“Equitable Growth in Conversation” is a recurring series where we talk with economists and other academics to help us better understand whether and how economic inequality affects economic growth and stability.

[Editor’s note: This conversation took place on March 15, 2018.]

Heather Boushey: It’s a pleasure to be able to talk to you today about some of your research.

Janet Currie: Thanks. It’s great to be here.

Boushey: On the surface, when you look at the body of work you’ve done, it seems as though your research interests are kind of all over the map, from pollution to the effects of the social safety net on fetal development to a lot of different aspects of childhood. But my understanding is that a really important theme of your work is your interest in human capital formation. So, can you walk me through how you got from thinking about Head Start to thinking about fetal development, and how that’s related to human capital?

Currie: I was trained as a labor economist and so it was natural to think about human capital. But traditionally most economists had thought about human capital in a limited kind of way, captured by things like how many years of education did you have, did you go to high school, did you go to college? These, of course, are really important outcomes, but not the only ones of interest. That was one strand of my background that was important; another strand was talking about program evaluation.

Currie: I really got involved in this research as a labor economist thinking about the effects of welfare programs. Reading the literature on welfare programs, it struck me as really odd that although the programs were clearly aimed at benefiting children—since most of them were only available to families with children—no one had ever asked the simple question of, well, do they benefit children?

Instead, the research literature was all about whether these programs made parents work more or less and whether they have effects on family formation—the research was basically about the effects of programs on parents and not looking at the children at all. So, I decided I was going to look at the effect of welfare on children.

At that time, we still had significant cash welfare programs, so I thought, okay, I’ll look at the effect of Aid to Families with Dependent Children [or AFDC, the federal welfare assistance program created in 1935 and in place until 1996, when it was replaced by the Temporary Assistance to Needy Families program] on children. That was really my first paper on this topic. The problem was like looking for a needle in a haystack. What should you expect to see if you give a family $100 a month? What great outcome are you going to expect to find? You could have a lot of small effects on a variety of outcomes, but chances are, you’re not going to find a huge effect on any one outcome with amounts this small.

Around that time, U.S. policymakers started on healthcare reform, expanding Medicaid coverage for children and pregnant women. So, I thought, here’s an example where it’s much more obvious what’s supposed to happen. If you give people health insurance, you want them to get healthcare and be healthier. You want them to consume medical care in a different way. You don’t want them all showing up at the emergency room all the time. You want them to go to the doctor. Here’s a clear question: If you give somebody access to health insurance, does that improve their health outcomes?

So, I started working on that question. When thinking about health, it really matters what your initial conditions are. There’s a lot of evidence not just from economics but from the healthcare literature about the lingering effects of health at birth. So, I started looking at the health of very young children as a form of human capital.

Boushey: You started off by saying that it used to be that economists thought of human capital as years of schooling or these sort of indicators from late childhood or early adulthood. How has that shifted toward understanding the role of early childhood in how much an adult has in terms of human capital?

Currie: Economists have become much more aware of the importance of the prenatal period of early childhood. Economists were always interested in the intergenerational transmission of poverty and how that happens—there was a seed there—but there was a general lack of knowledge about the far-reaching consequences of things that happen in early childhood. The economics literature has moved along in tandem with a blossoming of the literature in developmental psychology, neuroscience, and also in medicine. There’s been a huge increase in the basic science of how socioeconomic factors “get under the skin.”

Boushey: So, it sounds like it was both a change in economics and also driven by our changing understanding of science.

Currie: There’s been a sea-change in the science as well. For example, understanding epigenetics as well as genetics—it’s only in the past 20 years that we’ve understood that the nature-versus-nurture dichotomy isn’t actually very helpful. Almost all the results you see reflect an interactions of some sort between nature and nurture. If you don’t capture the interactions, you’re leaving out most of the action. We just didn’t know that 20 years ago.

Boushey: That’s fascinating. All of this new science really pushes academics and policymakers to think about the health of mothers and the health of infants, which can also be important and probably what’s complicated is that these things often go hand-in-hand with universal access to education. So, in your research work showing how early childhood matters, how do you get at the issue of causality? Because it’s so difficult to know what’s causing what, can you tell us how you think about the effects of early childhood on later outcomes? How are economists thinking about the causality question?

Currie: An important feature of poverty is that bad things are more likely to happen to the poor than to others. So, if you’re trying to look at the effect of an event or a program, it’s hard to tease out the effect of that event or program from the effect of just being who you are: a poor person in a poor neighborhood.

