Background
U.S. Immigration and Customs Enforcement (ICE) needed a way to forecast its detention bed space requirements, better manage its resources, and justify requests for additional staff, funding, and construction projects. This need required a contractor with expertise in advanced analysis and modeling techniques and a comprehensive knowledge of illegal alien immigration and detention trends.
Project Summary
Fentress began working with ICE in 2000 to develop a simulation model that forecasted the number of illegal alien detainees based upon the number of apprehensions, the number of agents, and detainee length of stay. The simulation model examined additional parameters, including nationality, offense type, and case type. Constructing the simulation model required extensive knowledge of two federal databases. One of the databases contained information on illegal alien apprehensions such as nationality, facility, status, case, and relevant dates; the other is an ICE operational database that tracks personnel apprehensions at various levels, including district, sector, station, and port of entry. Developing the parameters of the model required Fentress to engage, organize, and manipulate data from these databases in a short timeframe.
Results
The simulation model reduced the complex operations of ICE detention into a manageable quantitative system flow by focusing only on those detention outcomes that were most probable and most applicable to a majority of detainees. The model takes into account the links between arresting agents (e.g., border patrol agents, criminal investigators, and immigration inspectors) and detention needs, providing increased operational flexibility and responsiveness. The result was a model that is intuitively reasonable, highly accurate, and cost-effective. These benefits made the model appealing to senior decision-makers within the organization. The model was essential to ICE in producing credible and defensible budget requests in support of its detention programs and resource needs.