Skip to main content

Interview with Anne Case

Princeton economist on the cost of AIDS in South Africa, “deaths of despair” in the U.S. and women in economics

December 12, 2017


Douglas Clement Managing Editor (former)

Article Highlights

  • Princeton economist on diverse work, women in economics
  • Research addresses child health, middle-age mortality
  • AIDS epidemic in Africa her focus
Interview with Anne Case
Anne CasePhoto by Peter Tenzer

“What I love about economics,” says Princeton economist Anne Case, “is that you can follow your nose. See where data and theory lead you. Discover things you never dreamed of.”

Case’s intellectual curiosity has led to a range uncommon in economics. The Princeton scholar has developed econometric theory, pioneered epidemiology, estimated technology diffusion models and explored political economy, opioid addiction and funeral expenditures. The work has taken her from India to Indonesia, from Albania to Africa, and now to Appalachia.

Some of her earliest studies involved the influence of neighbors, loosely defined, on teen behavior in Boston, sickle adoption for rice harvesting in rural Java and tax policies among U.S. states. Her focus turned to health and education in South Africa, particularly the role that AIDS has played in shaping that nation’s economic future. Another research agenda explores the extent to which one’s health as a child determines later-life outcomes, and the relationship between child well-being and consanguinity.

Most recently, Case has documented that middle-aged, non-Hispanic whites in the United States are facing rising mortality rates, a reversal of a decades-long trend for all U.S. demographic groups. Case, with co-author Angus Deaton, concludes that these are “deaths of despair”—people suffering economic and social anguish that drives them to substance abuse and suicide. “We don’t think people abuse alcohol or drugs, or pick up a gun, if their lives are going well.”

Case’s scholarship has garnered wide recognition. In October, she was elected to the National Academy of Medicine. Her work on U.S. mortality rates won the National Academy of Sciences’ 2015 Cozzarelli Prize for scientific excellence and originality. She is also a fellow of the Econometric Society and a member of the American Academy of Arts and Sciences and the American Philosophical Society.

Clearly a role model for women in economics, Case is critical of their unequal treatment in the field, something she has experienced since grad school. Economics is not “altogether a healthy discipline for women,” she observes. “Unfortunately, I don’t see that as a problem that is going away.”

Interview conducted October 3, 2017.

Selfish genes

Region: I’d like to ask about a full gamut of your work, from early days to current research, and I wonder if we could start with your papers on the distribution of benefits within families—benefits meaning health care, education, expenditures on food. You’ve compared benefit distribution in families that are related by blood versus stepfamilies, families formed through adoption or foster children. And you looked at families in both the United States and South Africa.

One of these papers is titled “How Hungry Is the Selfish Gene?” What did you find? Just how hungry is it?

Case: I think the first thing to note is that within a household it continues to be the case that mothers matter enormously for how well children do. Mothers tend to be the gatekeepers for health care systems, for social benefits. Mothers tend to be the people who make sure there’s a regular bedtime, that seatbelts get worn, that homework gets done; they regulate the kind and quality of food at home. We found in the U.S. that these outcomes for children were associated with the relationship between mother and child, and varied between biological moms, stepmoms, foster moms or adopted moms.

When the mother is absent, children tend to be at higher risk for worse outcomes along many dimensions, and the level of relatedness between the mother in the household and children in the household tends to be predictive of good or bad outcomes.

There are many stories in Africa about a woman running out of her hut, yelling at her boyfriend, because he’s found the place where she hides the money for the children’s school fees and he’s drunk the money. What we find is that the architecture of the household, the family structure, matters for children’s outcomes.

In Africa, where women are dying in middle age at high rates from AIDS, sisters take in a sister’s children upon her death. We found that a woman’s own biological kids were significantly more likely to be enrolled in school and that significantly more was being spent on their educations, relative to the nieces and nephews who had lost their mom. Children of the same age, living under the same roof, with the same household resources were being put on very different life trajectories.

We got a lot of hate mail for that work from families that have something other than two biological parents. We were just trying to shine a light on where some of the problems come up, not to point fingers or lay blame. But there was a lot of discontent the summer that those results were published.

Child health and later-life outcomes

Region: You’ve done a great deal of important work, much of it with Christina Paxson, about the relationship between child health and socioeconomic well-being later in life. In several of the articles, height is a significant metric. For instance, a 2010 paper said, “The associations between height and outcomes throughout adulthood indicate the childhood circumstances are important.”

