Chapter XIII: The Volume of Money and the Volume of Trade--Trade and Speculation (1)

In proving that an increase of money must proportionately increase prices, it is necessary to prove that the volume of trade is independent of the quantity of money and credit instruments by means of which trade is carried on. Money on the one hand, and quantity of goods to be exchanged on the other, are the two great independent magnitudes, whose equilibration mechanically fixes the average of prices. This notion, as to the essence of the quantity theory, finds expression in Taussig,[225] "The statement of a quantity theory in relation to prices assumes two independent variables: total money or purchasing power on the one hand, total supply of goods or volume of transactions on the other." Taussig, though he would maintain that this independence holds, so far as money and trade are concerned, admits that it breaks down so far as trade and elastic bank credit, bank-notes and deposits, are concerned. Trade and elastic bank-credit are largely _inter_dependent.[226] This concession on Taussig's part means virtually giving up the quantity theory for Western Europe and the United States and Canada, though Taussig still sees something left of the quantity theory tendency in view of the "irregular and uncertain" connection which he finds between money and bank-credit.[227] Fisher, however, makes no such surrender. He is quite as uncompromising as to the independence of _deposits_ and trade as he is with reference to the independence of _money_ and trade. He does, indeed, make the concession that increasing trade tends to increase deposits _indirectly_, by increasing the ratio of M' to M, by modifying the habits of the people as to the use of checks as compared with cash (p. 165),[228] but he denies stoutly that there is any _direct_ relation between them. (P. 168.) Trade acts only _via_ a modification of the ratio between M and M', and M still remains controlled, not by trade, but by quantity of money. As to any control over T by M', he repudiates it explicitly, (P. 163.) Increasing M', either through an increase of M, or through an increase in the normal ratio between M and M', will have no effect on T,--or, for that matter, on the V's. The introduction of credit, therefore, leaves the quantity theory intact: an increase of M, increasing M' proportionately, leaving the V's unchanged, and having no effect on T, must exhaust its influence on P, raising P proportionately, if the equation of exchange is to remain valid.

The argument set forth to prove that T is not influenced by M or M' is as follows: "An inflation of the currency cannot increase the products of farms or factories, nor the speed of freight trains or ships. The stream of business depends on natural resources and technical conditions, not on the quantity of money. The whole machinery of production, transportation and sale is a matter of _physical capacities and technique_, none of which depend on the quantity of money. The only way in which quantities of trade appear to be affected by the quantity of money is by influencing trades accessory to the creation of money and to the money metal.... From a practical or statistical point of view they amount to nothing, for they could not add to nor subtract one-tenth of 1% from the general aggregate of trade." (_Loc. cit._ p. 155. Italics mine.) Something similar is said on p. 62, where "transitional" influences of M on T are being discussed: "But the amount of trade is dependent, _almost entirely_, on other things than the quantity of currency, so that an increase of currency cannot, _even temporarily_, very greatly increase trade. In ordinarily good times practically the whole community is engaged in labor, producing, transporting, and exchanging goods. The increase of currency of a "boom" period cannot, of itself, increase the population, extend invention, or increase the efficiency of labor.[229] These factors pretty definitely limit the amount of trade that can reasonably be carried on. So, although the gains of the enterpriser-borrower may exert a psychological stimulus on trade, though a few unemployed may be employed, and some others in a few lines induced to work overtime, and although there may be some additional buying and selling which is speculative, _yet almost the entire effect_ of an increase in deposits must be seen in a change in prices. Normally the _entire_ effect would so express itself, but transitionally there will be also _some_ increase in the Q's." (Pp. 62-63. Italics mine.)

Fisher is here exceedingly uncompromising, even where transitional periods are concerned, and it is not necessary, in order to do his position full justice, to make much distinction between "normal" and "transitional" effects in my counter-argument. I shall, however, take account of the distinction as I proceed, in justice to other, more moderate, quantity theorists.

It is a familiar doctrine that the quantity of money is irrelevant, that things go on in much the same way whether money is abundant or scarce, the only difference being that in the one case prices are high and in the other, low; that, in particular, it is a gross fallacy to connect the rate of interest with the amount of money, since (as many writers would put it) the rate of interest depends on the amount of _capital_ rather than _money_. At the opposite extreme, we have writers like Brooks Adams (_Law of Civilization and Decay_), who see the fate of nations and the progress of civilization resting on the abundance or scarcity of money. Fisher takes the first position in its extremest form.[230]

