S5 Ep2: Why Is There Not More Spousal Insurance?

Welcome to Conversations on Transformation. My name is Tim Phillips. The 6th annual Steg conference has just been held at Strathmore University in Nairobi and we're going to be bringing you a first look at some of the most interesting new research from it. When one partner, usually the husband, loses their job, it's not unusual for the other partner in a household to go to work to stabilize the household income. We would assume this would be especially true if there was no other safety net available. But is this what happens in practice? New research examines what we call spousal insurance around the world. Todd Sherman of the Minneapolis Feds, one of the authors and Todd joins me now. Hello, Todd. Good to be here. Todd, how risky is it to have wage employment in a low income country because this story all starts with unexpectedly losing a job? Yeah, that's right. And it's just vastly riskier than in rich countries. So what we typically look at, because this is what is really easily available in the data is if I take a wage worker, what's the chance they still have a wage job three months, so one quarter later. And then almost every developed country, you'll see that less than 5% of wage workers will lose their job in three months. So 95% of people still have a wage job three months later. For perspective, like the United States is considered to be a very dynamic and unusual rich country because they're six to 7% of people will lose their wage job. And in developing countries, as we go down the income spectrum, it's between 10 and a few countries are over 20% of wage workers will lose their job in three months. Wow. And then on the flip side, they're no likely to find them. Rich and poor countries, wage workers about equally easy to find. And so what that tells us is that it's just much riskier in developing countries. It's not easier to find, but you're much more likely to lose it. So as you say, this happens more often in low income countries, why in these situations does spousal labor supply matter more than in high income countries? We know that having a head or a husband lose a job is a really serious economic event for most families and most households. And we know that they need to find some way to deal with this, right? Like it's the person loses their job for three months. How do we handle this shock? In developed countries, I think we kind of have some idea. We all know somebody who's lost their job and we kind of know how the family deals with it. And it's some combination of you try to live off your savings. The government may be providing some unemployment insurance, check. Maybe there's some other social safety net type benefits you can access. And we just know that those things don't really exist in developing countries. One thing that we're really interested in is if we go out there and look and develop in countries without all these other options available or all these other margins available, how do families cope with the loss of income from a head losing a job? What do we already know about this about? Well, I mean, the presence of benefits and savings and how they changed that response to unemployment. There's a big and active and now old literature that mostly deals with the United States and Europe, which examines exactly this question. So to fix ideas and to be really precise, think about a situation where I see a bunch of married couples, the husband's working and the wife's not. If the husband loses his job, how much more likely is the wife to start working in the next few months? We know that that's a pretty low rate. If you can zoom out and look at households in total, we just don't see a lot of people joining the labor force. People who dug into that said, well, that's exactly what I just said. It's because you have a really pretty generous unemployment insurance and social safety net in every developed country. We can debate about differences among them. And I know there's a lot of fighting and disagreement there, but every country has something. And even the United States, again, was considered to be this crazy outlier. It's because our unemployment insurance is only six months long. And so what people will say is, well, without kind of insurance, you can take your time, find another job and the family can get through this shock. And the research is mostly, as you say, from the US and Europe, which seems odd because you're not discovering much happening there. Is that because of data availability? Yeah, that's exactly what it is. You need fairly specialized data to investigate. The labor market shocks that families face and how they can ensure that because, of course, you need to be able to see a whole household or a whole family at one point in time. You need to be able to see how they're doing in the labor market. But that's not enough to talk about risks and shock and insurance because you need to see them multiple times. And so historically, that data has been mostly available for developed countries. The United States, as what's called the current population survey, it goes back to 1947 actually. And then Europe from the 70s and 80s, various countries started building these data sets. And now under the European Union labor force survey, they've built this and it really exists through most of Europe. It's relatively easily accessible, high quality data. These are the easy places to look. So what are you and your co-authors trying to discover? So what we would really like to know is if we leave these, I guess, safe confines, I don't know exactly how to describe it, this world where at least we have some unemployment trends. And if we go around the world to other countries with different income levels, different government programs, different labor market institutions, and different social norms, does this pattern carry on? Or are there other countries where actually the household and the family are really important, you know, women and maybe children or elderly parents, anybody joining the labor force is a really important form of insurance for the household. We'd really like to know that. And I think there's just no way to do that before. So this doesn't sound like the easiest thing in the world to do. When you're doing this detective work, where'd you get your data from? Well, how many countries can you get data from what sort of countries? The starting point for this all is work I've been doing with Kevin Donovan, my long-term co-author. Yeah, and it's been like 10 years now. I'm getting old. About 10 years ago, we started to notice that it was not just rich countries that had this data anymore. Developing countries had started to copy these rich country data sets and build