In 1996, Congress added Section 287(g) to the Illegal Immigration Reform and Responsibility Act, authorizing local law enforcement agencies to enter into agreements with the Department of Homeland Security to enforce federal immigration laws. Since then, over 900 officers in 62 law enforcement departments have been trained and certified to enforce federal immigration laws through the 287(g) program. While most literature on immigration control focuses on policy creation at the national level, much less attention has been paid to the growing trend of subnational immigration legislation. In addition, relatively little attention has been given to what takes place between policy adoption and outcomes: the process of immigration policy implementation. This paper addresses 287(g) implementation in Nashville, Tennessee. Since its inception, 287(g) has been controversial. Numerous policy reports suggest that granting immigration enforcement authority to local officers encourages these officers to engage in racial profiling and harms public safety by making immigrants afraid to report crimes. The problem with these characterizations, however, is that they fail to acknowledge the variation in the types of agencies with 287(g) programs (police departments, sheriff's department, state troopers), and the types of 287(g) agreements in which agencies can enter (the task force model or jail model). For example, police officials most frequently criticize the 287(g) task force model, which trains field officers in immigration enforcement, but remain silent on the more commonly implemented jail model, where individuals are screened for immigration violations after they are arrested. This distinction is important because in some cases, the agency that screens individuals for immigration violations is not the agency that arrests them. In Nashville, for example, while the police department is the only agency that can arrest people in Davidson County, it is the Sheriff's Office that screens individuals for immigration violations. Drawing on interviews with employees in the Davidson County Sheriff's Office this paper seeks to address how local employees implement the 287(g) program. Theoretically, I draw from literature on bureaucracy that suggests that people on the front lines of policy implementation essentially make policy through their practices. How individuals implement policy depends on the preferences of actors in the policy arena, and the institutional characteristics of the bureaucracies charged with implementing the policy. I show that while local police have the discretion to avoid arresting immigrants for minor crimes, Sheriff Deputies possess no discretion in their authority to screen individuals for immigration violations, regardless of the severity of their alleged crime. In contrast to police officers, who encounter immigrants while they are in the community, once individuals are arrested, Sheriff Deputies see immigrants as "criminals" ready for processing. ..PAT.-Unpublished Manuscript [ABSTRACT FROM AUTHOR]