A History of the Federal Reserve, Volume 1

by Allan H. Meltzer

  RESERVE REQUIREMENTS AND OPEN MARKET POLICY, 1935–37

  By the time Eccles joined the Federal Reserve, Roosevelt’s economic program was about to change. The Supreme Court soon declared the NRA and the AAA unconstitutional. Gold and silver purchases continued routinely, but hopes for reflation to the 1926 or 1929 price level were no longer widely held.156 Agricultural prices (measured at the time) had increased absolutely and relative to other prices, as Warren had predicted, but they remained 25 percent below the 1926 average. The consumer price index was about 30 to 35 percent above its low but 25 percent below the 1929 level.

  Roosevelt had not yet abandoned his hopes for a balanced budget, but the low level of activity and relief expenditures kept the hope unrealized. To finance the deficit while keeping interest rates from rising, Morgenthau bought bonds for the new Exchange Stabilization Fund and the Treasury trust funds—Postal Savings, Railroad Retirement, and others. The Federal Reserve, as fiscal agent, made purchases for the Treasury, limiting its own operation for most of 1934 to exchanges of long-term for short-term debt. The open market portfolio remained below $2.5 billion, about 25 percent larger than in March 1933.

  Morgenthau recognized that using the stabilization and trust funds not only freed him from dependence on the Federal Reserve but gave him an opportunity to influence its decisions. Despite the legislative changes that had increased de jure Federal Reserve independence, the Federal Reserve was less independent of the administration from 1934 to 1941 than in any other peacetime period.

  155. The Board also used Eccles’s 1938 letter to Senator Vandenberg to respond to proposals for nationalizing Federal Reserve banks or repaying the government debt by issuing currency. On the latter issue, Eccles makes the extraordinary claim that issuing currency to buy back the federal debt would not raise prices or increase prosperity (Eccles to Senator Arthur Capper, Board of Governors File, box 141, June 5, 1939, 3). The claim is that the currency would return as excess reserves and remain idle. In the 1938 letter, this is followed by a contradictory claim that inflation would result (Eccles to Vandenberg, Board of Governors File, box 141, June 14, 1938, 5, 7).

  156. Fisher continued to argue for a higher gold price, an increase to $41.34, the maximum permitted under the Thomas amendment. Roosevelt listened but did not act (Barber 1996, 81). The experiment had not worked in 1933 as Warren and Fisher promised, so Roosevelt had moved on.

  Policy Issues, 1935–36

  In January 1935 the Board approved reductions in discount rates to 2 percent at Philadelphia, Atlanta, and St. Louis and to 2.5 percent at Richmond, Minneapolis, and Dallas. These were the lowest rates at these banks up to that time, but further reductions in discount rates, open market rates, and deposit rates soon followed. The proximate cause of lower rates was the gold inflow in response to the $35 price. The Federal Reserve had not made any net open market purchases for more than a year.

  the gold clause On January 25, Eccles told the FOMC’s executive committee about the Treasury’s concern that prices of bonds carrying the gold clause had increased absolutely and relative to the prices of other bonds. On January 11 the price of Treasury bonds with the gold clause was 0.75 percent above bonds without the clause. The difference remained throughout the month. The price difference reflected the impending Supreme Court decision in Perry v. United States, known as the gold clause case.157

  The New York directors responded to the spread in rates by authorizing sales of gold clause bonds in exchange for other bonds. The FOMC followed. On February 5, it approved purchases or sales of up to $250 million (FOMC Minutes, Board of Governors File, box 1451, January 25 and February 5, 1935).

