moderate activity, like those from 1874 to 1878, always show large reserves, and therefore little need for money; while prosperous years, like those from 1879 to 1882, or from 1887 to 1889, show moderate or rather low reserves, i.e., much need for money. (See Diagram I, 2, b-f.)

II.

Assuming it to be sufficiently established that the money need is subject to very considerable variations at different seasons and in different years, it is now incumbent upon us to consider whether the failure of the volume of money to adjust itself to these variations results in harm. That such harm does result, at first thought seems too evident to need proof. "Surely," one argues, "no one can doubt that if possible the supply ought to correspond to the demand; and that if it does not, some degree of evil must follow." But a second thought raises doubts. Variations in the need for a given commodity are not uncommon in the industrial world, and they do not cause serious inconvenience, though the total stock does not expand or contract in accord with these variations. For example, the street-car system of any large city is called upon to do extra work at certain hours of the day; but of course no one demands that there should be some process by which the number of cars in existence can be suddenly increased or diminished. The variations in need are met by having large stocks, which are idle during the quiet hours, and are brought into requisition when the special need arises. Just so, it may easily be argued, should it be in the case of money. A good banking system ought to furnish all really necessary elasticity. The process described above, whereby the idle money of the country flows into the bank reserves, first, of the small towns, then of the larger, and finally, of the central reserve cities, is just what ought to be. And the banks in the central reserve cities ought to remember that the money they receive is only temporarily idle that it will be wanted again in a few months; and, remembering this,

they ought so to manage as to have the money ready when it is needed. Doubtless there is much force in this presentation of the case; and doubtless banks do reckon upon such expansions of need and do more or less fully provide for them in advance.1 Nevertheless, bankers are, after all, private citizens in quest primarily of their own interests. To the mind of a bank officer, a reserve does not present itself in the first instance as a part of the equilibrating reservoir of the currency. To him the reserve is rather the bank's stock in trade the store from

which is drawn the cash, and the foundation that at once sustains and limits the credit which the bank loans for a consideration. Bank officers will always be tempted to treat excessive reserves as so much extra loaning power, and so will be tempted to inflate the bubble of credit. Further, there will always be danger that in thus treating their surplus reserves they will leave themselves insufficient resources for the time when the need for money has expanded. In a word, under normal conditions a banking system as such provides against excess or deficiency in the ordinary circulation, but not against excess or deficiency in the bank reserves themselves.

It would, perhaps, be begging the question to say at the outset that our system is subject to alternations of excess and deficiency in the bank reserves; since "excess" and "deficiency" imply injuriously high or injuriously low bank reserves. But that each year shows alternations of decidedly high and decidedly low reserves needs little proof. The data already used to show the variations in money need, show the existence of such alternations during the years 1886-92. My own studies have covered the period from 1871 to 1895; and of those twenty-five years there is only one which does not show within its limits a variation of more than fifteen millions of dollars in the surplus reserves, while a variation of twenty-five millions is not uncommon. In like manner, whole years and groups of years compared with one another exhibit similar alternations of high and low reserves. For example, 1885 and 1894 were years of extraordinary plethora; 1874 and 1895, of

1 Kinley, The Independent Treasury, Appendix vi.

decided plethora. (See Diagram I, 2, b-d.) On the other hand, outside the panic years, at least three were years of pretty constant stringency-the reserves being below the legal minimum during nine or ten weeks of each year, and never averaging for a month above a ten-million surplus. Are such variations in any degree harmful?

In the first place, it is plain that whatever harm may be caused by alternations of high and low bank reserves, must come, in the first instance, from their influence on the loan market. In that market they will evidently tend to cause alternations of ease and tightness. Of such ease or tightness, two sorts need to be noted. We may mean by ease a state of things where the borrower finds no difficulty in getting advances on ordinary securities; and tightness would of course be the reverse. On the other hand, we may mean by ease a condition of things where a low discount rate prevails, and by tightness, the reverse. In minor markets the one will often be present without the other; in a great center like New York they are usually found together: but, in any case, they are to be considered distinct forces as regards their power for good or evil.

That alternations of high and low reserves do give rise to alternations of ease and tightness in the first sense, scarcely needs proof. The sort of credit and the sort of security which lenders demand, vary constantly with the abundance or scarcity of the funds at their disposal. If they have abundance, they cannot afford to be particular. In the opposite condition, they must reject some applications, and will, of course, reject the least desirable. If one opens a copy of the Chronicle for some date in the fall of almost any year, his eye will fall on sentences like these:

Lenders carefully scrutinized the collateral, declining to loan upon a certain class of stocks and bonds; and they required unexceptionable security well margined [Oct. 3, 1891].

Loans on such collateral as will sell in the market are now the principal method of employment of capital. As none but those who have the least need for money have the collateral, this usually cuts off a great part of the usual loans by banks [Oct. 6, 1883].

But alternations of high and low reserves no less certainly give rise to alternations of ease and tightness in the second sense, that is, to low and high rates of discount. The contrary opinion of some eminent economists cannot possibly stand against the overwhelming evidence of experience. In determining the short-time rate, the state of the bank reserves is undoubtedly one of the most important elements. To put to rest any lingering doubts on this matter, I have made detailed comparisons week by week of the surplus reserves and of the rates of discount in New York City for seven

years, and have found the usual opinion on this point entirely confirmed. As the phenomena do not readily admit the use of averages, I have taken the year 1889 as typical and have embodied the facts of that year in the diagram accompanying. The connection of high bank reserves with low rates, and of low reserves with high rates is here very evident.

Accepting as established, then, the proposition that alternations of high and low reserves will almost invariably be followed by alternations of ease and stringency in the loan market, the question still remains: Are these alternations of ease and stringency productive of harmful consequences? In the first

place, it would be natural to argue that these conditions of themselves increase greatly the risk and uncertainty of all industrial operations. Unexpected tightness in the loan market means that thousands of men who are doing business on borrowed capital suddenly find themselves unable to get their usual supplies at any price. This will inevitably cause some losses, considerable curtailing of business, and not infrequently bankruptcy. A mere rise in the rates is much less serious, since it is less likely to cause failure; but it will certainly entail notable losses. How great these losses are there is no method of ascertaining statistically; but some idea of their extent can perhaps be obtained by reflecting that in September of 1892 the different banking institutions of this country had out, on demand or short-time paper, loans aggregating about three billions of dollars, and that half of this sum was loaned in nineteen large cities, where rates are subject to more or less constant fluctuation.

According to a common view, there is another way in which alternations of ease and stringency in the loan market increase the risks of business; namely, by causing abnormal fluctuations in prices. Such an effect, it is held, is particularly clear in the case of securities and of those staple commodities which are speculated upon in the great exchanges of the world. The reasoning upon which the above conclusion is based is comparatively simple. An easy money market means that it is easy and cheap for the prospective buyer to get the funds with which to make his purchases. Demand is therefore excessive, and prices rise. A tight money market, on the other hand, means that it is difficult and expensive for the buyer to get funds. Demand consequently falls off, and prices decline.

Without doubt the above reasoning is extremely plausible, but a somewhat extended investigation of the facts has failed to furnish any considerable confirmation from experience. I am not prepared to say that there is no effect of this sort, but I feel sure that the amount is in common opinion greatly exaggerated. A high level of prices seems perfectly compatible with the average fall stringency, and a low level with the