What relationships are at work here? Also, why do you hypothesize in a related 2005 paper that a child’s health may have “intergenerational” influence?

Case: Any question about height is really a loaded question. I should start from the beginning; we know that in adulthood there’s a really strong relationship between resources—whether it’s income or earnings—and health.

Photo by Peter Tenzer

If you’re a social epidemiologist, you tend to point the arrows from the direction of “if I don’t have income, if I don’t have control at work, that makes me sick.” If you’re an economist, you think about the fact that “if you’re not healthy, then you’re not in a good position to earn money.” So there could be two-way links between health outcomes and income or earnings.

Chris Paxson and I decided that one way we could try to cut into this question would be to look at income and health in childhood, because children in the U.S. don’t generally work for money. A sick child could keep parents at home, but we can control for that.

We found that while most children are born into good health, children in poorer households are more likely to fall into worse health as they age. They have more chronic conditions. Their chronic conditions aren’t addressed until they’re more severe. They spend more days in bed. They miss more school. They have more hospitalization episodes.

It’s not a step function; it seems to be that the wealthier the household, the healthier the children. Using the Panel Study of Income Dynamics, we could look in the years prior to the child’s birth and ask, “What does the long-run income for this household look like?” That was the strongest predictor of how healthy those children were going to be.

What happens then is that the children from a poorer household arrive at the doorstep of adulthood in poorer health. With poor health, and having lost more school, they are less able to have a healthy life or to earn a good living as an adult—when they become parents themselves.

So we think part of the intergenerational transmission of poverty is coming through the body: People in poorer health earn less as adults, and a lot of the poor health came with them from poor childhoods. We also know that wealthier, better-educated people tend to have better prenatal care, so the wealth-health correlation probably starts in utero.

Height at age 3 is correlated to height in adulthood. The correlation coefficient is about 0.7 for both men and women; it’s very strongly predictive. We see height as a marker for health and good nutrition in childhood. Inflammation, for example, that comes from gastrointestinal problems or from infection can actually have an effect on children’s physical development—for which height is a marker in adulthood.

We also find that in the U.S. and the U.K., on average, children who are taller score better on cognitive tests from the beginning of the time such tests can be administered to children. Even at age 9 months, a baby can be tested to see whether or not it recognizes a toy or something else that’s been placed in the crib. And we find the longer (taller) the baby, the more likely it is to recognize what’s being placed in the crib.

None of that came as a surprise to our friends in developmental psychology. They’ve known that for a long time. But economists came late to that game.

Now, all it means is that if you reach your physical potential, you’re more likely to reach your cognitive potential, so, on average, taller people score more highly on cognitive tests. They’re more likely to have reached their cognitive potential.

We got a lot of mail that said, “Obviously, your research is flawed because Einstein wasn’t tall.” Well, Einstein may have reached his physical potential, and so we don’t think that all by itself is evidence. Or if someone is short and their parents were short, they may be reaching their physical potential regardless of how tall they got to be in the long run.

There are two very famous birth cohort studies in this line of work that have followed all children born in one particular week in England, Scotland and Wales from the time they were born through middle age. They took cognitive tests along the way. It turns out that if you control for their scores on cognitive tests at ages 5, 7 and 11, that can explain the fact that in adulthood they earn 2 percent more for every inch of height. So, yes, if you use Occam’s razor, it looks like height is a marker for cognitive ability.

Middle-aged despair

Region: In a 2015 PNAS [Proceedings of the National Academy of Sciences] paper and a 2017 Brookings paper, you and Angus Deaton—a frequent collaborator, and your husband—document that at the turn of this century, after decades of steady decline, morbidity and mortality rates began to rise for middle-aged, non-Hispanic whites in the U.S., while continuing to decline for other U.S. demographic groups and other industrialized nations.

This work has—for good reason—received extensive coverage, so perhaps we don’t need to talk about it as much as we otherwise would, but I wonder if you could review some of the highlights. Also, you refer in your 2017 paper to a “preliminary but plausible” story behind this trend and a “profoundly negative” policy implication.

Could we please hear the highlights, the explanation and the implication?

