Dados do Trabalho

Título

COVARIANCE PREDICTION IN LARGE PORTFOLIOS

Resumo

Many financial decisions such as portfolio allocation, risk management, option pricing and hedge strategies are based on the forecast of the conditional variances, covariances and correlations of financial returns. The paper shows an empirical comparison of several methods to predict one-step-ahead conditional covariance matrices. The methods are assessed in terms of out-of-sample minimum variance portfolio return measures on portfolios constructed by stocks from the S&P 500 index traded from January 2, 2000 to November 30, 2017. The results show that the DCC method using composite likelihood provides the best results according the standard deviation measure, the Risk Metrics 2006 method with non-linear shrinkage is the best in terms of information and Sortino's ratios and the Risk Metrics 1994 with non-linear shrinkage is the best according to an overall criterion.

Palavras-chave

Minimum variance portfolio, risk, shrinkage, S&P 500