Predicting future stock returns at short and long horizons

• Main Author: Ekundayo, S. (Sulaimon)
• Format: Dissertation
• Language: English
• Published: University of Oulu 2015
• Subjects: Finance
• Online Access: http://urn.fi/URN:NBN:fi:oulu-201505211593; http://nbn-resolving.de/urn:nbn:fi:oulu-201505211593

It has been established in a vast number of financial and econometric literature that financial and macroeconomic variables such as dividend-price ratio and term spread forecast aggregate stock market returns at both short and long horizons in many developed and emerging economies. Using financial ratios and macroeconomic variables, Rapach and Wohar (2006) report in-sample and out-of-sample evidence of predictability of future excess stock returns. However, contrary results are reported in the literature such as Welch and Goyal (2008). It is argued that regressions analysis produced poorly predicted in-sample and out-of-sample results. Following the methodology of Hjalmarsson (2010), I apply the OLS regressions analysis, where aggregate stock market returns are predicted by financial ratios and macroeconomic variables. I conduct the analysis for each of the countries: Australia, Denmark, Finland, France, Germany, Japan, Norway, Sweden, United Kingdom, and the United States using data from Morgan Sterley Composite Index (MSCI). I find that financial ratios such as dividend-price and earnings-price ratios accept null hypothesis of no predictability of stock returns at short horizons for all the countries except Norway and United States where predictability is evident at 3 months horizons, but there exist predictability at the long horizons. This is a contrary view to what previous literature reported. The overall findings of this thesis suggests that dividend yield, earnings yield, term spread, and short interest rate predict stock returns at long horizons in all the countries with an exception of short interest rate and term spread that fail to forecast stock returns at all the horizons for Denmark. Book-to-market ratio predicts stock returns at short and long horizons in 80% of the countries that are covered in my sample.