The Depression of 1920–21: The Deflation America Let Run

On 1 June 1920 the Federal Reserve Bank of New York raised its discount rate to 7 per cent. No Reserve Bank had ever charged that much, and none would again until 1969. Benjamin Strong, the governor who had built the New York bank into the operating centre of the young system, was not in the room. Tuberculosis had taken him out of the office for most of that year, and the decision belonged to a Federal Reserve Board in Washington chaired by a different Harding — William Proctor Gould Harding, a banker from Birmingham, Alabama, who would spend the next eighteen months explaining to Congress why farm prices had halved.

Wholesale prices in the United States were then at their highest level in the country's history. Twelve months later they had fallen by more than a third. Industrial production dropped by something close to a third as well, roughly a fifth of the industrial workforce lost its job, and just over 500 banks suspended payments. It was the sharpest deflation the American economy has recorded, and almost nobody outside the farm belt remembers it happened.

The Price of Winning

The Federal Reserve entered the 1920s as an institution that had never once set policy on its own account. Created in 1913 after the Jekyll Island meeting and the legislation that followed, it spent the war as an arm of the Treasury. Liberty Loan bonds had to be sold at par, which meant interest rates had to stay where the Treasury wanted them, and the Reserve Banks obliged by lending at preferential rates against government paper. Carter Glass, Treasury Secretary from December 1918, saw no reason to change an arrangement that had financed a war.

Money supply and prices responded as one would expect. Consumer prices roughly doubled between 1913 and 1920. The boom did not end with the armistice in November 1918; after a brief pause it accelerated, because European demand for American goods survived the fighting and domestic producers were restocking. Wholesale prices rose again through 1919 at an annual pace above 20 per cent, and the dollar's purchasing power was disappearing fast enough that ordinary people noticed it weekly.

What finally moved the Board was not the price level but the gold cover. Reserve Banks were required by statute to hold gold equal to at least 40 per cent of their note issue, and through the spring of 1920 the system's reserve ratio slid toward that floor. An institution that cannot legally expand further has to contract, and contraction in 1920 meant the discount rate. New York went from 4.75 to 6 per cent on 23 January and to 7 per cent on 1 June, where it stayed for eleven months.

Elmus Wicker's study of early Federal Reserve policy argues that the Board understood the gold constraint far better than it understood the transmission of a rate rise into a real economy, and that nobody in Washington had a usable estimate of how much output would have to fall to restore the ratio (Wicker, 1966). Allan Meltzer reaches a similar verdict in his history of the institution: the 1920 tightening was late, and once it began there was no framework for calibrating when to stop (Meltzer, 2003).

The Fall

Business activity peaked in January 1920 by the National Bureau's later dating and reached its trough in July 1921, an eighteen-month contraction. Within it, the collapse of prices is what distinguishes the episode from every other American downturn.

Between the 1920 and 1921 annual averages, the wholesale index fell 36.8 per cent. Nothing in the American record matches it, including 1930–33, when the equivalent three-year decline was spread across three years. Stanley Lebergott's unemployment estimates put the 1921 rate at 11.7 per cent, a figure Christina Romer later revised down to roughly 8.7 per cent on the grounds that the pre-war series exaggerated cyclical swings (Romer, 1986). Either number describes a labour market in serious trouble.

Equities did what equities do. The Dow Jones Industrial Average had peaked at 119.62 on 3 November 1919 and bottomed at 63.90 on 24 August 1921, a decline of 47 per cent. That August low is the floor from which the bull market of the 1920s begins, which is why it appears in most narratives only as a starting point for the story of the boom that ended in October 1929.

The Farm Belt Pays

American agriculture had expanded through the war to feed Europe, financing the expansion with mortgage debt taken on at wartime land values. When European farms came back into production and the Fed raised rates into the same season, the two shocks arrived together.

Cotton had sold above 40 cents a pound in 1919 and traded below 10 cents by mid-1921. Wheat, which averaged more than $2.50 a bushel in 1920, fell under a dollar by the end of 1921. Corn in some Iowa counties was cheaper as fuel than as grain, and farmers burned it. Land values in the Corn Belt fell by roughly a third between 1920 and 1923 and kept sliding for most of the decade, taking with them the small unit banks whose loan books consisted of mortgages on that land. Rural bank failures ran at several hundred a year for the rest of the 1920s, a slow-motion collapse that was well advanced before the stock market ever wobbled.

Congress noticed sooner than the Board did. In January 1921 it revived the War Finance Corporation's agricultural lending powers over President Wilson's veto, putting Eugene Meyer back in the business of financing farm exports. A Joint Commission of Agricultural Inquiry spent much of 1921 taking testimony on whether the Federal Reserve had deliberately broken the farm economy, and W.P.G. Harding's defence — that the Board had faced a legal reserve requirement and a speculative credit expansion, and had no instrument that could distinguish a farmer's note from a stock promoter's — did not persuade the agricultural bloc. His term as Board governor ended in 1922 and the Agricultural Credits Act of 1923 added a farm representative to the Board, a small monument to a large grievance.

