I vividly remember Earl’s call. It was early in 2009. The U.S. faced a growing economic crisis as a global recession continued to unfold. President Obama’s American Recovery & Reinvestment Act (ARRA) had just been signed into law and I was at the Office of Management and Budget (OMB), which was tasked with distributing the nearly $800 billion Congress green-lighted. Earl Devaney, chair of the ARRA oversight body (lovingly known as the “RAT” board) called to tell me he was standing in a newly rented empty office on Pennsylvania Avenue, chosen for its direct line of sight to the White House. His message was not subtle: The overseers were going to watch every move we made as we distributed the funds.
Earl and I became close colleagues. In addition to helping us prevent fraud and ensure critical funds made it into the economy, his oversight board built the Recovery Operations Center (ROC), which proved to be a paradigm shift in how big data and forensic analytics are used to track government program spending. Now that our nation is facing another unprecedented crisis, we need another shift in how government leverages digital, AI, and analytics to monitor spending.
As the Coronavirus Aid, Relief, and Economic Security (CARES) Act is implemented, oversight committees should draw upon the lessons learned from ARRA and launch the next-generation ROC.
By 2009, advanced data and analytics practices had expanded from the private sector into the intelligence community; however, advanced data analytics had not been broadly deployed into government, least of all in tracking government expenditures. Recognizing the potential of bringing cutting edge analytics into monitoring $800 billion in spending, Earl and his team built the ROC to be the most analytically advanced federal operation outside the intelligence community. The ROC rapidly improved the speed and specificity of fraud detection, flagging suspicious patterns in fund disbursement data that normally eluded federal agencies racing to spend ARRA money to stimulate local economies. For example, addresses flagged as suspicious after receiving multiple stimulus checks could be cross-referenced on an early-stage Google Earth to assess the integrity of their claims without deploying a team to the site.
Today, there are far more advanced technological capabilities available than the Excel spreadsheets and Google Earth of 2009. Artificial intelligence (AI) and new advanced data analytics techniques can circumvent the tradeoff of speed vs. rigor in monitoring stimulus spending. Private sector banks and other fraud-prone industries already use multi-layered approaches to identify fraud by differentiating transactions based on risk; think of when credit card companies flag unusual transactions for your validation and approval. AI can also analyze tax filings and other SEC documents to identify suspicious businesses, transactions or hidden relationships between companies. CARES Act funds will roll out more quickly and with integrity if overseers make proactive investments to enable these technologies to come online this summer.
Distributing $800 billion in ARRA funding taught several lessons critical to overseeing CARES Act spending — speed, transparency, and impact.
In 2009, everyone’s first question was, "how quickly will this stimulus hit the economy?" In answering this and future questions, we learned the importance of establishing mechanisms for continuously updating the country on how funds were distributed. CARES Act oversight should sustain clear channels of communication into program leaders and overseers from a manageable number of stakeholders.
Transparency was achieved by making the program for distributing funds public, from initial application to final documentation and post-award audit standards. The three CARES Act oversight bodies — Special Inspector General for Pandemic Recovery, Congressional Oversight Commission, and the Pandemic Response Accountability Committee — should coordinate oversight with clear responsibilities and shared compliance standards.
When measuring the impact of funds spent, we needed to openly discuss how to define “impact.” The CARES Act needs a plan to measure and track the number of lives and jobs saved because of CARES Act funding to health care systems and businesses. Members of the public are a key ally in fighting fraud and error. Providing a clear channel to report concerns to the government can uncover schemes that unfortunately emerge when new, large government programs are launched.
The CARES Act is a highly complex pillar of our recovery — its implementation should leverage today’s advanced tools and the lessons from the ARRA.
And while advanced technologies are only one piece of the puzzle, they will be critical to CARES Act spending oversight (as the ROC was for ARRA). The government should partner with technology experts inside and outside of government to stand up an advanced analytical and AI-enabled control tower utilizing existing centers of excellence as appropriate. In 2009, Earl and the ARRA oversight committee led the federal government in digital capabilities to track spending. Today, the lessons and tools are here for today’s leaders to build the next generation ROC.
Danny Werfel served as Acting Commissioner of the Internal Revenue Service and as Federal Controller of the Office of Management and Budget. He is currently a Managing Director and Partner at Boston Consulting Group.