Mean estimates of annual earnings appear to be subject to relatively small levels of response error, however absolute differences indicate significant over- and underreporting at the individual level. This study investigates the existence of reporting bias patterns along gender lines in the context of optimism, social desirability bias and gender identity norms. This study uses the structure of the PSID (Panel Study of Income Dynamics) dataset, where either husband or wife completes the survey instrument on behalf of the household. Preliminary findings based on FE-IV estimations show that in a Mincer equation log real hourly income specification, the levels vary depending on the respondent. Although both genders report higher income than their spouses do on their behalf, the upward bias is larger and more significant for men.
|Publication status||Published - Jan 2018|