Credit Policies of the Federal Reserve System
XIV: The Federal Reserve System and the Treasury
CHAPTER XIV
THE FEDERAL RESERVE SYSTEM AND THE TREASURY
It is a generally accepted principle of modern central banking theory that central banks should be free from the control of the treasuries of the countries which they serve. It is the function of a central banking system to maintain a sound credit situation, and that objective is bound to conflict to a certain extent with the interest of the treasury in borrowing as cheaply, and at times as extensively, as it can. In this respect the situation of a treasury differs from that of other borrowers only in respect to power.
Compliance with the desire of national treasuries for cheap money having been the source of most of the recent disastrous inflations, the charters of the central banks which have been founded during the stabilization era have as a rule contained elaborate provisions designed to maintain the independence of the banks from government, and especially from treasury, control.1 I believe that this is a sound principle, but one which is certain in practice to be forgotten when war or other public emergency makes it necessary for treasuries to mobilize resources quickly.
At the time of the creation of the Federal Reserve system, comparatively little attention was given to the question of proper relations between the System and the United States Treasury. Federal government borrowing was of minor importance, and it was not anticipated that the Treasury as a borrower would have an important interest in the credit policies pursued by the Reserve system. Inclusion of the Secretary of the Treasury in the membership of the Board was suggested by the fact that the Treasury was expected to be a heavy depositor in the Reserve Banks, and also by the fact that the Treasury for many years had been exercising some of the functions of a central bank, putting funds into the market in times of seasonal strain or of crisis, and holding them idle as a reserve in ordinary periods.
The World War changed this situation completely. Credit policy, like every other field of public administration, became an instrument of warfare. Not only was the pyramid of bank credit expanded to facilitate war finance, but discount rates were kept low, Liberty Bond paper was given preferential treatment at the Reserve Banks, and member banks were encouraged to load themselves with government obligations and with customers’ paper secured by such obligations. Protests of Reserve Board members were met by intimations that unless Reserve authorities were willing to play the part assigned to them the Treasury would have to take over direct control of the Reserve system.2 Not until January 1920 was the Reserve Board formally free to shape its policies in accordance with commercial, agricultural, and industrial, rather than fiscal, necessities.
That this situation existed is not disputed. Critics of the System and apologists alike place the responsibility for policies pursued during the war and early post-war years on the Treasury rather than on the Board or on the management of the Reserve Banks.3
The situation during the years since 1919, however, is not so clear. It is obvious that discount policies have no longer been dominated by the exigencies of Treasury financing to anything like the extent that maintained during the war. Whether the System has been entirely free in its credit policies from Treasury interest in cheap money is a much more difficult question. It cannot be settled by examination of statistical and historical data, since only actions, not motives, are matters of record. All that we can do is to examine the public utterances of responsible public officials and others in a position to have an intelligent opinion concerning the motives back of Reserve system policy, and compare these utterances with the record of actual Treasury and Reserve Bank transactions. Moreover, Reserve officials in their public utterances have naturally been reticent on the question of Treasury domination of their policies. Our conclusion must be based more on the records of actual practice than on statements of policy.
In the hearings on the Strong bill, Adolph C. Miller, the senior member of the Federal Reserve Board, said: “There is a constant disposition not to work at cross purposes, but to let the Treasury’s program, whenever it is practicable, work in with the Federal Reserve’s.” Standing alone this statement might possibly be interpreted as an avowal of purpose to use the Reserve Banks as a means of manipulating the market to favor the Treasury’s program. In its context, however, it seems to mean rather the converse, namely, that the Treasury’s operations are timed so as to support the program of the Reserve system.
At the hearings held by the Senate Committee on Banking and Currency in 1931, Governor Harrison was asked concerning the influence of the Treasury on Federal Reserve policy and testified as follows:
The Chairman [Senator Glass]. Let me ask you one question right here. Does the Treasury undertake to influence the action of the New York Bank, or any other Federal Reserve Bank, in transactions of that sort [open market operations in government securities]?
