Empirical evidence of stock return predictability using macroeconomic variables

Sammanfattning: We investigate whether macroeconomic variables can predict returns of the OMXS30 index in the short run, and if an investor can generate abnormal profits from using the variables with significant predictive power. Granger causality tests, along with a predictive OLS regression framework show that the first difference of the repo rate and the log difference in exchange rates significantly Granger cause stock returns on the Swedish market. The findings confirm that changes in the repo rate affect stock returns in line with the transmission mechanism effect of monetary policy and supports that currency depreciation negatively affects future stock returns. We also show that an investor could have generated abnormal returns using macroeconomic variables by deploying a regression based rolling window trading strategy, that yielded statistically significant four-factor alpha between 1998-2008 (without considering transaction costs) - although generating lower returns going forward. The strategy's worsened performance is further linked to the negative interest rate regime, creating difficulties to estimate its recent linear interdependency with stock returns that is used to trade on - indicating that the effectiveness of the repo rate as a monetary policy tool affects strategy returns.