XXV Edition

1-2 December 2016"

Portfolio Selection with Mental Accounts. An Alternative Approach Combining Loss Aversion and Hyperbolic Discounting

Martelli Duccio, University of Perugia
Burchi Alberto, University of Perugia
Ghisellini Fabrizio, Italian Ministry of Economy and Finance

The identification of a risk measure coherent with the investor behavior has always been a prominent theme in the asset management industry, since standard deviation does not seem to be an appropriate risk proxy. Compared to other papers present in literature, our study introduces two major elements: i) the use of different risk measures which take into account the real behavior of loss-averse investors; ii) in a context of adaptive expectations, the way in which hyperbolic discounting affects mental sub-portfolios which are characterized by different time horizons. The aim of the paper is threefold. Firstly, to identify risk measures which take into account the investors behavior in practice. Secondly, to estimate the impacts that hyperbolic discounting has on portfolio choices. Finally, to jointly evaluate the two effects, in order to identify a portfolio model consistent with the real preferences of individual investors. To achieve these goals, different methodological approaches need to be used. In particular, we compare the effects of different optimization models, using alternative risk indicators, such as the maximum drawdown, the time required to recover the loss and how often adverse events occur. Moreover, the use of hyperbolic discounting allows the identification of target returns consistent with the real investor preferences. Preliminary results show that optimal portfolios differ significantly from previous evidence present in literature. This is mainly due to the fact that our models try to take into account the real perception of risk and return as formulated by investors and described in the behavioral portfolio theory.

Area: Risk management

Keywords: Investor behavior; Mental accounting; Loss aversion; Hyperbolic discounting

Paper file

University Network