Asset pricing models an application of FF6F model in BIST XUSIN index

The purpose of this research is to test FF6F model's capability to capture excess returns in BIST-XUSIN index. To that end, data set has been collected, refined and cultivated to construct portfolios and Fama-French risk factors. Based on 2x3 sorting method, 456 portfolios have been constructed. By using monthly data set, July 2005-December 2023 period has been taken into consideration. In order to employ the right methodology, pre-tests such as stationarity, normality and multi-collinearity tests have been conducted. Ordinary least squares (OLS) methodology has been employed after pretest results. Post-tests results also suggest that OLS is the right methodology. To evaluate the models as a whole, Gibbons, Ross and Shanken (1989) GRS-F test has been employed. The findings suggest that FF5F and FF6F models have performed well in terms of GRS-F test probability results at 0.05 significant level. When the models have been ranked in terms of performance, FF5F model has leaded the line based on GRS-F test probability results. Furthermore, when other GRS-F test family members such as Sharpe ratio and adjusted R-squared have been taken into account, FF5F and FF6F models are still working well and they are better than the previous nested models at 0.05 significant level. Finally, HML factor has been tested in terms of redundancy both for FF6F and for FF5F models. Factor spannning test results suggest that HML factor is redundant for the both models.