Governor Harrison. Never, now.
The Chairman. I know it did when I was Secretary of the Treasury.4
Governor Harrison. That is the reason I said “now.”
The Chairman. And I thought it was a pretty vicious thing to do and was only done under war necessities, or immediate post-war necessities rather, and I wondered whether it were continued or not.
Governor Harrison. I think there was a time when as a result of the pressure of war necessity the interest of the Treasury was a very strong factor in certain Federal Reserve policies. I think that was really not a matter of very severe criticism in the circumstances, but in recent years, and since that period has terminated, there has never been any effort on the part of any of the Treasury officials that I know of, as far as the Federal Reserve Bank of New York is concerned, to influence our rate policies or our operations in government securities.5
It has been stated repeatedly by H. Parker Willis, whose background of experience as the first secretary of the Federal Reserve Board gives his statements especial interest, that the practice has been for the Reserve authorities to shape their credit policies more or less continuously with a view to helping the Treasury to carry through its borrowing operations.6 It is of interest therefore to compare the record of the Federal Reserve rediscount and open market operations with the record of the more important financial operations of the Treasury in order to see whether the facts support this charge.
Since 1920 changes in rediscount rates do not appear to have been timed so as to facilitate Treasury borrowing. Let us examine first the record of the long-time borrowings from January 1, 1922 to September 15, 1931. The accompanying table shows the dates and amounts of the issues of more than five years’ maturity, and the rediscount rates which were charged at New York at the time the bonds were issued.
New York Rediscount Rate at Time of Long-Term Treasury Borrowings, 1922-31
| Date | Borrowings | Rediscount Rate (Per cent) | ||
|---|---|---|---|---|
| Amount (In millions of dollars) | Maturity (In years) | Yield (Per cent) | ||
| Oct. 16, 1922 | 764 | 25-30 | 4.250 | 4 |
| Dec. 15, 1924 | 757 | 20-30 | 4.000 | 3 |
| Mar. 15, 1925 | 290 | 20-30 | 3.940 | 3 1/2 |
| Mar. 15, 1926 | 495 | 20-30 | 3.710 | 4 |
| June 15, 1927 | 495 | 16-20 | 3.360 | 4 |
| July 16, 1928 | 359 | 12-15 | 3.375 | 4 1/2 |
| Mar. 16, 1931 | 594 | 10-12 | 3.375 | 2 |
| June 15, 1931 | 821 | 15-18 | 3.125 | 1 1/2 |
| Sept. 15, 1931 | 800 | 20-24 | 3.000 | 1 1/2 |
At the time of the first issue, in October 1922, the New York discount rate had been unchanged for nearly four months, and it remained at 4 per cent until February 23, 1923. The second issue was brought out in part on December 15, 1924, when the rediscount rate had stood at 3 per cent for over four months. The remainder of this series was brought out on March 15, 1925, just two weeks after the rate had been advanced from 3 to 3 1/2 per cent. The issue of March 15, 1926 was brought out less than a month before a one-half per cent reduction in the rediscount rate, and that of 1927 came out about six weeks before a similar reduction. The issue of 1928 furnishes the most plausible basis for the charge that rate changes were timed to aid Treasury needs. The subscriptions were closed on July 7 and the rediscount rate was advanced to 5 per cent on July 13, three days before the bonds were actually floated. The issues of 1931 came in periods of very low rediscount rates, but the case for low rates in terms of the business situation was so strong that it is hardly likely rates would have been higher in the absence of any consideration of fiscal needs.
In summary, leaving out the three issues of 1931, one issue was brought out just before a rate advance; one just after a rate advance; two came just before rates were lowered; and in two cases there was no rate change at a date near that of the issue. This record certainly creates no presumption that the Treasury has controlled the rates in its own interest.
The record of short-term note issues likewise gives no indication of a policy of lowering rates at the particular times when they are being floated. The record of issues is shown in the table on the following page.
