A Discussion of Unemployment's Impact on Family Welfare
Tom Heintjes: Hello, and welcome back to another episode of the Economy Matters podcast. I'm Tom Heintjes, managing editor of the Atlanta Fed's Economy Matters magazine, and today I'm sitting down once again with Julie Hotchkiss, a research economist and senior adviser in the Atlanta Fed's research department. Welcome back to the podcast, Julie.
Julie Hotchkiss: I'm glad to be back, Tom. Thanks for inviting me.
Photo: David Fine
Heintjes: Julie recently published some new research that we're going to talk about today. Julie, your paper's title is "Family Welfare and the Cost of Unemployment," and ever since you mentioned this research to me some time ago, I've been eager to get you into the studio to talk about it. So I'm delighted it's happening today. Oh, I also wanted to mention your coauthors on the paper: Robert Moore of Georgia State University and Fernando Rios-Avila of Bard College. Let me just jump right in, Julie—what led you to undertake this research?
Hotchkiss: Well, Tom, the Federal Reserve Act of 1913 was amended in the 1970s to explicitly provide for policy goals of the Federal Reserve System. One of those goals is to maintain full employment.
Heintjes: Well, now, as a labor economist, you can tell me what full employment looks like, right?
Hotchkiss: Well, actually it's really one of those things that we know it when we see it. But the implication of the policy mandate is that high levels of unemployment are undesirable. High levels of unemployment are typically a signal that our economy is not growing at its full potential, and—in the extreme—that we’re in a recession. So what we wanted to do with this research was to take a macroeconomic event—an increase in the aggregate unemployment rate—and assess what the implication of that would be for families at the microeconomic level.
Heintjes: It sometimes seems, Julie, to the casual observer at least, that monetary policymaking is something of an abstraction. But this research of yours reinforces the fact that policymaking does in fact have a definable impact on people's lives. Was that part of your intention behind your work in this case?
Hotchkiss: No doubt. The discussion surrounding monetary policy decisions is sprinkled with statistics like the unemployment rate, the inflation rate, GDP [gross domestic product] growth—these are macroeconomic terms that describe the condition of our economy as a whole. What's behind these statistics, of course, are people—someone losing their job means that there's a higher unemployment rate. A higher rate of inflation means that you and I, for example, will pay more for a loaf of bread this year than last year, and slower GDP growth translates, perhaps, into a smaller economic pie to be divvied up among our children than what we've enjoyed.
Heintjes: I guess it seems sort of obvious that a rising unemployment rate would come with a cost to affected families. But what does your work show us, apart from what seems sort of intuitively obvious?
Hotchkiss: Well, that's the beauty of economics, Tom, is that it's so logical. In economics, if we observe someone working, we assume that that person wants to work—that the job provides them more welfare, or utility, or even happiness, I guess you could say, from the income that they earn than the welfare that they lose from having less leisure or time to do other important things in their lives, like grocery shopping.
Heintjes: So if they lose their job...?
Hotchkiss: Well, if they lose their job we expect them to suffer a cost, a loss in that welfare or happiness. And of course, that's what we find.
Heintjes: So you looked at some nonfinancial aspects of this kind of situation?
Hotchkiss: Well, rather than just measure the cost of unemployment by the income that they've lost, our economic model also takes into account the gain in time to do other things—like exercising, or watching TV, or even doing housework.
Heintjes: All exciting.
Hotchkiss: [laughs] Very much. Obviously, the welfare gain from more time not working is less than the welfare lost from income, and that's why we estimate a cost to rising unemployment.
Heintjes: Well, in this case, how did your methodology differ from other researchers' attempts to quantify the effects of the unemployment rate?
Hotchkiss: Our approach to measuring the cost of unemployment differs from what others have done in two primary ways. First of all, as I just mentioned, we don't just take into account the income earned from having a job as increasing a family's welfare, but we also account for the value of time spent not working. So the income loss from losing a job is somewhat offset from a gain in time that can be spent doing other things.
Heintjes: So that's one way.
Hotchkiss: The other way our methodology differs is that in calculating the cost to a person or to a family of the unemployment rate increasing, we don't just assume someone loses their job. We take into account the probability of the person losing their job, so what we end up with is the expected cost of the unemployment rate going up.
