Charles Davidson: Welcome to another Economy Matters podcast. I'm Charles Davidson, a staff writer with the Atlanta Fed's digital magazine, Economy Matters, and today I'm going to be talking with David Wiczer, a research economist here at the Atlanta Fed. David, welcome—and thanks so much for your time.
David Wiczer: Yes, I'm really happy to be here.
Davidson: So what we're going to discuss is a forthcoming paper 
—it's forthcoming. Is that right?
Wiczer: Yes—it should hopefully be out soon.
The Atlanta Fed's David Wiczer. Photo by Stephen Nowland
Davidson: OK. A paper on Social Security Disability Insurance—SSDI—and the effects of it, with a drilling down on the geographic areas where big shares of the population tend to be disability recipients, and some of the implications of that. And David, I think you wrote this paper along with Amanda Michaud of the Federal Reserve Bank of Minneapolis, and another colleague—Timothy Moore at Purdue University. First off, David: again, welcome. Thank you for your time. What drew you guys to this subject matter to begin with? And what's important about understanding these sorts of nuances in the workings of the Social Security Disability Insurance program?
Wiczer: I think the answer to the question is in the second part of it—that it's hard to understate how important Social Security Disability Insurance is. Along with its sister program, SSI [Supplemental Security Income], it's the only program that we have that provides long-term income replacement to people in the US. It's this vital part of the social welfare program that we have—and also, it's just gigantic. At any given time, there are more people getting SSDI than there are getting unemployment insurance. So when we think about the sorts of programs that help people when they're down, Social Security Disability Insurance plays this unique and exceptionally large, important role in the world. And then I guess, just some personal background: I'm kind of an economics-y child of the Great Recession, which was the peak of this two-decade rise in the program. And so, it was when Amanda, my collaborator, and I were graduating from grad school, it was like a really important question: What was behind the rise of this program, and what was its significance?
Davidson: That was during the Great Recession, so presumably lots and lots of people were out of work, so supports were in great demand. So, we're talking, what—8, 9 million people at a given time, receiving that disability?
Wiczer: Absolutely.
Davidson: So from the work you guys did here, David, what do you view as the most significant findings from this, or insights that you came away with?
Wiczer: Well, this is part of a larger program of research on it, and I think some of the biggest outcomes are just to see how concentrated the recipiency is—which then leads one to think that this is very important for some targeted set of places. Doing the further, even more in-depth research, it points out that it's even more important than you would think on face value. And that's because of this basic idea that economists have—that when you don't have that much, having a little bit more is particularly important. And so disability insurance is really important in places that are impoverished—places that have very high need for it. And so then, some of the tools that we put onto this question, let us say not only that it's very concentrated there—it's particularly important and well appreciated there. And so it's this vital program that's even more geographically important in some places.
Davidson: Speaking of geographic concentrations: I think I read a term in the paper from earlier researchers, I guess, that identified what they termed a "disability belt," and that, I guess in a way, sadly, includes a lot of our district. So, your paper suggests that that's still the case, right—this disability belt is still intact?
Wiczer: Absolutely. And some other work that we did showed kind of the evolution of the geographic dispersion, and basically the summary of that is that places that start with high levels of disability, stay with high levels of disability. So it's very persistent in terms of ranks. The highest will consistently be the highest. But then as the number of people on the program has risen, it's risen the most in the places with the most recipiency. So this "disability belt" has become more and more concentrated, actually.
Davidson: Why is that? I mean, do we know the causes of that? Are these places that have really suffered from deindustrialization, or that kind of...
Wiczer: Yes, yes.
Davidson: A fascinating nugget I've come across was this idea that over the past several decades, as the economy has evolved to become much more service-oriented, that's accrued really disproportionately to metropolitan areas where lots of people concentrate who do that kind of work. And therefore, other areas are, to some degree, left behind. Is that the kind of phenomenon that is behind this?
