Charles Davidson: Welcome to another Federal Reserve Bank of Atlanta podcast. I'm Charles Davidson, staff writer with Economy Matters, the Atlanta Fed's digital magazine. I'm talking today with Stuart Andreason—Stuart is director of the Atlanta Fed's Center for Workforce and Economic Opportunity—and Mels de Zeeuw. Mels is a senior analyst in the Atlanta Fed's Community and Economic Development department. Guys, thanks for your time today.
Mels de Zeeuw: Thanks, Charles.
Stuart Andreason: Thanks for having us.
Mels de Zeeuw (left) and Stuart Andreason of the Atlanta Fed
Davidson: Yes, sure thing. This is one of our first "pandemic-era" podcasts, and I'll admit the first I've done with a dog in the room, so we'll see how it goes as we push into a little bit of a new frontier here. But what we're going to talk about today is a really cool new online instrument that these guys created called the Unemployment Claims Monitor. So first off, guys, can we talk a little bit about what prompted the launch of this tool? I know that you put it together very quickly and got it out in early May, I believe. So what spurred the idea, and how did you guys bring it to fruition so fast?
Andreason: Thanks, Charles. You're right—we conceived of this tool right at the beginning of the pandemic. I think that it actually was born in the first weeks of the country's shutdown, and what we saw happen with unemployment claims shooting up really quickly. In that first week we saw just huge variation in terms of what was happening state to state, and what was happening across the country. The national numbers were incredibly high, but then there were some states that had incredibly low numbers. We knew that there were certain people that were going to be feeling the economic shocks of the pandemic more than others, and we wanted to try and understand what that meant.
And so a number of people across the Bank started just really exploring everything that we could learn about unemployment insurance, and what it told us about what was happening week to week. And I think that we actually started to get interested in building the tool for a couple of reasons. As we talked to people who administered the unemployment insurance system, we learned that with the passage of the CARES Act the world was going to be really different. There have always been a number of special unemployment programs that touch people that don't get reported in that headline number.
I should mention that unemployment insurance has traditionally been for traditional employees of a firm that pays unemployment insurance premiums to cover their workers. Now, those employees are ones that receive a W-2 every year. Someone who works as a contingent worker or a contract worker—people might call them "gig workers"—or someone that received a 1099 at the end of the year rather than a W-2 would not have traditionally been covered. But with the passage of the CARES Act, there was a new program that covered those 1099 workers to receive compensation for lost work, very similar to the unemployment insurance system.
The CARES Act, also through that program, extended coverage to a number of other workers—workers that hadn't worked in their job long enough to be eligible—there's usually a waiting period for eligibility for unemployment insurance—or they hadn't worked enough hours to actually qualify for the program. And we saw states across the country doing very different and interesting things with unemployment insurance to help people weather the challenges of the pandemic and weather the challenges of losing their job because of the pandemic. But we knew that just looking at that traditional number as things unfolded was not going to be enough. And it was really the work of Mels to find ways to take a lot of the data that was coming out from a number of different places and turn it into something that we really wanted to get out to the public to be a service to researchers or to policymakers—to kind of get a fuller picture both of everyone that was experiencing disruptions, but also some insight into who among the entire set of workers was experiencing that. So we've been able to look at what industries unemployment claims are coming from, and the demographics of unemployment claimants, and to understand how this pandemic has been really affecting the economy.
Davidson: Do we have a good sense yet for who's using the tool?
Andreason: Well, I'll say a couple of things: We know that the tool has been very popular, and we've been getting a lot of feedback. We've heard from national associations that work on unemployment insurance issues, we've heard from researchers so we know that researchers are getting it. I've had calls from all types of researchers—those that are in the applied world, thinking about how to focus strategies on economic development and economic resiliency, to people in the academic world. It's also getting used by practitioners. We've shared it with people who are coming up with their COVID recovery strategies, to understand who's been impacted. So we know that state and local policymakers are using it, and to some extent we know the workforce ports are using it to figure out how they can target who has lost their job. We think that it will be useful for that as people can figure out where there are skills overlaps between those who have lost their jobs and where there's going to be some more hiring and economic activity happening.