Even things like accidents are not random at all. Accidents are more likely to happen to the poor than to people who are not poor. So, how can we separate the effect of the accident from the effects of the factors that made the accident more likely to happen? Addressing this problem of causality is fundamental in all of this research.

The main strategy for dealing with that problem is to find a comparison group. So, a lot of studies, including some of my own, look at siblings as a comparison group. That has strengths and weaknesses, notably that siblings aren’t exactly the same. Another approach is to look for some kind of exogenous change, what people like to call a natural experiment. That might be the implementation of a program or the ending of a program. In some cases, you might even see a program begin and end, so that you could see whether the effects are equal and opposite. Researchers may also use actual social experiments with randomly assigned treatment and control groups.

What I like to tell my students is that often, these types of studies are like a collage. You have different pieces of evidence that all point in the same direction. But other social scientists sometimes criticize economists and say it’s not possible to definitively determine causality. That’s true. It isn’t possible to determine causality with 100 percent certainty. All that it’s possible to do is to assemble enough evidence to establish a relationship that is beyond reasonable doubt.

Boushey: That’s a good caveat, too, because we at the Washington Center for Equitable Growth work with a lot of policymakers, and they want to know what the outcome will be if they implement a program. We can’t always say for certain, but we can say that if you have a number of studies that are looking at the question in a variety of ways, then you can get closer to saying, well, our best guess is based on all of this evidence. That’s really important for us to bear in mind.

One of the things that you’ve already mentioned a little bit is the nature/nurture question. One of the things that the economics literature tends to talk about is this distinction between what economists call cognitive and noncognitive skills. My understanding is that other disciplines don’t use those terms, so maybe you can unpack these two terms. How do these so-called noncognitive skills lay the foundation for future learning? What do we know about that, and how is that related to this nature/nurture distinction that you were talking about a few minutes ago?

Currie: The idea of a sharp distinction between cognitive and noncognitive skills, and the use of those labels, is falling out of favor in economics as well.What people have in mind when they say cognitive skills is something like a score on an IQ test. But if I were to give you an IQ test when you were really hungry and tired, you would get a different score than if I were to give it to you when you had just eaten and were well-rested. Which would be the correct score?

It’s also not clear exactly what is meant by a noncognitive skill. Do you mean somebody’s mental health status? Do you mean something that’s thought to be relatively constant over time, like a personality trait? Do you mean something that can change depending on whether you’re hungry or tired? The way that the concept of noncognitive skills has been used in economics so far is a sort of first cut at some really complicated measurement issues, and it’s not clear that we should stick with that first cut going forward as the right way to characterize children’s outcomes.

Boushey: Would you have a better recommendation?

Currie: Well, a lot of my work has focused on trying to integrate the literature on different types of human capital and arguing that they’re not completely separate things. The way economics is structured is if you’re working on education, then you’re a labor economist. If you’re working on health outcomes, you’re a health economist. But then what about kids who are sick and so don’t do well in school? Where does that fit into the dichotomy? I’d like something a little bit more holistic in terms of thinking about child outcomes.

Human capital is what helps people to succeed, and that has different aspects, certainly. But I’m not sure that you can identify policies that only affect one aspect of human capital and no other. By their nature, there’s going to be a lot of spillover and overlap in the effects of policy on different child outcomes. So, its important to capture that richness but also not to be too hung up on categorizing outcomes.

Boushey: That seems really important, especially with younger children and early childhood care. There’s a distinction made between childcare versus early pre-Kindergarten programs, but it seems like actually that child, no matter if you call it a day-care center or pre-K, is still learning and they’re learning lots of things—not just how many blocks they’re playing with, but also how to share those blocks with the other kids and all the other great things that kids are learning. That’s important to bear in mind, but it also makes it just a little bit more complicated for policymakers and even for academics who are studying this to know exactly what you’re measuring if we’re broadening that distinction.

Boushey: Tell me a little bit more about your research on the prenatal period and how economic inequality may mean that there’s different inequality in mothers’ lived experiences and in the things that children are exposed to in the womb, so that they will have different outcomes when they become adults.

Currie: I’ve done a lot of research on things that happen while mothers are pregnant and how those impact child outcomes. A nice thing about pregnancy is that it’s self-contained. You have nine months when someone is in-utero and generally no more, no less. So, you can identify something that affected a person during that narrow time window. Then you can look later after birth, sometimes many years after birth, and see what the effect of the event is. From a research-design standpoint, it is conceptually easier to begin at the beginning of life and study subsequent outcomes. Studying later life events, it can be hard to know what the right time window is, and there is also the possibility that the effects of a later life event will vary with the child’s previous history.

