It’s been less than five years since the COVID-19 pandemic broke out in March 2020, but it’s become difficult (at least for me) to recapture in my mind the fear and uncertainty of those first few months. The unemployment rate jumped from 3.5% in February 2020 to 14.8% in April 2020, just two months later. It was unclear how many employers would function or how many people would make a living. In short, it was a time when rapid legislative response was considered essential. But although no prominent person pointed this out in March 2020, it is very likely that even hasty plans to spend trillions of dollars will end in excessive proportions of waste and abuse.
Cecilia Elena Rouse makes this point in the following extensive discussion: “Lessons for economists from the pandemic” Presented as the annual Martin Feldstein Lectures by the National Bureau of Economic Research (NBER Reporter2024, no. 3). She describes the scale of federal spending in response to the pandemic:
In response, in 2020, Congress passed two bills: the Families First Coronavirus Response Act (providing $192 billion for COVID research, enhanced UI, and health funding) and the Coronavirus Aid, Relief Act on March 18, 2020, and President Trump signed this. , and introduce the CARES Act within 10 days (providing over $2.2 trillion in economic stimulus). CARES alone was the largest stimulus package in U.S. history. This was followed by the Coronavirus Response and Relief Supplemental Appropriations Act of 2021, signed in December 2020, providing $900 billion in additional funding and stimulus. And in 2021, Congress passed and President Biden signed the American Rescue Plan. The plan adds $1.9 trillion in stimulus and recovery funding. In total, this amounts to over $4.5 trillion in stimulus, compared to just over $2 trillion during the 2008 global financial crisis (both in 2022 dollars).
Rouse focuses on some of the key aspects of the CARES Act passed in March 2020. As she pointed out, the law was ‘good enough’ but sloppy. She points out:
This lesson is based on the fact that federal data, computer, and human resources infrastructure was and still is not up to the task of providing precise and rapid support to the economy. Components of the CARES Act highlight this reality. For example, the Paycheck Protection Program (PPP) provided unsecured and forgivable loans to small businesses (generally those with fewer than 500 employees). These loans can officially only be used to retain workers (with several disclaimers), cover payroll and health insurance costs, and pay mortgage, rent and utilities. If these conditions are met and businesses meet their hiring goals, the loans will be fully forgiven after the pandemic. The Economic Injury Disaster Loan (EIDL) program provided low-interest loans of up to $2 million payable over up to 30 years. The loan also included an option to defer all payments for the first two years while businesses and nonprofits get back on their feet after the pandemic. And finally, the scope and generosity of the UI has been dramatically expanded. Benefits were increased to $600 per week, and people not normally covered, such as gig workers and contractors, also temporarily became eligible.
You might say it was “good enough,” but it was sloppy. Meanwhile, nearly 1 million businesses received PPP loans (valued between $150,000 and $10 million), and 3.9 million businesses received EIDL loans. On the other hand, this support has been delivered rather inefficiently. Waste and poor targeting were the problems. David Autor and his co-authors estimate that the cost of a PPP loan is between $169,000 and $258,000 per job saved. This is more than double the average salary of these workers. They also estimate that more than two-thirds of total spending on the program accrued to business owners and shareholders rather than employees.
Outright fraud was also a big problem. The Government Accountability Office (GAO) estimates that PPP fraud accounted for about $64 billion of the total $800 billion in loans. This means that approximately 8% of all PPP loans may be fraudulent. Under EIDL, some borrowers used falsified names or business information to request loans, often running away with the cash. Ultimately, GAO and the Small Business Administration estimate that EIDL fraud is much more prevalent (in dollar terms) than PPP fraud. That’s more than $136 billion. UI fraud has also surged during the pandemic. GAO estimates that fraud costs between $55 billion and $135 billion.
Adding this up, the CARES Act’s $2.2 trillion economic stimulus could have been accompanied by over $300 billion in outright fraud, as well as extremely high cost savings per job. Of course, adding in the fraud and waste elements of the rest of the pandemic response bill will add to the total fraud and waste.
Was the trade-off between speed of response and the almost inevitable fraud and waste that comes with it worth it? Rouse notes that the U.S. GDP recovery through the end of 2021 was much stronger compared to other high-income countries. She also points out that this increased federal support also played a role in sparking a resurgence in inflation. She argues that while the pandemic response was certainly not perfect, it could at least be classified as “good” given the stresses of the time.
I’m willing to give Congress and President Trump a pass of sorts on the pandemic aid that was passed like crazy in March 2020. It seems to me that in December 2020 and December 2020, about $2 trillion worth of additional stimulus was passed into law during President Trump’s term. In March 2021, President Biden probably deserved more scrutiny than they received.
An obvious follow-up question is whether it would be possible to reduce the trade-off between rapid response on the one hand and waste and fraud on the other (a trade-off that has occurred in other high-income groups) when similar crises occur. The same goes for countries. Rouse points out that minimizing fraud and waste means proactively investing in information systems and technology that allows managers of these programs to fulfill their role as frontline fraud and waste detectors. But the United States is lagging behind in these investments. Rouse focuses on the example of information technology in unemployment insurance programs operated at the state level, but her points may apply more broadly to many government information systems. She points out:
As of 2020, less than half of states have modernized their UI. [unemployment insurance] system. Some state machines still run in COBOL. In most states, it is nearly impossible to submit an application from a mobile device, and some states require workers to physically receive a password in the mail to log into their UI account. Due in part to these issues, only about 57% of unemployment claims nationwide had been paid through the end of May 2020. This created a dual crisis that led to more fraud as overworked employees did not have the resources needed to rigorously verify their claims, while truly eligible employees were forced to wait weeks or months before receiving their benefits.
Moreover, if the goal during a pandemic is to provide income to those experiencing immediate disruption, the unemployment insurance that exists is an imperfect tool. The program does not apply to many people who engage in “alternative” work arrangements, such as contract workers, the self-employed, temporary workers, or students who work while taking classes at the same time.
We are already seeing high-profile prosecutions for fraud during the pandemic. But while it requires effort to identify offenders years later, it is also a costly and ineffective way to limit illegal activity. What we are not seeing is serious thought (and investment) being put into reducing the risk of fraud and waste in these programs during current and future crises.