Davidson: Are there things that we've learned during this bout of this huge surge in unemployment—things that have been surprising, or that maybe are particularly valuable pieces of knowledge?
de Zeeuw: Yes, I would say so. First of all, the data that's being displayed in the tool really helped us confirm some of the trends that a lot of other media and researchers had already picked up nationally: how this crisis is hitting certain groups like women or young workers particularly hard, and the by now famous—or infamous—images of the huge spike in unemployment claims. But I think one of the interesting things that our tool has added to the discussion is to show how some of these trends differ across the country, differ by state. And that there are some pretty striking geographic differences, which could point to a variety of factors like differences in timing of lockdown policies, or in when the virus really hit certain states, or differences in states' ability to process unemployment insurance claims, or in the industry mix of a state.
And in terms of what groups are hit harder…I mentioned already women and young workers, but there are some pretty striking geographic differences. For instance, we look a lot at the southeast of the U.S. Nationally about 52 percent of UI [unemployment insuance] claimants are women. That's more than their share of the labor force, which is about 47 percent. But if you zoom in—and it's pretty easy to do that using the tool—if you zoom in on certain individual states from the southeast, if you look at Alabama, Florida, Louisiana, Mississippi, women's share of UI claimants is 60 percent—so quite a bit higher than the national average. So in these states, this trend is particularly severe.
Davidson: Do we know why that might be, Mels?
de Zeeuw: That's a good question. It could depend on the industry mix in states. We really need to do some more research to figure out what's driving that.
Davidson: Yes. So speaking of the geographic differences: Florida—and certain pockets elsewhere in our region, too—is obviously heavily reliant on tourism, leisure and hospitality, so have we seen Florida get hit especially hard here?
de Zeeuw: Yes. Particularly if you look at the accommodation and food services industry, that's been particularly hard hit in Florida. About a quarter of UI claims are centered in that industry, which that has an effect on a certain subset of workers. So you definitely see that in the data, and that's the same in Nevada as well, which has been particularly hard hit in this sense.
Davidson: How does this feed through to help inform Fed policy? I know that's kind of a large question, and this thing is still pretty new. But how does this tie to the formulation of monetary policy?
Andreason: In terms of what the tool can tell us about how our leadership—and Fed leadership across the country—sets monetary policy, it's doing what we hoped it would do, which is to provide another view of what's happening with unemployment. And ideally—I'll be honest—hopefully another view of what's happening with recovery. This tool is updated every Thursday. New data is available every Thursday, and we get to see a view of what's happening week to week. Now, weekly data can be messy, so we don't want to draw too many conclusions from any one week, but we can certainly start to see some trajectories in what's happening when we start looking at it over time. Now we know, for example, that in some of the early weeks of the pandemic, 6 million people got on a computer, or found some way—over a phone system, or even potentially some in person—to apply for unemployment insurance.
Now, not all of those people ultimately were granted unemployment insurance. Not all of them were eligible—for some there were challenges. There may have been more, because some people may have delayed filing or some people may have made a mistake and had to refile a week later. So we know that there's a ton of information that comes from that, and that's to some degree a good indicator that that many people are self-reporting some disruption in their job. Now, it's not the only view. We know that there are other things that we use a lot that feed into what you would think of as the "headline unemployment rate"—which is a different survey and a different set of data that pulls that number. It comes just from a whole different category. But the tool helps us do a few things. It helps us see some of the things that are happening on a more frequent and timely basis, since we're getting stuff weekly, rather than that unemployment rate number, which comes out once a month.
But the tool also helps us really understand, and in pretty close to current time, what's happening to specific populations. Now, we know that different demographic groups have long experienced the economy differently. Minorities have dealt with much more challenging economic conditions. Their unemployment rate tends to be significantly higher. One of the things that we're interested in is seeing how that pans out in this pandemic. I'm particularly interested because this is the first time basically ever that we have information on how contingent workers and gig economy workers are recovering or dealing with job loss, and so we're interested in tracking that. Now, I don't have any conclusions to say on it because we have to do a lot more digging to understand that. But those are things that we want to consider, or I hope that we consider, like thinking about who is affected and how they're affected. And that helps us make decisions and is one of the points that I think gets weighed in monetary policy decision making, or could influence monetary policy decision making.
Davidson: Stu, is that where the richness of the data that this tool captures comes into play? You can go look at the tool online, of course, and you can see that it just includes numerous breakdowns—demographically, by industry, and so on. So is that why that's important, because it's basically just offering up—potentially, at least—new insights that may not have been at least readily available before?
