Riefler presented the reluctance view as a central element of his theory. The importance of bank indebtedness in the transmission of policy reflected the banks’ inability to control borrowing and their unwillingness to remain in debt.
33. A complete statement of the Riefler-Burgess framework is part of Brunner and Meltzer 1964a. This section is based partly on that paper.
34. As already noted, Strong and Stewart contributed independently. Strong read and commented on an earlier edition of Burgess’s book. The framework evolved to reflect major changes, notably the large increase in excess reserves in the 1930s. It remained as a guide to policy into the 1950s.
35. Riefler (1930, 21–22) compared the behavior of borrowing under the “reluctance” and “for profit” motives. Banks could have borrowed to equalize rates during the 1920s. When open market rates were above discount rates, banks would have brought them down if they borrowed for profit. This argument ignored risk elements.
36. “When the member banks find themselves continuously in debt at the Reserve banks, they take steps to pay off the indebtedness. . . . Conversely, when most member banks are out of debt at the Reserve banks, they are in a position to invest their funds; and money rates, including commercial paper rates become easier. The relationship rests largely on the unwillingness of the banks to remain in debt at the Reserve banks” (Burgess 1936, 220).
The most obvious theory is that member banks, on the whole, borrow at the reserve banks when it is profitable to do so and repay their indebtedness as soon as the operation proves costly. The cost of borrowing at the reserve banks, accordingly, is held to be the determining factor in the relation between the reserve bank operations to money rates, and the discount policy adopted by the reserve banks to be the most important factor in making reserve bank policy effective in the money markets. At the other extreme, there is the theory that member banks borrow at reserve banks only in case of necessity and endeavor to repay their borrowing as soon as possible. According to this theory the fact of borrowing in and of itself—the necessity imposed by circumstances on member banks for resorting to the resources of the reserve banks—is a more important factor in the money market than the discount rate . . . open market operations . . . contribute more directly to the effectiveness of the reserve bank credit policy than changes in discount rate. (Riefler 1930, 19–20)37
Chart 4.1 shows that the relation between discounts and government securities is negative in the 1920s. The bivariate relation is much less than one-to-one, however. On average, open market purchases reduce discounts by less than the amount of the purchase. A more complete analysis in appendix A allows for other relevant factors and casts doubt on the posited relationship.38
The discount rate has an ambiguous role in Riefler-Burgess. At times its role is modest; open market operations drive banks to borrow and repay at the prevailing rate. More often, open market operations prepare the way for discount rate changes. Strong testified in 1926 that the Federal Reserve continued to study and learn but had reached some preliminary conclusions:
If speculation arises, prices are rising, and possibly other considerations move the Reserve banks to tighten up a bit on the use of their credit, and we own a large amount of Government securities, it is a more effective program, we find by actual experience, to begin to sell our Government securities. It lays a foundation for an advance in our discount rate.
If the reverse condition appears, . . . then the purchase of securities eases the money market and permits the reduction of our discount rate. (House Committee on Banking and Currency 1926, 332–33)
37. The acceptance market differed from the market for borrowed reserves. The Federal Reserve announced a price, the discount charged on acceptances. Banks sold to the Federal Reserve, at their initiative, only if it offered a price above the going market rate.
38. The points at the upper right of chart 4.1 are for second quarter 1928 to third quarter 1929.
The Riefler-Burgess doctrine was compatible with the real bills doctrine and the Federal Reserve Act, but it permitted activist policymaking. Open market operations could be conducted so as to accommodate agriculture and commerce, as the act prescribed, but they could also be used for other purposes. However, nothing in either the Riefler-Burgess or real bills doctrine distinguished between real and nominal interest rates, a major reason for later misinterpretation of policy.
The “reluctance” view of borrowing is the weak link in the Riefler-Burgess doctrine. Banks borrowed heavily in 1920–21, when it was profitable. The Board’s annual reports and statements of members during the next few years seem intended to inform banks of the “tradition” against borrowing or to impose it on them through the administration of the discount window.39
39. In 1922 Strong discussed borrowing in a talk at the Harvard Economics Club. He placed considerable emphasis on profitability. Banks repaid borrowing when other opportunities were less attractive: “Now, in the long run, it is my belief that the greatest influence upon the member bank in adjusting its daily position is the influence of profit or loss” (Strong 1930, 181). Possibly Strong and others revised their earlier opinion. An alternative explanation is that policy changed, and the reluctance theory of borrowing reflected constraints that reserve banks imposed on borrowers. The “reluctance theory” failed later in the decade when the Board tried to reduce borrowing by exhortation.
Being able to control borrowing without large changes in discount rates had strong appeal. If interest rates could be held in a narrow range without jeopardizing control of inflation, System policy would be effective and criticism would be muted. Until the end of the decade, discount rates stayed within a narrow range, 3.5 percent to 4.5 percent at New York.