Case: Sure. We had started to work on the relationship between suicide and people’s reports of their own well-being. People’s life satisfaction or life evaluation is a metric that’s getting a lot of traction now—for example, in Britain. We wondered whether some marker for not being well that seems way out in the tail of the distribution, very extreme, like suicide, was related to life evaluation. In other words, did we find that in the United States, geographically, in those places where people’s life evaluation was lower, was it the case that suicide rates were higher? But we didn’t find any correlation between those two.

Region: To quote your 2015 study: “With some exceptions, suicide has little to do with life satisfaction. Correlations were absent or inconsistent.”

Case: Absolutely. But that started us on the trail of looking at suicide. We found that suicide rates were rising. That opened another door: What about mortality in general? That’s when we found that mortality among whites in middle age was actually increasing.

That came as a surprise to us, so we tried it out on friends of ours at medical schools. We thought this result must be known to them, or others, but it came as a surprise to basically everyone we showed these results to.

So we drilled into it and found it wasn’t that heart disease was increasing, it wasn’t that cancers were increasing. It turned out that the largest increases were for suicide, alcohol-related liver mortality and drug overdose, which we combined into what we called “deaths of despair.” Because to us it all seems like suicide; it’s either quickly with a gun or it’s slowly with drugs and alcohol.

Region: And it’s within this narrow demographic.

Case: A narrow demographic in terms of being white, non-Hispanic, yes, but it is throughout middle age. It’s happening for people in their 20s and their 30s, 40s and 50s.

Region: The other crucial difference is education.

Case: That’s right. The group that’s being hammered is people with less than a bachelor’s degree, more specifically, people with a high school degree or less. Those groups are at highest risk of those deaths.

None of that would’ve come to light if we had continued to make progress against heart disease in the U.S. Unlike other rich countries where mortality rates from heart disease are falling, they flat-lined in the U.S. We don’t really understand why.

Some people are saying, “We’ve been telling you for two decades that you’re all getting fat and that eventually obesity will catch up with you.” But that’s a little bit premature, we think, because in Britain, the obesity rates are almost as high as they are in the U.S., and yet their heart disease rates continue to fall.

Photo by Peter Tenzer

So it’s something of a mystery. If heart disease progress had continued, it would’ve overwhelmed the increases we’ve seen in drugs and alcohol and suicide. But given that it’s flat-lined, it’s allowed those deaths to actually cause mortality rates to rise. We don’t fully have a picture of why this has happened even though the National Institutes of Health are spending $30 billion annually on health research and as a country we are spending $3 trillion a year on health care. We decided that part of the mortality turnaround was probably due to the circumstances into which people in the U.S. without a B.A. were finding themselves.

This transitions to the Brookings paper, which looks into what is going on under the surface here. We don’t think people abuse alcohol or drugs, or pick up a gun, if their lives are going well. We think that to understand what actually happened, we’re going to have to reach back in time to the 1970s.

It was the early 1970s when it became less and less the case that people with just a high school degree could get on a job ladder that had on-the-job training and benefits, and the potential to make a middle-class life. We think it’s not a coincidence that the median wages for men in the U.S. have not grown since 1972. The opportunities available became less and less good, which meant men became less marriageable.

We know from sociologists that women don’t want to marry men unless they think those men are a good prospect. They may live together and even have a child together.

Region: Cohabitation and children, but not marriage.

Case: Exactly. Cohabitation is more common than it was a generation ago, but those cohabitations tend to be fragile, unlike in Europe where cohabitation is a lot more like marriage.

So you’ve got a fragile home life, a job without prospects, and you put those together with a change that also happened in the U.S. where people, though they don’t report themselves to be less religious, are less likely to report belonging to the “legacy” churches of our youth, and instead they’re moving to more evangelical churches, which are much more individual-oriented. The kinds of pillars that used to help people in community are absent.

It can be a perfect recipe for suicide. A lot of this is still speculative, and we’re going to look much more deeply in the next tranche of this work at the differences between the U.S. and Europe on these dimensions, because Europe also saw jobs disappear and go to the Far East. Europe also saw rapid technological change.

Region: Germany, for example, has faced similar transitions, but their tax system is very different, right? And that makes the difference?

Case: Yes, their tax system and redistribution are different. The German case is interesting; we want to look at it in much greater detail because of the kind of opportunities that are available to people to learn trades and to have a protected job.