  The administration also prepared for the Court’s decision. Harrison was told to stabilize the foreign exchange and gold markets by keeping the French franc within the gold points. The plan was to use the Exchange Stabilization Fund to buy francs by selling sterling “violently” if necessary (Harrison Papers, file 2012.5, February 18, 1935). The Court’s decision, favorable to the government, required no action.158

  157. There were several cases. The Court issued separate opinions for private bonds and pubic debt. The plaintiffs in the private bond cases asked to receive compensation for the 59 percent devaluation of the dollar against gold by payment in dollars at the old exchange rate of dollars for gold. They did not question the right to devalue or withdraw gold. The government claimed the right under its explicit power to coin money and regulate its value. The decision, expected in early February, was delayed until February 18. The Court found for the government by five to four, with Justices Hughes, Stone, Cardozo, Brandeis, and Roberts in the majority and Butler, Sutherland, Van Deventer, and McReynolds in the minority. Citing earlier decisions in the Legal Tender Cases (1871) and the Court’s opinion following the 1834 6 percent reduction of the gold content of the dollar, Hughes’s opinion found that the plaintiff had not been damaged and placed the constitutional power to regulate the value of money above the obligations of private contracts. Stone’s decision, in the case involving government bonds, Perry v. U.S., concluded that the plaintiffs had not suffered a loss. McReynolds’s dissent denied that the Constitution gave Congress power to repudiate contracts. He found that Congress had acted to “destroy private obligations, repudiate national debts and drive into the Treasury all gold within the country in exchange for inconvertible promises to pay, of much less value” (Krooss 1969, 4:2865). The gold clauses in contracts did not prevent Congress from regulating the value of money. Justice McReynolds is reported to have said, “The Constitution is gone” (Pearson, Myers, and Gans 1957, 5618). The court ignored the higher market price of bonds with the gold clause, clear evidence that the option was valuable as protection against future inflation.

  delay and inaction The System’s inaction in 1935 was not accidental.159 As excess reserves rose, members of the FOMC became concerned about potential credit expansion and uncertain what to do. A background memo prepared for the March 21 meeting addressed the issue by asking, What is the duty of a central bank in the present situation? (Excess Reserves and Federal Reserve Policy, Board of Governors File, box 1449, March 21, 1935).

  The memo had two parts. The first traced the increase of excess reserves and discussed the reasons for their continued growth. It found that, initially, excess reserves were expected to pressure banks to expand private loans by pushing down the yield on government securities. This could happen, but there was little evidence so far. One reason given was that government deficits supply bonds that the banks bought (ibid., 2–3): “If this process should continue, should we not expect on the basis of the experience of other nations that eventually a point will be reached where the banks will be unable or unwilling to absorb the government debt, so that the government will be forced to expend its stabilization fund . . . or request the reserve banks to purchase more government securities. . . , or to borrow directly from the Reserve banks” (4).160 The memo concluded that neither past experience nor central bank theory gave any guidance in present circumstances.161 Previous inflations abroad had occurred with rising activity and government borrowing directly from the central bank.

  158. Issues about the gold clause did not end with the cases. The decision for the government was based in part on the finding that, since prices had fallen, bondholders had not been harmed. This suggested that the decision might be reversed at a later date. In March 1935 Robert A. Taft, acting for the Dixie Terminal Company, demanded payment on a $50 bond with the gold clause at the value of gold in 1918, when the bond was issued. The Treasury refused, so Taft sued in the Court of Claims on behalf of Dixie Terminal and other clients. In November 1936 the Court of Claims rejected these suits. A year later, the Supreme Court agreed with the government (Patterson 1972, 152–54). I am indebted to Leonard Liggio for this reference.

  159. The Federal Reserve was not alone in its inactivity and hesitancy. Harrison reports on a meeting with Roosevelt in late May. The NRA had been declared unconstitutional on May 27. Har
rison describes Roosevelt as “harassed and stumped and for once I thought he had no definite plan and seemed quite hopeless and helpless” (Harrison Papers, memo to personal files, June 3, 1935, 3). The meeting came about after Morgenthau told Harrison that the president was concerned that Harrison might be angry about the banking bill and, for that reason, no longer called on him. Harrison made an appointment. The president “chided me for not having called on him and rather expected me to explain why I had not called” (3). They agreed that Harrison would call and visit when he was in Washington.

