On Critical Values of Tests against a Change in Persistence*
发现两篇文献对同一均值得分检验统计量给出了不同百分位数，差异源于样本模拟的系统性不同，并澄清了均值指数检验统计量的正确样本版本与临界值对应关系，同时修正了文献中的极限分布。
Abstract Kim, Belaire‐Franch and Amador [ Journal of Econometrics (2002) Vol. 109, pp. 389–392] and Busetti and Taylor [ Journal of Econometrics (2004) Vol. 123, pp. 33–66] present different percentiles for the same mean score test statistic. We find that the difference by a factor 0.6 is due to systematically different sample analogues. Furthermore, we clarify which sample versions of the mean‐exponential test statistic should be correctly used with which set of critical values. At the same time, we correct some of the limiting distributions found in the literature.