exactly what we needed, which is these panel labor force servers where you can track households over time. And we started collecting a few for Latin America, and then we kept digging and digging, and we still keep finding more and more of them. So now we've collected this data for 54 countries around the world and harmonized it. And that includes most of Latin America, as I said, but also there are countries in Africa and more coming every year. There are countries in Asia that have made this data available. And so now we have not just the US and Western Europe, these developed high protection countries, but we have countries with very different labor market institutions, insurance levels, income levels, all sorts of things. So what does this data need to be able to tell you? What are you looking for? We see a household multiple times, and so we're looking for two things. First, we're looking for households where we can see something that looks like a shock. And that would be the first time I see the household, the husband, or the household head, or the primary owner, you know, we have different things we can look at. He's working, and the second time we come back, usually three months later, he's lost his job. That's the first piece we're looking for. And once we find households where they've had that shock, and then what we're going to do is look at everybody else. We start with a wife, but we can look at everyone else in the household or in the family and ask, have they responded to this shock in any way? So the most classic thing is to take a couple of the husband's working on the wife's not. If he loses his job, does she start working? But you can do all sorts of variations on that theme. Imagine the husband and the wife are both working. And he loses his job. How can she help out? The easiest thing she could do in that case is just make sure she keeps her job. We were trying to be really creative and really examinous as thoroughly as she could. So I think in the end, we have like 15 or 20 different figures of different ways to look at this data and think about this question. Are the other household members providing insurance and how they work? When you look into this and you find all these different ways that households are responding, how do spousal labor responses differ across all these income levels that you have now? The short answer is they don't. Wow. Much of my continued surprise. It's been like a year of me just being surprised and thinking I did something wrong. But to give you an example, I said, so husband loses his job. What's the probability that a wife starts working in a typical rich country? It's about 1%. But a developing country? That's about 1%. Basically the same. If you look at a couple where they both work, the husband loses his job. Is the wife more likely to keep her job? No, not in rich countries and also no, not in developing countries? I really wanted to make sure before I started presenting this paper and telling people this and we don't see much evidence that the household labor supply provides insurance anywhere. You don't really want to be presenting it and someone to say, oh hang on, did you check for this and you think, oh that's what I missed? Because you must be really surprised to find that there is no difference. Yeah, this is one of these fun things about presenting research and being an academic because you go out there and you tell people what you're thinking about. And so I've learned two things. One is I was surprised, but not everybody else was. Surprise isn't understatement. I was shocked. Yeah, yeah. Then as I talked to other people, I got two reactions. One is definitely people tell me, oh you didn't think about it right? So when I say we tried 20 things, well the first time we tried 10, but I've gotten many excellent suggestions, some other folks about different ways to look at the data, different ways that wives or other household members could provide insurance. And so that list keeps growing. That's fun. That's fantastic. The other thing though is that people did do a really nice job of giving me different ways to do this problem. They said, well I'm not surprised. And then I would ask, well why not? What did you expect or what did you see that made you think differently? And that's been a lot of fun. And so for example, some people would tell me, well we know that Actually, a lot of this is handled through informal insurance. Right. If you're a local area, you're relatives, you're neighbors, you're friends, can give you a little bit of money to help tie you up. And so they thought, well, maybe they just didn't need it. And that's interesting. That's something you need to think about controlling. There are lots of candidate explanations for this, aren't there? I just want to ask you about a couple of them. The idea that a shock might be correlated, how would that affect the response? Yeah. And one of the things that really popped out in the data to us and that we need to explore a lot going forward. So, you know, one of the surprising findings is if a husband and a wife both work and he loses his job, she's actually more likely to lose her job too. About 10% of wives whose husband lose their job also lose their job. And we really tried to think hard about why is that because you would think they would be desperate to keep the job. Absolutely. I would think it would go down the possibility that they would lose a job. And what we found is that it's particularly likely if the husband and the wife are both in the same industry. So if they both work in agriculture or manufacturing or construction, in those cases, if the husband loses his job, the wife's actually 25% more likely to lose her job. Now, there's a good question of why they're both working in the same industry. Maybe that's what's available locally, but there are times when there's just a local economic shock that hits both of them. Got it. Both of them have lost their job because there is limited availability of work, there's very limited availability for them to explore any outside options, I guess. Exactly. The product typical example would be, let's look at a rural area of the developing country. What's the main job that's available there to agriculture? Yeah. To lean season or it's just a bad year and the husband's trying to find insurance outside of agriculture. Typically, when we're with the wife thing, it's mostly agricultural jobs. We just said those jobs aren't paying right now. They're not hiring. And so I think in those types of cases, you could easily imagine why it's hard for the wife to provide insurance. Now, as an economist, you look for economic explanations, and as you say, they've been hard to find. If the first set of explanations