Wiczer: This is such a good question, and it's really one of the formative questions of the research agenda. And so, when you asked earlier, "What are the insights?"—that was one of the big things that we set about trying to find out, is what was driving it. And it's because, just like you said: these are places that are rural. Southern America is experiencing all sorts of economic shocks—the deindustrialization, very well documented—but they're also relatively sick places. That important work by [Anne] Case and [Angus] Deaton about these "deaths of despair" are really happening to a great degree there—and I guess I should say "here," rather. And that's also pretty well documented and measurable. And so one of the things that we're doing in this research is trying to kind of disentangle what's driving the dispersion, whether it's the health need, the economic need, or actually there are some interesting nuances of the program that also kind of shape it, such that the—we call it the "replacement rate"—the amount that you get paid relative to the average earnings that you could get by working in a place, varied tremendously. Just to give a ranked ordering of these, the most important, actually, driving geographic dispersion is health differences. Secondarily to that are these economic differences, and behind that are differences in program design that kind of unintentionally favor some places over the other.
Davidson: Right. So I guess it's kind of a whole other research agenda, maybe, to figure out all the connections between these economic issues and health.
Wiczer: Absolutely. That footnote was happening in my head, too.
Davidson: Clearly, they can't be completely separated, but I think that's a different podcast. But what I did want to do was to try to get at maybe how significant the impact of the program is, and in order to try to get at that I want to read a passage from the paper, and I want to ask you to give me the upshot, in laypeople's terms, if you would: "The overall size of DI"—disability insurance—"and the degree of dispersion results in estimated local welfare effects that are much larger than for specific, place-based policies or state transfer programs." Now I want to ask you what the welfare effects are, and what those other programs may be. And now reading again, "These results highlight that federal programs focused on protecting individual risk can have large, economically important effects, that are amplified by geographic heterogeneity and cross-subsidies." So can you just kind of take that, and unpack that a little bit? What are we getting at there?
Wiczer: I mean, first, we should just appreciate the poetry behind those words, right? [laughter]
Davidson: It just rolls off the tongue.
Wiczer: I know. Usually I read "heterogeneity," but when you say it out loud, you really see how much of a mouthful it is.
Davidson: It's a tricky word. It's a roller coaster, there.
Wiczer: Exactly. You picked really the crux of both the motivation for the work and what it's about. So first, it's that the welfare dispersion associated with this program is very large. So what we're basically saying is that this concept that I was talking about earlier—that the dollars spent in poor, rural counties are particularly needed, particularly appreciated, and particularly important. And so, that's what we're summarizing when we say the "welfare effect." It's a slightly different use of the word "welfare" than I think we often think of.
Davidson: Right. So, $10 in a county in South Georgia that meets some of these criteria is far more impactful than $10 spent in an affluent part of the city.
Wiczer: Exactly—especially with this program in particular. It's not just $10 dropped out of a helicopter. It's $10 given to somebody who's going to have a difficult time working, and also who's...you don't get on this program unless you have health problems. So it's $10 given to a person in need and a place in need. And so that's where you create very large welfare impacts. We sometimes play it, being macroeconomists—we were then trying to say, relative to a lot of the other things that the government does, which actually are very geographically different. One of these is like the American Recovery and Reinvestment Act, during the aftermath of the Great Recession. It didn't go evenly to every place. The line was like "shovel ready," so places with more roads got more money, etc., etc. The difference in magnitude between places that didn't get much and places that got a lot in that program are just dwarfed by the difference in magnitude comparing a high disability recipient county to a low disability recipient county. This is both the largest way in which we create geographic transfers from rich places to poor places, and just by an order of magnitude, it's bigger.
Davidson: I guess this question stems, to some degree, from that answer, and that is, is the program basically serving the intended purpose of improving the general welfare of these places that apparently do need some help? And stop me if this is beyond the scope of the work or beyond what you're comfortable addressing, but did you guys find places where the program could be tweaked and made better?
Wiczer: Yes. This is a really interesting question, and my common answer when somebody asks me, "Is it a good program or not?" is to say, "Don't ask me—that's above my pay grade." [laughter] But since it's my podcast, I'll speculate a bit.
Davidson: You're the expert in the room, so it's closer to your pay grade than mine—have at it.