The truth, I think, is intermediate. The effects of the New World discoveries of gold and silver after the voyage of Columbus on trade and industry were tremendous. Trade was enormously increased. Walker, in his _Inter__national Bimetallism_,[231] asking, from the standpoint of a quantity theorist, why prices only increased 200% while money increased 470%, admits that the chief reason was the increase in trade, due in large part to the very increase in money itself. Sombart, in his _Der Moderne Kapitalismus_,[232] finds in this influx of money a tremendous source of capitalistic accumulations, (a) for the Conquistadores, (b) for the handicraftsmen whose prices rose faster than their costs, (c) for tenants whose rents were fixed in money, (d) for landowners, whose rents were fixed in kind [a point not obviously true], and (e) for bankers, as the Fugger. An increase of capital, savings that would otherwise not have been made, must have profoundly modified the whole industrial system, and greatly increased both industry and commerce. If it be objected that effects of this sort are not usual, that they came in a world which had been starved for money, and which, by means of the enormous increase in money was able to pass from a "natural" to a money economy, I reply that the difference between such a case and the usual effects of an increase of money are in degree rather than in kind. The world of Columbus' day was in part on a money economy, and the world to-day, despite Professor Fisher's emphatic denial,[233] still employs a great deal of barter, or equivalents of barter. I shall revert to this point later. But even this consideration would not rob Sombart's points of their significance for modern conditions. Further, we have an even more striking case, on Walker's own showing, in the effects of the Californian and Australian[234] gold discoveries in the 19th Century on trade, industry, and speculation.[235]

Nor is the tremendous agitation over bimetallism, involving a literature so great that no man could dream of reading it all, involving great political movements, Presidential campaigns, great Congressional debates, repeated legislation, international conferences, etc., for twenty years, to be explained on any other ground than that the world felt practical, important, and unpleasant effects on industry and trade from the inadequacy of the money supply.

The view of Hartley Withers[236] is interesting here. He says: "any such great addition to currency and credit would have a great effect in stimulating production, and so would lead to a great addition to the number of real goods which humanity desires and consumes when it can get them.... Trade would be more active." On p. 23 he speaks of the enormous expansion of trade made possible by paper representatives of gold. On p. 83 he speaks of the attitude of the money-market toward gold, which the orthodox economist is apt to think of as a survival of Mercantilism. Withers thinks that the money market is right in a large degree.

As illustrating Withers' statement about the views of "practical men" on this point, the following extract from a recent address by Theodore Price, quoted with approval in a "market letter," written by Byron W. Holt,[237] is interesting: "The fact seems to be that the exigencies of war in Europe are leading to an extension of credit such as would not have been possible in peace, because the hesitant conservatism of bankers would have then prevented it, and we are finding that instead of working harm it is doing good, because huge masses of fixed capital are thereby made productive, and are circulating with the increased velocity that always quickens enterprise and accelerates the wheels of industry.... All the precedents of history indicate that accelerated activity will come with peace and continue until the exuberance of success has led men to build faster than the world has grown and to demand credit upon the basis of future rather than of present values."

What is the essential causation in the matter? Well, viewed merely as a matter of mechanical equilibration, the quantity theory view is not strictly true, by any means. For a given country--and Fisher's quantity theory is always a theory for a given country, and, indeed, for any separate market, even a single city[238]--an increase of banking credit means an increase in non-monetary capital, because, to a greater or less extent it dispenses with the use of gold, which goes abroad, bringing back wealth in other forms in exchange. Adam Smith saw this clearly, and phrased it strikingly, likening gold and silver coins to the wagon-roads of Scotland, which are necessary for transportation, but which none the less prevent the use of the roadways for raising grain; whereas bank credit is like a wagon-road through the air, which restores the roadbeds to cultivation. Increased non-monetary capital, other things equal, should mean increased trade.

But, more fundamentally, an increase in gold itself within the country, if not bought by the export of an equivalent amount of other goods, _is an increase of capital_. Not all capital is money, but standard coin is capital. Money is a tool of exchange, and exchange is part of the productive process. More money means more exchanging. That is what money is for. Part of the mechanism is in the money rates, which go down as money becomes more abundant, making it profitable to effect exchanges which would not have been profitable had the money rates been higher. Granted that the money-rates and the general rate of interest tend, in the long run, to keep--I will not say at the same figure[239]--a certain fairly definite relation to one another, it still does not follow that the new "normal" equilibrium will give us an interest rate which is the same as the general rate of interest was before the influx of gold. On the strictest static theory, this is not to be expected. Because the total amount of capital in the country is increased, and this means a lowered interest rate all around, in the marginal employment of capital. The margin of the use of capital will be lowered everywhere, including the margin for the use of money. This means permanently lowered money rates in the country, even though the permanent level be higher than the initial money rates immediately following the access of new gold. I have put the argument in terms that suggest the productivity theory of interest, because it is more simply stated that way. I do not accept the productivity theory, as a fundamental explanation of interest, but for many purposes, the results to be obtained by it coincide with the psychological time theories,--which also, in their present form, seem to me imperfectly developed. I need not try to construct a theory of interest here, however, as the familiar theories lead to no trouble at this point. It is enough to point out that the increased amount of capital, meaning better provision for present wants--wants concerned with gold in the arts and with money for productive exchanges, as well as goods generally since part of the new gold will be exported for other things--will lessen the pressure of present as compared with future wants, and so lessen the rate of interest on the time-preference theory. The final outcome will be an extension of the marginal use of money, and a greater volume of exchanges. Of course, the increase in the supply of any kind of capital good, apart from a prior increase in the demand for its services, will, on the mechanical view of economic causation, necessarily lead to some fall in its capital value. Gold money will be no exception to this rule. As to how much the increase in its quantity will lead its capital value to fall, however, we are unable to say. For the quantity theory, the fall will be in proportion to the increase. For the theory just outlined, the fall will depend on the elasticity of demand for gold in the arts, and on the elasticity of "demand" for money, meaning by demand for money simply the demand for the short-time use of money as a tool of exchange, a demand which governs _directly_, not the capital value of money, but rather the "money-rates." The relation between the money rates and the capital value of money will best be discussed at another point.[240] We have no reason at all to suppose that either of these demands[241] exhibits the tendency to obey the law of proportional variation which the quantity theory requires of money.

