195. The report showed the peak in the index of industrial production in June 1929, up at a 19 percent annual rate for the first six months. Automobile production peaked in April, 67 percent above its 1928 average. Agricultural prices continued to fall in 1929, at a 4.5 percent annual rate from December 1928 to September 1929 (Board of Governors File, box 2461, January 15, 1930, 10–11).
196. Allyn Young was a leading economist of his time. He finished college at seventeen. He was the first American to be president of the Royal Economic Society. He was also president of the American Economic Association and the American Statistical Association. He served as an adviser to President Wilson at the Versailles conference.
197. Warburg (1930, 1:506–7) blamed the Board for “the most anomalous rate structure ever devised by any powerful central bank.” This refers to the refusal to raise the discount rate in 1929. In the annual report of his bank, published in March 1929, Warburg accused the Federal Reserve of “tossing about today without its helm being under the control of its pilots” (ibid., 826). Kindleberger (1986, 96) recognizes that prices were not at extraordinary levels relative to profits. Like Warburg, he located the problem in the financing on the call money market: “The danger posed by the market was not inherent in the level of prices and turnover so much as in the precarious credit mechanism that supported it.” Warburg did not criticize Strong or the Federal Reserve for helping Britain remain on the gold standard. He criticized the failure to promptly reverse policy (raise interest rates). He did not speculate on whether a prompt reversal would have reversed the gold flow.
By the spring of 1929, recession had started abroad. It was probably too late to stop a worldwide recession, but there was ample time to stop the severe deflation that followed. The National Bureau of Economic Research marked a cycle peak in April for Germany and in July for Britain. March was the peak month for production in Belgium; Canada’s peak came in the spring. By fall, financial and business failures had increased in Britain, Germany, and elsewhere (Kindleberger 1986, 102–4). The Federal Reserve’s production index, available at the time, peaked in June. By October it was 8 percent below the peak. Monthly peaks in the stock markets in the United States, Canada, and France came in September 1929, but markets in Germany, Sweden, and Switzerland reached peaks in 1928, and in Britain the peak came in January 1929198 (Kindleberger 1986, 110–11, based on League of Nations data).199
Kindleberger (1986) and Galbraith (1955) propose that causality went from the stock market crash to the economy. Kindleberger wrote: “It is hard to avoid the conclusion that there is something to the conventional wisdom that characterized the crash as the start of a process. . . . The stock market crash is less interesting for the irony it permits the historian, bemused by the fables of greedy men, than the start of a process that took on a dynamic of its own” (116).
Charles O. Hardy, a contemporary observer, was skeptical about arguments of this kind.
In my judgment, the case for the campaign against speculation was weak. It is easy now to see the evidence of over-optimism in the judgment of those who made the stock prices of 1929—though today’s appraisals may look just as absurd three years hence. . . . There was no evidence in 1928 or 1929 that business and agriculture were suffering from the competition of the stock market—there was only apprehension that such suffering might ensue. . . .
198. Beckhart (1972, 227) listed as proximate causes of the October decline: reports of smaller corporate earnings, the flooding of the market with new security issues, the rise of London bank rate to 6.5 percent, the Hatry failure in London, and a decline in business activity clearly evident by October.
199. These data suggest that between January 1926 and September 1929, Canada and France experienced a stock market boom greater than in the United States. United States stock prices rose 112 percent, Canadian prices 243 percent, and French prices 156 percent (Kindleberger 1986, 110–11).
There was no evidence in 1928 or 1929 that brokers’ loans were too high for safety, except that they were higher than a few years before. (Hardy 1932, 177–78)
Friedman and Schwartz (1963, 306–7) described the stock market crash as partly “a symptom of the underlying forces making for a severe contraction in economic activity. But, partly also, its occurrence must have helped to deepen the contraction.” They suggested tentatively that the decline in velocity and interest rates in 1929–30 was consistent with a desired reduction in spending and the ownership of financial assets and a desired increase in the demand for money. This effect, they said, was dwarfed over the next two years by the decline in the money stock.200
A puzzling aspect of Kindleberger’s argument is that the association between falling stock prices and recession differed across countries and time. Kindleberger showed that the stock price index he used for the United States had an initial decline of 28 percent between September 1929 and February 1930. It then recovered 10 percent, with evidence of economic recovery in early spring 1930. The French index dropped by 25 percent between September 1929 and October 1930. France did not experience the severe recession for more than five years. Nor did the United States experience a recession after a similar fall in 1987. The difference in these experiences resulted in part from the policies followed at the time.201
Far from being the expansive agent that Miller and others alleged, the Federal Reserve followed a generally deflationary policy from mid-1927 on. Growth of the real value of the monetary base increased briefly in spring 1927, but it reversed quickly. Assistance to Britain produced a modest increase in the fall, but that too reversed quickly. Growth of the real base was negative for more than a year before the stock market crash. Chart 4.5 shows these data. The data suggest that if the Federal Reserve had used the monetary base as an indicator of policy thrust instead of the interest rates or bank credit, it would leave increased monetary growth.
