AN EMPIRICAL STUDY OF THE MEAN-VARIANCE AND MEAN-ABSOLUTE DEVIATION PORTFOLIO OPTIMIZATION MODELS

Abstract

In this project, we study various portfolio optimization models under the modern portfolio theory. We begin by tracing out the general formulations of mean-risk models. Following which, two mean-risk models, namely the mean-variance model and the mean-absolute de viation model are discussed in detail. The two models are further investigated empirically on real historical data from two global indices, CNX-100 (India) and DAX-100 (Germany) over a period of five years. Extensive computational results are provided to compare the financial performance of the optimal portfolios generated from the two models along with the equally weighted naive portfolio. We observe that the Markowitz mean-variance model records better out-of-sample performance in terms of risk measures as well as performance ratios.