Investigating the effectiveness of model-based, out-of-sample predictors regarding the predictability of stock (excess) returns in the US

Abstract

This thesis investigates the effectiveness of model-based, out-of-sample predictors in determining the predictability of stock returns in the US market, focusing on the S&P 500 index. The research examines the impact of inflation and interest rates on stock returns. Significant results indicate that Year-over-Year (YoY) inflation has predictive power for one-, three-, and twelve-month returns. Detrending data significantly improves these results. Multivariate regression incorporating both interest rates and inflation provides better forecasting than univariate regression. This study aims to offer insights for investors and researchers. Keywords: Inflation, Interest Rate, Predictability, Excess Stock Return, Detrending, Univariate Regression, Multivariate Regression, In-sample, Out-of-sample, S&P 500, U.S. market.