Wiczer: We can stab at that in a couple of directions. There's this overarching question of: What's the purpose of the program? You get an award because you have a health problem—but also, it's serving this dual purpose of providing income replacement for somebody who has a difficult time working. And that gray area is really tricky for the program to disentangle, and I think even for a program designer to disentangle. A fun fact along those lines is that early in the program, in the ‘80s—it's actually much older, but it was redesigned in the ‘80s—10 percent of cases were then adjudicated after a rejection by a work specialist. So this work specialist would say, "Oh, David hurt his leg, and in order to be a research economist, he has to have a working leg." I'm not a very good work specialist, but they would say something like that, right? And nowadays that's 90 percent, so a vast majority of the cases that go into adjudication are being adjudicated because of some dispute over, what's the interaction between a health problem and the work requirements? So is this a well-designed program? You might imagine clarifying, we have some income replacement for industrially affected regions, and we have a health replacement. Instead, we have something that's kind of a hybrid. And so I think that adds some complication to "is it a well-designed program?" Is it doing its intention? Kind of yes, and kind of no—it has these broad intentions, and it's doing something to that effect. The second part of your question is also really interesting, and to that we actually answer it within the context of a fairly sophisticated macroeconomic model, which is: could you tweak it? And the way that we ask that is: What's the marginal value of the next dollar spent?
Davidson: So "marginal" meaning "above the last dollar spent"?
Wiczer: Exactly. We're spending a lot more of Social Security Disability program dollars in rural Arkansas and rural Georgia than we are in metro Atlanta. I spent a lot of time already saying that there's a lot of value in these rural, poor places, but is it the right level? And actually, one of the things that we show is that you could redistribute a little bit away from them and get a little bit more value. That's not saying that we should have the same dollars going everywhere, because clearly the need is higher in rural Georgia than metro Atlanta. But it might be a little bit more skewed than it needs to be.
Davidson: Got it. A related question: This program is clearly a critical component of the social safety net, and badly needed. I think that seems beyond dispute. And I guess the average recipient—correct me if I'm wrong here—is about $1,200 a month. That's what the average recipient receives.
Wiczer: Yes.
Davidson: So nobody's out there getting rich off this, clearly. But at the same time, I wanted to ask if this is a plausible question, whether the preponderance of recipients in these kind of rural, impoverished places—is it fair to ask whether that might be contributing to keeping those places mired in this cycle, generation after generation, of people on disability, and the people—as well as the place itself—not maybe coming along a little bit better economically?
Wiczer: This is a really interesting, important point that we kept on kind of walking up to, and then deciding not to hit. I'm now going to speak a little bit beyond the research, but there are program design reasons that would lead you to say exactly what you're saying. So, for instance, when somebody is receiving disability insurance, they're not really allowed to work anymore. There's this cap on how much money they can get from work. So just imagine a place where, at the top end of the disability receipt distribution, 10 percent of the working-age population is getting disability. So, you're talking about pulling 10 percent of the labor force out, and that has to have a large effect. The other thing is that the 10 percent who are not working are spending less. There is, as you said, it's $1,200—that's less than they would be making if they had a job. And so I think a reasonable person could absolutely come to the conclusions that, for the local economy, if you were to somehow wave a magic wand and put people from disability insurance back into employment, that that would be a stimulative effect.
Davidson: But presumably, they're on the program because they need to be, so in a way that's a purely academic question, to some degree.
Wiczer: Exactly—and that's why it's a magic wand. And a really interesting line of work is saying, for the set of people who get it, what would be their alternative? And just as you said—not speaking academically now, speaking of the real world—the people who get this program don't have that many other options.
Davidson: I know people who are on it, and they need to be, for sure.
Wiczer: Yes, exactly.
Davidson: So, another point in the paper that I find fascinating—and not to be morbid here, but there seems to be some sort of connection between mortality rates for, say, a county, and the portion of disability insurance recipients among that county's population. So, can you explore that a little bit, David? Am I right to find that interesting?
Wiczer: Certainly, you're right to find that morbid. [laughter]
Davidson: What about interesting, though?
Wiczer: Yes, and also very interesting. And I particularly like that question, because I thought it was a clever thing that I did in it, actually. [laughter] So there's not that much county-level data on health outcomes that we could use for this paper. There's a lot of data in hospitals, but it's difficult to really link a hospital to a county, etc. So we used mortality to be kind of a summary of the overall health of the place.
