Returns, Risk, Portfolio Selection, and Evaluation

We present additional evidence on the risk and return of stocks in the USA and globally in the 1997–2009 period. We use a stock selection model incorporating fundamental data, momentum, and analysts’ expectations and create portfolios using fundamental and statistically based risk models. We find additional evidence to support the use of multifactor models for portfolio construction and risk control. We created portfolios for the January 1997 to December 2009 period. We report three conclusions: (1) a stock selection model incorporating reported fundamental data, such as earnings, book value, cash flow, and sales, and analysts’ earnings forecasts and revisions and momentum can identify mispriced securities; (2) statistically based risk models produce a more effective return-to-risk portfolio measures than fundamentally based risk models; and (3) the global portfolio returns of the multifactor risk-controlled portfolio returns dominate USA-only portfolios.