Market efficiency, margin requirements, and risk-taking

Abstract

This dissertation consists of three essays that study the evolution of market efficiency and the impact of margin requirements on market efficiency and investors’ risk-taking behavior. In the first chapter, we examine how margin requirements affect the market’s efficiency in incorporating firm-specific and market-level public news. Combining the Fed’s 22 changes in margin requirements with a hand-collected sample of earnings announcements between 1934-1975, we show that higher margin requirements induce greater delay in incorporating earnings information into prices. We draw similar conclusions when we analyze the Hou and Moskowitz (2005) price delay measure, as well as indirect measures of leverage constraints over recent years. Further tests suggest that, despite the Fed’s expressed intent to curtail excess speculation, higher margin requirements restrict trading by arbitrageurs more than noise traders.

The second chapter studies the impact of margin requirements on investors’ risk-taking behavior. Investors bid up the riskier stocks in the long legs of hedge portfolios associated with the market, HML, and SMB factors relative to the less risky stocks in the short legs, when they anticipate the Fed increasing margin requirements. Following such a policy change, the returns on these hedge portfolios decline, implying lower subsequent compensation for bearing the risk associated with these three factors. In contrast, margin requirements are unrelated to returns on the momentum factor. Our evidence suggests that investors adjust their risk exposures to the market, SMB, and HML factors when leverage constraints are changed, but not momentum.

In the third chapter, we document a striking U-shaped pattern in the evolution of market efficiency over the period, 1934-2018. In terms of investors’ response to both firm-specific and market-wide news, markets are more efficient during the early and late years in this extended sample, while they become less efficient in the middle decades. We argue that this U-shaped pattern in the degree of market efficiency over time has been driven by two distinct economic dynamics. While the recent evolution in information-processing technology has led to more efficient markets in the later periods, the surprisingly high degree of market efficiency in the 1930s and 1940s reflects the greater relative importance of earnings announcements as a critical source of information that commanded investor attention, at a time when there was less overall information to process and fewer alternative information venues to consider. Overall, these results highlight that the evolution of market efficiency has not followed a linear path, but rather, divergent economic forces have caused the U-shaped pattern in market efficiency over time.