Davidson: High mortality—or a lot of young mortality—means the place is not especially healthy.
Wiczer: Exactly. I think that's interesting, both because academically, that's a useful way of very fine grained-ly measuring the health in a local area—but you're also right that this is pointing to the program fulfilling one of its stated purposes, which is that in places with high mortality, there's bad health associated with that. And this is also where disability concentrates.
Davidson: Yes. So, David, you noted earlier, I think, something about this paper in the larger context of a bigger body of research. Can you talk a little bit about that larger body of research, and how this is meant to advance that kind of research enterprise?
Wiczer: Yes, absolutely. I think it's part of two big literatures. One of those is just thinking more about disability insurance. What we're trying to put forward is, I think, a fairly new take on the importance of its geographic dispersion. Disability insurance is not just something that's happening more now than it used to be, or to some set of people rather than others, but it's happening in places. And because of our focus on places, we're actually able to talk to some of the earlier studies. Some of the really pioneering work looked at what happens when a lot of people get disability in, for instance, West Virginia coal-mining towns. One of the things that we point out is that, actually, that's not particularly representative. That's an area where it's very high receipts, and high-receipt areas actually look really different, and respond to disability very differently, than low-receipt areas. So Washington, DC—if somebody waved a wand and said, "Let's put more people on disability in Washington, DC," it has a different effect than in West Virginia. And then the other body of research that I think is really important that we're trying to speak to is this place-based policy. And that's, again, to say that programs from the federal government are federal, and they have different effects in different places, and maybe we can think about that.
Davidson: Yes. David, what you just said, I think, gets at something that often strikes me about not just this, but a lot of economic research, and that is that in a lot of cases, you turn up some stuff that's not especially surprising. In this case, a larger share of the population in impoverished counties with a lot of bad health outcomes are on disability insurance—OK, makes perfect sense. But then, far more interesting, I think—in my view, at least—below the surface, maybe, you come up with other stuff that's maybe a little more nuanced and interesting. Talk a little bit about those kinds of findings that came out of this work.
Wiczer: Yes. Well, I think one of the first nuanced findings is just this idea of disentangling. Are these impoverished places, which also happen to be sick—are they getting more disability insurance because they're sicker, or because they're more impoverished? And that, we can use kind of an interesting variation where you actually see places that are healthier but poorer, or richer but sicker, and we can kind of disentangle that. The other bit that I think is a really maybe nerdy sort of a point, is that we call this "insurance." I actually got into a fender bender recently, because it's Atlanta. And for that, I had insurance against something that I did not expect to happen, and it helped me in a time when I needed it. But one of the features of this program is that it's not just insurance at the individual level. There are places where you're much more likely to get that payout than others. And yet, in terms of insurance, the premium doesn't change. So, in that sense, it's also just a transfer, on average, from one place to another. Understanding the bit of the program that is insurance—bad thing happened, health-wise or employment-wise, and this program stepped in—versus the part of the program, which is just on average if I live in this area, I don't pay more for this coverage, but I'm more likely to get it. I think that's one of the important nuances of the research that you probably don't need to do quite as many computer programs as we did to figure out, but I think takes a bit of thinking.
Davidson: David, again, a little bit along the lines of the larger body of literature—now you've, in the past, I think, looked a little more broadly at other types of fiscal transfers, stuff like what we often refer to as stimulus payments to households, like the CARES Act, in the thick of COVID. Correct me if I'm wrong on this, but from my reading, you and your co-authors found that most households—and especially lower income, lower wealth households—tended to use that money not to go out and buy stuff, but rather to pay down debt.
Wiczer: Absolutely.
Davidson: And I think, to a degree, at least, that sort of belies the conventional economic wisdom that lower-income people have what economists call a higher propensity to consume. You get a little extra money, "Why am I going to use it to go out and buy stuff, because I need the money"—versus a more affluent person who thinks, "Oh, a little 100 bucks extra: big deal. I'll just put it in the bank. I don't need it." So I wanted to ask you: Have I got that right in general? And how might a finding like that inform future research or policy discussions?