It is further important to note that as a country gets more abundant capital, there seems to be a tendency to extend the use of money rather more than the use of many other capital goods. Where the interest rate is 10 and 12%, as in Arizona and New Mexico, money, even when brought in, tends to leave in large degree to bring in other forms of capital which the situation calls for more imperatively. The early American colonies, needing money pressingly, and making shift with a great variety of substitutes for good metallic money, thoroughly acquainted with the advantages of a money-economy from their European experience, and having "habits" as to the carrying and using of money which they had brought with them from Europe, still found it impossible to keep a great deal of metallic money, in view of the still greater importance of other forms of capital. It is in the most highly developed commercial communities, commercial centres, and _par excellence_, in the speculative centres, that the demand for the money-service is most elastic.[242] A country where the rate of interest is low, loses other forms of capital, and gains money, in the process of reequilibration, as compared with a new and undeveloped section, although the new section also extends the margin of the money service, in effecting a greater number of exchanges, when money is increased.

And this leads to a vital distinction, which quantity theorists almost always lose: the distinction between the volume of _production_, and the volume of _trade_. Even in the mechanical system of causation which they describe, it is true only of production and transportation that _technical_ and _physical_[243] factors are of primary significance, and that money is of minor significance. For trade and commerce, money is always highly important. To the extent that a region is primarily given over to the primary productive activities, mining, and agriculture, such trading as is necessary can be done by means of a small amount of money, supplemented by barter and long-time book-credit. A region or a city whose chief business is _commerce_, however, needs a large part of its capital in the form of money, and of banking capital, which is largely invested in money for banking reserves. _Trade_, as distinguished from industry (and it is after all trade that is under discussion), is helped or hindered as its tools are more or less abundant. These considerations would suggest that the elasticity of the demand for the use of money is greater than the elasticity of demand for the use of capital in almost any other form. Production is, indeed, limited by labor supply and natural resources, in considerable degree. _Trade_,[244] however, even from the standpoint of mechanical causation, is limited chiefly by the relation between the profits to be made in commercial transactions, and the "price" that must be paid for the money and credit that are required to put them through. There are enormous numbers of transfers that could be made to advantage if there were no cost at all involved. They are not made, because exchanging requires pecuniary capital. Let the pecuniary capital increase, however, and sub-marginal exchanges become worth while, the general margin is lowered. Commerce is the most highly flexible and elastic portion of the whole productive process. The elasticity of demand for commercial capital is, thus, greater than the elasticity of demand for any other form of capital.

