The U.S. Census Undercount of Native-Born Children: Estimates, Correlates, and Implications

Janna E. Johnson , University of Minnesota

With the approach of Census 2020, the census’ accuracy in measuring the country’s population has received increasing attention. Young children are known to be particularly poorly counted. I estimate undercount of native-born children age 0-5 using a variant of Demographic Analysis, a technique long employed by the Census Bureau to evaluate census measurement of the full population. Using PUMS data, rather than the full 100 percent counts used by Census, I show that infants are undercounted in the 1990 PUMS by 20 percent for non-blacks and 30 percent for blacks. I also compute undercount by state of birth, a statistic not reported by Census, and show rates vary widely by state and race. By adjusting the PUMS weights for the undercount of children and performing several empirical analyses using these undercount-adjusted weights, I demonstrate the implications the undercount of children has for the measurement of child poverty.

See extended abstract

 Presented in Session 20. Methods for Evaluating Population Programs