CHAPTER IX

THE VOLUME OF MONEY AND THE VOLUME OF CREDIT

John Stuart Mill, who first among the great figures in economics gives a realistic analysis of modern credit phenomena, thought that credit acts on prices in the same way that money itself does[155] and that this reduces the significance of the quantity theory tendency greatly, and to an indeterminate degree. The quantity theory is largely whittled away in Mill's exposition of the influence of credit. In Fisher we have a much more rigorous doctrine. The quantity of money still governs the price-level, because M governs M'. The volume of bank-deposits depends on the volume of money, and bears a pretty definitely fixed ratio to it. Just how close the relation is, Professor Fisher does not say, but the greater part of his argument, especially in ch. 8,[156] rests on the assumption that the ratio is very constant and definite indeed. At all events, the importance of the theory, as an explanation of concrete price-levels, will vary with the closeness of this connection, and the invariability of this ratio. It is not too much to say _that the book falls with this proposition_, to wit, that M controls M', and that there is a fixed ratio between them. We would expect, therefore, a very careful and full demonstration of the proposition, a care and fullness commensurate with its importance in the scheme. But the reader will search in vain for any proof, and will find only two propositions which purport to be proof. These are: (1) that bank reserves are kept in a more or less definite ratio to bank deposits; (2) that individuals, firms and corporations preserve more or less definite ratios between their cash transactions and their check transactions, and between their cash on hand and their deposit balances.[157]

If these be granted, what follows: the money in bank-_reserves_ is no part of M! M is the money in circulation, being exchanged against goods, not the money lying in bank-vaults![158] The money in bank-vaults does not figure in the equation of exchange. As to the second part of the argument, if it be granted, it proves nothing. The money in the hands of individual and corporate depositors is by no means all of M. It is not necessarily the greatest part. The money in circulation is largely used in small retail trade, by those who have no bank-accounts. A good many of the smallest merchants in a city like New York have no bank-accounts, since banks require larger balances there than they can maintain. Enormous quantities of money are carried in this country by laborers, particularly foreign laborers. "The Chief of the Department of Mines of a Western State points out that when an Italian, Hungarian, Slav or Pole is injured, a large sum of money, ranging from fifty dollars to five hundred or one thousand, is almost always to be found on his person. A prominent Italian banker says that the average Italian workman saves two hundred dollars a year, and that there are enough Italian workmen in this country, without considering other nationalities, to account for three hundred million dollars of hoarded money."[159] I do not wish to attach too great importance to these figures, taken from a popular article in a popular periodical. It is proper to point out, too, that these figures relate to hoarded money, rather than to M, the money in circulation. But in part these figures represent, not money absolutely out of circulation, but rather, money with a sluggish circulation. And they are figures of the money in the hands of poor and ignorant elements of the population. Outside that portion of the population--larger in this country than in any other by far[160]--which keeps checking accounts, are a large body of people, the masses of the big cities, the bulk of rural laborers, especially negroes, the majority of tenant farmers, a large proportion of small farm owners, especially nominal owners, and not a few small merchants in the largest cities, who have no checking accounts at all. A very high percentage of their buying and selling is by means of money. Kinley's results[161] show that 70% of the wages in the United States are paid in cash, and, of course, the laborers who receive cash pay cash for what they buy. (Not necessarily at the _time_ they buy!) Money for payrolls is one of the serious problems in times of financial panics.[162] To fix the proportion between money in the hands of bank depositors and non-depositors is not necessary for my purposes--_a priori_ I should anticipate that there is no fixed proportion. But it is enough to point out that money in the hands of depositors is not the whole of Fisher's M. Of what relevance is it, then, to point out, even if it were true, that an unascertainable portion of M tends to keep a definite ratio to M', when the thing to be proved is that the _whole_ of M tends to keep a definite ratio to M'? Fisher's argument is a clear _non-sequitur_. If it proves anything, it proves that a sum of money,[163] not part of M, and another sum of money, an unknown fraction of M, each independently, for reasons peculiar to each sum, tends to keep a constant ratio to M'. This gives us _l'embarras des richesses_ from the standpoint of a theory of causation! Two independent factors, bank-reserves and money in the hands of depositors, each tending to hold bank-deposits in a fixed ratio, and yet each moved by independent causes! By what happy coincidence will these two tendencies work together? Or what is the causal relation between them? And if, for some yet to be discovered reason, Professor Fisher should prove to be right, and there should be a fixed ratio between M as a whole and bank-deposits, would it not indeed be a miracle if all three "fixed ratios" kept together? Bank-deposits, indissolubly wedded to three independent variables[164] (independent, at least, so far as anything Professor Fisher has said would show, and independent in large degree, certainly, so far as any reason the present writer can discover), must find their treble life extremely perplexing. May it not be that Professor Fisher has pointed the way to the real fact, namely, that bank-deposits are subjected to a multitude of influences, no one of which is dominant, which prevent any fixed ratio between bank-deposits and any other one thing? At a later point, I shall maintain that this is, indeed, the case.

