Steve Levitt believes in the power of incentives and he has a new proposal for how to make widespread testing successful: set up a testers’ lottery and give gigantic cash prizes. And, Eric Zwick takes a closer look at PPP lending to small businesses under the CARES Act and how many of these loans missed the target.
If you’ve listened to Pandemic Economics this season, we’d appreciate your feedback. Please take a minute to complete our survey: pandemiceconomics.org/survey.
TESS VIGELAND: As the weeks go by, life in the time of coronavirus can feel unending and permanent. Yet, we’re also seeing a lot of coverage and speculation on where we might look for a way out of this crisis.
EDUARDO PORTER: We’ll hear how a $200 million prize to encourage healthy people to get tested for the virus could help us move out of lockdown.
TESS VIGELAND: This is Pandemic Economics, the podcast about the global impact of COVID-19 from Stitcher and the Becker Friedman Institute for Economics. I’m Tess Vigeland.
EDUARDO PORTER: And I’m Eduardo Porter. We’ve been invited to have this series of conversations with University of Chicago economists.
TESS VIGELAND: In today’s episode, we speak with Steve Levitt and Eric Zwick about some of the solutions being proposed and implemented as we struggle through the crisis. We’ll hear where the rescue money for small businesses actually went as opposed to perhaps where it should have gone.
EDUARDO PORTER: But first, I speak with Steve Levitt about why mass testing may not be as straightforward as it seems, and his radical solution to the problem. The idea to pay people to get tested for COVID, I think is going to strike a bunch of people as a little bit counterintuitive. And I was wondering if you could just walk us through that idea of yours, why do you think this is necessary and what you think this could achieve?
STEVE LEVITT: So I think this idea is anything but counterintuitive to economists, because this is the simplest thing in the world, which is that people respond to incentives. And COVID is a particularly interesting case, because it seems to be quite infectious, even before I feel symptoms, or even if I’m asymptomatic all the way through my having COVID, I can spread the disease. And so that leads to this crazy situation where if we’re going to successfully get out of lockdown, many of the modelers who’ve been modeling it suggests that the only way it’s going to happen is with intensive testing.
So my co-author Paul Romer suggests we have to test something like 20 million Americans a week to have any hope of coming out of lockdown without big resurgence of the disease. Now it’s pretty easy to see why if you’re feeling pretty crappy you might go get tested for COVID. I don’t think we’re going to have a hard time incentivizing sick people to go get tested for COVID, but I’m perfectly healthy, why would I want to go get tested for COVID?
What does it involve? It involves a somewhat unpleasant procedure, where they stick a six inch long swab through my nose into the back of my head. And it’s not that comfortable. And I probably got to go spend a half an hour, an hour waiting in line at the local pharmacy with a bunch of people who probably are sick, because they’re there to get the test, because they’re sick.
I just think it’s not realistic to think that a person who has no symptoms and is being asked to be tested 25 times a year or maybe 40 times a year is going to keep on dutifully showing up at the pharmacy having tested negative 29 times in a row or 26 times in a row. If you’re realistic, you’ve gotta believe it’s going to be hard to get people to get tested. And yet, the scientists say it’s really, really important.
The crux of our proposal, which strikes people as odd, or comical, or silly is that we actually think that it’s important even if you’re not sick we want you to be happy to go and get tested. And so that’s where this idea we call pandemillions, of a lottery. So you get a lottery ticket. So there’d be some kind of algorithm that would suggest who would be appropriate to test in a given week. And you’d get that call and something like jury duty. You get a message saying that it’s your turn to get tested. And we would incentivize you by if you go and you get tested, you would get a lottery ticket. And we could put a lot of money into this lottery and still have it worthwhile. So we proposed something like $200 million a week to go in the lottery.
