Distinctively black names and mechanisms of discrimination: Evidence from the early 20th century

Abstract

What were the effects of having a distinctively African American-sounding name on educational attainment, occupation, income, marital status, and longevity in early 20th century America? How did those effects differ for people based on their phenotypical race/ethnicity? The findings of contemporary research have shown racialized names to be related to negative outcomes such as job interview callbacks, birth outcomes, and teacher expectations. Furthermore, previous research has shown that the consequences of race-specific names explain as much as 10% of the historical between-race mortality gap (Cook et al., 2016). Theoretically, we argue, names should have been less of a mechanism for racial discrimination in earlier eras of American history. Using a sibling comparison design and linked administrative records, we hypothesize that there was little racial discrimination based on people's names in early 20th century America. We find that men with more African American-sounding names do no worse (or better) with respect to education, wages, occupation, or longevity than their brothers with less African American-sounding names; this finding holds for white and black men. This does not imply the absence of race-based discrimination in early 20th century America. Instead, it implies that people in this era discriminated based on something other than names and the race implied by those names.