Wiczer: I'm so happy you asked about that project. You summarized that perfectly. We looked during the COVID period at the checks that went out, asked people what they did with it, and also asked them about their asset and income position. And what we found was that people with more debt—lower net assets—disproportionately said that they used these checks to pay down their debt. And that was surprising. We spent a good couple of weeks trying to figure out where we made the typo, because this conventional wisdom is that the less that one has, the more likely they are to go out and spend it because they need things. And the kind of conclusion that we started to draw is that first of all, this fact is correct that, indeed, people with more debt pay down debt. And the reason that they pay down debt is because it's very expensive to have a lot of debt. If I'm carrying a large balance on a high-interest rate credit card, it's really good for me to use an extra $1,200 to pay that off, because I save on debt service for a long time into the future.
Davidson: So that's a smart sort of calculation, really.
Wiczer: Absolutely. So, I think it highlights the financial savvy that you have to imbue people with, when you think about their behavior. I think it also changes a lot about how we think about targeting programs. In the past conversation, I've talked a lot about how a dollar in a place that's impoverished, to a person who needs it, is worth a lot. And in a lot of the prior thinking about, "When and where do you spend stimulus payments?" there was this beautiful coincidence where the dollar spent by the government was also the highest impact, in that it was the most likely to get turned around. If you give somebody very much in need of a dollar, they'll immediately spend it on a haircut, and then the barber spends it—we get this multiplier effect. What we're saying is—not to be too economist-y, but—it's not so nice, right? The value is still there. The person who needs it still gets the most value out of that dollar, but they're not necessarily the most likely to turn around and spend it. So you don't get to go to sleep at night thinking that you did everything perfect. You actually have to think about that trade-off, as a policymaker. And so I think that changes a lot about how we think about targeted policy during downturns.
Davidson: Is there a way to sort of weigh—that's an awkward phrasing, but—is there a method by which to weigh, say, if people are going to pay down debt, that's obviously a very responsible, fiscally wise thing for them to do. It's going to better their welfare. So, to balance the benefit of that versus if they get money and go out and spend it, as you mentioned, and it's going to have a knock-on effect, they're going to buy stuff and then the recipient of that, the barber, is going to buy, and so on and so on. Is one better than the other? I suppose the answer is, you'd have to ask each person, they might have a different—you know what I'm getting at, I think. So, if you're going to design these policies, do you shoot for one or the other? How do you make sense of that?
Wiczer: I think that's a really interesting question, because before I said, "no trade-off, just easy answer"—and now I say there's a trade-off. So, think about the trade-off. I need to weight these—at least within the kind of quantitative work that we do, it's better still to target at people with the highest welfare gain, even if they're going to pay down debt. And the reason actually has kind of a nice intuition, which is that in the past, the past thinking was, "If I give a person in need this dollar, they'll go out and spend it and get a lot of value from it." Now, what we're saying is actually revealed preference—to use another economist-y term—they would have gotten a lot of value from spending it right away, but they chose not to. They chose to pay down debt. They chose to save it, in a way. They must have gotten even more value from it. So whatever answer you used to say about how important it was to target at the bottom, is actually now even more valuable. These ancillary, cascading effects of multipliers, those tend to be fairly short lived, and there's a lot of conflict within the profession about how large those are. So the initial impact is even more important, and then these ancillary ones are ancillary.
Davidson: Yes, all right. Well, David, this is great stuff. I could go on and on, but I think we probably need to come to a conclusion here. And I just wanted to note that you're going to be presenting this paper at the Minneapolis Fed in—it's October, is that right?
Wiczer: That's right.
Davidson: October at the Opportunity and Inclusive Growth Research Conference at the Minneapolis Fed's institute that focuses on those kinds of matters. But again, David, thanks so much—this was a really interesting chat. I enjoyed it.
Wiczer: Yes, thank you. I had a lot of fun. Thanks for listening to me.
Davidson: Okay, great. And thank you folks for listening. And please come back often, and visit our website at atlantafed.org for more material like this and lots of other economic goodies. Thanks so much.