But again, if that’s not enough, the country could happily put a billion dollars a week into this lottery. And it still is so much cheaper either than staying and lockdown or of coming out of lockdown and having a resurgence, because a bunch of people are being infected because asymptomatic people haven’t been tested. So from an economic perspective, there could not be anything simpler, easier, more straightforward.
What’s a little different than the typical economic solution is usually we use a stick, rather than a carrot. We punish people, but in this setting. I think it’s a much better idea to use the carrot to positively reward people for doing the kind of behavior that will have a huge community value.
I see, but couldn’t the government just make testing mandatory?
I think it potentially could, but think about the apparatus you’d have to build around that. So we’re talking about 22 million Americans a week who the government is somehow going to compel to do something they don’t want to do. And I just think the government’s got a real challenge in that regard of how you would enforce that. But you can’t just put out a program and say 22 million people will be tested.
And if we have enough tests, then we’re fine. But really I think we’re trying to get at is not necessarily the way to solve the problem but to bring attention to what is a fundamental problem, which is just having testing capability and telling people to get tested. I guarantee you that is not going to solve our problem. If we don’t get the incentives right, everything else is going to go for naught.
EDUARDO PORTER: Is there a clever way to figure out how much we should pay to provide the right incentive for people to get tested?
STEVE LEVITT: Keeping the economy locked down costs us something like $500 billion a month. And so if this ends up being the difference between locking down the economy and not, you could imagine you’d be willing to pay a few hundred billion dollars a month. Our plan, pandemillions, will cost a few billion a year.
So it’s a trivial share of the costs relative to the total cost. And if you actually really felt like you needed to pay more to get people to do the tests, you should be willing to pay something like $25, $50 per test to get people to do this test. Our pandemillions idea is maybe the payouts are $5 per person for doing the test. So way below the social value, but then really as an economist, you’ve got to think about the elasticity of supply.
So how much more responsive will people be to a relatively small payment versus a big one? And our guess is that you can make the payments big enough to get people to comply without spending nearly what the social value are. I can’t say that we’ve thought very carefully about the exact trade-off there. Our hope has been really something much less ambitious, which is, could we just get this idea out into the ethos? But without the right incentives, this is just going to fail, this whole escape from lockdown and how to implement these incentives.
EDUARDO PORTER: You estimate the cost that an infected person imposes on everybody else. I wonder if you could unpack how you reach your estimate of that. What goes into that number?
STEVE LEVITT: So there are two very different ways that we approach coming up with the value of this kind of test. And the first one is a bottom up approach. Well, we know that the experts claim that this disease has what they call a R0, a reproductive factor of 2.5.
So if I get the disease and I’m going about my daily life, I’m likely to spread it to 2.5 more people. And then you just calculate, well, what’s the cost to those 2.5 extra people having the disease in terms of mortality and health care costs and morbidity? And if you go through that, to keep someone from spreading the disease to 2.5 more people, you’d be willing to pay about $20,000.
That’s not a perfect approach, though, because if you think about it, it’s not just that I spread the disease to 2.5 more people, you also gotta worry about the fact that those 2.5 people are going to spread it to 2.5 more people each. And so that’s really a very incomplete accounting of what goes on. So another way to think about it is from a much more macro perspective and just say, well, let’s say you could estimate if we don’t do anything, if we come out of lockdown, and we just say, look, we’re going to take our lumps and we’re going to let people die and we just think the economy is more important than all these deaths, then some reasonable model suggests that we’re going to have something like a million people die in the US from that model.
So then you say, well, how many people will die if we do pandemillions and it actually works and then say the estimate comes to maybe 200,000 total? So we’ve already had 65,000 70,000 deaths and maybe will go up to 200,000. So in essence, we’ll have saved 800,000 lives by doing the pandemillion approach, as opposed to doing one where we just let stuff go. And we just accept the fact that eventually we’ll get herd immunity.