Chart 4.5 shows that monetary growth remained in a narrow range during 1927 to 1929. The same is true of real interest rates; they too remained in a narrow range until summer 1929, then fell after industrial production peaked and the economy moved toward recession. These data deny that a rise in (ex post) real interest rates ended the stock market boom. Rather, the evidence suggests that deflationary monetary policies in all the gold standard countries induced reductions in output abroad that later spread to the United States and other countries. The stock market responded to the actual and anticipated decline in corporate profits. The fall in stock prices lowered the price of existing capital relative to replacement cost, reducing investment, output, and income and increasing the demand for money. (See Brunner and Meltzer 1968b, 1993.)
200. The decline in the demand for money in 1929 and 1930 is largely consistent with contemporary changes in interest rates and wealth. Annual estimates for these years do not show large residuals. Relatively large residuals are found in 1926 to 1928. See appendix to chapter 5 below and Field 1984.
201. After reading an earlier version of this chapter, Michael Bordo referred me to Sirkin 1975. Sirkin used a valuation model to show that, although individual companies may have been valued optimistically, “the marked overvaluation of stocks was not general” (231).
POLICY ACTIONS AND EFFECTS
Federal Reserve policy remained deflationary in the 1920s when judged by growth of the nominal monetary base.202 Accelerations of the base were typically short-lived, followed by renewed declines. In the two years ending June 1929, the monetary base fell 2 percent.
Growth of the money supply, M1, currency and demand deposits, generally moved with the growth of the base during the years 1927 to 1929. Chart 4.5 above shows these series. Real GNP rose 9.2 percent during these years. Rising output with slow or falling money growth produced deflation.203 As noted earlier, Balke and Gordon’s (1986) GNP deflator started to fall in third quarter 1928, five quarters after the local peak in the monetary base.204
202. Appendix A shows the statistical relationships discussed in this section.
203.
Narrow monetary aggregates such as the base and M1 are more revealing of the deflationary policy than broader aggregates (M2) in this period. M2 rose 3.6 percent in the two years ending in second quarter 1929 (based on quarterly averages) (Friedman and Schwartz 1963, table A-1). One reason for the difference is that, as noted several times, banks encouraged depositors to shift from demand to time deposits, reducing the average reserve requirement ratio and permitting bank assets to rise relative to M1. I interpret the shift of deposits as a response to the restrictive policy. Banks could not obtain desired reserves from the Federal Reserve, so they “innovated,” sharing some of their gains with their customers by paying higher interest rates on time deposits.
Actual and Predicted Inflation
The Federal Reserve was responsible for sterilizing gold inflows and for the deflationary policy in the United States. Chart 4.6 compares predicted and actual inflation, quarterly, from 1916 to 1930. The predictions come from estimates of the response of current inflation to past inflation and past money balances.205
Predicted and actual values generally move together, most notably in the quarters preceding the depression. Actual or measured inflation is much less variable quarterly than in the years after World War I. The data predict deflation or price stability at the end of the decade.
204. GNP data were not available at the time. The Bureau of Labor Statistics (BLS) index of wholesale prices declined also. By 1928 the BLS index was 6.6 percent below its peak in 1925, a larger decline than Balke and Gordon’s (1986) index for the same period. The BLS index declined an additional 1.5 percent in 1929. These data were available at the time and are at times cited in the minutes. The annual rate of change of consumer prices lies between 0 and –3 percent from July 1926 to June 1929.
205. Appendix A reports the equation used for these estimates.
Fiscal Actions
Budget surpluses from 1921 to 1929 averaged 20 percent of tax revenues a year; for 1927 to 1929 the average is 23 percent. These data suggest that the government followed fiscal policies usually described as “restrictive.” This characterization ignores the stimulus to enterprise from debt retirement and tax rate reduction. The Treasury retired almost 30 percent of the debt outstanding in 1921 as part of Secretary Mellon’s program to reduce the wartime debt. The large surpluses permitted the Mellon Treasury to reduce income tax rates in 1922, 1923, 1924, 1925, 1928, and 1929, the last a temporary reduction. The maximum tax rate on $1 million or more declined from 66 percent to 23 percent; the rate on incomes of $2,000 to $3,000 fell from 2 percent to 0.1 percent.
REORGANIZATION OF THE OPEN MARKET INVESTMENT COMMITTEE
As open market operation increased in importance and discount rate policy declined, the Board lost influence. Control of policy shifted to the five reserve banks represented on the OMIC and particularly to Strong, its chairman. Once Strong left the New York bank, the Board undertook to change the balance of influence.
