar Shrinking? (Reference Fisher1914). In these works, the chapters dedicated to “transition periods,” in contrast with Chapter 4 of The Purchasing Power of Money, no longer presented the lag in interest as the central cause of credit cycles.Footnote ²⁰ In the upswing phase, whereas sequences 1, 3, 4, and 5 remained unchanged, sequence 2 now read as follows: “(2) ‘Enterprisers’ … get much higher prices than before, without having much greater expenses (for interest, rent, salaries, etc.), and therefore make much greater profits” (Fisher Reference Fisher1914, p. 77; see also Reference Fisher1912, p. 187).

The concluding line was amended accordingly: “Rise of prices generates rise of prices and continues to do so as long as the enterprisers’ profits continue abnormally high” (Fisher Reference Fisher1914, p. 78; italics in original; see also Reference Fisher1912, p. 187). Henceforth, Fisher would typically include production costs in general (such as “rent, salaries and wages—not to mention raw materials”), along with interest, as responsible for the “lagging of total expenses behind total receipt” (Fisher Reference Fisher1925, p. 180; see also Reference Fisher1920, p. 66; Reference Fisher1923, p. 1025; Reference Fisher1926, p. 787; Reference Fisher1932c, p. 30; Reference Fisher1936c, p. 496). However, he did not make such a change in the second edition of The Purchasing Power of Money published in 1913, for a reason he revealed in the preface;Footnote ²¹ and, on some occasions, he would again assign a primary importance to the maladjustment of the rate of interest (see Fisher Reference Fisher1914, pp. 84–86; Reference Fisher1923, p. 1024; Reference Fisher1930, pp. 43, 411–416). It is therefore difficult to assess the extent to which he really adjusted his views on this matter. In any case, after he developed his debt-deflation theory in 1932–33, Fisher would consider both the rigidity of production costs in general, and the lagging behind of interest rates in particular, as insufficient explanations, in and of themselves, of booms and depressions.Footnote ²² From then on, he would turn his attention to another sequence of the “debt–money–prices” triangle.

IV. FISHER’S DEBT-DEFLATION THEORY OF GREAT DEPRESSIONS (1932–33): SHIFTING THE FOCUS TOWARD THE D-TO-M′-TO-P CAUSALITY

The Great Depression of the 1930s made Fisher tackle the issue of monetary instability from a different angle. Until then, as discussed, he had held the view that credit cycles originated with a disturbance of the general price level (P)—“whatever the first cause may be”—from which the whole cumulative interplay between P, D, and M′ was set off. In the boom years preceding the Great Depression, however, both the total money stock and the commodity price level had remained fairly stable, largely as a result of the countercyclical actions undertaken by the Federal Reserve System during those years (see Fisher Reference Fisher and Cohrssen1934, pp. 250–251). Whatever rise in prices occurred at that time was especially concentrated in the stock market.Footnote ²³ Possibly for this reason, Fisher’s analysis evolved. Noting that the Great Depression had originated with “a state of over-indebtedness” (Fisher Reference Fisher1932c, p. 25), he would thereafter take the nominal debt-volume (D) as the starting point of his credit cycle explanations, and focus on the D-to-M′-to-P sequence of the triangle.

Fisher’s new theory of booms and depressions—which, when applying it to the downswing phase, he termed the “debt-deflation theory”—was presented in a series of works, including his book Booms and Depressions (Reference Fisher1932c) and an article published in the first volume of Econometrica (Reference Fisher1933b).Footnote ²⁴ This theory, Fisher (Reference Fisher1932c, p. vii) maintained in the preface of his 1932 book, sought to highlight how some “nine main factors” combined to explain business cycles. However, he quickly focused on the first three of them in particular, which he singled out in a specific chapter: the debt volume, the deposit currency volume, and the price level (pp. 8–28). We find here, once again, the “debt–money–prices” triangle—D, M′, and P. The “remaining six factors”—which included “net worth,” “profits,” “production, trade and employment,” “optimism and pessimism,” “the velocity of circulation,” and “the rate of interest” (Reference Fisher1932c, pp. 29–43)—were presented as rather secondary factors, mostly reacting to a falling price level, and were treated, as a reviewer observed, “in a cursory manner” (Arakie Reference Arakie1933, p. 485). There was nothing really new in this list, as all nine factors already appeared more or less explicitly in Fisher ([1911] Reference Fisher and Brown1913, pp. 67–70, 335).

