Portföy Yönetiminde Bayesci Yaklaşımlar: BIST30 Endeksi Üzerine Bir Uygulama

Abstract

Markowitz's Mean-Variance portfolio is a breakthrough in finance because it leads the use of mathematics in financial applications. Moreover, with the model one period portfolio decisions can easily be taken and this is also important feature of the model. On the other side, the studies in the literature show that the performance of the mean-variance portfolios are very poor due to the estimation errors. Therefore, to overcome this problem several new theories and methods/approaches are proposed. The bayesian portfolio methodology is one of them. In this study we aim to compare three bayesian approaches, i.e. Jorion (1985, 1986), Kan and Zhou (2007) and Tu and Zhou (2011), with the Markowitz's mean-variance model. In addition, as being a base model the equally weighted portfolio is also compared with those models. For this purpose, in the study the daily return series of 30 assets traded in ISE30 during the period of 2010-2017 are considered and the performance of the portfolios of different dimensions which are constructed with the help of the simulations generated from the real return data are compared. As a general overview, it is concluded that Kan and Zhou's bayesian approach performed best among others.