Andreason: Well, I'll answer that two ways. I think that some of the new new insights are those around the contract workers, the 1099 workers, and those that are not eligible for unemployment insurance as it existed. So that's not exclusively gig workers, but people that are not eligible. Those are going to be new insights. We have not had data on that before in the same way. There have certainly been people who have approached research on that, but not in the same way with weekly, readily available data. The other data is publicly available, but we're hoping to make it a lot more accessible and usable—to not only experts, but to a much broader audience—and so that's what we really hope to do with this.
Davidson: Well, Stu—or Mels—this may be getting ahead of things a little bit, but is it possible to mine this data and see any signs at all for how likely particular jobs are to come back? There's speculation we've heard from dire predictions of 40 percent or so unlikely to come back, to more optimistic projections. But is there anything in this that that helps to inform those kinds of predictions? Or maybe not.
de Zeeuw: Well, it's difficult because these data are backward looking, and so, for instance, one clue that you could get on that to answer your question is we have data on where claimants have worked in terms of industry, where claimants have worked in terms of what their occupation was. Unfortunately, with the exception of some states those data are mostly monthly, so there's some delay on that. There's just a handful of states that have May data available. So that gives you some clue on where people are continuing to claim unemployment insurance, and where there's actually a downward trend—in which occupations, and in which industries. But it's backward looking, so you need to be a little careful with interpreting those data.
Davidson: Yes.
Andreason: Can I add on to that?
Davidson: Yes, sure.
Andreason: I think that in work that we see closely aligned with the tool—but not directly in the tool—is looking at strategies that the communities and states have come up with to help reskill and rapidly reengage people with work, because often one of the best things that you can do for someone who's gotten on unemployment is to reengage them with work. We're looking to build some frameworks to understand. How do things transfer easily? How do people move from one industry to another, and where are there quick and easy overlaps? We've seen some really wonderful programs in many states that have created those exchanges, and so we're hoping to broaden that and share some of those ideas soon.
Davidson: Stu, could you cite maybe just one quick example or two of programs like what you've described, that seem to be working so far?
Andreason: Sure, I'll talk about one that grew relatively quickly. It went from nonexistent to over 2,000 participants in just a matter of weeks during the pandemic in Tennessee. It's called the Tennessee Talent Exchange, which was run by their Department of Labor. It really looked to actually move people from hospitality industries to transportation and logistics and retail positions that were not affected, like grocery. Now, these are often lateral moves for people, but it keeps them engaged in work, keeps them earning money, having benefits that they need to live their life. So that program has grown really quickly, and they're looking at future phases of it to help people transition to health care and to IT and other industries that we think will be, or the state of Tennessee particularly, sees as strong during this. So that's one example. We've seen that in similar programs in New Jersey and other states that are relying on some of their labor data to do that, and those are very encouraging. So for people that are interested in the Tennessee Talent Exchange, we actually had a webinar with the team from Tennessee that ran and stood up that program in our "Ask Us Anything" series.
Davidson: Cool. So, Stu and Mels, there have been widespread reports of a lot of people who have lost their jobs that had difficulty getting through on the phone, getting through websites, as these state offices have been really just deluged with people. And they, I guess, really were not designed to handle this kind of sudden, huge influx. Do we get a sense from our tool as to how successful people have been at actually getting their unemployment benefits, to which they're entitled?
de Zeeuw: We can get some indication of it. For instance, we added some data on what share of the workforce in any given week has filed an initial claim for unemployment insurance, and you see that some states are still seeing relatively high percentages of their labor force still filing initial claims. Georgia is an example of that. So that might be related to states clearing a backlog of cases. There is data out there on the length of time it's taking states to process, how long it takes states to go from a claim being filed to the first payment. We're hoping to add those data to the tool soon.
Davidson: Yes. Mels, I think this is probably a question for you, just sort of the mechanics of putting this tool together. We don't want to delve too deeply into the weeds here, I think, but it sounds like it was quite an undertaking to pull this thing together so quickly and to set it up so that you had so many streams of data feeding into it. Can you talk just a little bit about what that was like, and how you guys went about building this thing?
de Zeeuw: Sure. I think one of the main benefits of our tool is all of these data are publicly available—most of it is coming from the U.S. Department of Labor—but that doesn't mean that it's easily available. These data are dispersed over multiple files, and some of the data is only available in PDF documents, and so what we're really doing is we've created a program to capture all of these data from the various data files and to put it in one place, and that feeds into the tool that you're seeing.
