Adaptive Market Hypothesis and Return Predictability: A Hidden Markov Model Application in Borsa Istanbul

Abstract

The adaptive market hypothesis (AMH) has recently attracted significant interest in the financial literature. The AMH has started to be considered an alternative to the efficient market hypothesis. In this respect, this study, first of all, examines the AMH for the BIST100 index of Turkey’s Borsa Istanbul stock exchange market by testing the return predictability. The applications are performed via automatic portmanteau and the generalized spectral (GS) tests using daily closing price data between January 1988 and December 2017. Secondly, the results of these tests are utilized for a hidden Markov model (HMM) application to examine the periods that yield return predictability. According to the results, it is observed that there is strong evidence for the validity of AMH within the scope of Borsa Istanbul’s BIST100. Additionally, the results of the HMM application confirm the periodic predictability regarding the determinants of the index.