A lower bound on the distribution function of the likelihood ratio test of portfolio efficiency is derived. An empirical application demonstrates that the bound may sometimes be used to infer rejection of the null hypothesis without appeal to asymptotic statistical approximations. A procedure for incorporating partial information about the zero‐beta intercept, in the multivariate framework, is also developed and applied.

This paper extends Kandel's [3] analysis of the testability of the mean‐variance efficiency of a market index when the return on some component of the index is not perfectly observable. In addition to information about the mean and variance of the missing asset, considered by Kandel, we explore the usefulness of information about the beta of the missing asset on the observed sub‐portfolio in an economy with a riskless asset. The results are somewhat more supportive of the notion that mean‐variance efficiency is testable on a subset of the assets.

This paper challenges the view that the Arbitrage Pricing Theory (APT) is inherently more susceptible to empirical verification than the Capital Asset Pricing Model (CAPM). The usual formulation of the testable implications of the APT is shown to be inadequate, as it precludes the very expected return differentials which the theory attempts to explain. A recent competitive‐equilibrium extension of the APT may be testable in principle. In order to implement such a test, however, observation of the return on the true market portfolio appears to be necessary.

Many claim that GAAP financial information has become largely irrelevant to explaining valuations. Core et al. compare financial information's value relevance for the New Economy stocks with other stocks. We supplement their analysis with new evidence on the economic determinants of the time-series variation in the coefficients mapping financial information into prices. We document significant variation in the coefficients related to proxies for changing market growth expectations and discount rates and additional variation consistent with time-varying correlated omitted variables. Such findings make it difficult to draw unambiguous inferences about the relevance and reliability of financial information from value-relevance regressions.

We find reliable evidence that both book-to-market (B M) and dividend yield track time-series variation in expected real stock returns over the period 1926 91 (in which B M is stronger) and the subperiod 1941–1991 (in which dividend yield is stronger). A Bayesian bootstrap procedure implies that an investor with prior belief 0.5 that expected returns on the equal-weighted index are never negative comes away from the full-period B/M evidence with posterior probability 0.08 for the hypothesis (0.14 with the impact of the 1933 outlier tempered). Although this raises doubts about market efficiency, the post-1940 evidence is consistent with expected returns always being positive.