Electoral discrimination operates indirectly through clientelism targeting observable traits. This study finds skin color predicts vote buying across 7 countries with large visible populations of blacks and indigenous groups, despite race-neutral criteria disproportionately affecting dark-skinned voters. Using AmericasBarometer data (2010-2014) and conjoint experiments, it demonstrates a robust clientelism gap based on racial appearance. Wealth disparities partially explain this effect, but political engagement, trust levels, partisanship, and geography also vary significantly across nations like Peru, Bolivia, Colombia, Ecuador, Brazil, Guyana, and Suriname shaping the relationship differently. Importantly, even after accounting for these factors collectively – which include dark-skinned voters being more likely to receive offers – skin color remains an independent predictor of clientelism targeting.