Stationary Components in Stock Prices: An Exact Pointwise Most Powerful Invariant Test
开发了一种精确的小样本点最优势不变检验，用于判断股票价格是否包含平稳且可预测的成分。检验结果一致表明，股票价格在所有样本期内（包括或不包括1930年代高波动年份）都存在一个驼峰形、缓慢均值回归的平稳成分，并解决了文献中的三个难题。
This article develops an exact small-sample, pointwise most powerful invariant test to determine whether stock prices contain a stationary and therefore predictable component. This test generates consistent evidence that stock prices contain a hump-shaped, slowly trend-reverting stationary component over all sample periods tested, including and excluding the high return variance years of the 1930s. The empirical evidence in this article addresses three prominent puzzles in this literature—the negative and positive autocorrelations found in stock returns, the role of the 1930s, and the very low reported power of previous statistical tests that find a stationary component.