Be it noted further, however, that even if we grant a fixed ratio, on the basis of Fisher's argument, between M and M', Fisher has offered no jot of proof that the causation runs from M to M'. He simply assumes that point outright. "Any change in M, the quantity of money in circulation, _requiring as it normally does a proportional change in M'_, the volume of deposits subject to check." (_Ibid._, p. 52, Italics mine.) For this, no argument at all is offered. A fixed ratio, so far as causation is concerned, might mean any one of three things: (a) that M controls M'; (b) that M' controls M; (c) that a common cause controls both. Fisher does not at all consider these alternative possibilities. I shall myself avoid a sweeping statement as to the causal relations among the factors in the equation, because I do not think that any of the factors is homogenous enough, as an aggregate, to be either cause or effect of anything. But if a generalization concerning these magnitudes were required, I should be disposed to assert that the third alternative is the most defensible, and that to the extent that M and M' vary together it is under the influence of a common cause, namely, PT! That is to say, that the volume of bank-deposits and the volume of money tend to increase or decrease in a given market--and Fisher's theory is a theory of the market even of a single city[165]--_because of_ increases or decreases in PT (considered as a unitary cause rather than as two separate factors) in that market. But I shall not put my proposition in quite that form, as I find the factors in the equation of exchange too indefinite for satisfactory causal theory.

So much for the validity of Fisher's argument, assuming the facts to be as he states them. Are the statements correct? Do banks tend to keep fixed ratios between deposits and reserves? Do individuals, firms, and corporations tend to keep fixed ratios between their cash on hand and their balances in bank? Regarding this last tendency, Professor Fisher says in a footnote on p. 50, "This fact is apparently overlooked by Laughlin." I think it has been generally overlooked. I have found no one who has discovered it except Professor Fisher. Certainly no depositor whom I have consulted can find it in his own practice--and I have put the question to "individuals, firms, and corporations." The further statement which Professor Fisher adduces in its support does not prove it, namely, that cash is used for small payments, and checks for large payments.[166] It would be necessary to go further and prove that large and small payments bear a constant ratio to one another, and further, that velocities of money and of bank-deposits employed in these ways bear a constant relation. If Fisher has any concrete data, of a statistical nature, to support the doctrine of a constant ratio between bank-balance and cash on hand in the case of individual depositors, he has failed to put them into his book. Nor is there any statistical evidence offered in the case of banks. It should be noted here that finding a general average for a whole country or community would not prove Fisher's point. General averages give no concrete causal relations. Fisher's argument, moreover, starts with individual banks and individual deposit-accounts (pp. 46 and 50) and generalizes the individual practice into a community practice. He would have to offer data as to individual cases.

While general averages could not _prove_ the contention of a constant ratio between reserves and deposits for individual banks, general averages can _disprove_ the contention. A constant general average would be consistent with wide variation in individual practices, on the principle of the "inertia of large numbers." But if the general average is _inconstant_, it is impossible that the individual factors making it up should be constant. This disproof is readily at hand, both for the ratio of deposits to reserves in the United States, and for the ratio of demand obligations to reserves among European banks (most of which do not make large use of the check and deposit system).

For the United States, from 1890 to 1911, taking yearly averages, we have a variation in the ratio of reserves to deposits of over 73% of the minimum ratio. The ratio was 26% in 1894, and 15% in 1906. "The juxtaposition of these extreme variations shows how inaccurate is the assumption that the deposit currency may be treated as a substantially constant multiple of the quantity of money in banks."[167] For New York City, the annual average percentage of reserves of Clearing House banks to net deposits varies from 24.89% in 1907 to 37.59% in 1894.[168] The extreme variations[169] in weekly averages are (for the sixteen years, 1885-1900) 20.6% in August, 1893 and 45.2% in February, 1894. These figures are extreme, since the number of occurrences is small for them, but there are numerous occurrences of deviations from the mean as wide apart as 24% and 42%.[170] The yearly fluctuation in all these ratios is very great.