The novelty of Fisher’s debt-deflation theory was to be found, first of all, in its analysis of the interrelations between the three main factors, D, M′, and P. In contrast with his former theory of credit cycles—which, as we saw, focused on the P-to-D causality—he now focused on the D-to-M′-to-P sequence, emphasizing how a decrease in D led, through a contraction of M′, to a decrease in P (and conversely in the upswing phase):

Then, with the ensuing P-to-D sequence, the process would come full circle and repeat itself cumulatively, thus feeding booms and depressions.Footnote ²⁵

Above all, Fisher came to stress another new feature of his analysis—which he would later term the “Debt Paradox” (Fisher Reference Fisher1936a, p. 407)—pertaining to the case of particularly great booms and depressions: the “theory that when over-indebtedness is so great as to depress prices faster than liquidation, the mass effort to get out of debt sinks us more deeply into debt” (Fisher Reference Fisher1933b, p. 350). That is, the fall in the price level—spelling an increase in the real value of each dollar of debt outstanding—could happen faster than the decrease in the nominal debt-volume, so that the real debt-volume would actually increase:Footnote ²⁶

Thus, whereas Fisher had until that time considered price-level variations as tending to be “self-corrective” (Fisher [1911] Reference Fisher and Brown1913, p. 70), he now argued that under certain circumstances, if left to themselves, they could become self-defeating instead.

Fisher (Reference Fisher1932c, p. 15) further specified that, owing to their special role in the money-supply mechanism, “commercial bank debts” were “the only kind of debts directly involved” in the debt-deflation process, although other debts played an aggravating role. Yet, however much the tie between money and debt was central to his theory, he stopped short, at that time, of putting the monetary system into question. He even sometimes pushed M′ into the background of his analysis, turning the spotlight on D and P:

The D-to-M′-to-P sequence was thus sometimes treated simply as a D-to-P sequence, with M′ apparently accorded a secondary importance. This was especially the case in Fisher’s 1933 Econometrica article, in which he referred to “two dominant factors, namely over-indebtedness to start with and deflation following soon after,” on the one hand, and seven “secondary variables,” into which he now relegated the “circulating media,” on the other hand (Fisher Reference Fisher1933b, p. 341; italics in original). This was in contrast with all of his other writings from that period (e.g., Fisher Reference Fisher1932a, Reference Fisher1932b, Reference Fisher1932c, Reference Fisher and Fisher1933a, [1933] Reference Fisher and Fisher1934), in which M′ was explicitly given a central role. Could this be because, writing in an academic journal, he perhaps sought to find a receptive audience among the opponents of a monetary interpretation of business cycles?

This leads us to the question of whether Fisher’s debt-deflation theory should really be regarded as monetary in nature. That this was the case was obvious to Ralph Arakie (Reference Arakie1933, p. 485), who reviewed his 1932 book. Joseph Schumpeter (Reference Schumpeter1954, p. 1122), however, although discussing it in a chapter covering the “monetary theories” of the cycle, argued that the debt-deflation theory was “in essence not monetary at all,” basing his contention on the fact that Fisher identified real factors—such as “new investment opportunities” (Fisher Reference Fisher1933b, p. 350)—as important debt starters.Footnote ²⁷ Fisher, it is true, held that any kind of cause, whether monetary or not, could produce an initial variation in D (just like he had held, in his earlier credit cycle theory, that any kind of factor could produce the initial change in P). He especially stressed the fact, however, that any such variation would then spark off cumulative variations in all three of the key variables (D, M′ and P), and it is this monetary amplifying mechanism that was crucial to his analysis.Footnote ²⁸ The whole interplay, moreover, could as well be started by an initial variation in P. Footnote ²⁹ Although it is true that Fisher’s debt-deflation theory was not entirely monetary, money (and more precisely bank money) definitely played a key role in it.

An “important corollary” of the debt-deflation theory, on which Fisher (Reference Fisher1933b, p. 350) put strong emphasis, was that “great depressions are curable and preventable through reflation and stabilization.” Indeed, the Federal Reserve System could always forestall the contraction of M′ by acting on the monetary base and using “credit control” instruments, such as open market operations, the rediscount rate, or reserve requirements, in a countercyclical way (Fisher Reference Fisher1932c, pp. 121–131). Other measures, such as “gold control” or “velocity control,” could also be used (Reference Fisher1932c, pp. 136–141). Fisher (Reference Fisher1932b, p. 690) therefore believed that booms and depressions could be prevented: “It can always be done, because the price level is the one thing that is easy to control.”Footnote ³⁰ The stability of P would still be better achieved, however, if only the cumulative interplay between P, D, and M′ could be prevented from occurring in the first place. In this respect, Fisher’s analysis was soon to reach yet another stage.

V. FISHER’S MONEY-DEBT TIE ANALYSIS AND 100% MONEY PROPOSAL (1935): FINALLY FOCUSING ON THE D-TO-M′ CAUSALITY

With his debt-deflation theory, as we saw, Fisher had come to focus on the D-to-M′-to-P sequence of the “debt–money–prices” triangle. He occasionally stressed the decisive importance of the specific D-to-M′ causality but without, at the time, questioning the dependence of the money supply upon bank loans.Footnote ³¹ This step would finally be taken in his book 100% Money, first published in 1935, in which he proposed “to raise reserve requirements against checking deposits from 10%, or thereabouts, to 100%” (Fisher [1935] Reference Fisher1945, p. xi). He explained that “the quest for non-dependence of money on loans was what started the present writer on the 100% system” ([1935] Reference Fisher1945, p. 58).Footnote ³² From the Chicago Plan memoranda of 1933, Fisher ([1935] Reference Fisher1945, p. xiii) claimed to have obtained “many of the ideas” embodied in his own proposal. He had embraced this reform idea by the end of 1933, and was already campaigning for it in early 1934, as William Allen (Reference Allen1993, pp. 707–708) recounts.

