Race Discrimination in Internet Advertising: Evidence From a Field Experiment

TL;DR

This study provides empirical evidence of racial bias in Meta's advertising platform, showing darker skin models are less favored, leading to higher costs for advertisers and reinforcing societal biases.

Contribution

It offers the first field experiment quantifying racial bias in online advertising and highlights how platform tools perpetuate these biases.

Findings

Darker skin models are penalized in ad engagement.

Advertisers spend 15.9% more to reach similar engagement with darker skin models.

Meta's budget tool favors lighter skin images, reinforcing bias.

Abstract

We present the results of an experiment documenting racial bias on Meta's Advertising Platform in Brazil and the United States. We find that darker skin complexions are penalized, leading to real economic consequences. For every \$1,000 an advertiser spends on ads with models with light-skin complexions, that advertiser would have to spend \$1,159 to achieve the same level of engagement using photos of darker skin complexion models. Meta's budget optimization tool reinforces these viewer biases. When pictures of models with light and dark complexions are allocated a shared budget, Meta funnels roughly 64\% of the budget towards photos featuring lighter skin complexions.