The ratio of money held by the banks and money held by the people also shows wide variation, and considerable yearly fluctuation. There is a further complication, for the United States, of varying proportions of the total monetary stock held by the Federal Treasury. As between the banks and the public, the banks held about a third in 1893 (average for the year), and nearly half in 1911.[171] Whatever may be the relations between money in the hands of the people, money in banks, and volume of deposits, in "the static state," there is no statistical evidence whatever to justify the notion of fixed relations among them in real life.[172] We shall later show that there can be no static laws whatever governing the relations of credit and reserves.[173]

For European banks, the case is equally clear. European bankers deny any intention of keeping any definite reserve ratio. This appeared very clearly in the "Interviews" obtained for the Monetary Commission with leading European bankers.[174] The Banque de France increased its gold reserves, between 1899 and 1910, by 75%, but increased its discounts and advances during the same period by only 5%.[175] J. M. Keynes[176] points out that the reserves of the great banks of the world, and of Treasuries which act as central banks, have absorbed an enormous part of the gold produced in the fifteen years before the War, increasing their holdings from about five hundred million pounds sterling in 1900 to one billion pounds sterling at the outbreak of the War. "The object of these accumulations has been only dimly conceived by the owners of them. They have been piled up partly as the result of blind fashion, partly as the almost _automatic consequence_, in an era of abundant gold supply, of the particular currency arrangements which it has been orthodox to introduce.... The ratios of gold to liabilities vary very extremely from one country to another, without always being explicable by reference to the varying circumstances of those countries.... The contingencies, against which a gold reserve is held, are necessarily so vague that the problem of assessing the proper ratio must be, within wide limits, indeterminate. It is natural, therefore, that bankers, who must act one way or the other, should often fall back on mere usage or accept _that amount of gold as sufficient_ which, _if they are chiefly passive, the tides of gold bring them_. [Italics mine.] At any rate, the management of gold reserves is not yet a science in most countries. There is no ideal virtue in the present level of these reserves. Countries have got on in the past with much less, and under force of circumstances could do so again."

It will be noticed that Keynes, in the passage cited, is speaking of _gold_ reserves, while Fisher's contention relates to all kinds of money available for reserves, which in this country would include gold, silver dollars, greenbacks, and, for many State banks, the notes of national banks. He is also talking of the relation of reserves to demand _liabilities_, which for most great European banks are primarily notes, rather than of reserves to deposits. But as an exposition of the theory of the ratio of reserves to deposits (the chief liability of American banks), it is applicable to American conditions, and as a statement of the facts, it of course gives a basis for testing Fisher's doctrine generally. I do not think that Fisher's fixed ratio, as between reserves and deposits, or even the ratio which more moderate quantity theorists might seek to find between gold and demand liabilities, will find any justification in the facts of banking history.[177]

A factor which has developed on a grand scale in recent years has tended still further to weaken any tendency that may be supposed to exist toward a fixed ratio between money-reserves and demand-liabilities. I refer to the gold exchange-standard, in India, the Philippines, and elsewhere, and to the practice of the great banks of the continental countries of Europe, particularly the Bank of Austria-Hungary, of holding foreign gold bills, rather than gold exclusively, as reserve to cover note issue. In the case of the Austro-Hungarian Bank, which has carried this practice to the extreme, all possibility of a fixed ratio between gold reserves and demand-liabilities has vanished. The ratio is highly flexible. When bills are cheap, _i. e._, when the exchange is "in favor" of Austria-Hungary, the Bank buys bills with gold; when bills are high, when the exchanges have turned "against" Austria-Hungary, the Bank sells bills for gold. Commonly, the holder of a note of the Austro-Hungarian Bank does not ask for it to be redeemed in gold, but in foreign exchange. The reason for this practice on the part of the Bank is primarily economy. A large holding of gold would represent idle capital--a heavy burden for the Bank of a debt-ridden and poorly developed country. Foreign bills, however, serve equally well for maintaining the value of the bank-notes, and at the same time bear interest.[178] A similar practice has been employed by the Reichsbank, by the National Bank of Belgium,[179] by virtually all the debtor countries of Europe, and the great trading countries of Asia.