that you're looking for, they don't come up with an explanation for what you found, there are other candidates. Yeah, we hunt for and why, and that's one of the fun things about presenting this research and talking to people about this research. You get lots of ideas. We started this by saying, well, we know that in developing countries, there's a big informal labor market. And so we think this is an area where also it would be very easy for wives or women or other household members to jump into the labor market because they can do these little informal jobs that pay a little money. And people said to me, no, you're thinking about this wrong. In the world where informal jobs are really easy to find, that just tells you the husband will find it easy to find another job. So after we didn't find women working, we went back and said, well, do the husband just find another job extremely fast? And we found actually the answers, yes. In rich countries, if someone like take a husband or a household head, he loses his wage job. Mostly he just remains unemployed until he finds another wage job. Yeah. But in a developing country, between about 60 and 80% of them depending on the country will take a self-employed job, at least temporarily. And they don't pay as well as wage work, but they're not terrible. They're taking like about a 15% pay cut on average. One thing people said as well, they don't take up insurance because they can just insure themselves. Not the household, the husband, the head himself can do that. I could tell you about all the things that didn't work. I guess that doesn't make for such an interesting podcast, doesn't it? It would make a very long podcast by the sound of it. Oh, well, there we go. Do you have weeks for this one? No. Although I'm sure that anyone that's listening to this who wants to go along to see you presenting this paper, well, they can have the conversation about that later. One thing that's present in all of this is that gender lens. We're talking about the male head of the household losing a job and the female in the household fill again. Well, what does this research? Does it tell us anything about labor market shocks and gender inequality? I would say much to my surprise, not so much. Really? When we entered into this project, we really thought that there would be an important gender angle to all of us. And in particular, what we know is that even among developing countries, there are really wide differences in the social norms towards women working. So there are countries like India and Egypt where acceptance of women working are much lower than other developing countries in their somewhere. It's almost to develop country levels. And so I told you about the results that household labor supply isn't important insurance in developing versus adult countries. We also spend a lot of time looking at other angles like countries where it's okay for women to work. It's widely accepted versus countries where it's not. We didn't find big differences there either, much to my surprise. And again, I really thought that we could and maybe would, just not from there. As this detective hunt goes on, one thing that doesn't change is that unemployment insurance, it is an expensive policy in a low income country. What does this research tell us about whether policymakers in low income countries can should rely on households to self-insured? We're saying it happens. Are you saying that this is successful? That's a tough question to answer so precisely. I would say it's been much more successful than I expected it. And I say that based on two things. One is that we see these heads or husbands jumping back into work fairly quickly and earning maybe 15% less than they were earned in their last wage job. The second thing is we don't see any sign that anybody else is joining the labor force. If you thought that these households were really being crushed by this risk, then I would expect some reaction from, again, we looked at wives, we looked at children, we looked at parents, we looked at everybody else in the household, and we're just not seeing any of them react very much. That says that there's at least some margin of success here. I want to be a bit cautious about that because in developed countries, a lot of the conversation is not about just finding a job, but about finding a good job. We want to give people some time with unemployment insurance so they can look for the right work. And when we see people taking all these very low-end self-employed work jobs, I don't think anyone thinks those are great jobs, maybe unemployment insurance could have a role in helping them find better work. The other thing that I would caution developing countries about unemployment insurance is expensive, and we know that they have really severe budget constraints. They have so many wonderful ways that they could potentially spend money and they have a hard time collecting taxes. So if they pass a policy like this and they cost a bunch of money, what are they going to be tempted to do? Well, the main way they can raise more revenue is by cracking down on the informal sector. Yes. So you can get more people out of the informal sector and into the formal sector than they pay taxes and you get more revenues, but the informal sector is exactly the jobs that people are currently using for all of this wonderful self-insurance. It's the jobs that they're taking up to ensure themselves against the shocks. And so I would be a bit nervous because I think you could easily add up a situation where you provide a very expensive government unemployment scheme, and you do that by eliminating exactly the jobs that they were using to ensure themselves. I don't know how far you'd come in ahead in that decade. This is an ongoing research project. I don't know. Maybe I'll try to answer that question. There's some interesting questions there about how good this is and how good the government run alternative. It is one of the most interesting null results that I've ever discussed because it leads to this conversation that we're having about how people really cope, under stress and adversity, that is of course very productive. Thank you very much for talking about it. Thank you so much for having me. As Todd says, it's a work in progress. So look out for it, working title, spousal insurance around the world, and the researchers and the backer Kevin Donovan, Philip Grubener, Lucas Nord and Todd Schoen. You have been listening to conversations on transformation, featuring research presented at Steg's annual conference, 2026. Steg stands for Structural Transformation and Economic Growth, and you can find out more about it by visiting CEPR's Growth Research Platform at grp.cepr.org/steg.