Problems in the Use of Historical Data in Estimating Economic Loss in Wrongful Death and Injury Cases: Comment
评论Hosek关于在过失致死和伤害案件中使用历史数据预测未来收入增长率和利率的论文，指出历史时间序列的非平稳性问题，并质疑依赖历史平均值和关系的做法。
In a recent issue of this Journal, Hosek states that the purpose of his paper is to question [the] common usage of historical data in making projections or extrapolations about future earnings growth paths and future interest rates. '1 For past data to be used to project future interest and earning growth rates, the historical time series must be stationary -invariant with respect to time. Autocorrelation functions 'for levels of interest rates and for percent changes of wage rates are presented. Hosek finds that these autocorrelation functions . tend to behave more like those for nonstationary series rather than stationary series. 2 His major conclusion . is that the uncritical use of historical averages and relationships provides a weak basis for estimating future growth rates in income and future rates of interest .3 Further, he states that the nonstationarity problem is apparent in the simplified approach which relies on the close historical relationship between nominal earnings increases and nominal interest rates to calculate future lost earnings.4 Hosek transforms the nominal wage and interest rate time series data into ordinary first differences and states that their autocorrelation functions look like those for stationary time series. I Box-Pierce Q statistics are also calculated in order to test, at the 5% level, the hypothesis that autocorre-