A Comparison of New Factor Models: Evidence From Turkey

Abstract

The purpose of this paper to compare the performances of new factor models with the former models in Turkey. Inthat aim, newly proposed q-factor model and Fama-French five factor model are compared with Fama-French threefactor, Carhart four factor and Pástor-Stambaugh factor models. The performance metric is chosen as maximumsquared Sharpe ratio which gives a better understanding in comparison of two or more models accordance to Barillasand Shanken (2017). As per the measure of maximum squared Sharpe ratio, the q-factor model outperforms of allbetween July 2009 and June 2017. After that, Carhart four factor model follows as the second best performing model.It is considered that this result may be due to the portfolio formation frequency of profitability and momentum factors.Thus, it can be inferred that the higher the data frequency, the better the explanatory power of the model. AlthoughFama-French five factor model is similar to q-factor model, the considerable outperformance of q-factor model canbe attributed to the way of factor construction and calculation. Consequently, it seems as though the performanceof the model is sensitive to the way of factor construction and calculation.