infonet-economy

We present a new methodology for decomposing the (risk-adjusted) performance in empirical finance. Our technique offers the same straightforward economic intuition and all the statistical benefits of the portfolio sorts approach, in particular robustness to cross-sectional correlation, and in addition resolves the major drawbacks of portfolio sorts. Most importantly, our regression-based methodology handles multiple dimensions and continuous firm, fund, or investor characteristics. Moreover, the technique allows for relying on standard Wald-tests as an alternative to the popular Gibbons, Ross, and Shanken (1989) test. We illustrate our methodology with an asset pricing application and a long-horizon event study.