How widely the volume of trade differs from the volume of production, and how great is the element of speculative transactions in trade, will best appear, I think, from an analysis of the figures which Fisher gives[245] for the volume of trade in the United States. His figure for the volume of trade in the year 1909 is $387,000,000,000.00, three hundred and eighty-seven billions of dollars! This figure is reached by equating the figures he has reached for MV plus M'V' to PT, and assuming P to be one dollar, by making the "unit" of T, arbitrarily, a dollar's worth of each sort of commodity, at the prices of 1909. I have already commented on the legitimacy of this method of summarizing T,[246] and need not say more here, beyond calling attention to the fact that "volume of trade," as commonly used, does in fact mean, not T alone, but PT. Fisher for years other than 1909, however, makes use of a different method of getting at T: he takes certain indicia of _relative_ amounts of trade, compares them with the same indicia for 1909, and estimates the trade for other years as being such a percentage of the trade for 1909 as their indicia are of the indicia of 1909. The indicia chosen are: (1) quantities of certain commodities, cotton, fruit, cattle, etc., _received at_ principal cities of the United States, taken as typical of the variations of the internal _commerce_ of the United States; (2) quantities of 23 articles of import and 25 articles of export, for each year, taken as typical of variations in the foreign trade of the United States; (3) sales of stocks. These three indicia, weighted in a manner to be described in a moment, are then averaged. There is a second element in the index, made up by taking the figures for railroad _tonnage_, and the figures for _receipts on first class mail_, which are averaged. The first average and the second average are then combined into a third average, which is the final index. The relation between this index for every year other than 1909 and the same index for the year 1909 determines the amount of T for each year--the two indicia, together with the figure, $387,000,000,000.00, giving the required amount by the "rule of three." I shall not go into details with the method of constructing these averages, but I wish to make clear the comparative _weight_ given to each element in the final index: The first three elements count _twice_ as heavily as the last two, and so constitute the biggest factor. In the first average, based on the first three elements, the item taken as typical of internal trade is _weighted by 20_, the item taken as typical of foreign trade is _weighted by 3_, and sale of stocks _by 1_. It appears from Fisher's figures (p. 479), that the one really big _variable_ among all the indicia is the sale of stocks, but the weight given it is so small that it makes virtually no difference in the final result. Thus, as between 1898 and 1899, stock sales increased over 50%, but total trade, as shown by Fisher, increased only 5%. In the following year, stock sales _decreased_ over 21%, but total trade, on Fisher's figures, _increased_. The following year, 1901, stock sales virtually doubled, but Fisher's final figure shows only an increase around 13%. Two years later, in 1903, stock sales fell off about 40%, from the figures for 1901, but again, as compared with 1901, total trade on Fisher's figures shows an appreciable gain. The influence of stock sales on Fisher's index is, virtually, negligible. The dominating factor is the _receipts_ of selected staples, cattle, cotton, rice, pig iron, etc., in the principal cities of the United States. There is not a _single year_ in which his final figure for T does not move in harmony with this factor (p. 479). He gets, thus, for the volume of trade through the fourteen years under consideration, a surprising steadiness, and a pretty uniform progressive development.

In defence[247] of his method of weighting, Fisher says, simply: "These weights are, of course, merely matters of opinion, but, as is well known, _wide differences in systems of weighting make only slight differences in the final averages_." (Italics mine.)[248]

Are these figures valid? Well, first one is struck with the absolute magnitude assigned to T. The figures seem vastly greater than would have been anticipated. The method of calculating it, for 1909, I shall discuss in detail in the chapter on "Statistical Demonstrations of the Quantity Theory." For the present, it is enough to note that the absolute magnitude is derived from figures collected by Dean David Kinley for the National Monetary Commission,[249] of deposits, exclusive of deposits made by one bank in another, made in about 12,000 banks (out of 25,000) on March 16, 1909. These deposits were classified as (1) money (with subdivisions) and (2) checks and other credit instruments. A cross-classification divided them into (1) retail deposits; (2) wholesale deposits; (3) all other deposits. Kinley's object was to determine the extent to which checks are used, as compared with money, in payments, particularly in wholesale and retail business. Fisher's total, briefly, was obtained as follows: Kinley's figures, for the one day, were increased to make an allowance for the non-reporting banks; they were further increased on the assumption that March 16 was below the average for the year; the figure finally obtained for the day was then multiplied by 303, assumed as the number of banking days in the year, and the product, 399 billions, was taken as representing the total circulation of money and checks in trade. For some reason not made clear, this total was subsequently reduced to 387 billions. Counting the average price, P, as $1, T was considered to be 387 billions.[250]

In the statistical chapter to follow, it will be shown that this estimate is a very decided exaggeration. Deposits made in banks greatly overcount trade. Very many payments represent duplications, loans and repayments, taxes, etc., and are in no sense trade. This is true of all classes of deposits, wholesale and retail, as well as "all other." But for the present, I am concerned with the question, not of the absolute magnitude of the volume of trade, but rather, the questions of its character, of the elements that enter into it, and, above all, of the extent to which it is physically determined by technical conditions of production, and the extent to which it is flexible, a matter of speculation, etc.

We may approach this question from the angle of several bodies of statistical information. First, the question may be raised: what is there in the country which could be bought and sold enough in the course of a year to give us anything like so great a total? The subtractions which we shall find it necessary to make will still leave us an enormous total.

The United States Census Bureau[251] in 1904 reached the conclusion that the _total wealth_ of the country was only $107,000,000,000. Of this, over $62,000,000,000 was in real estate; $11,000,000,000 in railroads; street railways, over $2,000,000,000; telephone, telegraph, water and light, and similar enterprises total nearly $3,000,000,000 more. None of these things enter into ordinary wholesale and retail trade. The items that one would ordinarily think of are agricultural products, $1,900,000,000; manufactured products, $7,400,000,000; mining products, $400,000,000. Can these things be exchanged often enough in the course of a year to account for $387,000,000,000!

These figures are for 1904,[252] whereas Fisher's figures are for 1909. If the Census Bureau had taken an inventory in 1909, the figures would doubtless be larger. The inventory for 1912 made by the Census Bureau does show a very considerable increase, the largest item being due to a rise in real estate values. The figures for agricultural, manufacturing, and mining products are, also, figures for a given time rather than for total production through the year. But, making all the allowance one pleases, it is quite incredible that one should reach a figure of $387,000,000,000 by taking only the exchanges necessary to bring raw materials through the various stages of production to the consumer. The greater part of the $387,000,000,000 is to be explained in another way!