Issues of Treasury Notes7
| Date | Amount (In millions of dollars) | Yield | Maturity |
|---|---|---|---|
| Feb. 1, 1922 | 602 | 4 3/4 | 3 yrs. 1 1/2 mo. |
| Mar. 15, 1922 | 618 | 4 3/4 | 4 yrs. |
| June 15, 1922 | 660 | 4 3/8 | 3 yrs. 6 mo. |
| Aug. 1, 1922 | 487 | 4 1/4 | 4 yrs. 1/2 mo. |
| Dec. 15, 1922 | 469 | 4 1/2 | 2 yrs. 6 mo. |
| Jan. 15, 1923 | 367 | 4 1/2 | 4 yrs. 11 mo. |
| May 15, 1923 | 1,336 | 4 3/4 | 3 yrs. 10 mo. |
| Mar. 15, 1927 | 1,360 | 3 1/2 | 3-5 yrs. |
| Sept. 15, 1927 | 619 | 3 1/28 | 3-5 yrs. |
| Jan. 16, 1928 | 607 | 3 1/2 | 3-5 yrs. |
Occasionally there have been cases when the sequence of events was such as to support such interpretation; at other times precisely the opposite relationship appeared. In 1922 the rediscount rate at New York stood at 4 1/2 per cent from January till June 22; for the rest of the year and for the first six weeks of 1923 it was 4 per cent. During the 4 1/2 per cent period a billion and a half of three-to-four year notes were issued, and during the 4 per cent period, a little over a billion and a quarter. A 4 1/2 per cent rediscount rate was maintained at New York from February 23, 1923 to May 1, 1924, during which period about $668,000,000 was borrowed on short notes. During the period from May 1, 1924 to February 27, 1925, when rates were being progressively lowered, no notes were issued. In the remaining ten months of 1925, with New York rediscount rates at 3 1/2 per cent, about one billion dollars was borrowed in this way.
In 1926 no borrowing of any sort took place during the four-month period (April 23 to August 13) when the New York rate was 3 1/2 per cent, but during the ensuing period when the rate was 4 per cent, 1,360 million dollars of notes were issued. In the period of 3 1/2 per cent money, from August 1927 to February 1928, note issues amounted to about 1,200 million dollars. In February, in May, and in July 1928, the rate was advanced one-half per cent, the dates of the changes bearing no obvious relationship to the Treasury borrowings.
On the whole the record fails to substantiate the suggestion that during the past decade the Federal Reserve rediscount policy has been controlled with reference to Treasury needs. Let us examine next the record of open market operations.
The timing of open market purchases in relation to Treasury borrowing also indicates that the Reserve system has not been an instrument of Treasury finance. Obviously, if the Treasury sales of new issues tend to coincide with Reserve Bank purchases, the effect of the coincidence is to lessen the influence of Reserve Bank operations on the money market, and to enable the Treasury to sell its bonds at high prices. On the contrary, if Treasury sales and Reserve Bank sales coincide, the effect of the Reserve Bank operations is reinforced, and the Treasury has to sell on less favorable terms. The evidence is not conclusive one way or the other, but on the whole suggests that the co-operation between the two organizations has been such as to make the Treasury operations reinforce those of the Reserve Banks, rather than the reverse. The facts are these:
The long-term issue of October 1922 came at a time when the Reserve Banks were reducing their open market holdings. So did the issue of 1924-25. The 20-30 year loan of 1926 came at a time when the Reserve Banks were neither increasing nor decreasing their holdings to any considerable extent. The 1,360 million dollar sale of March 15, 1927 was carried through some months before the Reserve Banks began the extensive series of security purchases which was described above (pages 47-48). The issue of July 1928 came at a time when Reserve Bank holdings were at the lowest ebb in more than four years; the issues of 1931 came when they were extremely high.
In short, in 1922, in 1924-25, and in 1928, large Treasury issues were made at times when the Reserve system was actively operating to tighten the market, and in 1927 a still larger issue was made several months before a great expansion of holdings started. There is nothing in this record, any more than in the history of rediscount rate changes, to suggest co-operation between the Reserve authorities and the Treasury to enable the Treasury to do its borrowing at low rates. Indeed, it might more plausibly be argued from these cases that the Treasury has become an instrument of Federal Reserve policy.
