Will the “True” Productivity Series Please Stand Up? Identifying Inconsistencies Across Productivity Series

Abstract

We conduct a systematic empirical investigation into the consistency of competing industry-level total factor productivity (TFP) series in the United States, including the crop and animal production or farm sector. We apply an algorithm based on time-series methods to three widely-used databases frequently utilized by researchers to study productivity. Our analysis reveals a substantial proportion of inconsistent industry-level TFP series across these databases, including those for the farm sector. We further investigate the potential causes of such discrepancies by examining the consistency of the individual output and input variables used in the calculation of the TFP series. Individual input and output series pairs show different patterns of inconsistency across databases. Our findings highlight the need for publishing agencies and researchers to explore opportunities for methodological standardization.