Behavioral Finance applied to Portfolio theory: An empirical challenge to Markowitz’ framework

Over the past decades, assumptions related to modern portfolio theory such as investors‟ rationality, market efficiency and Gaussian returns have been challenged, namely through the emergence of behavioral finance. This paper offers a literature review of these shortcomings and presents the alternatives given by behavioral portfolio theory. The main contribution of this paper relies on an empirical study, largely based on Bourachnikova et al. research (2016), that aims at presenting the comparison between Markowitz‟ framework and both the behavioral model (Shefrin and Statman, 2000) and its mental account extension (Das et al., 2010). Our objective is to determine whether or not the optimal behavioral portfolio is located on the efficient frontier (i.e. as understood by Markowitz) and, in the positive case, to discuss the relevant implications of such location in terms of risk. Based on 2500 simulations, defined by the daily historical prices of 387 stocks stemming from the S&P500 from 2005 to 2015 (Bloomberg), we find that in the case where an optimal solution exists (i.e. 60%), 60% are located on the efficient frontier. More importantly, this solution is almost always (i.e. 95%) located on the upper right side of this frontier, meaning that it is characterized by a high risk profile. We state that while typical standard investors would see this investment opportunity as too risky, a behavioral portfolio investor would rather see it as optimal. We argue that it is the consequence of the different approaches used to assess risk in the respective optimization program of the two models (i.e. standard deviation and safety first constraint).