Aside from the question whether in specific instances the market has been rigged to help the Treasury, there is the more elusive question whether, regardless of year-to-year changes, the whole level of money rates fostered by Federal Reserve policy from 1922 to 1928 was so low that it must be regarded as a concession to the Treasury’s interest in cheap money. This is a question on which opinions are likely to differ, partly in accordance with one’s views as to what constitutes a high or a low rate level, and especially in accordance with the weight one attaches to other factors which influenced, or might have influenced, Federal Reserve policy. Advocates of a policy of keeping rediscount rates so high as to make rediscounting only an emergency procedure can easily believe that the failure to follow this course was due to Treasury domination. On the other hand, those critics who condemn the Reserve authorities’ policy as one of deflation, of stabilizing gold imports so as to prevent them from raising prices and lowering interest rates, will scoff at this idea.9
To me it seems very unlikely that Treasury policy has been a major factor in the decision to keep the general level of rates either as low or as high as it has been kept. It would be rash to say that this factor has been given no consideration, but at the times when low rate policies have prevailed there have been other powerful forces working in the same direction, and at times when they have not prevailed the Treasury’s interest in low rates has been as great as at other times. Certainly if there has been a dominant purpose to help the Treasury borrow cheaply, the policy has been carried out very ineffectively.
The Reserve system has co-operated with the Treasury’s program of keeping afloat a large volume of short-dated debt. Before the war the debt of the United States, other than current bills, was all funded. During the war there was inaugurated a practice of issuing short bills to meet current expenses, later refunding these bills from the proceeds of successive issues of Liberty bonds. At the close of the period of war finance, the refunding process was left incomplete, and ever since that time there has been a definite policy of keeping from one-half to one billion dollars of the Treasury’s outstanding obligations in the form of short-term certificates, which are owned chiefly by banks. By leaving the proceeds of sale of these certificates on deposit with purchasing banks, the Treasury has made them more attractive investments for banks than for investors. Moreover, the fact that member banks can borrow from Reserve Banks, using the certificates as collateral, has made them an ideal secondary reserve.10 Finally, the Reserve system’s practice of buying and selling these certificates as a routine method of influencing the credit situation has made a market for them and thereby given them some added value.
It is beyond the scope of this book to estimate the fiscal results of the Treasury’s policy of keeping the government debt in the commercial banks in the form of short-term certificates. Our interest is in the part played by the Reserve Banks in supporting this policy and in the resulting changes in the credit structure of the country. Is the sale and purchase of short-time government paper by Reserve Banks an effective method of carrying out a policy of easing or tightening the money market? Does the use of government paper for this purpose, or as collateral for member bank borrowings, interfere with the attainment of other objectives of Reserve policy?
To the first question the answer must be in the affirmative. For the purpose of regulating the pressure on the Banks, through changing the form of Reserve credit from open market purchases to discounts and vice versa, the practice of buying and selling short-term government securities suits the convenience of the Reserve Banks admirably.11 Such securities are not essential for the purpose, however. If they were not available, acceptances would probably be available in greater volume, and if the supply of these were not sufficient, Liberty bond purchases and sales could be made to serve the same purpose.
The policy of lending freely to member banks on the collateral of government securities has been criticized as making access to Reserve Bank funds too simple and too convenient, an objection which to me does not seem valid. If member banks are to be discouraged at any time from using Reserve credit, the proper method is to make it more expensive or, if necessary, to ration it—not to make it more cumbersome and awkward to utilize. Even now the use of the Reserve system by country banks is restricted by the fact that city correspondents sometimes make it convenient for them to borrow from the city banker rather than from the Reserve Bank. The fact that borrowings can be made on the security of government paper without the labor and red tape involved in using acceptances or rediscountable commercial paper is a gain.
