Which Economic Freedoms Contribute to Growth? Reply

A number of studies have recently tested the connection between the degree of a nation’s economic freedom and its economic performance based upon a single, aggregated index of economic freedom derived from a weighted summation of several measures of underlying components of economic freedom. In a recent article in this journal, we argued that existing aggregation procedures are inappropriate since they assume all underlying freedom components are positively correlated with growth (Heckelman and Stroup 2000). We have shown this is not a correct statistical assumption through bivariate and multivariate analysis of the underlying components of the most typically used freedom indexes from Gwartney, Lawson and Block (1996, hereafter referred to as GLB). Subsequently, Sturm, Leetouwer and de Haan (hereafter referred to as SLH) have criticized our methodology on several grounds. SLH agree with our main point that the standard weighting procedures are ad-hoc and potentially misspecified. Specifically, we were concerned that any weighting procedure which forced all the components to positively contribute to the aggregate freedom index (and therefore positively contribute to a country’s predicted economic growth) would yield a biased estimate of the aggregate freedom index-economic growth relationship. As such, the lack of significance when we regressed economic growth on the GLB aggregated freedom index was not surprising. Our stated goal was to develop a more statistically sound methodology for constructing an aggregate economic freedom index that would be meaningful for predicting economic growth. The procedure we developed was based on hedonic regression analysis, where economic growth was regressed directly on each of the GLB freedom components to determine the weighting of each com-