Taking a Bite Out of Illegal Food Stamp Traffic

It's sad but true—when millions of dollars change hands, you can bet that a criminal element will try to get a piece of the pie. This phenomenon is manifest even in programs designed to benefit needy and impoverished people.

By visualizing data on a map, state employees can quickly spot anomalies in the data so they can make informed decisions.

In the state of Louisiana, more than 600,000 people receive food stamps each year. This represents approximately $600 million in United States federal funding. Unfortunately, according to the United States Department of Agriculture, at least 4 percent of those funds are fraudulently received or trafficked. To help reduce this illegal activity, the state of Louisiana Department of Social Services deployed technology from Information Builders and Esri to identify patterns that reveal fraudulent activity.

"We are on the lookout for people who are taking advantage of the system," said Raymond Pease, assistant director of the Fraud and Recovery Section of the Louisiana Department of Social Services, Office of Family Support. "Our agency needed a powerful business intelligence tool to analyze trends among millions of transactions. Other reporting tools could not accurately handle our requirements. Information Builders' WebFOCUS delivers results above and beyond our expectations."

Trafficking in food stamp benefits is also known as discounting. Typically, food stamp recipients sell their benefits to dishonest grocers for cash rather than food. These grocers pay recipients considerably less than the face value of the stamps and pocket the difference. Although detecting patterns of fraudulent activity involving millions of transactions occurring over a large geographic area might be difficult using other methods, it is an ideal problem for GIS.

The Louisiana Department of Social Services uses ArcIMS to add GIS capabilities to WebFOCUS reports. The Department of Social Services can input data from its mainframe databases, transform this data into valuable business information, and interpret complex relationships that might otherwise be difficult to detect. "The beauty of this technology is how it can visually display report data on a map," said Pease. "In our case, it makes it much easier to detect trends that indicate fraudulent activity. The data speaks very loudly and clearly with this system."

Spotting Trends

Years ago, recipients of food aid were mailed food stamp coupons. In 1997, the Louisiana Department of Social Services adopted an electronic system for delivering and tracking food stamp benefits in response to federal mandates for Electronic Benefit Transfer (EBT) technology.

"Increased state involvement in the investigation and prosecution of food stamp traffickers was one of the things that motivated the change to EBT," Pease explained. "EBT was a catalyst for us to increase the efficiency of the information systems we use in fraud detection as we work with various law enforcement agencies to investigate suspected recipients and retailers."

Better reporting technology went hand in hand with the change to EBT. Previously, the department used a Paradox database from Borland Software to generate reports, but this relational database environment lacked the business intelligence capabilities needed for effective data analysis. "Our old system was not providing us with the capabilities we needed to do a first-rate job," said Sherwood Lemoine, an internal management consultant in the IT department at the Louisiana Department of Social Services. "When we looked at WebFOCUS, we realized that it could give us the ability to easily spot trends and patterns in food stamp transactions."

With help from Information Builders Consulting, the department deployed WebFOCUS as its primary reporting tool, setting up an architecture for both real-time and staged reporting. EBT transaction data is periodically loaded into a data warehouse that allows field investigators to easily access and mine information about transaction amounts, times, and locations.

Visual Analysis

Once the basic reporting infrastructure was in place, the state of Louisiana began extending its fraud detection capabilities even further by incorporating geographic information into the system. By visualizing this same data on a map, state employees can quickly peruse anomalies in the data so they can make informed decisions.

The WebFOCUS/ArcIMS system can visually depict information in a variety of ways. Any geocoded data can be mapped. Geocoding involves correlating names and addresses to locations on the map. "Once you geocode your database, all of the applications that use that database have access to that information," Pease explained. "GIS has taken hold in government agencies because it is so valuable. Combining it with a powerful reporting environment puts two complementary technologies together. The potential of the WebFOCUS/ArcIMS system is limited only by our imagination."

For example, one WebFOCUS report reveals the average number of transactions per client in the state of Louisiana. Once that data is displayed geographically on the map, Pease can focus on particular regions. Simply by drawing a circle around a portion of the map, he can highlight the customers and stores in that area and use analytical capabilities of ArcIMS to examine activity in the area.

Extending the Reach

"Ultimately, we hope to make the WebFOCUS/ArcIMS system the primary reporting and geographic analysis tool for the entire Department of Social Services so they can discern trends in everything from child welfare protection cases to demographic analysis," said Pease. "This will help us better serve the public."

In the past, reporting against flat files on the mainframe involved time-consuming data conversions. But because WebFOCUS can read these files natively, the department can easily combine data sources to uncover new areas of inquiry. "The sophisticated data access architecture of WebFOCUS eliminates a lot of effort by our data processing personnel," confirms Lemoine. "For example, our EBT data can be combined with the U.S. Department of Agriculture database of retailer information. This helps us draw a clearer picture of suspicious activities."

Typical users include field investigators and supervisors in the Fraud and Recovery section. Most users need only a few hours of training because the system is so easy to use and the team already understands the EBT transaction information. However, there is so much functionality in the system that it takes users a while to completely understand the full extent of the capabilities.

Pease believes that the investment in the system is money well spent. With at least 4 percent of $600 million being fraudulently received every year, the system will more than pay for itself if his team can reduce illegal activity by even a small fraction of 1 percent.