Objection has been raised in some quarters to the policy which has been followed jointly by the Treasury and the Reserve system because of its effects on the development of the acceptance market and on the practice of financing through commercial paper. The injection into the banks of a large volume of government paper, nominally short-time but really of investment character, is regarded as one more step in the process of tying up the funds of the banking system in assets which can be liquidated only through sale in the market. The decline in the volume of eligible paper held by member banks is in large part attributed to the fact that eligibility is of little importance from the bankers’ standpoint. If banks could not borrow on government paper, they would have an added interest in encouraging business men to obtain capital through short-term “commercial” paper and acceptances, rather than through the security markets.
The policy with reference to short-term government securities conflicts with the policy of encouraging the use of acceptances. As was noted in Chapter XII the Reserve system has professed great interest in the development of the bank acceptance and has given it almost continuously a rate preference over rediscounted paper. The System has not, however, given the acceptance a preference over government paper in its purchases. As was pointed out on page 261, one prime cause of the reluctance of commercial banks to buy acceptances is the fact that under present conditions Treasury certificates are a better investment, because they are tax-exempt and readily salable, and carry with them a preference in the securing of government deposits.
The issue as to the desirability of funding the short-term debt reduces itself, in so far as it is a banking and not a fiscal question, to the same one which was discussed in Chapters XII and XIII, that is, whether bank funds should be put out through the medium of self-liquidating instruments in the old sense, or whether banks obtain sufficient liquidity by buying and accepting as loan collateral securities which have an open market salability but cannot be liquidated outright.
If we hold fast to the ancient tradition of sound banking, the policy pursued jointly by the Treasury and the Reserve Banks must be condemned; it has been a powerful factor undermining the commercial paper market and obstructing the growth of the bill market. As is indicated more fully in Chapter XVII, my judgment is that these results have not been harmful to the credit structure of the country. The marketability of the United States Treasury certificates has made the development of a market for commercial paper and acceptances difficult, but it has done it by providing what seems on the whole a very satisfactory substitute.12
- 1The distinction between “banking policy” and “credit policy” is taken from Annual Report of the Federal Reserve Board, 1928, pp. 9-10.
- 2Compare pp. 132-40.
- 3This is, of course, in substance the position of the Banking School in the British banking controversy of the early nineteenth century.
- 4“. . . When increased credit demands can be met only by recourse to the Federal Reserve Banks, the volume of Reserve Bank lending is a sensitive indicator of credit conditions. . . .” (Federal Reserve Bulletin, 1923, Vol. 9, p. 411).
- 5Annual Report of the Federal Reserve Board, 1923, p. 10 (quoted below, pp. 79-80); Federal Reserve Bulletin, 1928, Vol. 14, p. 6.
- 6The omission of reference to price changes in the text is intentional and without prejudice. I believe that there is need of a thorough reexamination of the orthodox statement of the influence of gold movements on price changes and the extent to which in pre-war days changes in price levels were caused by gold movements and operated to check the gold movements.
- 7a Compiled from Annual Reports of the Federal Reserve Board and from Federal Reserve Bulletins, Vol. 18, pp. 186, 352, 358, 400.
- 8b Bureau of Labor Statistics index; in points (average for 1926=100).
- 9For fuller discussion of these points, see Chap. IX; compare also B. H. Beckhart, The Discount Policy of the Federal Reserve System, Chap. II.
- 10The cash balance carried in the United States Treasury was occasionally drawn upon to meet seasonal or crisis demands, either through the depositing of funds in banks or through the purchase of bonds by the Treasury, but the device was very crude and there was no continuity of policy in its management.
- 11Compare chart, p. 35.
- 12The possibility of temporary nullification of a central bank’s control through the mobilization of fresh resources is illustrated by an action of the Midland Bank, which in 1927 bought securities so heavily as to lower its own reserve ratio by several per cent. The amount involved was small enough so that the Bank of England was able to offset it. The net result was simply to increase the earning assets of the Midland Bank and decrease those of the Bank of England; not to ease the market.