Davidson: So now going forward, let's assume that at some point things are going to get better. Do we envision this tool being around for the long term, or is it more something that we're going to use as long as the unemployment situation remains kind of dire and then maybe put on the shelf?
Andreason: Well, here's my personal view on unemployment claims: a lot of times that can be a pretty good leading indicator of things that are happening. While even backwards looking, I think it's going to be interesting to watch—particularly some of the specific information in terms of where we see claims starting to decrease and to drop off in terms of industries. That's going to give us some indication of the early signs of some improvement. This pandemic's been so unique. Plenty of people have used the word "unprecedented," but if you think about the definition of the word it's probably true that it's unprecedented. Had this played out like previous recessions, having this tool up and running gives us an idea of things that are starting to unfold before there's some leading indication that things are happening. So I think that the tool is useful both in terms of the current situation, showing what's happening, but also helping us monitor what's happening in the economy going forward. So I imagine that we'll keep it running going forward, so that we can use it to track the recovery and keep an eye on the future.
Davidson: I understand that the tool has proven to be pretty popular and has gotten pretty widespread notice. Do you guys mind talking a little bit about maybe some of the folks you've heard from about it?
Andreason: Sure. Mels, do you want to start off?
de Zeeuw: Yes, I think just a variety of researchers in academia, for instance, have reached out to us and are hoping to use these data. Stu was talking about data points that we make easily accessible, as this data on the Pandemic Unemployment Assistance recipients, data on Pandemic Emergency UC program that's really hard to find or capture, and so we're making it pretty easily available, and we've heard from various researchers that are hoping to do work with those data.
Andreason: And I think that is a conversation that I've had regarding the tool: it's helped national media to understand some of the things that are going on. I've had the opportunity to talk with a number of reporters from national media who are using the tool to understand what's happening, to track what's happening with the new programs, and where there are disparities. And so it's gratifying to know that it's getting use. I'd also say that it's gratifying when we've had the opportunity to talk about the tool with local workforce programs. I got the chance to talk about it with groups that were navigating training programs in a number of different situations, and just hearing some of their thoughts about the things that mattered to them were really important. And we're actually taking a lot of that feedback and thinking about the features that we can add to the tool that help answer some of the questions that we're hearing from them.
Davidson: Very cool. So is it possible that this tool can help to inform not just monetary policy, but also the formulation of policy in real time—let's say, "As Congress or others debate whether the additional unemployment benefits or traditional unemployment benefits are going to expire at a certain point, does there need to be more funding channeled into those?" Could a tool like this help to inform that debate?
Andreason: It certainly could. We don't know exactly what's going to unfold on the policy side—or on the economic side—in the coming months, but we hope that making the data accessible, whether it's at a federal, state, or local level, helps inform those decision making processes. We think it could be really helpful.
Davidson: Okay. Well, guys, anything I've left out? Anything that you guys think is especially important or interesting that you want to bring out?
de Zeeuw: Yes. I just want to add on one thing that Stu was just mentioning about the disparate effects, or differences in how minorities are experiencing this economy. We have data on demographics of UI claimants, and we're seeing some evidence that particularly Black workers are making up an outsized share of those, and definitely in the Southeast. So we were looking at Alabama, Louisiana—where, for instance, 40 percent of UI claimants are Black, though they make up just 25 percent of the labor force. So they're really experiencing some outsized hardship that I think is pretty interesting to policymakers.
Davidson: Do we have any inkling as to why that might be? Does that mostly have to do with what types of jobs are involved there?
de Zeeuw: That's what I think. It depends on the industry mix and on the subset of workers in those industries, and certain industries have been harder hit in this economic crisis. So yes, I think that's what's going on.
Davidson: Yes, wow.
Andreason: And I think those things are the exact types of things that policymakers and organizations that are supporting workers need to think about. They need to realize and understand where there are specific economic pain points and make sure that they're reaching out to the communities that especially need support.
Davidson: Great point. I think that's a nice, though not especially cheery, a nice way to kind of wrap things up. So, Stu, Mels—thanks so much for your time. It was really interesting. I appreciate it.
de Zeeuw: Thanks, Charles.
Andreason: Thank you.
Davidson: All right, thanks for listening. Please go to frbatlanta.org, there's plenty more on this and other topics related to economics and monetary policy, as well as additional podcasts. Thanks so much.