The Midas Paradox
Financial Markets, Government Policy Shocks, and the Great Depression
By Scott Sumner

Highlights | Synopsis


  • Although economic historians have made great progress in unraveling the causes of the Great Depression, The Midas Paradox offers the first attempt to provide a comprehensive explanation of both monetary and non-monetary causes of that cataclysm. There are two major themes. First, a series of shocks to the global gold market created a deflationary contraction during 1929–33 and again during 1937–38. Second, after March 1933, a promising recovery was repeatedly cut short by President Roosevelt’s attempts to artificially raise wage rates. The combined impact of a dysfunctional gold standard and a counterproductive wage policy produced the greatest macroeconomic disaster in American history, a depression lasting twelve years.

  • Numerous studies have examined the role of monetary policy in the Great Depression, but The Midas Paradox is the first to successfully integrate the monetary policy of individual central banks with the powerful forces shaping the global gold market. This approach allows us to explain for the very first time why the Great Depression began in late 1929 and not at some other date.

  • Contrary to widespread belief, individual central banks were not innocent bystanders: they actively destabilized the gold market by hoarding vast quantities of gold. Ironically, the governments that hoarded gold the most aggressively typically ended up suffering the most when the rising value of gold led to deflation. In addition, persistent (and often justified) fear of currency devaluation led to four key episodes of private gold hoarding, each of which destabilized global assets markets and plunged the economy deeper into depression. Hence the book’s title: The Midas Paradox refers to a world economy impoverished by the hoarding of gold.

  • President Franklin Delano Roosevelt’s devaluation of the dollar in the spring of 1933 produced the fastest economic growth in U.S. history, but the record upswing lasted only four months due to the destructiveness of his National Industrial Recovery Act. This Orwellian-sounding law was supposed to help revive the economy, by encouraging farmers to grow less and by restricting the number of hours worked in industry. Although it succeeded in reducing output, the Act failed to spur the recovery that Roosevelt had hoped for. Not to be deterred, he tried to raise wage rates on four subsequent occasions, and each attempt led to a sharp slowdown in the recovery, if not an outright relapse into depression.

  • New Deal legislation led to five separate nominal wage shocks that aborted promising economic recoveries, but wage shocks were only part of the cause of the economic volatility of the 1930s. The other key cause was monetary policy, a factor that most economists have measured incorrectly. Under an international gold standard, the most reliable indicators are changes in the gold reserve ratio and the price of gold—not the domestic money supply, interest rates, and gold flows. The gold market approach helps to explain price-level volatility from October 1929 to March 1933. Surprisingly, it is even more useful for understanding the first five years after the United States departed from the gold standard.


Since the early 1960s, economic historians have made great progress in explaining the causes of the Great Depression. Despite their prodigious efforts, however, none have offered a convincing account of the multitude of twists and turns the economy took from 1929 to 1940—until now.

In The Midas Paradox: Financial Markets, Government Policy Shocks, and the Great Depression, Bentley University economics professor Scott Sumner offers a novel approach to this vexing topic, one that answers previously unsolved mysteries of the 1930s by focusing on gold, wages, and financial markets. This approach enables him to explain all of the economy’s fits and starts in output, including the timing of its initial decline, something that had eluded a cogent explanation.

Drawing on financial market data and contemporaneous news stories, Sumner reveals that the Great Depression is ultimately a story of incredibly bad policymaking—by central bankers, legislators, and two presidents—especially mistakes related to monetary policy and wage rates. He also shows that macroeconomic thought has long been captive to a false narrative that continues to misguide policymakers in their quixotic quest to promote robust and sustainable economic growth.

The Midas Paradox begins with a crash course in the history of research about the Great Depression. Sumner shows that although economic historians have plowed important ground, they have left many questions unanswered because they have neglected a crucial factor: the international gold market. In Part II he utilizes a gold-market approach to deepen our understanding of the Great Contraction (1929–33), paying particular attention to the hoarding of gold by the world’s central banks. Part III focuses on monetary policy in one particularly fateful year, 1933. Part IV looks at a key consequence of dubious monetary policies: private gold holding. Part V shows how a misreading of the events of 1932–33 has profoundly distorted twentieth-century macroeconomic thought. Along with offering his conclusions, Sumner provides a technical discussion of theoretical issues related to modeling the Great Depression.

Twenty years in the making, The Midas Paradox is a landmark treatise that solves mysteries which have long perplexed economic historians and corrects misconceptions about the true causes, consequences, and cures of macroeconomic instability. Like Milton Friedman and Anna J. Schwartz’s A Monetary History of the United States, 1867–1960, it is one of those rare books that will shape all future research on the subject.

Gold, Wages, and the Great Depression

In Part I, Sumner sets the stage by examining the issues to be addressed and the strategies to be used. U.S. industrial output didn’t plunge only once during the Great Depression: it rose and fell more than a dozen times. Why was the economy so volatile? Economic historians have studied this issue for decades, but just as they have appeared to get a handle on one aspect of the puzzle, another emerges. One neglected resource for solving the puzzle—a treasure-trove of information rarely utilized in books about the Depression—is financial market data.

From 1929 to 1938, financial markets were unusually erratic and moved closely in response to news stories related to gold and/or wage legislation. The correlation is so strong that financial market data enable us to explain most of the output volatility of the 1930s. Using a simple aggregate supply and demand framework reveals that the demand shocks to the economy were triggered by gold hoarding (or changes in the price of gold), and the supply shocks were caused by policy-driven changes in hourly wage rates.

The Great Contraction, 1929–1933

Economic historians have rightly blamed banking panics and a collapse of aggregate demand for the economic decline known as the Great Contraction. Sumner, however, argues in Part II that they’ve overlooked a more fundamental factor: the surge in the demand for gold holdings. Understanding why gold hoarding soared requires that we cast our gaze afar.