Heintjes: OK, you're going to have to help me a little with this angle.
Hotchkiss: Well, think about it this way: the cost of maybe losing your job is less than the cost of definitely losing your job.
Hotchkiss: So when the economic environment is such that the unemployment rate is rising—say, during a recession—everyone has a higher probability of losing their job. We estimate the probability for each person separately based on the person’s age, education, and so forth. And we believe that this gives us a more realistic assessment of the expected cost of a rise in unemployment.
Heintjes: OK, well, that helps, so thank you for that. Julie, what was the source of the data you used in your research?
Hotchkiss: We use a very well-known monthly survey called the Current Population Survey . This is the same survey that the Bureau of Labor Statistics uses to calculate the official U.S. unemployment rate, and we use the data for single and married households. Our data doesn't know really how to handle same-sex couples, so they aren't included in our analysis. But we hope to include that category of family into future analyses.
Heintjes: Right. Does a declining unemployment rate bring equivalent benefits to families that a rising unemployment rate brings in negative impact—in other words, is there a sort of inverse relationship at work here?
Hotchkiss: Our model isn't really designed to consider a reduction in the unemployment rate. It's easy to conceive of the probability that someone who has a job might lose it as the unemployment rate goes up, but conceptually the model can't really consider forcing someone who doesn't have a job to have a job when the unemployment rate falls—which would be the inverse equivalent to what we are doing here.
Heintjes: Your paper states that the annualized welfare loss generated by a 1 percentage point shock to the unemployment rate would equal $1,156, on average, across all families. That's a pretty specific figure—how did you arrive at that?
Hotchkiss: Well, the exact figure may be a bit misleading in its precision. Of course, all statistical estimates are just that: they're estimates, with a plus or minus associated with them. But this average cost is what we calculated based on the probability that a person loses their job, the income that would be lost, and the value the person places on income relative to the value they place on the time spent not working. So it is important to know that this figure does not necessarily represent any one family's experience, but is an average across what each family might experience.
Heintjes: Right; so it's not a "one size fits all" figure.
Heintjes: Well, generally speaking: how would this level of impact or this type of impact differ across families with, say, different incomes and different levels of education?
Hotchkiss: There are two driving forces behind the assessment of how a rising unemployment rate affects families. The first is the probability of actually being hit by unemployment, and the second is the amount of income that would be lost if the person becomes unemployed. These forces are sort of moving in opposite directions as education rises, so someone with a higher education level will lose more income on average than someone with a lower education level when they become unemployed. But also, someone with a higher education level is also less likely to become unemployed.
Heintjes: Julie, do we know how the significance of this lower likelihood of unemployment compares to the potential income loss?
Hotchkiss: As it turns out, it looks like the higher income is more important than the lower probability of unemployment, since we find that the average cost is higher for those with higher levels of education. So again, on average—I feel like I'm repeating myself on that point, but—on average, across all families, those with a college degree face an annualized welfare loss that is two and a half times greater than those with just a high school degree.
Heintjes: That's a great point—one worth repeating. In your paper, you look at different types of families when you assess the impact of unemployment on married families and single people. What differences did you find in how they are affected by rising unemployment rates?
Hotchkiss: We estimate that married families experience a significantly greater annualized loss in welfare than single families, in the face of a percentage point rise in the unemployment rate.
Heintjes: I guess I would have expected the opposite finding—if you're the sole breadwinner and become unemployed, the cost is enormous versus half of a couple losing a job. Am I misreading your findings here? [laughs] That's always a possibility.
Hotchkiss: No, I understand exactly what you're saying. Again, remember the two forces driving the welfare loss from the rise in unemployment rate: there's that probability of being hit by nonemployment, and then there's the income loss if you are hit. So the average probability of at least one person in the household being hit by nonemployment increases more for a married family when the unemployment rate rises than it does for a single family. So the probability increases by nearly a percentage point for single households, and by just under 2 percentage points for a married family. You can kind of think of it roughly as a married household having twice the chance of being hit by nonemployment.
Also, on that same point—married individuals earn more income on average than single individuals. So with both a higher income to potentially be lost, and a higher increase in the probability of being hit by nonemployment when the unemployment rate rises, it makes sense that we should see a higher welfare cost for married families.
