A History of the Federal Reserve, Volume 1
Page 76
Harrison fought a rear-guard defense, urging that action not be taken solely to prevent Treasury action. He preferred to continue the policy of shifting maturities without changing the total portfolio, but once he recognized that he had little support, he favored giving the executive committee authority to prevent disorderly markets (Minutes, FOMC, April 4, 1937, 5).204
Only Governors Davis and McKee favored canceling the May 1 increase in reserve requirements. The dominant view was that inflation remained a threat. A majority supported open market purchases, some to prevent Treasury action, some to correct so-called disorderly market conditions.205
203. Eccles supported Morgenthau, agreed on the need for purchases, and at one point threatened to resign if the FOMC did not support him (Harrison Papers, file 2140.2, April 14, 1937).
204. Williams’s memo conveys the intense feeling, even animosity, between Eccles and Harrison (Williams to Harrison, Harrison Papers, FOMC, April 14, 1937).
205. By a vote of nine to two, the FOMC agreed that the disorder in financial markets was not caused by the Board’s policy action. Only McKee and Davis, both Board members, blamed the Board. Neither had voted for the increases. Despite Eccles’s April 3 statement, quoted above, he continued to absolve himself and the Board of responsibility.
With Harrison abstaining, the rest of the committee voted to adopt Eccles’s motion. Purchases began the next day. This was the first increase in the open market account since November 1933.
The committee’s action was mainly political.206 Only Eccles expressed strong support for purchases. He summarized the views of the other members as acting “on grounds of expediency, to avoid a break with the Treasury” (Harrison Papers, Supplementary Memo, Williams to Harrison, file 2140.2, April 14, 1937). Goldenweiser agreed with Eccles but regarded the decision as “not mainly an economic but a political question.”207
The executive committee met again the following day, April 5, and voted to purchase up to $5 million of Treasury bills. Harrison voted no. Bond rates fell, so no purchases were made until the next day, when rates again rose. Burgess purchased $29 million for the week. No one objected that the account manager (Burgess) exceeded the authorization for the week’s purchases.208
Bond yields reached a local peak of 2.80 percent at the end of the week. In all, rates increased 0.34 (14 percent) from the January low. The executive committee had set a limit to the open market account of $2.53 billion. The System continued to purchase until the portfolio reached $2.525 billion at the end of the month, an increase of $95 million for the month.209
The third increase in reserve requirements took effect on May 1. Banks had prepared adequately, so there were no additional repercussions. Bond yields reached 2.80 percent again, then declined. The Federal Reserve did not undertake additional purchases.
206. Harrison opposed the commitment to purchase a fixed amount in the next week, but he lost on a vote of eight to three. Only Szymczak and Sinclair (Philadelphia) supported him. The choice of a week reflected Morgenthau’s warning that he would judge their actions after a week. Harrison subsequently changed his vote to support the motion.
207. Williams summarizes the difference in economic outlook between Eccles and Harrison. Eccles believed the economy had been hurt by the rise in interest rates, citing the virtual standstill in new issues on the capital market. Harrison (and Williams) saw the economy acquiring “increased momentum.” They were more concerned about inflation, the budget deficit, wage settlements, and the beginning of armament demand (supplementary memo, Harrison Papers, file 2140.2, April 14, 1937, 2).
208. Eccles twice asked Harrison to reduce the acceptance-buying rate. Harrison took the issue to his directors but expressed his view that the reduction was not justified. The directors agreed (Sproul Papers, Open Market Policy, April 8, 1937).
209. Most of the purchases were made in periods of market breaks on April 6–8 and April 22–24. On the latter dates, the Federal Reserve was the principal buyer. The account also sold bonds when the bond market rose, for example, on April 10 (memo, Harrison to files, file 2012.7, April 10, 1937). Morgenthau was annoyed by the purchases on April 14 because he had to sell bills to continue gold sterilization and bill rates had increased a bit (Harrison Papers, file 2012.7, April 14, 1937). The sales were offset within the week by bill purchases so that the account would not decline.
