Pursuant to a congressional request, GAO reviewed the extent to which individuals residing in long-term care facilities are improperly included as members of food stamp households, focusing on: (1) how many individuals were included as members of food stamp households while they were residing in long-term care facilities and the estimated value of the overpayments to those households; and (2) whether computer matching is a practical means for identifying such overpayments.

GAO noted that: (1) in the seven states GAO reviewed, GAO identified about 4,500 individuals who were potentially improperly included as members in households receiving food stamps while residing in long-term care facilities; (2) these households could have received an estimated $500,000 in food stamp overpayments during calendar year 1997; (3) these potential overpayments represented a very small percentage of the $8.5 billion in benefits distributed in the seven states during fiscal year 1997; (4) GAO has provided the states with its computer match results for their use in eliminating or recovering the overpayments; (5) in view of the relatively small amount of potential food stamp overpayments made to households that included residents of long-term care facilities compared to the cost of computer matching, routine computer matching may not be practical for all the states included in GAO's review; (6) none of the seven states GAO visited were using computer matching to identify such overpayments; (7) officials in California and Kansas, which had the smallest amount of potentially improper benefits--in one case less than $25,000 and in another about $1,800--said that computer matching for these types of overpayments would not be practical or cost-effective to them; and (8) officials in the remaining five states said they would assess the potential benefits of computer matching, either as a tool for routinely identifying overpayments or as a means for periodically evaluating the effectiveness of procedures used to prevent such overpayments.