So then you start multiplying the costs of those deaths and the healthcare costs. And again, you get to some really, really large number, at least $20,000 per infection along the way. And, honestly, though, I would say in the world we’re living in with COVID, cost-effectiveness is really not the issue. What we’re doing now is so inefficient that even a really bad version of the incentives I’m offering, as long as they work, as long as they get people tested, as long as the people who are tested and tested positive quarantine, and as long as the models are right and that’s enough to avoid re-sparking the epidemic, that is just– the value of that economically and in terms of lives saved is so astronomical that we’re really talking about pennies in the cost of getting the incentives right.
TESS VIGELAND: Policymakers are attempting to keep this crisis from undoing millions of Americans, especially small business owners. After the break, we talk with Eric Zwick about his research into where the money went from the $350 billion dollar paycheck protection program.
EDUARDO PORTER: There’s more pandemic economics. So hang tight.
TESS VIGELAND: Now a conversation with Eric Zwick, whose research tried to suss out where all that federal rescue money went that was supposed to keep America’s small businesses from going under in this pandemic. Now you looked at two central questions about the Paycheck Protection Program, the PPP. Did the money go to where the need was greatest? And what role did banks play in deciding where the money went?
So let’s get to both of those questions in a moment, but first can you describe for us how this loan program worked? It wasn’t like the checks that Americans got in the mail, right?
ERIC ZWICK: Yeah, that’s right. So it’s a small business loan program up front, where firms applied through banks. There was a formula that was basically– the maximum amount they could apply for was something like 2 and 1/2 times monthly payroll. They had to provide documentation for the bank to agree to that loan amount. And then the provision was only available for firms that had less than 500 workers.
The maximum loan size, I think was $10 million. There was an interest rate that was much lower than the typical loan. And then there’s a forgiveness provision, which allowed for loans to be forgiven if they covered payroll during a specific period of time as long as at least 75% of the money being forgiven was spent on payroll. In that case, it would convert from a loan to a grant.
TESS VIGELAND: So something they wouldn’t have to pay back.
ERIC ZWICK: That’s right.
TESS VIGELAND: All right. Let’s start with where the money went. As you note in the paper, the goal was really here to prevent layoffs and bankruptcies by injecting all this cash, $350 billion worth, into small businesses. Where would that money have made the most difference? And what did you find out about whether or not it got to those places?
ERIC ZWICK: We found that generally there’s not really much of a correlation between the intensity of the pre-policy job losses and where the loans went. And if anything, areas that were sort of less hard hit prior to the program being rolled out, were the areas that got larger injections of PPP funding. Our overall average is something like one in five small businesses across the country received PPP support.
So it’s a pretty big program in terms of its coverage despite running out of funds pretty quickly, but really no relationship in terms of places that had sharper declines in hours worked, like in Massachusetts, New York, versus places that had less sharp declines, like South Dakota or Nebraska. You can see slightly increasing relationship where a larger share of businesses are receiving PPP in those less hard hit areas. In California, we see something like 15% of small businesses receiving PPP. In South Dakota, North Dakota, it’s more like 40% to 50% of small businesses. So a pretty big difference there.
TESS VIGELAND: That’s significant. Yeah. So why did this happen where this money wasn’t going to the places that needed it most?
ERIC ZWICK: It’s super important to understand, because the range of potential explanations has quite different policy implications. I think that there are probably three pretty important explanations.
One, areas that were less hard hit, are also areas where businesses face less of a shortfall in kind of demand for their services or product or less of an economic shock. That means that they’re more likely to be able to maintain payroll during the two months that the PPP was designed to cover. They have work for those workers to do. And so for that reason the forgiveness element of the program, which really depends on paying sufficient payrolls is more attractive to those types of firms. So in the less hard hit areas it actually– could be a better deal than the more hard hit areas, which is just the way the program was legislated.
TESS VIGELAND: Because of the requirement that they keep people employed in order for it to be more of a grant than a loan that made it easier for less hard hit areas, because they were probably already keeping more staff on anyway.