Summary: Reserve Requirements and Monetary Policy
The response to doubling reserve requirement ratios in 1936–37 remains controversial. The controversy began when Morgenthau blamed the Federal Reserve for the rise in interest rates and for the recession that followed. He did not mention the Treasury’s decision to sterilize gold inflows. Eccles and most Federal Reserve officials denied responsibility for both the increase in interest rates and the recession.210 Roose’s comprehensive study of the 1937–38 cycle includes monetary action as one factor affecting the decline (1954, 239). Friedman and Schwartz (1963, 526–31) argue for the importance of monetary policy acting on output and income by reducing the money stock. Calomiris and Wheelock (1996, 510) reject this explanation at least for the changes in reserve requirements. They give more attention to changes in reserve requirements than to gold sterilization, but they recognize both as factors affecting money growth.
Chart 6.3 compares the increase in the weighted average reserve requirement ratio in 1936–37 with subsequent changes in the years to 1953.211 The 1936–37 changes removed $3.1 billion of reserves as a base for monetary expansion in a period of nine months. The reduction is approximately 28 percent of the level of reserves on June 30, 1936. After subsequent changes in reserve requirement ratios, the Federal Reserve held interest rates constant, so banks could sell securities and restore desired reserve positions at unchanged interest rates. The principal effect of later changes in reserve requirement ratios was to raise (or lower) the tax on bank profits without any significant effect on the money stock.
In 1937 Morgenthau and the Federal Reserve agreed to prevent disorderly market conditions without pegging interest rates. Interest rates rose, and the effective monetary base declined. Banks did not restore the reserves absorbed by the changes in reserve requirements; total reserves, the monetary base and, beginning in second quarter 1937, the M1 money stock fell. In the four quarters of 1936, average M1 growth was 12.8 percent, propelled by the increase in gold. Growth fell to 5 percent (annual rate) in first quarter 1937. For the remaining three quarters of 1937, the average annual growth rate of money was –6.5 percent.212
210. McKee, who did not vote for the increases, is one exception, as noted earlier. The Board’s staff undertook a study of reserve requirements but did not study the effect of the 1936–37 changes. Their report reconsiders proposals made in 1931 to count vault cash as part of reserves, to make reserve requirements uniform for all classes of banks and types of deposit, and to put reserve requirements on deposit turnover (debits). The report gave a mixed review to these proposals, and none was adopted at the time (Board of Governors File, box 107, February 5, 1938). In March the Conference of Reserve Bank Presidents endorsed the proposal to count vault cash as part of required reserves up to 50 percent of required reserves (Board of Governors File, box 136, March 19, 1938). The Board made the change in vault cash beginning in 1959.
211. The chart is computed using deposits subject to reserve requirements (net demand deposits and time deposits) from the call reports published in Board of Governors of the Federal Reserve System 1943. See Cagan 1965, 198–99.
Interest rates on risky assets show relatively large increases. Table 6.5 shows the rates on Baa bonds and the spread between Baa and Aaa rates, a measure of the risk premium. The risk premium rose in 1937 and the first half of 1938. At its peak, the risk spread had returned to the level reached in third quarter 1931, when Britain left the gold standard.
The Federal Reserve’s error was the belief that excess reserves could be reduced without consequence. Its denial of the effect of doing so is puzzling in l
ight of the efforts that banks made to restore excess reserves, an effort Eccles and others commented on at the time. Since most short-term interest rates did not change, Harrison and others refused to believe that policy had tightened.
Table 6.6 shows the estimates of excess reserves at New York and other banks based on data available at the time. These data suggest that banks in New York and outside first restored, then increased excess reserve holdings, so that banks held more excess reserves at the end of 1938 than they did when the System undertook to eliminate them in August 1936. New York banks added more to excess reserves during this period than banks outside New York.213 The policy therefore did not achieve what the Federal Reserve set out to accomplish. It not only contributed to the recession but also failed to reduce the System’s fear that it could not prevent future inflation.
