Stock Market Forecastability and Volatility: A Statistical Appraisal
提出并实施了新的统计检验方法，避免了早期检验的严重统计问题，发现市场效率假说虽被拒绝但仅边缘显著，且回归检验在考虑有限样本偏差后未能提供传统估值模型被违反的有力证据。
This paper presents and implements statistical tests of stock-market forecastability and volatility that are immune from the severe statistical problems of earlier tests. It finds that although the null hypothesis of market efficiency is rejected, the rejections are only marginal. The paper also shows how volatility tests and recent regression tests are closely related, and demonstrates that when finite sample biases are taken into account, regression tests also fail to provide strong evidence of violations of the conventional valuation model. Copyright 1991 by The Review of Economic Studies Limited.