A recent report from the Government Accountability Office (GAO) found that almost $1.4 billion in COVID-19 stimulus checks were sent to deceased individuals. The fraud amounted to over one million erroneous payments. The data required to avert this loss of tax-payer money was readily available in government systems, but the data wasn’t accessible to the agency that needed it.
On its face, this sounds like a data-sharing problem. But even if the Social Security Administration (SSA) death index had been shared with the Department of Treasury (the agency that issued the stimulus checks), it wouldn’t have been enough. Why? Because there are 89 million records in the death index, and there are 143 million tax-payers in the United States. Detecting fraudulent applications amidst this volume of data cannot be done by employees alone. It requires data integration and software that looks for anomalies and alerts the employees before the checks go out.
There are many recent examples of this need in government. A lack of data sharing and data management has marred America’s ability to mitigate the impact of COVID-19 and hampered containment of fraudulent unemployment insurance claims. The data is just too voluminous for legacy data systems to handle.
Modern data management platforms offer the solution to this growing problem because they enable anomaly detection—sourcing actionable information from vast amounts of data. This thesis is confirmed by a Palo Alto Research Center study that looked at enormous healthcare datasets. Researchers determined that the key to detecting suspicious activity was establishing a system that can handle vast amounts of data, sorting through each claim while remaining aware of overall payouts. The study found a 23% increase in improper payments between 2007 and 2008 and a 60% increase by 2013. That equates to approximately $65 billion in fraudulent payments in healthcare alone.
How big is this problem for our government? According to McKinsey & Co., an estimated 20 percent of total government revenue disappears annually. That’s $5 trillion gone without a trace. With that much money vanishing annually, the return on investment for a data platform that can integrate disparate systems, handle structured and unstructured data, and provide anomaly detection capabilities more than justifies the investment.