Confidence in these conclusions is much increased by a study of the views of Professor Taussig.[180] Professor Taussig is, in his initial formulations of his doctrine, a quantity theorist. In a situation where only money is used, credit being excluded, in effecting exchanges, he would hold that the quantity theory correctly accounts for prices. He is fond of the old formulation, as a first approximation, even in dealing with the complex facts of modern banking. But he does not dodge the complex facts, and his theory becomes, substantially, first, a general formula, and second, an elaborate body of qualifications and exceptions, the latter making up the major part of the theory. His doctrine regarding the relation of money and credit is as follows: there is, in the long run, a real _limitation_ on elastic credit instruments in the quantity of _specie_. (This is very different from the assertion that there is a _fixed_ ratio between _deposits_ and _money_ in circulation, including paper, bank-notes, etc., in money. The present writer has no quarrel with the doctrine that the gold supply of the _world_ imposes _outside_ limitations on the _possible_ expansion of credit.) The limitation, Taussig holds, comes in two ways: (1), in the connection between prices in any one country, and prices in the world at large; (2), in various links of connection between the volume of deposits (and of notes elastic like deposits) and the quantity of specie. I shall consider at a later point the relation between prices in different countries.[181] I shall there maintain that the quantity theory, which explains gold movements on the basis of price-_levels_ in different countries, is inadequate; that not price-levels, but particular prices, of goods most available for international trade, are of primary importance, and that of these particular prices, one, namely the "price of money," or the short time money-rate, is most significant of all. For the present, I wish to analyze the linkages which Taussig finds between elastic credit instruments and specie, and to see how far they would go, not in proving Taussig's point (with which I have little quarrel) but in proving Fisher's contentions. The points involved are: (a) _Direct necessity_ constrains the bankers to keep _some_ cash on hand.[182] This fixes a _minimum limit_ (Taussig's contention), but does not at all suggest a "normal ratio" (Fisher's contention). (b) _Binding custom_, as to the proper amount of reserve that banks should carry, particularly important in connection with the Bank of England, but also in evidence in the Banque de France and the Reichsbank. Here again, however, minimal, rather than fixed, ratios are suggested. Limitations on the _expansion_ of credit these customs may impose, but they by no means determine a normal, or average amount of credit expansion--in England least of all, since there is so large a flexible element in the deposits of the Joint Stock Banks, whose reserves are largely secret. The statement _supra_ quoted from Keynes, together with the testimony of European bankers, may be considered in connection with this point, also, as to the factors determining the reserve policies of the great European banks. The extent to which custom really binds is doubtful. (c) _Direct regulation by law_, peculiar to the United States. Here again, a minimum, rather than a fixed ratio, is indicated. Some _limitation_ on credit expansion by the banks is caused by this at times, but Fisher's argument would require vastly more. (d) _The interaction in the use of deposits, notes, and other constituents in the circulating medium._ The point involved here is that different kinds of business call for different kind of media. Small retail business is not done with hundred dollar bills, nor are stocks and bonds bought with pennies. Limiting the size of bank-notes to five pounds in England compels the use of a large amount of gold for smaller transactions, and keeps a larger amount of gold in use than would otherwise be the case. Expanding business draws cash from the banks for circulation, trenching on reserves. That Professor Taussig has a point here is not to be doubted, but how closely it limits the expansion of credit will depend on the degree to which different kinds of media of exchange really _are_ thus specialized. In a country like the United States, where checks may be used for virtually any transaction of over a dollar, and where small change for less than a dollar will be increased by the Government to meet the demands of trade, the point would not seem to involve a practically serious limitation.

Finally, Professor Taussig recognizes a coefficient with the quantity of specie in the _temper of the business community_. Whether or not deposits are to expand, depends not only on reserves, but also on the attitude of borrowers.

Taussig concludes: "Thus there is only a rough and uncertain correspondence of bank expansion with bank reserves; much play for ups and downs which have no close relation to the amount of cash in bank vaults, _and still less direct relation to the amount of money afloat in the community at large_. Where bank media, whether in the form of deposits or notes, are an important part of total purchasing power, the connection between general prices and quantity of 'money' is irregular and uncertain." (Italics mine.)

This conclusion would be of little service in supporting Fisher's rigorous contentions! Our constructive theory concerning the relations of reserves and deposits, or reserves and demand liabilities, must wait for later discussion, in the chapter on "Bank Assets and Bank Reserves" in Part III. It will there be maintained that there are no "normal" or "static" laws governing the percentage of reserves to demand liabilities, or to deposits, that the reserve function of money is a _dynamic_ function, and that its whole explanation must be found in dynamic considerations. For the present, I am content to have analyzed two widely divergent views, one the extreme view of Professor Fisher, representing the quantity theory in its utmost rigor, and the other, the view of Professor Taussig, who virtually surrenders the quantity theory in complex modern conditions.

In between these two writers, verging more toward Fisher than toward Taussig, will be found, with great individual variation, the rest of the quantity theorists. The quantity theory, as an instrument of prediction, becomes important only to the extent that Fisher's view is maintained.