A detailed analysis of Kinley's figures, on which the estimate of total trade is based, leads clearly to the same conclusion. Kinley's figures for the banks that reported on March 16, 1909, are as follows:

Retail deposits 60 millions
Wholesale deposits 124 millions
"All other" deposits 502 millions

The "all other deposits" are vastly greater than retail and wholesale deposits combined! Notice, too, with reference to the question as to how often goods need to be turned over in getting to the consumer: wholesale trade uses only about twice as much money and checks as does retail trade. Goods are not, if these figures are in any way typical of actual trade, turned over many times in the process of reaching the consumer. The "necessary," or "physically determined" number of exchanges, in the routine of trade, is small, per item.

Retail deposits of 60 millions make up less than one-eleventh of the total. Retail and wholesale deposits together make up about three-elevenths. What is the other eight-elevenths, represented by the "all other deposits"? It will help if we see where these "all other" deposits are located. If we find them scattered evenly throughout the country, in rural regions as well as in cities, we might be at a loss. If, however, we find them bunched in the big speculative centres, we may conclude that speculation accounts for a large part of them. We do in fact find this.

The following figures show the different classes of deposits (1) in the South Atlantic States; (2) in reserve cities; (3) in New York City alone:

_South Atlantic States:_ _Per Cent._

Retail deposits $ 3,300,000 19.0
Wholesale deposits 4,900,000 29.0
"All other" deposits 8,900,000 52.0

_Reserve Cities (including New York City):_

Retail deposits $ 24,000,000 5.6
Wholesale deposits 78,000,000 18.2
"All other" deposits 326,000,000 76.1

_New York City:_

Retail deposits 9,000,000 3.7
Wholesale deposits 34,000,000 14.0
"All other" deposits 198,000,000 82.2

It is difficult, with Kinley's figures, to get figures which exclude returns from cities of substantial size, except for a State like Nevada, where the mining and divorce industries complicate the figures. As near an approach as can be made, perhaps, is to take the State of Louisiana, excluding New Orleans from the totals. Even here, however, we include five cities of over ten thousand, among them Shrevesport, with 28,000 people. The following figures are for the State and national banks in Louisiana, exclusive of New Orleans:

Retail deposits $179,915 24.1
Wholesale deposits 246,647 33.1
"All other" deposits 318,915 42.8

We cannot tell, in these figures for Louisiana, how many banks are represented, or what the average figures per bank are. For the whole State of Arkansas, however, including five cities of over 10,000, with two over 20,000, and one of 45,000, we can get an average for ninety reporting banks. Even here we do not know where these banks are located within the State; though it is probable that they are in the larger places, and so exceed the average deposits for the banks in the State as a whole, to say nothing of the average for the smaller places. The ninety banks are almost wholly State and national banks.

_Arkansas:_ _Per Cent._

Retail deposits $232,017 25+
Wholesale deposits 231,614 25+
"All other" deposits 456,544 49+

The average for all deposits, per bank, in Arkansas is $10,224; the average for all the 11,492 banks reporting for the whole country is, approximately, $60,000; the average for the 659 banks reporting from New York State is $502,136; the average for the banks in New York City alone is doubtless much higher, but cannot be stated, as Kinley's figures do not tell how many banks reported by cities.[253]

The "all other deposits" in Arkansas are 27.8% cash, and 72.2% checks; the "all other" deposits in the country as a whole are only 4.1% cash, with 95.9% checks; the "all other deposits" of New York City are only 1% cash, with 98.9% checks.

Several facts are very clear from these comparisons: (1) the proportion of "all other deposits" increases very rapidly as we get closer to the great centres of speculation, and is lowest in rural regions; (2) the great bulk of all the deposits is in the cities. The average for Arkansas banks, for example, is only one-sixth the average of the whole country, and is only one-fiftieth the average for the banks of New York State. It is a much smaller fraction of the average for New York City, but we cannot give an exact figure. The totals reported from the rural regions are trifling, as compared with the totals reported from the big cities. This, as will be made clear in the chapter on "Statistical Demonstrations of the Quantity Theory," is not because the country reports were less complete that the city reports. New York was probably less complete than the country as a whole. It is simply because the activity of country accounts is small, the amount of trading in the country districts small, and (as shown) the _average_ for country banks is small. (3) The character of the "all other" deposits in Arkansas differs substantially from that of the "all other" deposits in New York City, as indicated by the fact that the proportion of cash is high in Arkansas--substantially higher, in fact, for the "all other" deposits in Arkansas than for all deposits, or even for retail deposits, in the country as a whole. The percentage of checks in total retail deposits in the United States, in Kinley's figures, was 73.2; the percentage of checks in the "all other" deposits in Arkansas was 72.2. We may count these Arkansas "all other" deposits as, in considerable degree, deposits made by farmers. What were the "all other deposits" made in New York City?