The existing mixed monetary system, made up of both lawful and bank-created money, was now held directly responsible for the severe contraction of M′ (from twenty-three to fifteen billion dollars, Fisher specified) between 1929 and 1933: “This destruction of check-book money was not something natural and inevitable; it was due to a faulty system” (Fisher [1935] Reference Fisher1945, p. 7). Under this “10% system,” as Fisher called it, “some nine-tenths of the depositors’ deposits can be made out of their own promises, with the help of the bank” ([1935] Reference Fisher1945, p. 41). The resulting “constant trombone of expansion and contraction” of deposits subject to check, “tied, as they now are, to bank loans,” was, according to him, “the chief cause of both booms and depressions” (Fisher Reference Fisher1935, p. 522; [1935] Reference Fisher1945, p. xviii). He now insisted that the D-to-M′-to-P causality—which, as we saw, was central to his debt-deflation theory—was made possible only by this D-to-M′ connection:

As a solution, the 100% money proposal sought to “make money independent of loans; that is, to divorce the process of creating and destroying money from the business of banking” (Fisher [1935] Reference Fisher1945, p. xvii). To this effect, a sharp distinction between two kinds of deposits would be drawn. On the one hand, checking deposits (M′), serving as means of payment, could no longer be created or destroyed through bank loans. They would have to be fully covered by reserves in lawful money (Mo), the volume of which would be regulated by a legally mandated Currency Commission, independent of the government.Footnote ³⁴ On the other hand, savings deposits, fulfilling an investment function, would be freely used by the banks to finance loans and investments (and thus remain only fractionally covered by reserves) but would not be allowed to serve as means of payment.Footnote ³⁵ The banking business would therefore involve the handling of only pre-existing money, while the monetary authority alone could expand or contract the medium of exchange. In this way, the “mistaken tie between money and debt” (Fisher [1935] Reference Fisher1945, p. 177) would be severed, and the D-to-M′ causality ended—thus preventing the whole “debt–money–prices” cumulative interplay from occurring.

Henceforth, the volume of circulating medium (M) would no longer directly vary with bank loans and investments. To be sure, an increase in the real volume of trade (T), insofar as it exerted a downward pressure on P, would prompt the Currency Commission to increase M, supposing a price-level stabilization policy to be followed.Footnote ³⁶ But such changes would no longer be cumulative—thus providing, in Fisher’s view, the conditions for a true elastic currency:

The velocity of circulation of money (V), of course, might still vary. Fisher, however, held that the changes in V typically followed, rather than led, the changes in M. As he explained:

Fisher linked his new monetary analysis to his debt-deflation theory, providing, in Chapter 7 of 100% Money, an updated summary of his 1933 Econometrica article. This time, he gave the “money-debt tie” more emphasis, noting that “an underlying cause (or pre-condition) of great booms and depressions is the 10% system itself” (Fisher [1935] Reference Fisher1945, p. 120).Footnote ³⁷ Should the “debt–money–prices” interplay be prevented from occurring, he argued, overindebtedness would be much less likely to develop in the first place: “Moreover, under the 100% system, the depression could never get so big a start since the preceding boom and over-indebtedness would not be so great. This does not mean that, under the 100% system, there would be no booms and depressions whatever. It means simply that they would be vastly less severe” (Fisher [1935] Reference Fisher1945, p. 134).

Yet another argument made by Fisher was that, by separating the issuing from the lending of money, the 100% scheme would allow market conditions to prevail on the loan market, as the issuing authority would no longer need to manipulate interest rates in order to encourage (or discourage) the creation of money through bank loans:

Fisher was here referring to the countercyclical policy carried out by the Fed in the 1920s, under the leadership of Federal Reserve Bank of New York’s Governor Benjamin Strong—which, although it was successful in stabilizing the commodity price level, did so at the expense of interest rate distortions: “[It] is a dangerous thing to do, to interfere with the natural rate of interest, and it was largely because the rate of interest was abnormally low in 1925 that you had the speculation on the stock exchange, which had a great deal to do with this depression” (Fisher, Reference Fisher1937c, p. 288; see also [1935] Reference Fisher1945, pp. 139–140). Only under a 100% system, Fisher now held, was it possible to have both a stable price level and freely market-determined interest rates.

This money-debt tie analysis, underpinning the 100% money proposal, marked the last stage of Fisher’s evolving theories of monetary instability. The Yale economist would continue to advocate his 100% plan, and try to get it enacted into law, for the rest of his life.Footnote ³⁸