Data Types and Empirical Methods in 27 Top Economics Journals

This paper examines the relevance and frequency of econometric methods in 3,125 empirical papers published in 27 top economics journals during 2014-2018. There are three major findings. First, 93% of these empirical papers use microdata, while only 7% and 4% of them employ macrodata and time series data, respectively. Second, two-stage least squares and difference-in-differences are the two leading identification strategies, accounting for nearly half and a quarter, respectively, in empirical studies using microdata. Third, there is a clear divide in empirical studies using microdata and macrodata in terms of data collection, identification strategies and econometric methods.