Neither Riefler nor Burgess completed the framework to link money, credit, interest rates, and borrowing to income and the price level. Instead, they relied on the real bills notion that if productive lending expanded at about the same rate as production, prices would be stable. It was an easy, but invalid, inference to rely on member banks’ borrowing or a market interest rate as the proper measure of the thrust of monetary policy. If borrowing and interest rates were low, policy was easy; if the two were high, policy was tight. By the mid-1920s, high and low borrowing were defined as member banks’ borrowing above and below $500 million.
The System had adopted measures of tightness and ease that misled them at critical times. A principal problem was the failure to distinguish between an individual bank and the banking system. An open market sale removed reserves, but if banks were induced to borrow, reserve bank credit and the monetary base remained unchanged. The increase in borrowing may have induced some banks to repay, as Riefler-Burgess claimed. But unless all banks behaved that way, others borrowed at the unchanged discount rate.
During many of the cycles in Federal Reserve history, both member bank borrowing and the monetary base moved procyclically, rising relative to trend in expansions and falling relative to trend in contractions. The Federal Reserve interpreted increased (reduced) aggregate borrowing as evidence of restrictive (expansive) policy even if the monetary base and the money stock accelerated (decelerated).
Differences in regional discount rates and in the reserve position of member banks produced a market innovation. In 1921 banks with surplus reserves—reserves above current and near-term requirements—began to sell reserves to banks with deficient reserves. These sales (or loans) and purchases made better use of existing reserve balances and supplemented the correspondent banking system as a means of putting idle balances to work. The market also supplemented the discount facilities.
The new market was known as the federal funds market (Board of Governors of the Federal Reserve System 1959). Banks with surplus reserves exchanged checks drawn on their accounts at a reserve bank for checks drawn on the purchasing bank payable through the clearinghouse the following day (or later). The difference between the two checks included interest for the term of the sale. Most transactions were made in Ne
w York, and transfers occurred on the books of the New York Federal Reserve bank.
Once the banks established the market, its convenience attracted other users. Acceptance dealers, commercial paper dealers, and others settled transactions in federal funds—reserve balances at Federal Reserve banks. Brokers began to canvass regularly.
The market languished in the 1930s. Early in the decade, risk increased as bank failures rose, so far fewer banks were willing to accept the default risk. Later, gold flowed in and excess reserves accumulated. The market disappeared until after World War II (ibid., 29–30).
GOLD POLICY
Although the United States remained on the gold standard, Riefler and Burgess did not dwell on the role of gold and did not state a policy with respect to gold. The explanation may be that both authors sought to develop policy guidelines in place of the gold reserve ratio. Nevertheless, gold policy played a secondary, but important, role in the 1920s.
A contemporary reader has difficulty comprehending the strength of commitment to the gold standard by bankers, officials, and many economists. Federal Reserve officials were unanimous in their commitment to restore some form of gold standard. Strong and others took many trips abroad, motivated in part by efforts to restore fixed parities tied to gold.
Montagu Norman, governor of the Bank of England, expressed an opinion representative of the ideas of informed central bankers. Failure to restore the gold standard would mean “violent fluctuations in the exchanges, with probably progressive deterioration of the values of foreign currencies vis-a-vis the dollar; it would prove an incentive to all of those who were advancing novel ideas for nostrums and expedients other than the gold standard to sell their wares; and incentives to governments at times to undertake various types of paper money expedients and inflation” (Chandler 1958, 311).
The ruling orthodoxy of the period sharply separated governments and central banks. The decision to fix the exchange rate was typically taken by the government. Central bankers negotiated support operations among themselves, usually keeping their governments informed about their progress. Continuing prewar practice, Strong was the principal negotiator of these agreements for the United States.
It is convenient to treat gold policy in the 1920 as three separate topics: the monetary response to changes in gold; circulation of gold and gold certificates; and actions to foster or sustain the gold standard. The last of these raises the issue of international cooperation, about which much has been written (Nurkse 1944; Clarke 1967; Eichengreen 1992).
Gold and Money
The Federal Reserve has been both criticized and praised for not following gold standard rules during the 1920s (Brown 1940; Keynes 1930). To contemporary observers at the Federal Reserve, the rules did not apply in the circumstances of the period. These officials believed that the gold reserve ratio was not an adequate policy indicator as long as no international gold standard existed. New procedures had to be found while they waited for, and worked toward, convertibility of the principal European currencies into gold and elimination of embargoes and other impediments to gold flows.
The problem, as seen in the early 1920s, was that the United States gold stock had increased much more than expected. By the end of 1921, the System’s gold reserve ratio reached 72 percent; it continued to rise in 1922, and by midyear it had nearly doubled from its low point in 1920. The Federal Reserve did not want to monetize the entire increase, as required under gold standard rules, both from fear of a new inflation and from concern about subsequent deflation if gold should leave when foreign governments restored an international gold standard. The Federal Reserve had used the fall in the gold reserve ratio as a main reason for raising interest rates in 1920. Many in Congress and the public interpreted the rising gold ratio as a signal that the Federal Reserve banks should lower interest rates in 1922.