Miller especially resented Strong’s influence, but he was not alone. Other members of the Board, including Young, the Board’s governor from 1927 to 1930, also wanted to shift power from the banks to the Board.206
The strength of feeling and, at times, irrational character of the arguments came out in a conversation between Young and Harrison in 1929. Harrison began the conversation by referring to the dispute between Washington and New York over open market purchases after the 1929 market crash. The Board objected to New York’s decision to act on its own and, despite the crisis, required New York to stop purchases before it would agree to reduce the discount rate. He soon put the issue into a larger context, warning Young that failure to reach a compromise would have “very serious consequences” for the System. The Board had gradually but steadily restricted the role of the reserve banks by taking “not supervisory powers but the equivalent of operating functions.” Harrison cited conflicts over discount policy, acceptance policy, and open market policy to illustrate his point. The Board delayed for months raising discount rates at New York and other banks in 1929, despite repeated requests from the reserve banks and agreement of all reserve bank governors. The Board had increased its
206. Young’s view is more surprising because he had been governor at Minneapolis before moving to Washington.
role in setting acceptance rates so much that it was now difficult to make rates effective in the market. The New York directors did not want to be “a rubber stamp.” He recognized that the Board also should not be a rubber stamp. The path they were on would lead to “a central bank operating in Washington” (Harrison Papers, Confidential Memoranda, conversation with Roy Young, November 15, 1929).
Governor Young replied, “I think that is so.” Harrison was surprised at this frank expression. Young continued, “the Federal Reserve Board has been given most extraordinarily wide powers. . . . They would feel free to exercise them and Congress could determine whether they objected to having a central bank operating in Washington.”
The shift of control over open market operations had been under way for more than a year before the conversation between Young and Harrison took place. In August 1928 the Board voted to consider expansion of the OMIC to include all the reserve banks. Miller had long favored a further step, making the Board’s governor chairman of the reconstituted committee (OMIC Minutes, Board of Governors File, box 1437, August 16, 1928). In September the Federal Advisory Council approved a resolution to expand the OMIC to include the twelve governors but did not include any Board members. The resolution called for a five-person executive committee to carry out the policies, thereby reducing any discretionary action by the five members of the original OMIC.
A few days later the Board appointed Miller and Platt to draft a revision of the rules governing the OMIC. Their proposal followed Miller’s 1923 proposal (see above) to replace the OMIC with an Open Market Policy Conference (OMPC), chaired by the governor of the Board, to meet at the call of the Board. Purchases and sales would be made mainly in the acceptance market (ibid., October 30, 1928). The Board approved a slightly weaker proposal for consideration by the Governors Conference. The governors rejected the proposal and called on the Board to keep the 1923 resolution but expand the committee by including all reserve bank governors (ibid., December 3, 1928).
In January 1930 the Board approved a proposal replacing the OMIC with the OMPC, the latter consisting of twelve reserve bank governors. An executive committee could carry out only those decisions of the OMPC “as have been approved” by the Board. Atlanta, St. Louis, Minneapolis, and San Francisco approved the proposal.207 The members of the OMIC— Cleveland, Chicago, Philadelphia, Boston, and New York—either opposed the proposal or opposed the Board’s veto power over committee decisions. Chairman Gates McGarrah wrote for New York that “they would continue membership on the committee, provided it is not inconsistent with these [their] general views.” Most banks reserved the right to conduct open market operations outside the committee framework, as permitted by law. Only New York reserved the right of any reserve bank to “withdraw from the Committee procedure altogether, if it deems it advisable” (Organization of OMPC, Board of Governors File, box 1437, March 24, 1930). It would soon regret this insistence.
207. Richmond, Kansas City, and Dallas did not reply by the date of the meeting, so their positions are not included. There were objections also to other sections not discussed in the text. A summary of the responses by Hamlin is part of the minutes of a meeting of the governors (Open Market, Board of Governors File, box 1437, March 24, 1930).
The twelve-member OMPC began operations soon after the meeting. The Board gave way on the issue of control but did not abandon its efforts until the Banking Act of 1935 centralized power and control in Washington.
CONCLUSION
The 1920s began and ended with major recessions. The 1920–21 recession was the first test of Federal Reserve policy in recession. The depth of the recession, the belief that discount policy had not worked as expected, and the political response to h
igher interest rates encouraged Federal Reserve officials to search for new policy procedures. Suspension of the gold standard abroad reduced the usefulness of the gold reserve ratio as a measure guiding policy action. This too suggested the need for new procedures. By 1923 the Federal Reserve had developed a more activist policy stance. The new procedures seemed successful. Confidence rose in the Federal Reserve’s ability to moderate the business cycle and prevent inflation. These hopes or beliefs were reinforced by restoration of gold convertibility, within a gold exchange standard, in all major countries. The recession that began in 1929 destroyed these beliefs.
The new activist policy was supposed to achieve three ends: mitigate business fluctuations, prevent inflation, and restore the international gold standard. From 1923 to 1929 the United States economy experienced growth, with brief recessions and low inflation before 1925 and modest deflation thereafter. The apparent success of postwar policies in achieving the three main objectives and preventing financial panics increased the credibility of policies and the belief that a new more stable era had begun. The rise in United States stock prices relative to earnings in 1926 supports this interpretation.
In retrospect, we know that the years 1923 to 1929 were one of the best periods in the first eighty years of Federal Reserve experience. The good results were not permanent, however. A severe recession began in Europe in 1929. In August, the United States economy followed. The gold standard also showed signs of strain: Canada left the standard in January 1929; although it continued to maintain a fixed exchange rate against the dollar and the pound, it did not have an official gold price (Bordo and Redish 1988).
A History of the Federal Reserve, Volume 1 Page 39