Sex workers draw a premium for engaging in unprotected sex. We theoretically motivate a test of whether this premium represents a compensating differential for disease, thereby mitigating sex workers' propensity to use condoms. Using transaction-level data and biological STI markers from sex workers in Ecuador, we exploit within-worker variation across local disease environments. We find that locations with low disease prevalence exhibit a very low, insignificant premium for unprotected sex. A one percentage point increase in the local disease rate increases the premium for noncondom sex by 33 percent. Market forces may curb the self-limiting nature of STI epidemics. (Contains 9 footnotes, 2 figures, and 6 tables.)