Although U.S. monetary policy tightened in mid-1928, world monetary policy was stable from June 1928 to October 1929. During the twelve months that followed, however, central banks as a whole tightened sharply, and this policy switch caused (or substantially caused) a sharp drop in aggregate demand.

A key turning point in the Great Depression came in 1931. The German economic crisis and the British devaluation led to gold hoarding and a weakening of U.S. equity markets from mid-1931 to late 1932. The gold hoarding was the most important factor that depressed aggregate demand in the fall of 1931. The Federal Reserve’s increase in the discount rate in October 1931 had no impact.

Nor did the Fed’s open market purchases in the spring of 1932. Keynes suggested that this ineffectiveness might have been due to the existence of a “liquidity trap.” Friedman and Schwartz suggested that the modest upswing in late 1932 was a lagged response to the central bank’s efforts. The evidence supports neither view. The open market purchases were associated with massive gold hoarding,

a phenomenon that prevented any significant increase in the money supply. Stock prices, commodity prices, and economic output began to rise only when investors felt sufficiently confident to curtail their hoarding of gold.

Bold Experiments in Monetary Policy, 1933

In Part III, Sumner examines two of the most important policy shocks in U.S. history, both occurring in 1933. The first is Roosevelt’s dollar depreciation program. In April, at the president’s urging, the Federal Reserve took measures to raise the price level back to its 1926 level. This program, unique in U.S. history, helped cause a 57 percent surge in industrial output from March to July 1933. More precisely, the recovery resulted from the public’s expectations of future monetary expansion, not from the depreciation program alone.

The recovery was short-lived, a victim of the National Industrial Recovery Administration. In July 1933, the agency sought to sharply increase hourly wage rates—an unprecedented measure that led to a major stock market crash and helped lengthen the Great Depression by six to seven years. In a sense, there were two depressions: one related to the contraction of 1929–33 and a second, unrelated one caused by federal wage policy.

Economic historians have greatly underestimated the importance of the two policy shocks considered in isolation. Yet the importance of the dollar depreciation program can be inferred from the daily changes in the free-market price of gold, which after April 1933 became an excellent proxy for exogenous monetary shocks. The various market responses to dollar depreciation call into question many traditional theories of the monetary policy transmission mechanism.

Back on the Gold Standard

In Part IV, Sumner focuses on the level of private gold holding. Although the conventional view is that President Roosevelt took America off the gold standard, U.S. monetary policy became even more strongly linked to gold after 1934 than it had been before 1933. A recovery in the United States finally got underway when the Supreme Court declared the National Industrial Recovery Act to be unconstitutional in mid-1935. Because the demise of the gold bloc in 1936 reduced devaluation fears, its impact on gold demand and the broader macroeconomy was exactly the opposite of the British devaluation of 1931.

Actual and prospective gold dishoarding led to high inflation during late 1936 and early 1937. Expectations of future gold supplies were so high that tight monetary policies lacked credibility. Rapid inflation led to a “gold panic” in the spring of 1937 as investors worried that the buying price of gold would be reduced. During 1937, the expansionary impact of gold dishoarding began to be offset by wage increases, which reflected the resurgence of unions after the Wagner Act and Roosevelt’s landslide reelection.

Many economic historians have argued that the 1937–38 depression was caused by restrictive fiscal policy and/or by increases in reserve requirements. Neither view is persuasive. Instead, the rapid wage inflation (combined with the end of gold panic–induced price inflation) modestly slowed the economy during the summer of 1937. This slowdown led to renewed expectations of dollar devaluation during the fall, and as gold was again hoarded on a massive scale, expectations of future U.S. monetary growth declined sharply. It was this shift in expectations that triggered the precipitous declines in stock prices, commodity prices, and industrial production during late 1937.

From 1936 to 1937, commodity prices increased significantly. This boom was caused by instability in the world gold market, and was the mirror image of a more fundamental change—a sharp dip in the value of gold. Previous economic historians have overlooked the way that gold market instability triggered the boom and bust of 1936–38 and have mistakenly blamed the recession on tighter fiscal policy or higher reserve requirements. As with 1933, a complete understanding of the economy’s path during 1937 requires a subtle analysis of the interrelationship between gold market and labor market disturbances.

Gold, the Great Depression, and Macroeconomic Thought

Part V begins with Sumner’s discussion of how the misinterpretation of two key policy initiatives—the open market purchases of 1932 and the National Industrial Recovery Administration—profoundly shaped macroeconomic theory during the twentieth century. Because early Keynesian theory was based on a misreading of these policies, it could not survive the radically altered policy environment of the postwar period. By the 1980s, the original Keynesian model had been largely replaced by a (quasi-monetarist) “new Keynesianism,” featuring highly effective monetary policy and a self-correcting economy. This era may have ended in 2008.

Economic historians continue to debate whether or not the international gold standard was an important constraint for interwar central banks. It seems unlikely that this issue can ever be resolved, and the debate may have diverted attention from a much more important issue: how the world gold market affected contemporaneous policy expectations. At the deepest level, the causes of the Great Depression and World War II are fundamentally similar: both events were caused by policymakers moving unpredictably between passivity and interventionism.

The final chapter, which is geared toward Sumner’s colleagues in the economics profession, is a technical appendix on theoretical issues in modeling the gold market and analyzing real-wage cyclicality. If one defines real wages as the ratio of nominal wages and wholesale prices, then high-frequency fluctuations in real wages during the 1930s were highly correlated with movements in industrial production. Understanding real wage cyclicality is the key to understanding the Great Depression, and this requires separate analysis of nominal wage and price level shocks.

Buy The Midas Paradox