Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
5069628 | Finance Research Letters | 2016 | 8 Pages |
Abstract
The description of the dynamic behavior of multiple time series represents an important point of departure to obtain accurate forecasts both in economic and financial analysis. We provide a method for the comparison of the out-of-sample performance of portfolios, respectively, ignoring and exploiting serial and cross dependence in stock returns. The serial and cross dependence is modeled using both the classical linear and easy-to-use Vector AutoRegressive and more sophisticated models making use of copula functions. After deriving the classical and copula-based VAR conditional expected returns and covariance, we construct different portfolios and compare them in terms of Sharpe ratio in an out-of-sample period.
Related Topics
Social Sciences and Humanities
Economics, Econometrics and Finance
Economics and Econometrics
Authors
Giorgia Rivieccio, Giovanni De Luca,