A Monte Carlo Study for Pooling Time Series of Cross-Section Data in the Simultaneous Equations Model
通过蒙特卡洛实验研究联立方程模型中多种合并估计量的小样本表现，旨在为应用研究者提供合并时间序列与截面数据的操作指南。
Empirical studies utilizing time-series of cross-section data are constantly appearing in virtually every field of economics. This has been made easier by the increasing availability of panel data, and the increasing capability of computers in handling large data sets. Some of the earlier studies include Mundlak [1961, 1963] and Hoch [1962] in the production function literature, Kuh [1959] in the investment function literature, and Balestra and Nerlove [1966] in the energy literature. More recent studies include Chamberlin and Griliches [1975], Hausman and Wise [1979], Lillard and Weiss [1979] in the labor literature, and Griffin [1979] and Pindyck [1980] in the energy literature, to mention only a few. Efforts to develop the econometric theory for pooling time-series of cross-section data have concentrated largely on the single equation error components model.2 More recently, Avery [1977] and Baltagi [1981b] extended this error components literature to the seemingly unrelated regressions and the simultaneous equations model, respectively. Asymptotic as well as small sample properties of various pooling estimators received adequate investigation in the single equation model.3 However, the same cannot be said about pooling estimators in the simultaneous equations case. This paper is an attempt to remedy this situation. The small sample performance of various pooling estimators in a two-equation simultaneous model are studied by means of Monte Carlo experiments. The main purpose is to provide the applied researcher with some guidelines on how to pool time-series of crosssection data in the simultaneous model. Monte Carlo studies for the classical simultaneous model are plentiful,4 as are