Office of the Federal Detention Trustee's (OFDT) Cost Estimation Methods Reflect Features of Best Practices, but Processes Could be Enhanced

This letter formally transmits a summary of an oral briefing we gave to Congress on August 30, 2010 and subsequent agency comments. We gave this briefing in response to the Conference Report accompanying the Consolidated Appropriations Act, 2010 (H.R. Rep. No. 111-366, at 662 (2009) (Conf. Rep.)), which directed us to evaluate the methods the Office of the Federal Detention Trustee (OFDT) uses to project the federal detainee population and the per diem rates OFDT pays for detention services, as well as other factors that are translated into OFDT's annual budget request. In conducting this work, we analyzed OFDT's documentation of its detainee population and cost estimation model, compared OFDT's projections in the President's budget submission with actual values of detainee population and costs, and interviewed OFDT and Department of Justice (DOJ) officials to (1) assess how OFDT develops its detainee population and cost estimates; (2) determine the extent to which OFDT follows best practices for developing high-quality cost estimates; and (3) determine the reasons why OFDT required funds in addition to its annual appropriation in fiscal years 2005, 2008, and 2009.

In summary, OFDT develops its annual budget request for DOJ using three general steps: (1) estimating cost increases needed to maintain current service levels, (2) projecting the changes in the detainee population, and (3) estimating the costs associated with detainee population and other program changes. OFDT's process for estimating the federal detainee population and costs reflects some of the features of GAO's best practices for cost estimation, but OFDT has not adopted enough of the practices to meet the four characteristics that we associate with high-quality cost estimates. Specifically, we determined that OFDT's cost estimation processes substantially met the features that characterize accurate and comprehensive cost estimates and partially met the features that characterize welldocumented and credible cost estimates. OFDT documents the purpose, formulas, and data sources behind its cost estimates, but many of the technical aspects of OFDT's cost estimation model are fully understood only by the individual who runs the model. Further, the President's annual budget request to Congress for OFDT has begun to include estimates of how an increase in the number of individuals charged with drug, weapons, or immigration offenses may affect program costs, but, as a best practice, OFDT also could perform analyses that account for the risks and uncertainties in the model's assumptions and estimates. OFDT and DOJ officials attributed OFDT's need for additional funding to several factors, including unanticipated increases in the detainee population, higher per diem rates resulting from limited available detention bed space, and limited financial reserves at the beginning of the fiscal year. More specifically, law enforcement initiatives related to immigration have significantly increased the total detainee population. OFDT officials stated that it is difficult to incorporate such initiatives into its detainee population model because OFDT, as part of the federal budget process, must submit its projections for a fiscal year 18-24 months before that fiscal year and the law enforcement initiative begin. In addition, since OFDT has limited flexibility in how it spends its funds and its annual costs exceed $1 billion, even a small difference between projected and actual costs could result in OFDT facing costs that exceed its appropriated funds by tens of millions of dollars. As a result of our findings, we are recommending that OFDT improve its documentation of the calculations it uses in its detention population and cost estimation model. We are also recommending that OFDT quantify the extent to which costs could vary due to the inherent risks in key assumptions, data inputs, and other factors.