Agencies need a formal framework for detecting, preventing and managing fraud, waste and abuse.
Legislators may disagree over the amount the country should spend on government benefits, but one fact will always have bipartisan agreement -- the level of fraud, waste and abuse plaguing benefit programs is too high. Although some progress has been made, including a 50 percent decrease in the number of fraudulent tax returns reported by the IRS this past year, the federal government still has a long way to go. To illustrate the severity of the problem, the Government Accountability Office recently reported that the estimated amount of improper payments in fiscal 2015 across 121 government programs was $136.7 billion, up $12 billion from the previous year.
This negative trend is likely to continue as criminals steal more and more identities and open new, more sophisticated avenues of attack. Census data indicates 110 million Americans -- more than a third of the population -- are currently receiving means-tested federal assistance of some kind. Historically, more than half of all Americans have received benefits from at least one of the top six federal entitlement programs, according to the Pew Research Center. A lack of resources, the need to ensure citizens receive benefit payments as quickly as possible, and an increased transition to electronic platforms have combined to leave gaping holes in agencies’ abilities to mitigate fraud. Meanwhile, fraudsters have become more sophisticated in their approach, and are constantly evolving their methods to stay ahead of preventive measures. We need a more structured, formalized and governmentwide approach to combat fraudsters.
In addition to policy and protocols, such a framework would include analytical models for detecting, preventing and managing fraud, waste and abuse. This includes making full use of all relevant data spread across multiple, disparate government systems. Agencies need a way to enhance data integrity, effectiveness and accessibility, as a recent GAO report concluded.
By adopting a formalized fraud framework, agencies will be able to achieve a holistic view of the critical information driving their programs. This will result in an increased ability to quickly and effectively resolve identities and stop fraud before payments are made, as well as:
- Detect more fraudulent activity
- Decrease losses associated with fraud
- Reduce costs associated with detecting and investigating fraud activities
- Provide a consolidated view of fraud risk and losses
- Foster improved transparency and accountability
An effective fraud framework focuses on the data, using analytical capabilities to identify abnormalities, networks and patterns that may indicate fraudulent activity and must be applied throughout the entire benefits process -- from application through payment.
Many individual components must be considered when developing a fraud framework, all of which work together as a unified whole for preventing, detecting and recovering improper payments. An effective framework should include:
- Data collection, integration, deployment, management and business rule creation capable of extracting relevant data from different systems, external data sources, unstructured text and other sources.
- Advanced analytics and model testing, deployment and management that enhance the value of existing business rules by enabling the discovery of emerging suspicious activity that would otherwise go undetected.
- Detection and alert generation, which enable the systematic detection of suspicious activity using a fraud scoring engine that employs a combination of analytic techniques to determine the likelihood of the presence of fraudulent actors or activity.
- Alert management capabilities to assemble alerts from multiple monitoring systems, associate them with individuals and provide investigators with a more complete perspective on the risk of a particular individual.
- Social network analysis, an invaluable method that helps investigators detect and prevent organized fraud by going beyond individual transaction and account views to analyze all related activities and relationships in a network.
- Case management, which provides a systematic means for facilitating the investigation and capturing and displaying all information pertinent to a case.
Fraud continues to cost taxpayers billions of dollars every year. In order for the government to meaningfully combat this trend and ensure benefits are finding their way to individuals with a legitimate need, agencies require a more integrated, formal data collection and analytical approach. While there have been promising advances in capturing the low-hanging fruit, fraudsters are becoming bolder and more sophisticated, with more fraud networks causing increasing losses. With the greater sophistication and coordination that the more advanced networks have displayed, it is possible for one big hit to wipe out an investigative team’s effort for an entire year.
To ensure the continued integrity of government benefit programs and reduce the rising level of improper payments, more advanced methods are needed to make a measurable and meaningful impact. Agencies need to adopt a comprehensive fraud framework to have real, sustainable success. And given the fiscal survivability of our benefit system, we need these successes sooner rather than later.
J.R. Helmig is chief analytics officer for federal programs at SAS.
Photo: Pictures of Money, via Flickr