At the beginning of 1923, discount rates at New York, Boston, and San Francisco were 4 percent, 0.5 percent below the rates at other banks. Concerns about inflation prompted these banks to consider raising the discount rate, as they subsequently did, despite their gold reserves. Concerns about public interpretations of the gold reserve ratio prompted the Governors Conference to approve a resolution urging the Board to issue a statement about the diminished importance of the gold reserve ratio (Governors Conference, March 28, 1923, 379). Case (New York) expressed the dominant view: “The average person cannot understand why we should be thinking of high rates with that reserve ratio” (ibid., 768).40
Contemporary observers report a difference in the Federal Reserve’s response to gold movements before and after 1925 (Hardy 1932, 148). Table 4.1 divides the period 1923–29 at the second quarter of 1925, the date at which Britain returned to the gold standard. These data support Hardy; the monetary base more fully reflected the gold flow in the earlier period, before the Europeans returned to the standard. Chart 4.2 gives more detail, using quarterly data for the period.41
40. Once again, Adolph Miller had a different view. He asked: “Suppose it [gold] doesn’t presently flow back?” And he warned that it was risky “to predicate a policy upon a possibility that may or may not materialize” (Governors Conference, March 28, 1923, 769). The Board did not follow the governors’ recommendation. It did little to change opinion about the gold reserve ratio. Chandler (1958, 191) reports that this irritated Strong and other governors, who faced this issue when talking to bankers, businessmen, and farmers. It seems likely, however, that the Board produced the policy statement in the tenth annual report partly in response to these demands. Friedman and Schwartz (1963, 283) point out the Federal Reserve continued to cite the possible withdrawal of gold as a reason for sterilization after the gold standard was restored, when the argument was no longer valid.
If the Federal Reserve had followed strict gold standard rules, gold movements would be fully reflected as changes in the base, and changes in the base would reflect only changes in gold. All points in chart 4.2 would lie on a straight line through the origin with a unit slope. We know that the Federal Reserve allowed discounts and open market operations to change the base. The points in the upper left quadrant suggest that large gold inflows were more than offset at times; the lower right quadrant shows that the base could rise while gold flowed out, contrary to gold standard rules. Nevertheless, there is a weak but clear positive relation between current quarterly gold movements and current quarterly changes in the base for the period as a whole. The Federal Reserve did not follow gold standard rules, but it did not ignore them entirely in the short run.
Together the data in chart 4.2 and table 4.1 suggest that gold flows often affected the base on arrival. In this sense the Federal Reserve “followed the rules” to a degree, most likely as a result of fixing the interest rate on discounts and acceptances and allowing reserves to respond to unanticipated gold flows. After Britain returned to the gold standard, the long-term change in the base was independent of gold flows.42
41. Appendix 4B describes the monetary base and its relation to Federal Reserve policy operations. Appendix 4A analyzes the statistical relation between gold and the base.
42. Short-term and lagged responses are shown in appendix 4B. The long-term relation is positive but small. Criticism of gold sterilization was common abroad. Criticism of the United States and France is a main point of the League of Nations (1932) retrospective study of the interwar gold exchange standard. On the other hand, Keynes (1930, 2:258) praised the Federal Reserve for showing that “currency management is feasible in conditions which are virtually independent of the movements of gold.”
Gold as Currency
One aim of the Federal Reserve Act was to pool reserves by centralizing gold holdings in the reserve banks. In its first decade, the System worked to achieve this objective by replacing gold certificates with Federal Reserve notes.
Policy changed in the 1920s. Unwilling to allow prices to rise and concerned about the political pressures to expand as a consequence of a high reserve ratio, the governors
looked for ways to reduce the reserve ratio without inflating. Early in 1922 Secretary Mellon proposed to substitute gold certificates for Federal Reserve notes. Gold certificates had 100 percent gold backing instead of the 40 percent behind Federal Reserve notes, so they reduced the gold reserve but had no effect on money or inflation.
The Federal Reserve was at first reluctant to change its policy of centralizing the gold reserve. Miller proposed instead raising the 40 percent gold reserve behind currency issues. This proposal, like Mellon’s, seemed too transparent to defuse political pressure. A second alternative kept the gold in Europe on “earmark” and, by ruling of the Board, excluded from reported gold reserves. At first the amount earmarked was relatively small, $20 million or less. Earmarked gold rose to $50 million in 1924–25 and again in the first half of 1926. The maximum during the decade was $200 million, about 5 percent of the monetary gold stock.
Despite Mellon’s request, in May 1922 the Governors Conference approved a proposal that made issuing gold certificates to the public a last resort. Gold inflows continued. In August, New York began issuing gold certificates and sent a letter to the Board asking all reserve banks to do the same. The other large bank, Chicago, did not accept the policy until February 1924, after repeated requests from Undersecretary Gilbert at the 1923 Governors Conference and by letters. The two banks had issued over $700 million in certificates by the end of 1924. The gold flow then reversed, so after discussion with the Treasury, Strong changed policy. The new policy kept total gold certificates equal to $1 billion, the amount outstanding in 1925. This policy remained in effect until mid-1928 (Governors Conference, March 1926, 126–29).43
A History of the Federal Reserve, Volume 1 Page 25