ABSTRACT
Data quality in the Penn World Tables varies systematically across countries that have different growth rates and at different stages of development, thus introducing measurement error correlated with variables of economic interest. We explore the seriousness of this problem with three examples from the literature, showing that the problem appears to be minor in growth convergence regressions but serious in estimating the effect of growth volatility on the average growth rate and in a cross-country test of the Permanent Income Hypothesis. The results suggest, at the very least, a need for performing appropriate sensitivity tests before drawing conclusions from analyses based on these date.