212. Changes in short- and long-term interest rates on government securities show only modest effects of the policy action. Rates on four-to-six-month prime commercial paper increased from 0.75 percent to 1.0 percent in April 1937 and were otherwise unchanged. Monthly average rates on ninety-day banker’s acceptances moved steadily from 0.19 percent in December 1936 to a peak of 0.56 percent in April 1937 before declining again. As noted earlier, rates on long-term governments peaked at 2.80 percent in April 1937, then declined slowly. Although these relatively modest changes disturbed Morgenthau, they were much less than the annualized rate of inflation, 8.3 percent for the GNP deflator, in the first half of 1937.
The source of this concern is a slightly modified version of the Riefler-Burgess framework. The principle was unchanged. In the 1920s Riefler-Burgess suggested that once banks were out of debt, the Federal Reserve had little control. For it to exercise control, the banks had to be forced to borrow. Since borrowing had almost disappeared in the 1930s, the doctrine changed. Now excess reserves (negative borrowing) rendered the System incapable of preventing inflation. By reducing excess reserves below the size of the open market portfolio, the System believed it was in position to prevent runaway inflation; once excess reserves were smaller than the open market portfolio, open market sales could force banks to borrow. Under Riefler-Burgess, they would then want to pay off their indebtedness by contracting.214
213. The risk premiums in table 6.5 suggest the increase in uncertainty, as in Frost 1966. Despite the low interest rates on government bonds, Baa bonds were at about the same rates as in the 1920s, so the risk spreads were higher.
The System ran the experiment of reducing excess reserves three times. Each time banks responded by restoring excess reserves. Partly out of unwillingness to admit policy error and partly under pressure from the Treasury, the Federal Reserve ignored this contradiction of Riefler-Burgess, much as it had ignored contrary evidence earlier. It continued to cling to its theory.
THE 1937–38 RECESSION
Did the increase in reserve requirement ratios cause the 1937–38 recession? Changes in reserve requirements were part of monetary policy, and monetary policy was part of government policy. The data on interest rates, risk premiums, and changes in the monetary base and money suggest that the Federal Reserve did not offset the effects of the change. Monetary policy became more restrictive. The proximate causes of the monetary policy change were the increase in reserve requirement ratios, not offset by open market purchases, and the shift in December 1936 to gold sterilization.
Monetary factors were not alone.215 There were two large contractive changes in fiscal policy in 1937. One was the reduction of soldiers’ bonus payments and passage of the undistributed profits tax; the other was the beginning of Social Security tax payments. Passed in 1935, Social Security taxes became effective in fiscal 1936 (calendar 1937).216
Congress had insisted, over the president’s veto, on accelerating the soldiers’ bonus, so that veterans would receive payment before the 1936 election.217 Beginning in June 1936, the government issued $1.7 billion of bonds. By December veterans had cashed $1.4 billion of the bonds and spent the money. Balke and Gordon’s (1986) quarterly data show an 18 percent average rate of increase in real GNP for the final three quarters of 1936. The deflator rose, and profits reached a peak for the recovery in fourth quarter 1936.218
214. Concerns about membership appear to have been misplaced. The proportion of member banks among commercial banks increased annually from 1935 to 1939 and more rapidly in 1940 and 1941.
215. Romer (1992) estimates the effects of fiscal and monetary shocks using data for 1920 to 1937. She found no effect of fiscal shocks and attributed the 1937 recession to monetary shocks. Romer assumed a one-year lag of policy variables to recognize that the fiscal changes were known in advance. As the text shows, the Board gave advance notice of changes in reserve requirement ratios.
216. Concerns about the effect of Social Security taxes on the 1937–38 recession led to repeal of actuarial provisions and substitution of “pay as you go” or intergenerational transfer in 1939.
217. The bonus had been approved in 1924 for payment in 1945. Congressman Wright Patman (Texas) led the fight to have the bonus paid in 1936 (without discount). He proposed to finance the payment by printing greenbacks, and the bill passed the House and Senate with that provision. Roosevelt vetoed the bill but did not work to prevent an override after Congress omitted greenback financing. Bonds were issued to the veterans but could be sold immediately for cash (Blum 1959, 249–58).