Region: Their “dual-training” apprenticeship system is often held up as a model.

Case: Yes, and we want to see the extent to which that protects people in the way that isn’t available in the U.S. I think of it as being part of a different ethos: “In the European system, we protect each other,” in contrast to “in the U.S., I take care of my family and you take care of your family and we like it that way, thank you very much. We’re individuals who are going to make our own way. We’re all the Marlboro Man.” That, I think, can be a terrific thing for people who want to break free from where they come from, but it can also be a recipe for disaster.

Region: This brings to mind some of the studies that David Autor has done about the hollowing out of the middle class. His work, for instance, on marriage and jobs, voting patterns and so on.

Case: We think of that new work by Autor et al. as being absolutely consistent with what we’re finding. We also think of the [Justin] Pierce and [Peter] Schott paper on the China shock and the Autor et al. work on the China shock as being highly relevant. The China shock is one example of what’s happened in the U.S., and the outcomes being documented are certainly consistent with what we’re finding over a longer time period of erosion from many such shocks.

What we’re hoping to do in the next part of the work is to look at a much finer geographic grid, bigger than counties in general, because counties in many parts of the U.S. are not well-populated. We’re making geographic units that have at least 100,000 people in them, and we can break the country into about a thousand of them. We can attach them to PUMA (Public Use Microdata Area)-level data that come from the American Community Survey. We can attach them to the county-level identifiers we’ve got on the death records, and we want to see whether or not we can use those to tell us more than we currently know about the relationship between economic status, and morbidity and mortality.

Coming to the policy part of this, we think that, while it’s not an easy thing to do, it’s relatively easy to turn off the taps for the prescription opioids, which are out of control. I think that a consensus is building for this, despite a lot of push-back from medical doctors who will say, “You come to my clinic, you look at these people who are in horrible pain and you tell me not to prescribe them something that can actually relieve their pain.”

Well, we understand there’s a place for prescription opioids, but we think that they’re not currently being prescribed at an appropriate level. The fact that at current levels, every adult in the U.S. could be taking opioids around the clock for three weeks is more than probably makes any sense at all.

But when prescription opioids become too expensive or when the tap is turned off, that won’t entirely solve it. The same receptors in the brain respond to heroin, which is now coming in from Mexico, and it’s pure and cheap. Dealers don’t want to have to tangle with gangs in the inner cities, but they will deliver it to rural Maine and to Indiana, and that has made the illegal opioid crisis much worse as well. Talking about drugs is like trying to shoot at a moving target. In 2010, it was the prescription opioids; now it’s heroin and fentanyl.

Women in economics

Region: It seems that, at long last, serious recognition is being paid to the underrepresentation of women in economics. In their 2016 JPE [Journal of Economic Perspectives] article, for example, Amanda Bayer of Swarthmore, and your Princeton colleague, Cecilia Rouse, highlight the lack of diversity—gender and otherwise—in the profession. A recent SSRN [Social Science Research Network] paper garnered much attention by documenting the level of disrespect that some male economists pay to their female colleagues. CSWEP [Committee on the Status of Women in the Economics Profession] has long championed the promotion of women in economics, of course, but also continues to note their relative scarcity.

You have an extremely distinguished career in this field and a unique perspective as the frequent co-author and spouse of a male Nobel laureate.

What are your thoughts about women in economics?

Case: I wish I could say I thought that the problem was disappearing. Unfortunately, I think that economics, for multiple reasons, is still a discipline in which women are given a much harder time than men.

A lot of what happens is quite subtle, and I think a lot of men who certainly don’t see themselves as being in the least bit biased against women, when push comes to shove, repeatedly decide they like the male candidate better.

It’s a combination of things. One is that a lot of women work on topics that historically were not covered in economics, whether it’s in health or labor or development—those were fields that were not in the central square in economics. Another is that it’s still a discipline where sharp elbows are of great help. And I think women either by nature or by training don’t use sharp elbows as often as men do. And when women do use sharp elbows, they are seen as being overly aggressive.

Region: Where does that manifest? And how?

Case: In seminars. When I go to seminars in other disciplines, the tenor of the seminars tends to be a lot less about scoring points and a lot less about trying to nail the speaker to the blackboard. I think that women oftentimes don’t respond as well to that as men do. Now, some women do, obviously. There’s a distribution to both genders.

