This paper evaluates the predictability of monthly stock return using out-of-sample interval forecasts. Past studies exclusively use point forecasts, which are of limited value since they carry no information about intrinsic predictive uncertainty. We compare the empirical performance of alternative interval forecasts for stock return generated from a naïve model, univariate autoregressive model, and multivariate model (predictive regression and VAR), using U.S. data from 1926. It is found that neither univariate nor multivariate interval forecasts outperform naïve forecasts. This strongly suggests that the U.S. stock market has been informationally efficient in the weak form as well as in the semistrong form.

中文翻译：

---

股票收益可预测性：基于区间预测的评估

本文使用样本外区间预测评估每月股票收益的可预测性。过去的研究只使用点预测，因为它们不携带有关内在预测不确定性的信息，所以价值有限。我们使用 1926 年的美国数据比较了从幼稚模型、单变量自回归模型和多变量模型（预测回归和 VAR）生成的股票收益的替代区间预测的经验表现。发现单变量和多变量区间预测的表现都不优于幼稚模型预测。这有力地表明美国股票市场在弱形式和半强形式下都是信息有效的。