Sons, Daughters, and Parents’ Labor Supply: New Evidence from Matched CPS Data

Abstract

In a 2002 article, Lundberg and Rose (LR) claimed that men’s wages and labor supply increase more in response to births of sons than to births of daughters. However, underreporting of daughters in the PSID introduces nonclassical measurement error in men’s fertility histories, which biases LR’s fi xed-eff ects estimates. Th is paper uses matched CPS data to replicate LR’s analysis. Like the PSID data, the CPS data have a panel structure where respondents can be matched across years. Unlike the PSID, however, the fertility data in the CPS do not exhibit any sex-related biases. Th e results yield no evidence that fathers increase wages or hours worked by greater amounts in response to sons than to daughters. More generally, the results show that nonclassical forms of measurement error in fertility have potentially serious implications for microeconomic analyses of household behavior.