ERIC ZWICK: Correct. A second potential explanation in areas that were more hard hit, it’s not just the firms or the borrowers that are hard hit, but it’s also the banks. And so the banks are facing disruptions, branch closures.
Their workers are having to go home, or getting sick. And so it’s possible that the loan application process was more disrupted in those areas. And then a third explanation that’s important is it’s possible that some banks were unwilling to participate as aggressively for any kind of regulatory reasons or fear about how the program would be implemented or political reasons.
TESS VIGELAND: Let’s talk about where these loans went. I think everyone is familiar with the big national banks, Chase, Citi, Wells Fargo, B of A, but from your research, you said that those banks actually did not get a large share or did not hand out a large share of the PPP loans.
ERIC ZWICK: The way we were trying to think about this is if you just looked at how much PPP lending these top four banks did, because they’re much larger than the next set of banks, you would say, oh, OK, they distributed a significant number of loans per bank. And you would be impressed by that, but the thing is they’re really large banks.
And so if you actually compare that to their baseline rate of small business lending, those banks account for 36% of total small business loans. Those banks account for a quarter of small business lending by dollars. Their share of PPP activity is well-below their share of traditional small business lending in normal times.
TESS VIGELAND: So the total amount that they handed out might sound large, but compared to their portfolios that they had already, it’s not.
ERIC ZWICK: Exactly.
TESS VIGELAND: Eric, why does it matter that the four big banks didn’t have a larger share?
ERIC ZWICK: I think it only matters because relationships between borrowers and banks are sticky. And so it’s hard for borrowers to switch banks. These big banks have a lot of clients, a lot of borrowers, small business borrowers too. So their sort of slowness is problematic, as it means that a lot of those borrowers weren’t able to get the loans that they applied for. The pre-existing connections between the firms and banks means that those big banks are really important, because they’re connected to so many small firms.
TESS VIGELAND: So, Eduardo, you know it really seems to me like so much of almost everything we’ve been talking about is the law of unintended consequences.
EDUARDO PORTER: Oh, yeah.
TESS VIGELAND: And you certainly see that in the paycheck protection program. And Levitt– certainly, Steve Levitt certainly also talked about that as a possibility in testing.
STEVE LEVITT: Yeah, for sure. And Steve’s work is really about understanding human motivations and trying to figure out, OK, if I need this outcome, which in his case is people to go get tested, what kind of incentives do I need, and how can I best structure them? Whereas in Congress, it doesn’t really quite work that way. We need to get money out there, which I totally understand.
TESS VIGELAND: Clearly– I mean, this was legislation that they passed in a hurry in many ways for good reason. People are in pain. People are suffering. And they had to do something.
EDUARDO PORTER: So, yes, so much of this has been a crapshoot. And it’s now kind of interesting that Steve Levitt is actually deploying the idea of a crapshoot using a lottery, the ultimate crapshoot.
TESS VIGELAND: Right. Exactly.
EDUARDO PORTER: Economists can be useful in moments like this, because one of their jobs is to look for these unintended consequences. That’s their job description. Next time, Erik Hurst will sift through business payroll data to explore the devastating impact of the pandemic on the job market.
ERIK HURST: Since the first week of March, 22% of workers have lost their jobs, 22%. That is staggering. And if you look at every other recession in the US history, since World War II, the maximum employment lost during the first three months of the recessions that we’ve had since that time, the maximum was 1 and 1/2%. It just blows my mind when I look at it.
EDUARDO PORTER: Pandemic economics is produced by the University of Chicago’s Becker Friedman Institute. Our producers are Dana Bialek, Kaitlyn Nicholas, and Devin Robins. Our executive producer is Ellen Horne. Production and original music by Story Mechanics.
TESS VIGELAND: Pandemic Economics is part of the University of Chicago Podcast Network.
EDUARDO PORTER: I’m Eduardo Porter.
TESS VIGELAND: And I’m Tess Vigeland. Thanks for joining us.