This paper examines the efficacy of the general-to-specific modeling approach associated with the LSE school of econometrics using a simulation framework. A mechanical algorithm is developed which mimics some aspects of the search procedures used by LSE practitioners. The algorithm is tested using 1000 replications of each of nine regression models and a data set patterned after Lovell's (1983) study of data mining. The algorithm is assessed for its ability to recover the data-generating process. Monte Carlo estimates of the size and power of exclusion tests based on t-statistics for individual variables in the specification are also provided. The roles of alternative sizes for specification tests in the algorithm, the consequences of different signal-to-noise ratios, and strategies for reducing overparameterization are also investigated. The results are largely favorable to the general-tospecific approach. In particular, the size of exclusion tests remains close to the nominal size used in the algorithm despite extensive search.

We provide an accessible introduction to graph-theoretic methods for causal analysis. Building on the work of Swanson and Granger (Journal of the American Statistical Association, Vol. 92, pp. 357-367, 1997), and generalizing to a larger class of models, we show how to apply graph-theoretic methods to selecting the causal order for a structural vector autoregression (SVAR). We evaluate the PC (causal search) algorithm in a Monte Carlo study. The PC algorithm uses tests of conditional independence to select among the possible causal orders -or at least to reduce the admissible causal orders to a narrow equivalence class. Our findings suggest that graph-theoretic methods may prove to be a useful tool in the analysis of SVARs. I. The problem of causal orderDrawing on recent work on the graph-theoretic analysis of causality, we propose and evaluate a statistical procedure for identifying the contemporaneous causal order of a structural vector autoregression. *We thank Marcus Cuda for his help with programming and computational design, Derek Stimel and Ryan Brady for able research assistance, and to

We re-examine studies of cross-country growth regressions by Levine and , 1997b). In a realistic Monte Carlo experiment, their variants of Edward Leamer's extreme-bounds analysis are compared with a cross-sectional version of the general-to-specific search methodology associated with the LSE approach to econometrics. Levine and Renelt's method has low size and low power, while Sala-i-Martin's method has high size and high power. The general-to-specific methodology is shown to have a near nominal size and high power. Sala-i-Martin's method and the general-to-specific method are then applied to the actual data from Sala-i-Martin's original study.

No abstract

We thank Oscar Jorda, Judith Giles, Wayne Joerding, and participants in seminars at the University of California, Irvine, and the University of Victoria for helpful comments on an earlier draft. We also thank Orley Ashenfelter for his help in getting this project off the ground. AbstractThe work of Levine and Renelt (1992) and Sala-i-Martin (1997a, b) which attempted to test the robustness of various determinants of growth rates of per capita GDP among countries using two variants of Edward Leamer's extreme-bounds analysis is reexamined. In a realistic Monte Carlo experiment in which the universe of potential determinants is drawn from those in Levine and Renelt's study, both versions of the extreme-bounds analysis are evaluated for their ability to recover the true specification. Levine and Renelt's method is shown to have low size and extremely low power: nothing is robust; while Sala-i-Martin's method is shown to have high size and high power: it is undiscriminating. Both methods are compared to a cross-sectional version of the generalto-specific search methodology associated with the LSE approach to econometrics. It is shown to have size near nominal size and high power. Sala-i-Martin's method and the general-to-specific method are then applied to the actual data from the original two studies. The results are consistent with the Monte Carlo results and are suggestive that the factors that most affect differences of growth rates are ones that are beyond the control of policymakers. JEL Classification: C4, C8, O4 Truth and Robustness in Cross-country Growth RegressionsNo. Country