Trends, Random Walks, and the Expectations-Augmented Phillips Curve Evidence from Six Countries
发现主要工业国家的实际经济活动时间序列存在随机游走型非平稳性，且通胀和货币增长意外变化的影响估计严重依赖于非平稳性的处理方法。
Two major empirical results are reported. First, it is shown that time series measuring real activity in major industrialized countries contain nonstationarities of the random walk type. Second, it turns out that the estimated influence of unexpected changes in inflation and monetary growth strongly depends on the chosen approach concerning nonstationarity. The popular assumption of a deterministic time trend proxying for growth generally yields the hypothesized positive coefficients. The influence of nominal surprises, however, becomes indistinguishable from zero if the respective equations are estimated in stationary differences of real activity variables. Copyright 1988 by Ohio State University Press.