Anne Case
Photo by Peter Tenzer

There’s another paper you didn’t mention, which looks at gender and co-authorship in economics. So much work in economics is now co-authored, and if you look at the tenure rates of women versus men who have co-authored papers, the men appear to get full credit for the co-authored papers; whereas the women, if they co-author with a man, are given less than full credit for that work. I find that really interesting as well.

I think that women in economics oftentimes find that government work is one in which there’s more teamwork and their skills are fully recognized. A lot of women go into government instead of academia because when they look around they think, this is actually an environment that seems pretty healthy. And I think that it’s not the case that economics is altogether a healthy discipline for women.

My first graduate degree was in public policy, and all during my undergraduate years where I studied math and economics, and in public policy school, I never felt one whit of discrimination. But the minute I started a Ph.D. in economics, it was apparent. That was a long time ago now, but from that day until today, I have felt that if I want to be heard, I have to be willing to come in and be ready to fight, with sharp elbows.

Also, as is true in families, it can take a long time for something to be rooted out. In economics, if a generation ago there was a prejudice among faculty that women weren’t as good at what they did, the faculty then imparted that sense to the male graduate students, who then become faculty, who impart it to their graduate students, and it can take a very long time for that to work out.

Unfortunately, I don’t see that as a problem that is going away.

Region: The American Economic Association recently announced an ad hoc committee “charged with evaluating various aspects of professional conduct, including those which stifle diversity in Economics.” Its report is due in time for discussion at the AEA’s January 2018 conference.

Are you optimistic that this might help root out the prejudice and behavior you've just described?

Case: I’m very pleased to see the AEA has taken this on. I hope it helps, but I think it would be a stretch to say I’m optimistic.

The impact of AIDS on health services

Region: You’ve done so much work in Africa—South Africa, in particular. I’d like to turn to some of that research now. First, I want to ask about a 2011 paper you did with Christina Paxson about the impact of AIDS on health care in 14 sub-Saharan countries.

Can you tell us what you discovered and what the long-term consequences might be? Also, antiretroviral therapy [ART] has expanded throughout those regions since 2011. Has that effort succeeded?

Case: For several years, I volunteered time to the Economic Reference Group of UNAIDS, a group that’s a combination of the United Nations, the World Bank and the World Health Organization. In meetings in Geneva, it became clear that when people talked about the impact of the AIDS crisis on health care, no one had any idea what was going on. You could pull any argument you wanted out of your pocket. You could say it’s going to get worse because people ill with AIDS will take up all the beds. You could say the crisis would make health care better because more nurses and doctors will be available.

So Chris and I decided to look systematically at a part of health care with fairly uniform quality, such as prenatal care, care during delivery and vaccinations for children. And we found that in the places where the AIDS crisis got worse, medical care in those health domains worsened. Especially in the high-HIV countries, there appeared to be crowding out of services that would be helpful to pregnant women and during delivery. Also following delivery: whether or not the child was immunized against polio, DPT and so on.

ART looks like it’s beginning to have a real effect, although there are large parts of Africa where it’s still not available to the extent it’s really needed. But in a field site where I worked for a dozen years or so, the Africa Centre, which is now an arm of the University of KwaZulu-Natal, where we followed about 100,000 people through time, we found that from the year 2000 to the arrival of ART in about 2008, the fraction of children at any given age who had lost a mother went up and up and up. And after the arrival of ART, orphan rates stabilized.

ART obviously is not going to cure people, but it will keep mothers alive for longer, which is going to be enormously helpful to the children. If a woman is on ART, her viral load can fall to the point where she’s not a risk to her child. She’s also not a risk to her partners.

What can also happen if a child loses a mother at a very young age is that it puts them at risk for earlier sexual debut, which also then puts them at risk for HIV. So it’s another transmission from mother to child, not working through delivery or mother’s milk but from the fact that children, when they lose a mother, are at risk for many other kinds of bad outcomes. ART can help there as well.

AIDS in South Africa and funeral costs

Region: Your 2014 paper on the economic impact of AIDS mortality in South Africa between 2000 and 2011 found that the prevalence of AIDS is much higher in poorer households.

Can you elaborate on why that’s so?

