The Role of Ancillarity in Inference for Non-Stationary Variables
比较了回归方法与基于似然的推断，指出通过S-辅助性或强外生性概念，非遍历过程的渐近推断可与遍历过程相同方式处理，对统计学家和计量经济学家有参考价值。
Some examples of the regression method are compared with likelihood based inference. It is shown that although the asymptotic theory is distinctly different for ergodic and non-ergodic processes, the likelihood methods lead to the result that asymptotic inference can be conducted in the same way for the two cases by appealing to classical conditioning arguments from statistics using the notion of S-ancillarity or strong exogeneity. It is pointed out that the Fisher information can be considered a measure of the conditional variance of the maximum likelihood estimator given the available information in the sample.