Analysis of the Impact of Capital-Specific Policies or Legislation: Comment
评论了传统计量模型在政策模拟中的局限，指出当经济结构发生变化时，模型仅提供有根据的猜测框架，并以20世纪60年代末至70年代初的预测错误和1971-73年工资价格控制为例说明。
The conventional use of econometric models for policy simulation experiments involves the solution of a fixed structure of the economy under different assumptions about the paths of the exogenous variables (policy instruments). The shortcoming of such models is that they generally offer little beyond a framework for educated guessing in situations where the structure of the economy is believed or known to have changed from the regime under which the model was estimated. We are all familiar with such shortcomings of the products of econometric research. Examples abound of systematic forecasting errors during the late sixties and early seventies of models constructed on data bases ending in the mid-sixties under the (implicit) assumption that inflationary expectations were constant. The structure of such models has undergone a rapid evolutionary change in the past decade as have the implied policy multipliers that are conventionally cited. Another example of the limitations of models in cases of structural change is the various phases of wage and price controls during 1971-73. Model simulators and forecasters resorted to ad hoc roxy-variable hypot eses that should in principle be tested simultaneously along ith the theoretical spe ification. To round out the analysis, we need to know how ensit ve the authors' findings might be to reasonable al ernative ways of representing he real nd financ al rates of eturn n individual assets.