Behavioral evaluation of Fama-French factor model-based portfolios: A hybrid model application in Borsa Istanbul

Fama (1970), who proposed market efficiency on the assumption that people are rational, provided an innovative perspective on the field of finance by claiming that all information is now available on the market and therefore market players are unable to earn abnormal returns. his view is also supported by empirical studies. However, researches that imply there are anomalies in the markets, that it is impossible to reflect all information to the market simultaneously, and that investors can receive abnormal returns come to the fore in addition to research that support these claims. At the same time, it becomes apparent that the investor is acting irrationally rather than rationally. Investor behavior has been looked at from all angles while these notions are analyzed within the Behavioral Finance theory framework. These research' major focus is on cognitive biases and heuristics, and they are guided by sociological and psychological data. Aside from their rational decision-making behavior, people also like to make judgments with distinct views in the area of Islamic finance. The subject of Islamic Finance, which is governed by Islamic Law, is how investors' religious convictions affect their financial decisions. It is intended to identify which of the five factor proposed by Fama and French (2015) proves effective in which sort of portfolios within the scope of all this theoretical infrastructure. A hybrid model built using Artificial Neural Networks and Genetic Algorithm optimization is used to assess the given data. Possible distinctions between portfolios made from a conventional and an Islamic perspective are highlighted in this context. Through portfolios built using the quarterly data of the 159 companies in the Borsa Istanbul Whole Index and the 29 companies in the Borsa Istanbul Participation 50 Index for the period between 2014 and 2021, the validity of the Fama and French Five-Factor Models in both indices is first tested for this purpose. As a consequence of the examination, potential variations in the factors impacting the portfolios formed from the traditional and Islamic viewpoints have been identified in the Panel Data Analyses tests carried out with the Fama and French Five-Factor model. The diversity of effective gene parameters in portfolios constructed with various viewpoints also suggests differentiation, according to the assessment performed using the hybrid model based on Artificial Neural Networks and Genetic Algorithm optimization. As consequently, the behavioral aspects of portfolio difference between traditional and Islamic approaches have been examined extensively. Furthermore, it has been found that the hybrid model based on artificial neural networks produced by genetic algorithm optimization within the context of the two applicable models produces outputs that are more focused.