A Reexamination of the Rationality of the Livingston Price Expectations: Note
指出以往检验利文斯顿调查数据中价格预期合理性的研究存在误差项自相关和异方差问题，并采用更合适的数据和方法重新检验。
The Livingston survey data have been used often, although never entirely appropriately, to test the rationality of price expectations. John (1977) was the first to point out that the Livingston panel members' responses were properly viewed as 8 and 14-month forecasts sampled at six-month intervals. Brown and Maital (1981), using a technique suggested by Hansen and Hodrick (1980), investigated the rationality of the Livingston series while taking account of error term autocorrelation caused by the overlap of forecast intervals. Their analysis was flawed, however, by the use of Livingston's original published data rather than the more appropriate Carlson adjusted version.l A further difficulty is that the errors in panel mean forecasts are heteroskedastic due to year-to-year variation in the size of the panels and variance of panel responses.2 All previous tests of rationality have failed to properly account for at least one of these two features of the data. Here we remedy this deficiency.