Does regime type really matter for economic growth?

This article reports on an attempt to duplicate the results of "Regime, Polity, and Economic Growth: The Latin American Experience," published in the Winter 1995 issue of Growth and Change. The original article used the Parks method to analyze cross section time series data, and concluded that after controlling for other relevant variables military governments have lower rates of economic growth than their civilian counterparts. Using data provided by the original researcher, and using various regression techniques including the Parks method, we were not able to reproduce the original results. We attribute this to problems in both the original data and regression techniques.