A dynamic, globally diversified, index neutral synthetic asset allocation strategy

• 成果类型: Article
• 署名作者: Novomestky, F
• 刊物名称: MANAGEMENT SCIENCE
• ISSN/ISSBN: 0025-1909
• DOI: 10.1287/mnsc.43.7.998
• 发表日期: 1997
• 页码: 998-1016
• 关键词: Bayesian estimation; dynamic seemingly unrelated regression; PORTFOLIO SELECTION; Parameter Estimation; derivatives market; Markowitz mean-variance optimization; asset allocation; Capital asset pricing model; GLOBAL DIVERSIFICATION

摘要

An investor with the ability to assess the prospective return and risk structure of the global capital markets can construct portfolios that, over time, will not only outperform actively or passively managed domestic asset portfolios but will also outperform passively managed global portfolios. A Bayesian approach to dynamic seemingly unrelated regression (DSUR) is a robust and effective means to forecast the one-step ahead, conditional distribution of asset returns. This approach recognizes the time-varying nature of the global capital markets. The predictive moments are used to derive a single-factor return model once an index portfolio is specified. The index portfolio represents an investor's underlying portfolio. Given the factor model that assesses the relative attractiveness and risk of the assets in relation to the index, an index neutral portfolio is constructed as an overlay to enhance the returns of the index portfolio. This portfolio is mean-variance optimal, is notional neutral (i.e., the sum of the asset exposures is zero), and has returns that are designed to be uncorrelated with the returns of the index portfolio. The implementation of such an index neutral portfolio using derivative securities is simulated over the period January 1988 to December 1993.