Return Prediction with Time Varying Betas: a Research in BIST

İngilizce (OpenAlex)

In the present study, dynamic versions of beta, which is the risk measure of investment instruments, have been employed to predict daily return of 30 random portfolios made of 154 stocks transacted in BIST ALL between dates 02.01.2003 and 29.08.2013.BIST 100 Index has been employed as the market portfolio.The predictions have been made with rolling regression and MGARCH methods.The performance of return predictions of dynamic betas has been compared to the performance of return predictions of traditional beta.Dynamic betas have been estimated with rolling regression, MGARCH DVECH, MGARCH DBEKK, MGARCH CCC and MGARCH DCC.In the study, it has been identified that the return prediction made with dynamic betas performed better than the predictions made with traditional beta.However, the return predictions made with CCC betas have been superior to other dynamic betas in terms of beating the performance of traditional beta.