For more than a century, diversified long-horizon investments in America's stock market have consistently received much higher returns than investors in bonds: a return gap averaging 6 percent per year. An enormous amount of creative and ingenious work by a great many economists has gone into seeking explanations for the so-called "equity premium return puzzle," but so far without a fully satisfactory answer. We first review the facts about the equity premium and then discuss a range of explanations that have been proposed. We conclude that the equity premium puzzle has not been solved: it remains a puzzle. And we anticipate that the equity return premium will continue, albeit at a smaller level than in the past--perhaps four percent per year.

(Introduction, initial paragraphs) For more than a century, diversified longhorizon investors in America’s stock market have invariably received much higher returns than investors in bonds: a return gap averaging some six percent per year that Rajnish Mehra and Edward Prescott (1985) labeled the “equity premium puzzle.” The existence of this equity return premium has been known for generations: more than eighty years ago financial analyst Edgar L. Smith(1924) publicized the fact that longhorizon investors in diversified equities got a very good deal relative to investors in debt: consistently higher longrun average returns with less risk. It was true, Smith wrote three generations ago, that each individual company’s stock was very risky: “subject to the temporary hazard of hard times, and [to the hazard of] a radical change in the arts or of poor corporate management.” But these risks could be managed via diversification across stocks: “effectively eliminated through the application of the same principles which make the writing of fire and life insurance policies profitable.” Edgar L. Smith was right. Common stocks have consistently been extremely attractive as longterm investments. Over the half century before Smith wrote, the Cowles Commission index of American 3 stock prices deflated by consumer prices shows an average real return on equities of 6.5 percent per year— compared to an average real longterm government bond return of 3.6 percent and an average real bill return of 4.5 percent. 1 Since the start of the twentieth century, the Cowles Commission index linked to the Standard and Poor’s Composite shows an average real equity return of 6.0 percent per year, compared to a real bill return of 1.6 percent per year and a real longterm government bond return of 1.8 percent per year. Since World War II equity returns have averaged 6.9 percent per year, bill returns 1.4 percent per year, and bond returns 1.1 percent per year. Similar gaps between stock and bond and bill returns have typically existed in other economies. Mehra (2003) 2 reports an annual equity return premium of 4.6 percent in postWorld War II Britain, 3.3 percent in Japan since 1970, and 6.6 percent and 6.3 percent respectively in Germany and Britain since the mid1970s.

What forces led to rapid recovery of the U.S. economy after 1933? Why was recovery derailed by a severe recession in 1937? Since Friedman and Schwartz (1963), economists have emphasized monetary explanations. This dissertation shows that in three crucial instances other factors mattered. Methodologically, it demonstrates the value of using micro data to explore macro questions.

The first chapter considers the effect of the 1936 veterans' bonus. Conventional wisdom has it that in the 1930s fiscal policy did not work because it was not tried. This chapter shows that fiscal policy, though inadvertent, was tried in 1936, and a variety of evidence suggests that it worked. A deficit-financed veterans' bonus provided 3.2 million World War I veterans with cash and bond payments totaling 2 percent of GDP; the typical veteran received a payment equal to annual per capita personal income. This chapter uses time-series and cross-sectional data to identify the effects of the bonus. I exploit four sources of quantitative evidence: a detailed household consumption survey, cross-state and cross-city regressions, aggregate time-series, and a previously unused American Legion survey of veterans. The evidence paints a consistent picture in which veterans quickly spent the majority of their bonus. Spending was concentrated on cars and housing in particular. Narrative accounts support these quantitative results. A back-of-the-envelope calculation suggests that the bonus added 2.5 to 3 percentage points to 1936 GDP growth.

In my second chapter, I consider the causes of the severe recession that followed the boom year of 1936. The 1937-38 recession was one of the largest in U.S. history. Industrial production fell 32 percent and the nonfarm unemployment rate rose 6.6 percentage points. This paper shows that there were timing, geographic, and sectoral anomalies in the recession, none of which are easily explained by aggregate macro shocks. I argue that a supply shock in the auto industry contributed both to the recession's anomalies and to its severity. Labor-strife-induced wage increases and an increase in raw material costs led auto manufacturers to raise prices in the fall of 1937. Equally important, higher costs combined with nominal rigidity to make the price increase predictable. Expectations of price increases brought auto sales forward and thus sustained sales during the summer and early fall of 1937, despite negative monetary and fiscal factors. When auto prices did rise in October and November, sales and production plummeted. A forecasting exercise suggests that in 1938, this shock reduced auto sales by 600,000 and GDP growth by 1.2 percentage points.

In my third chapter, I consider the miraculous recovery of the U.S. economy during Franklin Roosevelt's first months in office. From March to July 1933, industrial production rose 57 percent, by far the most rapid increase since records began. The recovery was concentrated in durable goods industries, particularly autos. Using a novel dataset, I show that auto sales increased much more rapidly in farm than in nonfarm states in spring 1933. This suggests that dollar devaluation may have directly spurred recovery by raising farm incomes. The negative effect of higher crop prices on urban consumers may have been offset by the positive effect of higher crop prices on expected inflation.