Advances and Applications in Statistics
Volume 49, Issue 5, Pages 343 - 355
(November 2016) http://dx.doi.org/10.17654/AS049050343 |
|
PORTFOLIO SELECTION WITH CONDITIONAL COVARIANCE MATRIX AND NONLINEAR PROGRAMMING
Julio César MartÃnez Sánchez, David Sotres-Ramos and Martha Elva RamÃrez Guzmán
|
Abstract: This paper aims at selecting investment portfolios using the standard model, better known as the mean-variance (MV) model, against two portfolios with conditional covariance matrix: the first being the multivariate Exponential Weighted Moving Average (EWMA) model and the second is the Dynamic Conditional Correlation (DCC) model, the latter uses the univariate GARCH model for each series of returns. All the three models were applied to the main 15 assets of the Mexican Stock Exchange (BMV). To get the optimal portfolio, we use the method known as quadratic programming, which allows, unlike the method of Lagrange multipliers, to obtain positive weights. The results indicate that the use of the conditional covariance matrix reduces the portfolio risk by up to 47.78%, in the case of the Exponential Weighted Moving Averages model and 25.45%, in the case of Dynamic Conditional Correlation model, both compared with the unconditional covariance matrix model. Likewise, evaluation of portfolios shows an increase in performance in the long term, for both cases constructed by using conditional covariance matrix. |
Keywords and phrases: EWMA, DCC, MV, BMV. |
|
Number of Downloads: 541 | Number of Views: 1446 |
|