Controlling errors in allowance program administration

Analyzes error rates and error control methods in administering the experimental housing allowance program, based on records of the housing allowance offices (HAOs), which enroll clients, evaluate their dwellings, and disburse allowance payments. Interview protocols and documentation requirements help prevent errors in data on client income and household status. Most errors made are later corrected by reviewing enrollment forms and verifying a sample of undocumented information with third parties. Housing evaluators incorrectly pass or fail a dwelling in less than 2 percent of the cases, and independent auditors have found only negligible errors in HAO records. The net error rate in both experimental sites is estimated at less than one percent of the average annual allowance payment. Few clients complain about the rigorous error control methods. Further research will examine whether their administrative cost (now somewhat higher than the amount they save) can be reduced without significant increases in error rates.