Case: It was somewhat surprising to me that in South Africa, at least in KwaZulu-Natal where this demographic surveillance site is, that it was in the poorer households that people were more likely in the future to die of AIDS. And their poverty dates from before the people who died of AIDS would’ve become sick. One can imagine that if someone of prime age contracts HIV and becomes ill, the family would try all sorts of things, desperately trying to cure them of HIV, that that could impoverish a household.

Region: So the household loses that person’s wages and also spends more on health care.

Case: Yes, exactly, but the data we have on these households go back far enough that we can say this would’ve been before the person actually contracted HIV or if they had contracted it before they became ill, and we found that the people in poorer households were significantly more likely to be at risk of contracting HIV.

That was surprising to us because in other work we found that in the early years of the AIDS crisis, women who were better-educated (and so, presumably, wealthier) were at a higher risk of HIV, in part because they didn’t want to marry at an early age. They were more likely to strike out on their own and more likely to become sexually active. While we tend to think of education as being a cure-all in both developed and developing countries, we found in this particular case, before anyone knew anything about HIV, being better-educated was actually a risk factor for HIV.

Now we find that, as time has passed, it’s members of the poorer households of South Africa who are contracting HIV. It’s not uncommon for a disease’s burden to move from better-educated to less-well-educated people once the causes of the disease are discovered. But I had never seen it happen in real time before.

We think in part that women are now at greater risk of HIV if they participate in transactional sex. In southern Africa and parts of East Africa, transactional sex is just part of life. This guy gives me a ride in his car. He’s my minister of transportation. This guy pays my school fees so he’s my minister of education. And women will say, “I am not a prostitute. I know what a prostitute is, and I am not a prostitute. But I do need things, and things come from exchanging sex for the things I need.” That puts people at risk for HIV. That’s more likely to happen in poorer households. So we think that’s, in part, why people from poorer households in this part of Zululand are more likely to run the risk of becoming HIV positive.

Region: I’d also like to ask about your related research on funeral expenses in South Africa, documenting the tremendous financial burden they put on families. In one of your papers, you say funerals are “the most important rite of passage” and families will spend on average the equivalent of a year’s income on them, and often borrow to do so.

Would you talk about that research as well?

Case: Of course. In this part of KwaZulu-Natal and also in the Cape Flats outside of Cape Town, where we also had a field site, funerals are extremely expensive. KwaZulu was one of the old “homelands”—​the white government of apartheid forced people out to parts of the country where there wasn’t much to be done. That created a surplus pool of labor that could be drawn on to work in the mines or the fields when needed and sent back when they weren’t needed.

In this part of Zululand, probably the only business that was thriving was the funeral business. People were dying in middle age and needing funerals. When a baby dies, or a child at a very young age, there would be a traditional funeral. You would wrap them in a blanket and put them in the ground, and there’d be cookies and tea. When poor people died in old age, they would have an old-age funeral policy that would pay for their funerals; they pay in every month a little bit of money, and that takes care of funeral expenses.

But when people started to die in middle age, from AIDS, they didn’t have funeral policies to cover expenses. But the household felt compelled to bury them according to whatever their station was. For men, you would slaughter a cow. That’s a very expensive proposition. For women, you would slaughter a goat—less expensive but still a big chunk of change.

There would be food for the extended family, who would come from far and wide, and stay as long as they wanted to stay. And you would have to feed them for as long as they wanted to stay.

It was also impossible for local areas to say, “OK, we’re going to make a pact in this local area: No one has an expensive funeral,” because your family who is coming all the way from Johannesburg, they didn’t buy into that. They wanted to come and have this enormous party. It was very important to be able to do that.

We decided to document how expensive these funerals are because they are much more important than marriages, much more important than baby-naming ceremonies or graduations. They are a defining moment for the family, for the household. And if you couldn’t afford to bury your dead according to the status of that person, you would borrow money, and as a last resort you would borrow from a money lender at usurious rates.

So then these households would fall into a deep hole because they couldn’t get out of the loan they had taken out to pay for the funeral. We were able to document the extent to which further bad things happened in those households, whether or not children went hungry, whether or not children got to school, whether or not people reported more psychological stress than would’ve been caused just by the death of someone in the household. In those households that borrowed money, things got worse very quickly.

For further background on Anne Case, visit her website.