Heintjes: Sure. Julie, I know that researchers go into their work with an open mind regarding what the data will tell them, but even so, did any of your findings run counter to any of your own notions about the labor market, which you've studied for many years? I mean, you've spent your entire career examining the dynamics and the way the labor market works.
Hotchkiss: Well, that's a great question, Tom. One of the things I really enjoyed about this project is that it's an area that I haven't really done much research in before. My coauthors and I have estimated similar models to address different questions, but this is the first time we've applied this model to address the question of the welfare cost of unemployment. So methodologically there really weren't any surprises. But from a policy perspective, I learned a lot from this project, which I guess is why I like what I do so much—because there's really no end to what I can learn.
Heintjes: Right. So, did any of your findings jump out to you as especially remarkable?
Hotchkiss: One thing that first surprised me to see was the estimate of the average expected cost to rising unemployment of roughly $1,100. I guess to someone who's steadily employed this may not seem like a very high cost. But just about the time that we were doing our estimations there was a CBS MoneyWatch report, I think, that came out and said that 63 percent of Americans reported that they don’t have enough savings to face an unexpected expense of between $500 and 1,000. So it really put that into perspective for me, and a welfare equivalent of about $1,100 per year would actually be quite significant.
Heintjes: Well, Julie, there's a longstanding understanding of the relationship between the inflation rate and the unemployment rate—that is, traditionally the understanding was that when unemployment is low, inflation will rise. Does your work have anything to say about this supposed relationship, or its implications for family welfare?
Hotchkiss: Well, I'm really glad you asked that, Tom. Even though we don't have inflation per se in our model, one of the coolest things about the model is that we can simulate multiple types of shocks to see how they compare from a family-welfare perspective. So, specifically related to this unemployment/inflation welfare tradeoff question, one of the things that we do is to ask what price level change would result in the same welfare loss to families that a 1 percentage point rise in the unemployment rate produces.
So if we can assume that people feel similarly about inflation—which is a long-term, dynamic concept—as they do about a one-time, unanticipated shock to prices, then we can get something kind of like an unemployment/inflation tradeoff estimate.
Heintjes: Well, what did you find about this tradeoff?
Hotchkiss: We find—again, on average—that families suffer the same welfare loss from a 1.8 percent increase in prices that they suffer from a 1 percentage point rise in the unemployment rate.
Hotchkiss: But given that singles experience a lower welfare loss from a rise in unemployment, it also means that they wouldn't be willing to endure as large a price shock in order to avoid a rise in the unemployment rate as married families would.
Heintjes: Let me ask you this, Julie—why would your findings be of interest to monetary policymakers?
Hotchkiss: Even more important than the actual estimates of the dollar equivalent welfare losses, etc., or how much of a shock to prices families would be willing to endure to avoid an unemployment rate increase, is the finding that a rise in the unemployment rate imposes a differential cost on families depending on their demographics. Married families suffer a greater expected loss than single families, and those with higher education suffer a greater expected loss than those with less education. And as a percent of total income, higher-income families experience a greater loss from a price shock versus a shock to unemployment, and vice versa for low-income families.
There's a differential impact across multiple demographic dimensions, so just being aware that monetary policy, which can be expected to influence both inflation and unemployment, affects families differently will hopefully make for better-informed policymaking, which is always a good thing.
Heintjes: Indeed. Well, Julie, this has been a great conversation, and I really appreciate you spending some time with us today.
Hotchkiss: Thank you, Tom. It's been a pleasure—and you don't have to ask twice to get me talking about my research, so any time. [laughs]
Heintjes: You told me about some future work you have coming out, so I know we'll have you back on for that.
Hotchkiss: I'm looking forward to it.
Heintjes: I also want to note that we have a link to Julie's paper on our website, frbatlanta.org, and I encourage you to check it out. It's a very interesting and unique look at the dynamics of the market.
Well, we're at the end of another episode of the Economy Matters podcast. Please join us next month when we’ll sit down with Atlanta Fed economist Larry Wall to talk about the complex relationship between regulators and the agencies they regulate. I'm Tom Heintjes, managing editor of Economy Matters, and thanks so much for spending some time with us today.