Responding to criticism about deficit spending, and hoping to stimulate private spending, in March 1936 the administration promised to tax undistributed corporate profits (Eccles 1951, 260). The tax was based on the peculiar belief that corporations held funds idle instead of investing them. If these funds, like the excess reserves of the banks, could be put to work, the economy would expand faster.219 The Treasury expected the tax to raise $620 million, about 5 percent of the prospective deficit (ibid.).
Roose (1954, 238–39) adds some additional factors influencing investment spending, of which the most important is the increase in labor costs following strikes to organize major industries. The combined effect of higher interest rates, fiscal contraction, rising costs, and the growing belief that the Roosevelt administration had become more hostile—as shown by the undistributed profits tax and Roosevelt’s second-term rhetoric about “economic royalists”—raised current and prospective tax rates and costs of capital.220
The National Bureau of Economic Research ranks the 1937–38 recession as the third most severe in the years after World War I. Real GNP fell 18 percent and industrial production 32 percent in the thirteen months beginning June 1937.221 At its peak, the unemployment rate reached 20 percent, not much below the 25 percent maximum in 1932 (Zarnowitz and Moore 1986). It is no wonder that many feared the 1929–33 disaster had returned.
The Federal Reserve made no purchases until fall. The principal reason, again, was beliefs, not lack of information. John H. Williams recognized the beginnings of hesitation in the economy at the May 4 meeting of the FOMC, before the peak recorded by the NBER.222 He saw no reason for action, however, and he favored continuing the policy of preventing disorderly markets, if they should occur. Goldenweiser agreed there was no need for action. The economy had slowed, but “he did not see any possibility at this time of a new period of depression setting in” (Minutes, FOMC, May 4, 1937, 6).223
218. The bonus payment declined to about $15 million in 1937.
219. Eccles (1951, 260–65) opposed the Treasury’s bill on grounds that it discriminated against small companies with low retained earnings. Like the Treasury, he failed to recognize that the tax increased the cost of capital to corporations financing investment from retained earnings. Eccles’s public criticism, and proposals for a less regressive undistributed profits tax, was another reason for resentment by Morgenthau and his staff. The tax worked perversely. Dividend payments increased in advance to avoid the tax, then declined (Roose 1954, 236). Businessmen saw the tax as another example of the administration’s hostility toward business (Stein 1990, 87). I
t was repealed in 1938, effective January 1940.
220. In April, Roosevelt criticized high prices in the durable goods industry as a source of “excessively high profits” and ordered a shift in public works spending to avoid these industries (Roose 1954, 236). The statement reflected widespread concern in view of the rapid price rise. Currie had urged Eccles to consider using antitrust action to deter price increases (memo, Currie to Eccles, Board of Governors File, box 1433, December 16, 1936). This policy was adopted in 1938. Adolph Berle had urged it from the beginning.
221. The two more severe recessions are 1929–33 and 1920–21.
Not much had changed when the FOMC met again on June 8 and 9. The committee discussed the business situation and the continued gold inflow. Williams regarded the slowdown of business as “salutary.” He agreed with Goldenweiser that the gold inflows were the most serious problem of the moment (Minutes, FOMC, June 9, 1937, 3–5). Goldenweiser remarked that the System had to be in a position to offset gold imports when the Treasury stopped sterilizing, probably a reference to Morgenthau’s reluctance to continue borrowing to sterilize gold inflows (3).
Before Eccles left for summer vacation, he called a meeting of the FOMC executive committee to propose purchases of $200 million to $300 million to offset the seasonal increase in demand for base money.224 Harrison opposed “increasing our portfolio merely for the purpose of taking care of a seasonal demand for loans and currency. . . . [He] preferred to . . . have the banks borrow and show bills payable” (Harrison Papers, file 2140.2, August 27, 1937, 2–3). In making this argument, he showed the continuing influence of Riefler-Burgess—the need to get the banks in debt to the reserve banks. He argued that pressure on bank reserves in New York reflected the lower rates charged by correspondent banks. He proposed “reduction in discount rates at reserve banks outside New York.”