A regulatory agency charged with enforcing compliance with environmental and safety regulations but constrained by scarce resources should inspect facilities that pose the highest risk of non-compliance while considering the time required to detect a serious violation.  In most cases, however, agencies do not incorporate either measures of risk or estimates of time into their inspection decisions, or if they do consider these factors, they often use simple methods or rules of thumb.  This paper presents a general model of risk-based inspections and applies the model to the Gulf of Mexico, where the Department of Interior is responsible for inspecting 2,500 oil and gas production facilities for compliance with environmental and safety standards.  We use data from inspections in 2011 to develop a time-weighted risk-reduction score for each facility.  In particular, for each facility, we calculate risk based on the probability of detecting a shut-in violation (estimated based on facility and operator characteristics, using logistic regression) and the consequence of a violation (as measured by the production volume of the facility).  We then estimate the time required to inspect each facility (based on facility characteristics, using OLS regression), and we calculate time budgets (based on actual inspection time per inspector).  We rank facilities using these scores and we simulate the number of serious violations that would have been detected in 2012 using both risk-based targeting and conventional targeting methods.  We find that the risk-based targeting leads to significantly more detections of serious non-compliance than conventional methods.  As Department of Interior hires additional inspectors, under new authority it received from Congress after the Deepwater Horizon incident, this paper demonstrates that using a risk-based method to deploy these inspectors could result in significantly lower overall risk as compared to conventional methods.  The results also demonstrate that other agencies could also use risk-based inspections to maximize the impact of inspections despite scarce resources and limited budgets.