More on Monetary Policy and Stock Price Returns

Abstract

Recent research suggests a persistent empirical relation between U.S. monetary policy and stock returns since the mid-1980s. The findings seem questionable and incomplete, however, for at least three reasons. First, the results are sensitive to sample selection. Second, this research does not distinguish between anticipated and unanticipated monetary policy decisions. Third, such analysis does not satisfactorily consider that returns and policy are probably determined simultaneously because prices contain information about market expectations for the economy and, in turn, policy. Together, these issues suggest that investors are unlikely to profit from strategies based on past or anticipated Federal Reserve decisions.

No one would argue that Federal Reserve policy is irrelevant to financial markets. A more focused question is whether past data on the prevailing stance of monetary policy and stock prices imply that there is an anomaly that investors can exploit to realize excess returns. A cursory read of recent research that reports an empirical link between monetary policy cycles and stock returns since the mid-1980s might prompt one to question the efficient market hypothesis (EMH) and accept the existence of such an anomaly. At least three issues prevent me from drawing such a conclusion from the data.

First, the findings seem to be highly sensitive to sample selection. For example, exclusion of a single data point—the 19 October 1987 stock market crash—vitiates the findings. Also, the results vanish if the pre-1994 period (before the Federal Reserve began announcing its policy decisions) is excluded or the most recent easing cycle from 2001 through June 2004 is included. Of course, one can assail any seemingly robust statistical result with changes in the sample and torture the data until they confess. But these sample considerations, especially the increase in observations to include data past early 2001, should produce a more rigorous test of the hypothesis. At any rate, one might wish to know the effects of outliers and additional information before accepting the inference that the relationship between returns and monetary policy remains strong in recent periods.

Second, this recent research does not identify an anomaly that investors can exploit today, because the dichotomous measure of monetary policy that its authors use does not distinguish new from old information, which is absolutely necessary to detect an anomaly. In research reported in this article, I used federal funds futures to separate the two components and show that if monetary policy matters for stocks, it is new, not old, information that is potentially important. In other words, unanticipated rather than anticipated monetary policy is relevant, and the last 15 years of data suggest that an anomaly never existed, even within or between investment styles and sectors. Moreover, the impact of unanticipated policy actions is, on average, of only moderate magnitude in the context of overall stock market volatility.

Third, the recent research does not acknowledge that stock returns and monetary policy are probably determined simultaneously because stock prices contain information about investors' expectations for the economy and, in turn, Federal Reserve policy. Therefore, analysis based on the assumption that policy is exogenous should be viewed with suspicion.

So, is Federal Reserve policy still relevant for investors? No and yes. Old information does not move stock prices. New information seems to affect equity returns, but the effect is not substantial in the context of overall stock market volatility. If the data are any guide, portfolio managers are unlikely to profit from trading strategies based on past or anticipated Federal Reserve policy decisions.

I thank Stefania D'Amico, Refet Gürkaynak, and Jonathan Wright. My views do not necessarily reflect those of the Board of Governors of the Federal Reserve System or any other member of its staff.