Total expenditures on unemployment benefits in the United States have increased considerably during the recent recession, from $54 billion in 2008 to $108 billion in 2011 (both measured in 2011 dollars). These higher expenditures have brought more attention to the unemployment insurance system. As a result, fraud has been in the news recently. For instance, the Arizona Department of Economic Security discovered that several prisoners had received unemployment benefits, and similar cases were reported in Indiana. 1 This essay documents the recent data on unemployment insurance fraud in the United States and suggests that the overpayments due to these headline-grabbing types of fraud are small relative to the overpayments due to other types of fraud.
Unemployment insurance fraud occurs whenever an ineligible individual collects benefits after intentionally misreporting his or her eligibility. The table documents the total overpayments due to fraud from 2008 to 2011. Over this period, the average annual overpayment was $2.9 billion.
There are, however, several categories of fraud that should be taken into account. For instance, to be eligible to collect unemployment benefits, an individual must be able and available to work; thus, as implied above, prisoners are not eligible. Payments to such individuals belong to the fraud category of "Unable and Unavailable to Work." The table also documents the overpayments due to this fraud category. During 2008-11, such overpayments averaged only $152 million annually and amounted to roughly 5 percent of total overpayments due to all fraud.
During the same period, total expenditures on unemployment benefits (regular state benefits + extended benefits) averaged $110.8 billion. Overpayments due to all fraud are thus less than 3 percent of the total expenditures of the unemployment insurance program. Overpayments due to Unable and Unavailable to Work fraud, specifically, account for roughly just one-tenth of 1 percent of the unemployment insurance program's expenditures.
For a bigger picture, the chart breaks down overpayments into three types of fraud for 2008-11. "Concealed Earnings" refers to payments to individuals who were collecting unemployment benefits while they were employed and earning wages. The overpayments due to Concealed Earnings fraud account for almost two-thirds of the total overpayments due to all fraud.
In addition to being able and available to work, individuals are eligible for benefits only if they lose their job through no fault of their own. Overpayments to individuals who quit their job or who were fired because of poor performance account for 16 percent of total overpayments due to fraud. Although it is smaller than the amount due to the Concealed Earnings category, it is more than three times the amount due to the Unable and Unavailable to Work category.
Although fraud under the Unable and Unavailable to Work category has received recent attention, it accounts for a small fraction of total fraud. Again, the dominant category is Concealed Earnings. However, detecting Concealed Earnings fraud turns out to be costly. Less than 15 percent of cases are detected by state agencies via simple verification schemes such as cross-checking Social Security numbers of those collecting unemployment benefits with the state's directory of new hires and wage records.
Fuller, Ravikumar, and Zhang (2012) provide a mechanism that deters Concealed Earnings fraud by using two instruments: monitoring and benefits/taxes. In their mechanism, it is optimal to incur the monitoring cost and periodically verify the employment status of those collecting benefits. In between consecutive monitoring periods, it is also optimal to (i) gradually reduce the unemployment benefits for those who remain unemployed and (ii) levy a lower income tax on those who truthfully and quickly report transition back to gainful employment relative to those who are late in reporting the transition back to employment.