Do Race, Gender, and Age Differences Affect Manager-Employee Relations? An Analysis of Quits, Dismissals, and Promotions at a Large Retail Firm

Abstract

Using data from a large U.S. retail employer, we examine how demographic differences between manager and subordinate affect the subordinate’s rate of quits, dismissals, and promotions. We distinguish between two effects that demographic differences can produce: (1) the effects of dissimilarity per se, and (2) the effects of role breaking where the differences violate traditional social roles and status norms (e.g., non-whites managing whites). Our results suggest that both dissimilarity and role breaking can have statistically significant effects. Race: Dismissals and Promotions: Blacks and Hispanics with dissimilar managers are much more likely to be fired, and less likely to be promoted. We interpret these as dissimilarity effects. By contrast, white employees with non-white managers are less likely to be dismissed than whites with white managers, and more likely to be promoted. This suggests role breaking leads non-white managers to defer to white employees. Quits: While dissimilarity surprisingly has no effect on black quit rates, it does cause a moderate increase in Hispanic quits. Racial differences also cause a moderate rise in white quit rates. We expected the effect for whites to be larger because it is the sum of dissimilarity and role-breaking effects. Further analysis suggests the effect for whites is suppressed by pre-hire sorting; whites who dislike having non-white managers tend to avoid working for non-whites in the first place. Age: Age dissimilarity per se does not have effects. However, role breaking does; employees who are at least 20 percent older than their managers are much less likely to be dismissed, and more likely to be promoted. Gender: Gender differences have modest, adverse effects on all three employment outcomes.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 33 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.