Dean Kinley's list of the miscellaneous elements that enter into the "all other deposits," given on p. 151, contains only two that might be expected to bulk large in New York without appearing in Arkansas. These are: _brokers_, _and stock and bond financial corporations_. Of course, theatres, hotels, publishing houses, railroads, public funds, "those who have no specific business," and rich churches, will all be absolutely much larger in New York City than in Arkansas. But these things may be found in many places, scattered throughout the cities of the country, without making anything like such "all other" deposits as New York shows. It is not New York's foreign commerce that does it, because that is represented in New York's "wholesale deposits," which make up only 14% of New York City's total deposits for the day. It cannot be the supposed "clearing house" function of New York City,[254] whereby banks in different parts of the country pay their balances due one another in New York exchange, because such transactions would appear in New York chiefly in the figures for deposits made by one bank in another, and these figures are excluded from Kinley's totals. It cannot be the deposits of the "idle rich" for current expenses that swell New York's "all other deposits" so greatly--these could not equal the total retail deposits of the city, which are only 3.7% of the total in New York. Moreover, similar deposits are made in many other cities, without, in proportion to population, making any such totals. Figures, moreover, for the aggregate yearly income of the United States, and for the distribution of that income between rich and poor, make it clear that any such items must be bagatelles in comparison with these enormous figures. The only explanation that will really explain is the speculative and investment and financial transactions that centre in New York, and, in less degree, in the other great financial cities of the country.

This is Dean Kinley's opinion. In the "all other" deposits he makes a 50% allowance for speculative transactions. "A large proportion of deposits in this 'all others' class undoubtedly represents speculative transactions, all of which, or practically all of which, are settled with credit paper."[255] It is also the opinion of General Francis A. Walker, expressed concerning similar figures from earlier inquiries.[256]

Various kinds of evidence converge toward this conclusion. Thus, the evidence of clearings, total items presented by banks to the clearing houses of the country. New York clearings are usually nearly twice as great as total clearings for the rest of the country. New York clearings fluctuate in general harmony with transactions on the New York Stock Exchange. This has been commented on many times. The extent to which it holds has recently been carefully measured by Mr. N. J. Silberling, whose results appear in the _Annalist_ for August 14, 1916, under the title, "The Mystery of Clearings." Mr. Silberling applies the "coefficient of correlation" to the problem, getting in one significant figure a measure of the extent to which two variables, as share sales on the New York Stock Exchange and New York clearings, vary together. This coefficient has been used enough by economists not to require detailed explanation here. It is a figure always between +1 and -1. +1 indicates that the two variables in question are perfectly correlated, whereas 0 indicates no correlation whatever. -1 indicates an inverse correlation, such that two variables vary exactly and inversely with reference to one another.[257]

Mr. Silberling's studies show the following correlations: New York share sales (numbers of shares, not values) to New York clearings, using weekly figures, for the years 1909-10, r = .628. This is a high correlation. Limiting the observations to the middle weeks of the month for the same period, he gets r = .731(46). The reason for taking only middle weeks in the month is that thereby the disturbing factor of monthly settlements is avoided. The monthly settlements may be for stock transactions, or may be for other things, but as they are not dependent on the stock transactions _of the week_ in which they occur, their effect is to lessen the evident degree of connection between stock sales and clearings. Thus the middle weeks show a closer correlation between the two variables than do all the weeks taken as they come. If figures for the month were taken, this complication would be smoothed out, and a fairer result might be expected to appear. The middle weeks, eliminating monthly settlements, probably eliminate more other things than they do share sales (which are in large degree paid for in 24 hours[258]), and so exaggerate somewhat the relation between shares and clearings. Monthly figures avoid both complications, though they lose something of the concrete causation. An intermediate figure might be expected for the monthly correlation, and this we find: r = .718(23).

A striking single fact in connection with these figures, giving them point as less extreme variations could not do, is found in the behavior of clearings when the Stock Exchange was closed, during the crisis of 1914. At that time, New York clearings, which had been about twice as great as country clearings, fell suddenly _below_ country clearings. When the Stock Exchange was opened, the old proportions suddenly reappeared.

That speculation spreads far beyond New York, New York being the centre for dealings in securities, etc., which involve the whole country, is, of course, well known. The extent of this Mr. Silberling seeks to measure by correlating clearings outside New York with New York share sales. His weekly correlation for these two variables for 1909-10 gives r = .368(103), and the correlation for the mid-weeks gives a higher figure, r = .424(46). The monthly correlation shows r = .257(23), a lower figure, "which is perhaps due in part to the fact that the bulk of the outside monthly clearings show relatively moderate fluctuations, because of their diverse composition, and are less sensitive than the periods of shorter length."

