LPInformation Magazine November issue has a very informative article on Returns fraud and how Retailers are using Returns Management Systems to minimize the loss due to fraudulent returns. Going by statistics, Fraudulent returns total $9.6 billion a year by one estimate and around $17 billion by another. Measures being adopted by retailers include the following:

Laurie J. Sorensen, vice president of LP and shortage control for Macy’s Northwest, says company policy requires customers to provide proof of purchase in the form of a receipt or a customer return label (CRL) to receive cash or credit back. Otherwise, the system will only allow store credit in the form of an easy exchange card or merchandise-only certificate.
A CRL containing the original purchase information “is placed on the merchandise tag and scanned at point of sale,” Sorensen says. “When a customer returns an item without a receipt, we are able to scan the CRL and identify the original tender, which they will receive a refund to.”

...

Within its proprietary refund management systems, one major retailer applies an algorithm to specific receipts, creating a unique number associated with that receipt throughout its lifespan. Depending on the payment method the customer chooses, that number can often be tied back to individual customers.

...

Mike Keenan, director of LP for Hayward, Calif.-based Mervyns, recently completed the implementation of a returns management system. He was sold on the concept due to his experiences while working for another retailer.
“We reduced returns by approximately $20 million,” he says. “It drove the people who did nothing but fraudulent returns right out of our stores.”
Returns management systems vary, but basically use either the original sales transaction history (stored in a central database) or statistical modeling to determine whether a return is potentially fraudulent.

Integrated Solutions For Retailers Dec 2007 issue also has an editorial on Returns Management where it talks about two ways to tackle the menace.

Two popular approaches are software-driven analytics programs and item-level product tracking. Both approaches have merits, and depending on your merchandise mix, they can be used independently or in tandem.

An analytic software approach helps retailers maintain flexibility in their returns policies, applying different returns parameters to different customer profiles. The last thing you want your returns policy to do is make legitimate returns difficult for your best customers, which are often the customers who return the most.

Modern returns management software lets store associates authorize returns in real-time by screening for fraud through analysis of customer-specific return behavior. By cross-examining historical returns and customer-specific transaction data, analytical software helps retailers identify specific return patterns. This is a great management strategy that provides retail leadership with vivid insight into the successes and failures of its returns policies. Oracle, Newgistics, and others offer such solutions.