Population density and the accuracy of the land valuations in the 1798 federal direct tax

Abstract

The peculiar operation of the 1798 federal direct tax has led scholars to question whether tax officials reported the land valuations from their districts faithfully. Peter Lindert and Jeffrey Williamson argue that southern tax assessors systemically under reported the value of southern real estate, and they adjust their income estimates to account for the likelihood of corruption. This paper affirms the reliability of the tax returns by demonstrating that population density, rather than corruption or lax enforcement, can explain nearly all of the variation between the assessment districts. Accepting the tax valuations as accurate would lower Lindert and Williamson’s income estimates, imply slower growth rate between 1774 and 1800, and suggest a higher growth rate between 1800 and 1850.