Testing postmodern portfolio theory based on global and local single factor market model: Borsa Istanbul case

Abstract

This study makes a comparative analysis of the explanatory power of CAPM and downside CAPM based risk measures for stock returns in Borsa Istanbul. 22 risk measures based on mean-variance and mean-semivariance approaches using global and local single factor models are examined for 2005–2016 period in a panel data setting. Mean-semivariance approach (downside CAPM) based downside betas and downside standard deviations have significant explanatory power for stock returns whereas CAPM based local and global betas fail to explain stock returns. The mean-semivariance approach (downside CAPM) could determine cost of equity more accurately. Deviations of returns below the mean are better risk indicators than deviations of returns below risk free rate of return and negative returns. Borsa Istanbul is partially integrated with the global market index and the degree of integration is higher during periods of negative returns. Results suggest that USD/TRY relationship is the dominating factor compared to MSCI movements.