Seeking an index of the variations of that trade which is, in Professor Fisher's phrase, governed by "physical capacities and technique"--a law which Professor Fisher,[259] as we have seen, would apply to the great total of 387 billions which he has constructed--Mr. Silberling chooses the gross earnings of the principal railways as the best available test. Railways deal with all manner of other enterprises. He correlates this with clearings outside New York. "The question might arise at once whether changes in traffic are strictly concomitant with changes in payments involved by it, and therefore with the clearings resulting. The preliminary hypothesis that a 'lag' ensued between traffic and the bulk of the payments was first tested by correlating the railway figures with clearings of one month[260] and two months later, but no correlation was obtained. The direct month-to-month correlation yielded, however, a result r = .524(23)." This suggests that outside clearings are, in substantial degree, an index of physical trade, but Mr. Silberling calls attention to certain chance agreements between railway traffic and speculation in cotton and produce and grain, speculation in the crops which are in current movement, and regularly recurring concomitances between traffic and speculation in March, when the railway traffic revives after the February lull, and when there is a large mass of dealing in Spring deliveries in Chicago. In view of the facts later to be developed, with reference to the small actual value of the necessary physical exchanges (partially covered already) as compared with clearings, this query is well put. We may easily have here a "spurious" correlation. Taking it at its face value, however, and taking the correlation as indicating the influence of physical trade on bank transactions, we get the following results, when _total clearings for the country_ are compared with (a) New York share sales, and (b) with railway gross earnings: (a) r = .607(23); (b) r = .356(23). "Physically determined trade" is at best a minor factor in that total "trade" represented by bank transactions!

Mr. Silberling has buttressed his results with a consideration of various alternative possibilities which might give them a different interpretation. I need not, for present purposes, go further into his figures.[261] Taken in conjunction with the other data presented, and to be presented, together with the theoretical discussion of the nature of trade, and its relations to money and credit, which the present volume contains, they give the present writer abundant confidence in the thesis that the great bulk of trade in the United States is SPECULATION, rather than that sort of trade which is determined "by physical capacities and technique."

The figures given above, of the inventory of wealth at a given moment of time, by the Bureau of the Census, show only trifling magnitudes, as compared with the estimated 387 billions of deposits made in 1909, of items which could enter into ordinary trade, as distinguished from speculation and dynamic readjustments. An effort to calculate ordinary trade on the basis of figures running through the year may throw further light on the problem. Railway, gross receipts for the year ending June 30, 1909, were less than two and a half billions. This is six-tenths of 1% of the total. Receipts of the Western Union Telegraph Company were $30,451,073--less than one-hundredth of 1%. The Post Office in the fiscal year ending in 1909 took in $203,562,383. This is something over one twentieth of 1%. These are gigantic sums. But they are insignificant indeed in this computation. Millions of smaller items simply do not count at all--ten million items of $387 each would give only 1%. The total net income of the United States, as estimated by W. I. King for 1910, including all forms of income, dividends, interest, wages, rents, profits, salaries, etc., is $30,500,000,000[262]--around 7% of the 387 billions.

Let us sum up the major items of ordinary trade. From Kinley's figures, we may get some idea of the proportions of wholesale and retail trade to the total for 1909, assuming that the deposit figures indicate that total. Retail deposits make up less than one-eleventh of the total, and wholesale deposits about two-elevenths. The figures were: retail, 60 millions, wholesale, 124 millions, and "all other," 502 millions. But the "all other" deposits were lower than normal. New York City was, in the first place, probably less complete than the rest of the country, in the figures returned, and, in the second place, New York City, as shown by the clearings of March 17 (the next day, when checks deposited in New York would get into the clearings) was 28% below normal. The rest of the country was within 3% of normal.[263] Not to refine matters too much, we shall, on the assumption that the variable element in New York deposits is connected with the Stock Exchange (as shown by Mr. Silberling's correlations and other considerations), and on the assumption that deposits connected with the stock market appear in the "all other" deposits, add a little over 20% of New York's total of 198 millions, or 40 millions, to the "all other" deposits for the country, leaving the wholesale and retail deposits unchanged. What error there is in this is favorable to the wholesale and retail deposits. Our proportions, then, are: retail, 60, wholesale, 124, "all other," 542, total, 726. If the retail deposits correctly represented retail trade, we could then say that retail trade was a little less than one-twelfth of the whole, and wholesale trade about one-sixth. But there are many speculative transactions engaged in by wholesalers, and a good many by retailers. The writer knows a small delicatessen dealer on Amsterdam Avenue, in New York, who frequently speculates in eggs and canned goods. A colleague in the Harvard Graduate School of Business Administration is authority for the statement that speculation in canned goods and some other things is quite common among retailers, particularly "hedging" by the use of "futures," in canned goods. Speculation among wholesalers is very extensive. The same is true of manufacturers. The same authority cited some cotton manufacturers whose profits from cotton speculation are greater than their profits from manufacturing. We shall see reason to suppose that a very substantial part of manufacturers' deposits were included in the wholesale deposits. That the figures for retailers' deposits exaggerate the retail trade may appear from several considerations: (1) The proportion of checks to cash reported is too high: 73.2%. Dean Kinley allows 5% of the checks deposited to be "accommodation checks,"[264] cashed for customers, rather than taken in in trade. (2) If retail deposits are taken as exactly representative of retail trade, we should get a retail trade for the year of over 32 billions (1/12 of 387 billions), which would exceed the total income of the country as calculated by King for 1910. Dean Kinley reached the conclusion that the retail deposits reported in 1896 also exceeded the probable retail expenditures.[265] Of course, not all of retail trade is in consumption goods. Hardware stores, lumber stores, and some other retail establishments sell, not only to householders for domestic use, but also things which enter into further production, and so do not come out of annual income. If we include in retail trade various items which were not included there in Kinley's figures, such as hotels, theatres, newspaper receipts from subscription and street sales, physicians' fees, etc.--all those items which enter into the domestic budget, including domestic service, we should still not be justified in reaching a total as great as the total income of society, since there would then be no allowance for savings, which we should not count in trade, or for life insurance, which we shall count separately. The items sold at retail which enter into further production cannot make a great total, since large producers buy such things at wholesale. Total retail trade, therefore, and, in addition all the other items in the domestic budget, must be held below the figure for total national income. Suppose, to be very liberal, we allow 29 billions[266] for all these items, under the general head of "retail trade."

