The introduction of technology aimed at reducing the response times of emergency medical services has been one of the principal innovations in crisis care over the last several decades. These substantial investments have typically been justified by an assumed link between shorter response times and improved health outcomes. However, current medical research does not generally show a relationship between response time and mortality. In this study, we explain the discrepancy between conventional wisdom and mortality; existing medical research fails to account for the endogeneity of incident severity and response times. Analyzing detailed call-level information from the state of Utah's Bureau of Emergency Medical Services, we measure the impact of response time on mortality and hospital utilization using the distance of the incident from the nearest EMS agency headquarters as an instrument for response time. We find that response times significantly affect mortality and the likelihood of being admitted to the hospital, but not procedures or utilization within the hospital. Copyright © 2012 John Wiley & Sons, Ltd.