  160. This presumes without explicit recognition that the gold inflow is less than the deficit. Otherwise banks would continue to gain reserves and purchase Treasury bonds. Eventually prices would rise, reversing the gold flow.

  The second issue was the course to follow. The memo considered open market sales to absorb excess reserves but rejected this course on economic and political grounds. The economic argument was that sales might cut off an incipient expansion by overweighting future dangers of inflation and not encouraging expansion enough. The political argument was that the government could offset Federal Reserve actions by using the stabilization fund or resort to issuing paper money under the Thomas amendment. Further, with the banking act in Congress, the government could change the entire financial system, including the central bank: “It seems clear that we could act effectively only with the consent and cooperation of the administration” (ibid., 8).162

  The memo recommended no action for the present. The only policy change in the next two months followed a May 1 letter from the Board to the reserve banks, calling attention to discount rates and suggesting that the directors consider reductions. A week later the Board approved reductions at Dallas, Richmond, Cleveland, Minneapolis, and Kansas City. Discount rates were now 1.5 percent in New York and Cleveland, 2 percent at all other banks. Discount rates remained at these levels for the next two years.

  The volume of discounts fell below $10 million in January 1934 and, except for a small increase in the 1937–38 recession, remained there until the war. The acceptance portfolio reached $10 million in spring 1934, then gradually faded away. Open market rates remained below the discount rate and the buying rate on acceptances. Prime commercial paper was at 0.75 percent, banker’s acceptances were at 0.125 percent, and long-term Treasury bonds fluctuated around 2.75 percent.

  concerns about future inflation Propelled by gold inflows, the monetary base rose at an 18 percent annual rate for the first three quarters of 1935 and at a 25 percent annual rate in the fourth quarter. Chart 6.2 shows the very close relation between gold and the monetary base during the recovery. With the exception of a few periods, mainly in 1937–38, gold flows dominated changes in the base. The money stock rose and fell with the base.

  161. The memo has a rare acknowledgment of policy error. Looking back at September 1931, the memo commented that standard theory misled them following England’s departure from the gold standard. The Federal Reserve raised the discount rate, a classic response. “The rate increase probably served more to add to the deflationary movement of succeeding months than to check the gold outflow” (Excess Reserves and Federal Reserve Policy, Board of Governors File, box 1449, March 21, 1935).

  162. Governor Schaller of Chicago wrote on May 4 urging reduction of Treasury bill holdings, by allowing them to run off weekly, until the bill rate reached 0.5 percent. The Board responded: “Excess reserves should not be reduced until there is evidence of excessive borrowing or speculative expansion” (Board of Governors File, box 1451, May 4 and 17, 1935).

  The background memo for the October 1935 FOMC meeting noted the improvement in business conditions. But it noted also that member banks held almost $3 billion in excess reserves, an increase of $1 billion since March.163 The volume of excess reserves now exceeded the size of the open market portfolio.

  The memo was more anxious than the March memo. It asked whether the Federal Reserve should intervene to prevent further accommodation and the risk of inflation. Its conclusion: the System must coordinate with the Treasury. Monetary restraint without a reduction in the budget deficit would risk higher interest rates on Treasury financing (Memo for FOMC, Board of Governors File, box 1452, October 22, 1935, 11–12).

  The memo raised two questions that the governors discussed at length: What was the appropriate time to reduce excess reserves, and should it be done by open market sales or by an increase in reserve requirement ratios? The governors were divided. Some saw no reason to act; some favored an increase in required reserves; some wanted open market sales. Eccles favored an increase in reserve requirements, but he was concerned about how it could be presented to the public.

