Working Paper: NBER ID: w11732
Authors: Howard Bodenhorn; Christopher S. Ruebeck
Abstract: Black is not always black. Subtle distinctions in skin tone translate into significant differences in outcomes. Data on more than 15,000 households interviewed during the 1860 federal census exhibit sharp differences in wealth holdings between white, mulatto, and black households in the urban South. We document these differences, investigate the relationships between wealth and the recorded household characteristics, and decompose the wealth gaps into treatment and characteristic effects. In addition to higher wealth holdings of white households as compared to free African-Americans in general, there are distinct differences between both the characteristics of and wealth of free mulatto and black households, whether male- or female-headed. While black-headed households' mean predicted log wealth was only 20% of white-headed households', mulatto-headed households' was nearly 50% that of whites'. The difference between light- and dark-complexion is highly significant in semi-log wealth regressions. In the decomposition of this wealth differential, treatment effects play a large role in explaining the wealth gap between all subpopulation pairs.
Keywords: colorism; African American wealth; nineteenth century; economic history
JEL Codes: N3; J7
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| lighter-complected individuals (J15) | wealth (D14) |
| black-headed households (R20) | wealth (D14) |
| mulatto-headed households (J79) | wealth (D14) |
| skin tone (J79) | treatment effects (C22) |
| treatment effects (C22) | wealth (D14) |
| skin tone (J79) | wealth (D14) |