A1033
Title: Portfolio allocation: The advantage of using network approach
Authors: Fabio Vanni - University of Insubria (Italy) [presenting]
Abstract: The classical problem of portfolio selection follows a risk-return optimization approach and capturing the dependency structure between different asset returns is the main issue for a manager. The first portfolio selection model, proposed by Markowitz, uses the variance/covariance matrix in order to measure the dependence structure of the asset returns, which usually is estimated using the sample approach. Considering the drawback of the sample estimation, different robust estimators have been used to model the dependency structure. Recently, different works have focused on a new way of modeling the dependency between returns of different assets by means of the so-called market graph. The portfolio allocation problem is reformulated using the network theory. In particular, the random matrix theory (RMT) is used to analyze cross-correlation in financial time series. A comparison between the empirical correlation and the random matrices will be performed in order to identify non-random properties and underlying interactions. Moreover, deviations from RMT predictions will be analyzed in order to gain some insights into the statistical structure of multivariate financial data.
