The dynamic retention model (DRM) is a model that has been used to assess the retention effects of changes to compensation in both the military and civilian contexts. The DRM is part of a larger class of models known as structural models of individual behavior in which—in our context—people make retention decisions under uncertainty during their careers and have unique or heterogeneous tastes. That is, the DRM recognizes that individuals are forward-looking when they make their decisions (meaning they account for possible future events, such as future pay raises and how decisions might affect future events) and that people differ (meaning that people who face identical opportunities might make different choices). In the DRM, individuals make retention decisions in each year over their career and choose the alternative in each period—to stay or leave—that maximizes their utility, meaning they make the optimal choice each year given the information and opportunities available to them. The advantage of this approach to modeling retention is that the approach has a solid foundation in the literature and in theories of how personnel make retention decisions over their careers. In particular, the DRM is formulated in terms of the parameters that underlie the retention decisionmaking process rather than on a specific compensation system and retention outcomes. When such a model is estimated with empirical data on individual retention decisions, it can be credibly used for counterfactual analyses, permitting assessments of alternative yet-to￾be-implemented compensation systems. Furthermore, it includes a capacity to perform simulations of the retention and cost effects of relevant policy changes, including policies that have yet to be implemented.