This paper evaluates the specification errors of several empirical asset pricing models that have been developed as potential improvements on the CAPM. We use the methodology of Hansen and Jagannathan (J. Finance 51 (1997) 3), and the test assets are the 25 Fama-French (J. Financial Econom. 52 (1997) 557) equity portfolios sorted on size and book-to-market ratio, and the Treasury bill. We allow the parameters of each model's pricing kernel to fluctuate with the business cycle. While we cannot reject correct pricing for Campbell's (J. Political Econom. 104 (1996) 298) model, stability tests indicate that the parameters may not be stable. A robustness test also indicates that none of the models correctly price returns that are scaled by the term premium.

We document extreme bias and dispersion in the small-sample distributions of four standard regression-based tests of the expectations hypothesis of the term structure of interest rates. The biases arise because of the extreme persistence in short interest rates. We derive approximate analytic expressions for the biases under a simple first-order autoregressive data generating process for the short rate. We then conduct Monte Carlo experiments based on a bias-adjusted first-order autoregressive process for the short rate and for a more realistic bias-adjusted VAR-GARCH model incorporating the short rate and three term spreads. Conducting inference with the small-sample distributions of test statistics rather than with their asymptotic distributions provides a more consistent rejection of the expectations hypothesis. Plausible sources of measurement error in short and long yields do not salvage the expectations hypothesis.

Stocks with recent past high idiosyncratic volatility have low future average returns around the world. Across 23 developed markets, the difference in average returns between the extreme quintile portfolios sorted on idiosyncratic volatility is -1.31% per month, after controlling for world market, size, and value factors. The effect is individually significant in each G7 country. In the United States, we rule out explanations based on trading frictions, information dissemination, and higher moments. There is strong covariation in the low returns to high-idiosyncratic-volatility stocks across countries, suggesting that broad, not easily diversifiable factors lie behind this phenomenon.