Portfolio Construction and Management in the Barra Aegis System: A Case Study using the USER Data

Abstract

In this article, we demonstrate the effectiveness of the Barra Aegis system and Barra factor model to construct portfolios representing popular and well-established investment management strategies and attribute the performance of these portfolios to the Barra factors. Stock selection models often use momentum, analysts’ expectations, and fundamental data. We employ composite modeling of expected returns using fundamental, expectations, and momentum-based data for U.S. equities during the December 1979 to December 2009 period. We use Barra factor models for portfolio construction and risk control. We report two results: (1) a composite model incorporating fundamental data, such as earnings, book value, cash flow, and sales, with analysts’ earnings forecast revisions and price momentum variables to identify mispriced securities; and (2) the returns to a risk-controlled portfolio allow us to reject the null hypothesis that results are due to data mining.