The relationship between inflation and stock return revisited: A new empirical and narrative investigation.

摘要

This dissertation studies the relationship between unanticipated inflation and stock returns to explain the correct sign of the correlation and detect any parameter instabilities. To achieve our research objectives we use two methods. First, we use a narrative approach to obtain historic evidence on the behavior of the inflation-return relationship over time. Second, we use the Flexible Least Square (FLS) regression method to estimate three equations. The first equation examines the effect of unanticipated inflation, the bond default premium, the bond horizon premium, the Federal Funds rate, and industrial production on total returns of large stocks. The second equation estimates the impact of unanticipated inflation, the Federal Funds rate, and industrial production on total returns of long-term Government bonds. The third equation examines the effect of various independent variables including unanticipated inflation on returns to gold. We use FLS because it is an exploratory technique. It does not require ad hoc prior assumptions about the distribution of disturbances for the estimated equation or about the structure underlying the variation of coefficients. FLS has also the ability to account for the expected coefficient instability over time. Conclusions of both our narrative and our empirical research confirm our thesis that the correlation between inflation and stock return always has a negative sign and is time-varying. The non-ambiguous existence of parameter instability is strong evidence that the FLS statistical methodology is very useful in describing the dynamic behavior of the inflation-returns relationship over time. Results of this study show that there is a disparity in how the US stock and long-term bond returns respond to unanticipated inflation during the past forty years.