The Paycheck Protection Program (PPP) handed out $753 billion to 4.5 million entities, ranging from LLCs to corporations. The purpose of a PPP loan was to allow a company to retain their workforce during these trying times. However, the available PPP data is questionable. More than 800,000 entities claimed they retained negative or no jobs. The fact that the average PPP loan came in at $155,000, per request. Therefore we, the American taxpayer, spent roughly $124,000,000,000 to retain zero jobs. Granted, the people applying for the loans must have kept their job, but that’s not why the program was created in the first place. The application process and publicly-available data appear to be in disarray. The poor quality of the data, and lack of record keeping, indicates fraud and abuse of the SBA’s PPP loan program.

Getting into the data

The publicly-available data, released by the Small Business Administration (SBA), keeps being used to report how many jobs were saved by PPP. However, with closer examination of the data, it’s clear that the data can’t be correct.

Here are some reasons why:

There were exactly 878,275 loan requests that reported zero or negative jobs retained. Even the SBA website says the purpose of the loan is to “provide a direct incentive for small businesses to keep their workers on the payroll4.”
The Washington Post interviewed a few loan recipients who reported to the SBA that they had 500 employees retained. Those recipients told WaPo that they only have 10-20 employees. Therefore, the SBA data is wrong3.
There are over 12,000 loans that were granted through Wells Fargo, which reported zero jobs retained. All of these loans were requests for over $150,000. It seems to indicate that Wells Fargo didn’t gather this information from their applicants. Rather than leave it empty, they chose to put in zeroes.
North Dakota received the most PPP loans per a state’s population, coming in at $2,410.99/person5. This is very unusual, as North Dakota is renowned for its low cost of living6. While it’s likely that a company reports that they’re based in ND for tax reasons, the SBA data should also capture where the majority of their employees live. This way, the data reported can be verified. (see heat map below)

There are some industries that retained as many or more jobs as they had available in 2018, due to PPP loans. These include, but are not limited to: the performing arts, personal services, agriculture and forestry 2,7. It’s also assumed that not every entity in these industries applied for a loan. Even if every entity did apply for a loan, there must have been over a 10% growth in 2019 to get these numbers. Even in the booming economy of 2019, a 10% growth in jobs is not very likely.
Among the highest loan value requests, loans over $150,000, there are over 20,000 addresses recorded that aren’t valid. It seems unusual that a bank would grant a business a loan for over $100,000 and not have a viable address on file for the company.
There was an address in Idaho that was associated with about 15,000 entities “921 S ORCHARD ST STE G”, and it appears to look like a business building like any other.
Many businesses applied to multiple lenders, using the same business name and address. It’s not clear if these are duplicate entries or if duplicate loans were granted.

Key takeaways

So, there is some good news. According to the SBA data, Houston Texas retained the most jobs out of any city. It’s an unprecedented level of transparency for this administration to make all the SBA data publicly available. This allows for experts to review and give feedback. Most of the PPP loans were rushed out in April, when the process was still in its infancy. However, it has been confirmed that loans can still be applied for until August 8th, 2020. They might even be granted further into 20204.

Hopefully, some changes will be implemented in the application process. That will gather more accurate information, so we can measure the effectiveness of the PPP loans. It’s worth mentioning that this likely won’t be the last pandemic in our lifetime. Providing a roadmap to economists on how to properly navigate the “next time” will fall directly on gathering accurate data.

Due to the job retention metric, there seems to be an implied association with unemployment and the PPP loan. However, more money was given to states who had unemployment increase between January and May 2020 (as seen in the graph below). This image seems counterintuitive. The whole purpose of the loans given in April was to stop citizens from collecting unemployment benefits, and to retain their job. Only time will tell if the disbursement of the funds truly impacted the unemployment rate.

Looking at the graph, we can see there’s a correlation between PPP loan and unemployment. Delving deeper into this, the x-axis is normalized by taking the square root of each state’s total amount of PPP loan money ($ USD) received. The y-axis is of the difference between each state’s unemployment rate, from January 2020 to May 2020. When you do an ordinary least squares regression, this results in an overall P-value for the model of 0.06. The significance for the square root of the PPP loan is at 0.06 as well. While this doesn’t pass the hard science alpha of 0.05, keep in mind that we’re using real-world socioeconomic data— the correlation between the two variables does check out.

Anyone acquainted with the 2008 economic collapse is aware that unemployment is a complex measurement. It will take more than the amount of loans handed out in a state to explain its behavior. However, it’s disheartening to see that states with higher increases in unemployment are receiving more PPP loans. After all, the loan aims to keep people employed and not collect unemployment. Perhaps the primary location that the loan is applied for is not where the majority of employees live?

The process of applying for the PPP loan was rushed. In doing so, the publicly available SBA data became very messy. This indicated that fraud may have taken place. The loans were granted to entities without an address listed, and many addresses appear to be very wrong. This paves the road for the loans not being repaid or tracked appropriately. There needs to be some form of verification in the number of employees that a company has, and the number it can retain because it received the loan. There should be some guardrails in the application, with regards to the fact you can not retain more jobs than there are employees currently within the company. Different lenders gathered different fields of information, and hopefully in the future, business loans will become more standardized. This will ensure that their impact can be more easily tracked.

Where to go from here?

The publicly available PPP loan data is very messy, more than is reasonably expected when almost 8 billion dollars is sitting on the table. This money is coming from the American taxpayer, and the purpose is to keep people employed in their current roles so that our economy will keep going forward. We are currently dealing with an economy that is trying to cope with COVID-19, which isn’t an easy task. It’s not clear, to economists, what the best path is to retain a healthy, capitalist economy. The PPP loan program was very quickly implemented to deal with the uncertainty of 2020.

Going forward, the loans will continue to be granted until August 8th. Hopefully, the application process becomes more standardized, and the lenders will do their due diligence to gather the most correct information. This will allow them to implement some discernment in granting the funds. The lack of clean data for this amount of money being handed out points to fraud. The question is, at what level is the fraud occurring? The government? Lender? Individual business? Until we discover the source, it stands in the way of effectively keeping small businesses afloat.