A History of Commercially Available Risk Models

Multi-factor risk models have been used in portfolio selection since the 1960s and early 1970s. The work of Barr Rosenberg led to the creation of Barra, the first major commercially available portfolio selection software. Rudd and Clasing brought Barra to the academic audience as professional managers were embracing it in industry. It is important to see how quantitative analysis was developed in investment research and analysis. Harry Markowitz, Bill Sharpe, Jan Mossin, and Jack Treynor pioneered capital market equilibrium and the creation and estimation of the Capital Asset Pricing Model. In the 1970s, multi-factor risk models were developed and estimated by Barr Rosenberg, Andrew Rudd, and their colleagues at Barra; John Blin and Steve Bender at APT; and Sebastian Ceria at Axioma. Multi-factor models pushed out the efficient frontier relative to a single factor risk model. McKinley Capital Management developed its Horse Race risk model philosophy to assess the most effective risk models for maximizing the Geometric Means, Sharpe Ratios, and Information Ratios. Several special issues of practitioner-oriented applied investment research featured the McKinley Horse Race results and reported that properly constructed multi-factor risk models created portfolios in which tilt variables that have statistically significant ICs produced statistically significant Active Returns. One must remove the market effect and extra-market covariances to properly estimate the contribution of a variable to the creation of efficient portfolios. APT and Axioma have reigned supreme in those tests. The integration of Axioma and AAF and ITG transactions costs on the FactSet Investment have led MCM to implement an Axioma portfolio.