  The resolution adopted at the meeting recognized the risks of action and rejected taking any immediate steps. It called for action “as promptly as possible” to reduce excess reserves, and it provided for purchases of up to $250 million in the event of a disturbance following an increase in required reserve ratios, the decision to be taken by telegraphic vote. The members generally preferred a change in the requirement ratios to open market sales, because open market sales had previously been used only to tighten credit. This was not the intention. The effect of increased reserve requirement ratios, they said, would depend on the distribution of excess reserves by class of banks and by geographical location. The governors recommended that the Board learn about these distributions (FOMC Resolutions, Board of Governors File, box 1450, October 23, 1935).164

  163. This figure is preliminary. The final figure was $3.6 billion. The difference suggests the large changes occurring at the time.

  the role of excess reserves With borrowing reduced almost to zero, the key relation of the Riefler-Burgess framework was inoperative. Member bank borrowing could not be an indicator of policy action. Instead, the System focused on the level of excess reserves. A high level indicated potential credit expansion; an increase was a sign of increased potential expansion.

  This interpretation of excess reserves follows directly from the RieflerBurgess theory, if excess reserves are treated as negative borrowing.Instead of reducing borrowing when the credit market eased, banks added excess reserves. In the System’s view, beyond some point additional excess reserves were excess in the economic as well as in the accounting sense.

  There is no evidence of a study by the Board or the reserve banks to understand why banks held large excess reserves. With short-term interest rates below 0.5 percent, the opportunity cost of holding excess reserves remained low, but banks had other options. Interest rates on long-term Treasury bonds fluctuated around 2.75 percent. The Board appears to have made no effort to understand or explain this puzzle. The common presumption was that unless excess reserves remained concentrated in one part of the banking system, they could be absorbed without consequence.165 All Federal Reserve discussion at the time treated excess reserves as a redundant surplus.166

  164. After the meeting, Harrison prepared the first of many draft statements explaining that the increase in reserve requirements was a precautionary move, not a change in policy. (Harrison to Eccles, Board of Governors File, box 1450, November 4, 1935).

  165. There is remarkably little academic study of excess reserves. The best work (Frost 1966) attributes the increase in excess reserves to risk and the prevailing low level of opportunity cost. See also Brunner and Meltzer 1968a for conditions required for a liquidity trap in the banking system.

  166. A staff study showed that at the last call report on June 29 only 897 banks out of 6,410 would have to increase their deposit balances at reserve banks if reserve requirements increased by 25 percent. The Federal Reserve would have to provide only $99 million of additional reserves to offset the shortfall at banks with reserve deficiencies. All the banks had correspondent bank balances sufficient to cover the reserve deficiency. A 50 percent increase would require 2,041 banks to increase reserves by $528 million. All but 125 could meet the increase from correspondent balances. The memo makes clear that no further adj
ustment was expected following the increased requirements or the reduction in correspondent balances (Board Minutes, November 6, 1935, 5–6). Only James questioned whether some banks would adopt “less liberal lending policy” to restore excess reserves (6). The rest of the Board accepted the memo’s conclusion. Eccles used the memo in his discussions with Morgenthau and left a copy.

  The poor quality of the Board’s analysis shows also in the estimates of potential credit or monetary expansion. Their usual estimate is ten to twelve times the volume of additional reserves, but some estimates put additional lending potential at twenty times excess reserves. Emanuel Gold-enweiser’s book, written many years later, repeats these estimates (Gold-enweiser 1951, 175). To get these numbers, the staff used only the required reserve ratio, ignoring drains into currency holding and time deposits. A more accurate calculation, one that allowed for these concomitants of monetary and credit changes, would have estimated maximum credit expansion as seven or eight times the addition to reserves. And this calculation is almost certainly too high because, like the Board’s staff, it assumes that none of the excess reserves were held for reasons of safety based on experience. The result was a large overestimate of potential monetary and credit expansion and prospective inflation and an underestimate of the effect of higher reserve requirement ratios.

  A subsequent memo by the Board’s staff considered the pros and cons of a reduction in excess reserve achieved by raising reserve requirement ratios. The pro case claimed there was no question that the Board would have to act; it was “merely a question of timing” (Memo, Board of Governors File, box 1450, November 5, 1935). Prompt action, before banks expand, based on outstanding excess reserves, was best because delay might force loan liquidation. Also, reserves were “ample,” so increased reserve requirements would be less likely to lose members.