For wholesale trade, if we take the figures at face value, the estimate would be 65-3/4 billions (124/726 of 387 billions, or 17% of 387 billions). But we have seen that there is a great deal of speculation among wholesalers. Not all of their deposits, by any means, represent receipts from ordinary business. Moreover, there is much overcounting here, several checks being used for one transaction, especially where wholesalers have branch houses, and checks connected with loans and repayments, and transfers of funds from one bank to another. How much we should subtract for this there is no way to tell. In the case of retail figures, we have the additional check of the figures for total net income, but there is no such check here. We shall, therefore, make no subtraction, but shall content ourselves with pointing out that we are allowing many billions[267] to "ordinary trade" to which it is not entitled, which will much more than offset errors in the opposite direction which the reader may find in our computations.

Do manufacturers' receipts from first sales belong in the wholesale deposits, or must they be counted as a separate item? Dean Kinley does not say. In his list of items, as reported by banks, that go in the "all other" deposits,[268] he does not mention manufacturers, and the item is far too important not to have been mentioned by so careful a writer had he supposed that it belonged there. If manufacturers' first receipts belong, not in the wholesale deposits, but in the "all other" deposits, then we should expect manufacturing cities to show a high percentage of "all other" deposits as compared with wholesale deposits. The city of Pittsburg should be a good test case. The figures there, for State and national banks and trust companies, are:

_Per Cent._
Retail deposits $1,061,420 9.6
Wholesale deposits 3,368,004 29.7
"All other" deposits 6,672,378 60.6

For Pittsburg, the percentage of "all other" deposits is lower decidedly than the percentage for the country as a whole (about 75%), much lower than for cities where there is active speculation, as Chicago and St. Louis, to say nothing of New York, and is closer to the percentage of the South Atlantic States, 52%, than to the average for the country. The wholesale deposits of Pittsburg, however, rise to 29.7%, as against an average for the country of 17%. There is nothing in these figures to suggest that manufacturers' first receipts are exclusively in the "all other" deposits. I should think it safe to hold that a substantial part of them were included in wholesale deposits, and so already accounted for in our estimate. The total value of products manufactured in 1909 was $20,672,051,870. I shall allow $5,672,051,870 of this to have been already accounted for in our estimate of wholesale trade, and count 15 billions of it as a separate item. If there is an error here, it is very much more than offset by our failure to subtract anything from the wholesale figures for speculation. I think it probable that much more of the figures for manufactures should be assigned to the wholesale figures than I have assigned.

To these figures, we may add a number of other items, absolutely great, but insignificant, in comparison with the 387 billions not only, but also with the figures for retail and wholesale trade already reached. These are: total farm value of farm products (not nearly all of which is sold off the farm) $8,760,000,000; total mineral products, $1,886,772,843; total mill value of lumber, $684,479,859; total life insurance premiums (much of which is savings, and in no proper sense trade), $748,027,892; total fire, marine, casualty and miscellaneous insurance, $362,555,850; total wages and salaries, $14,303,000,000; total land rent, $2,673,000,000;[269] and the items for railway gross receipts, post office, telegraph, already mentioned. The total of these items, together with retail and wholesale trade and manufactures, is $141,860,618,000. This is only 36.6% of the total of 387 billions. It leaves over 245 billions unexplained. What can the 245 billions represent? There is really no way in which ordinary trade can make up more than a very few more billions, so far as I can see. There remain no items as big as 1% of the total, and, as we have seen, small items, of hundreds of dollars each, are like "infinitesimals of the second order"--they simply do not count at all when such staggering figures are involved.[270]