15% of Paycheck Protection Program Loans Could Be Fraudulent, Study Shows

When the Paycheck Protection Program began last year to help small businesses that were struggling during the pandemic, the federal government was determined to get the relief money out fast — so it waived much of the vetting lenders traditionally do on business loans.

The absence of those safeguards meant that fraud was highly likely. But just how much of the program’s $800 billion was taken illicitly?

A new academic working paper released on Tuesday contains an estimate: Around 1.8 million of the program’s 11.8 million loans — more than 15 percent — totaling $76 billion had at least one indication of potential fraud, the researchers concluded.

“There’s been a lot of anecdotes about fraud, but the tricky thing about anecdotes is that it’s very difficult to put them together and get at the scale of what’s going on,” said Samuel Kruger, an assistant professor of finance at the University of Texas at Austin’s McCombs School of Business and one of the paper’s authors. “We wanted to look for patterns in the data.”

The study pins blame for many of the questionable loans on one particular group of lenders: financial technology firms, known as “fintechs,” which focus on digital lending. Nine of the 10 lenders with the highest rate of suspicious loans fell into that group.

“Certain fintech lenders seem to specialize in dubious loans,” the authors wrote. Collectively, fintechs made around 29 percent of the program’s loans but accounted for more than half of its suspicious loans, the study concluded.

The Paycheck Protection Program, which ran intermittently from April 2020 to May 2021, relied on banks and other lenders to make the government-guaranteed loans, which are designed to be forgiven if borrowers followed the program’s rules. Government watchdogs have long warned of a high fraud risk on the rushed loans; the Justice Department has charged more than 500 people with improperly claiming hundreds of millions of dollars in borrowing.

Dr. Kruger and two other researchers at the university, John M. Griffin and Prateek Mahajan, identified a set of four primary and five secondary indications of a suspicious relief loan. Among the red flags: businesses that claimed they paid workers significantly more than their industry’s norm, and corporations and other formally structured businesses that lacked a state business registration. Then they combined the loan records released by the Small Business Administration, which managed the program, with other data sources, like registration records and industry wage data, to find loans with anomalies.

The $76 billion contains some false positives, the researchers acknowledged, because not every loan that raises red flags is improper. One of their indicators, for example, is multiple loans going to multiple businesses located at the same residential address. That’s often a warning sign, according to the researchers and to program lenders, several of whom have said they gave extra scrutiny to such loans. But there are also legitimate reasons a household could contain more than one home-based business.

A more restrictive calculation by the researchers, of loans with at least two suspicious characteristics, identified 1.2 million potentially fraudulent loans, totaling $38 billion.

“We were fairly conservative in the way we approached the whole analysis, so there’s also probably billions that we’re missing,” said Dr. Griffin, a finance professor at the university. “It seems like the fraud cost was high on this program.”

Dr. Griffin, the team’s lead researcher, is an owner of four companies that do consulting work on financial fraud investigations. None has any contracts related to the Paycheck Protection Program, he said.

In particular, the study cites two lenders, Capital Plus and Prestamos CDFI, as having fraud flags on roughly half their loans. Both of those lenders made nearly all of their loans through Blueacorn, a loan facilitator that drew in borrowers through a marketing blitz and steered them to its partners. Two other large online lenders, Cross River Bank and Harvest Small Business Finance, also had exceptionally high rates of suspicious loans, the researchers said.

All four lenders said they strongly objected to the study’s methodology, data and conclusions. At the same time, they emphasized that the populations they focused on — particularly solo entrepreneurs and tiny companies, including those without traditional business banking relationships — were inherently riskier.

“We made every effort to screen out ineligible applications,” said José Martinez, the president of Prestamos CDFI. “My team work day and night to support the smallest-of-small businesses, and we are proud of the work we did so the critical emergency funds reached all eligible applicants.”

Big banks mostly limited their Paycheck Protection Program lending to existing customers, a choice that reduced fraud but disproportionately excluded businesses owned by women and people of color. Online lenders and fintechs were frequently the only available option for those without business bank accounts and credit lines. Making loans to those with no prior relationship with the lender — especially in the absence of strict underwriting — significantly raises the risk of fraud.

“Cross River is a state-chartered, F.D.I.C.-insured bank with robust regulatory standards,” said Phil Goldfeder, a bank spokesman. “Unlike other lenders who prioritized their own customers, Cross River addressed the S.B.A. call to action and did not limit the program to existing customers.”

Before the study was released, Blueacorn sent a letter Jay Hartzell, the president of the University of Texas at Austin, objecting to the researchers’ approach. Blueacorn said that by relying on interim data released by the Small Business Administration before the P.P.P. ended, the study counted loans that its lenders initially approved but later canceled because of suspicious traits. Nearly 157,000 applications — about 16 percent of all of the loans Blueacorn’s lenders approved — were canceled by the lenders before they were paid out.

“As we reviewed increasing volumes of loan applications, we learned, adapted, and enhanced our fraud detection capabilities and protocols,” Barry Calhoun, Blueacorn’s chief executive, said in a written statement. “Along the way, we partnered with the S.B.A. and other authorities to ensure the integrity of the P.P.P. while providing a traditionally overlooked population with access to the funds they needed and deserved.”

The researchers said they hoped their work would help inform the ongoing policy debate about the Paycheck Protection Program’s effectiveness.

“Our evidence, along with evidence that the P.P.P. saved relatively few jobs at a high cost, provides growing evidence that the P.P.P. seems to have been a poor allocation of capital,” they wrote. “The sheer scope of the tens and hundreds of thousands of suspicious loans originated by many fintech lenders suggests that many lenders either encouraged such loans, turned a blind eye to them, or had lax oversight procedures.”

The Small Business Administration had no immediate comment.