Wages, Not Prices

The 1920–21 adjustment ran through wages with a speed that no later American downturn has reproduced. United States Steel cut wages 20 per cent in May 1921 and again the following year; money wages across manufacturing fell by something in the region of 20 per cent from their 1920 peak. Employers used the moment to break the closed shop, marketing the open-shop campaign under the name of the American Plan, and union membership fell from about 5 million in 1920 to roughly 3.6 million by 1923.

That flexibility is the empirical core of the argument made by those who admire the episode. Prices and wages fell fast, real balances recovered, and the economy turned up within eighteen months without fiscal stimulus or monetary rescue. James Grant's account presents the contrast with the 1930s as close to a controlled experiment, one depression allowed to liquidate itself and another managed into a decade of stagnation (Grant, 2014).

The comparison is not as clean as it looks. J.R. Vernon's work on the recovery points out that the rebound coincided with a rapid monetary expansion once the Reserve Banks began cutting — New York's rate came down in five steps from 7 per cent in May 1921 to 4.5 per cent in November — and that the money stock grew strongly through 1922 (Vernon, 1991). Milton Friedman and Anna Schwartz, who had no sympathy for liquidationism, treated 1920–21 as an early demonstration of the Federal Reserve's characteristic failing: acting too late, then acting too hard (Friedman and Schwartz, 1963). A deflation that is over in twelve months and one that runs for three years are different animals, and the difference has more to do with what the central bank did next than with the virtue of letting prices fall.

The Government That Did Less

Warren Harding — the President, not the Board governor — had campaigned in 1920 on a word he may have invented. Speaking in Boston that May, he told his audience that "America's present need is not heroics but healing; not nostrums but normalcy." By the time he took the oath on 4 March 1921 the country was ten months into the contraction, and his inaugural address made no offer of relief: "We must face a condition of grim reality, charge off our losses and start afresh."

What the administration did was cut. Federal spending had run to roughly $18.5 billion in fiscal 1919; by fiscal 1922 it was near $3.3 billion, a reduction with no parallel in American budgeting, though most of it was simply the war ending rather than a policy choice. The Budget and Accounting Act of June 1921 created the Bureau of the Budget and installed Charles Dawes as its first director, giving the executive branch a consolidated budget for the first time. Andrew Mellon's Revenue Act of 1921 cut the top marginal income-tax rate from 73 to 58 per cent and repealed the wartime excess-profits tax.

Trade policy moved the other way. The Emergency Tariff Act of May 1921 raised duties on agricultural imports, and the Fordney–McCumber Act of September 1922 lifted rates across the schedule, a protectionist turn that would find its conclusion in the tariff act of 1930. Herbert Hoover, newly installed at Commerce, convened a President's Conference on Unemployment in September 1921 to coordinate public works and local relief. It met, by the National Bureau's dating, two months after the economy had already turned.

The Argument in Europe

America's deflation was one national episode in a general post-war reckoning about what money was supposed to be worth. Britain was preparing for the same medicine on a longer timetable, and would take it in 1925 when Winston Churchill returned sterling to its pre-war gold parity — an experiment in deliberate deflation whose costs are set out in Britain's return to gold. Germany went the opposite way and discovered in 1923 what the far side of the monetary spectrum looked like.

John Maynard Keynes, writing his Tract on Monetary Reform in 1923 with the American and British experiences both in view, refused to treat inflation and deflation as symmetrical evils. Inflation robbed the saver; deflation destroyed employment and output. His conclusion was blunt: "it is worse, in an impoverished world, to provoke unemployment than to disappoint the rentier." The Tract was the opening statement of an argument about the cost of price stability that ran through the whole interwar period and has never entirely closed.

What 1921 Settled

Recovery was genuine and quick. Output regained its 1920 peak during 1922, industrial production ran hard through 1923, and the decade that followed produced the electrification, the automobile, and the credit expansion that made the 1929 break possible. For the generation of central bankers who lived through it, the lesson taken was not that deflation is harmless but that the Federal Reserve now knew it could move the price level. Strong returned to New York and spent the rest of the decade using open-market operations as a deliberate instrument rather than an accounting by-product, which is the practical birth of American monetary policy as a discipline.

The lesson the farm states took was different and longer-lasting. Between 1920 and 1933 the agricultural economy never fully recovered its wartime footing, and the political demand for cheap credit, price supports and debt relief that produced the New Deal farm programmes has its origin in the crop year of 1921, not in 1932. Deflation is only ever neutral in the aggregate. Its distribution ran from debtors to creditors and from the countryside to the cities, and the people it ran from spent the next decade voting about it.

Fifty-eight years later, when Paul Volcker set out to break a different inflation, he did it knowing exactly how much output a determined central bank could destroy, because the 1920 file was in the building — see the Volcker disinflation. The difference was that the Federal Reserve of 1979 chose the pain with its eyes open, and the Board of 1920 stumbled into it while counting its gold.