In terms of how inequality affects fetal outcomes, I’ve done a lot of work looking at the effects of pollution on fetal health. I have found that women of color and less-educated women are more likely to be exposed to most of the different sorts of pollution that we measure. Whether it’s air pollution, factory emissions, living near a busy highway, water pollution, or having lead paint in the house—almost anything that I’ve looked at, disadvantaged women are more likely to be exposed to it.

These exposures have negative effects on infants. So, infants can start life with a disadvantage because their mothers were exposed to pollution when they were in-utero. Another takeaway from this research is that exposures to levels that are less than official thresholds for conern are frequently dangerous. There may be many people out there who are not reaching their potential because of something that happened to them before they were even born.

Boushey: That’s striking and disheartening. I was talking to some scholar recently about this who was very adamant that looking at infanthood or prenatal is very fatalistic. If that happened, we’re sort of washing our hands, but you certainly don’t think that’s the case.

Currie: It is hard to talk about prenatal influences without sounding deterministic, but outcomes aren’t deterministic at all. One way that you can see that is that the same negative shock, such as a given level of air pollution, will typically have a much greater effect on a poor infant than on a richer infant.

What that observation tells you is that there is something that can be done to mitigate the effects of a harmful shock, and the richer parents are doing it, whatever it is. If you could find that and put it in a bottle, or put it in a program, then you would be able to mitigate the effects of these early-childhood insults.

And in fact, we have had reasonable success in mitigating the effects of some types of prenatal disadvantage through public programs. There is a great deal of evidence, some to which I have contributed, on the positive effects of the Supplemental Nutrition Program for Women, Infants, and Children during pregnancy on infant health outcomes. These positive effects are striking, given that the dollar value of WIC benefits is fairly small. It may well be that modest amounts of money combined with improved access to medical care and some psychosocial support can go a long way toward improving the health of mothers and their babies.

Boushey: You’ve done a lot of research on early-childhood programs such as Head Start and others that also document that there are impacts that go through life of those kinds of programs. Can you tell me a little bit about that and then what the scale of the different effects are across sort of looking at the prenatal period versus looking at later periods in life?

Currie: My own research suggests that Head Start closes about a third of the gap between the average Head Start child and average non-Head Start child. So, it is having a positive effect, although it does not completely close the gap. It is also reasonably cost-effective. Since it doesn’t totally close the gap, whether it is a success or failure depends on your perspective. It’s a question of whether the glass is half empty or half full.

Although we know that there are prenatal and early childhood interventions that improve outcomes and seem to be cost effective, I don’t think we should write off investments in older children. Some interventions, such as Big Brother/Big Sister, have also been shown to be cost-effective. I am not sure that the research community has actually done an apples-to-apples comparison of programs aimed at different age groups.

Boushey: But it’s certainly important, and I think that your earlier point about the shocks being bigger for poor families or poor children is really an important one. I have a specific question on the Head Start research that you did. My understanding is that you were re-evaluating previous studies of Head Start where people found that the effects of the program hadn’t continued over time. Part of your work was looking at that anew and saying that well, actually, if you had the right comparison group, then you would see a sustained impact. Can you tell me a little bit about that?

Currie: One of the questions about programs such as Head Start is whether there’s what people call fadeout in the effects. We know that you can improve people’s test scores or other outcomes with a relatively short intervention. Then, if you don’t do anything else, the effects often seem to gradually go away. In my work, I also found fadeout for some groups of children, though not all, by about 3 years after the Head Start program ended.

But several studies, including some of mine, have found that if you follow children long enough, positive effects of Head Start re-emerge and we see positive effects on outcomes such as high school completion and college attendance, as well as positive effects on health. It isn’t completely clear why one could get these “sleeper effects” of early intervention programs. It may have to do with the development of a broad range of skills that help the child succeed even if there is no difference in middle school test scores. Perhaps Head Start changes a child’s whole attitude toward school.

Another possibility is that perhaps we measure outcomes better in young adults than in middle schoolers, where there will be a lot of variation in stage of development within every grade. Every age has its own set of measures, and we try and make them consistent over time, but they can’t really be made completely consistent. You’re not measuring a 20-year-old the way you measure a 3-year-old. So, the answer could be that the way that we measure outcomes is not as sensitive for middle school children who are middle school age as it is for young adults.

Boushey: Well, the middle-to-teenage years are so different in the way that the brains are developing, and it’s such a different social experience, that you could think of reasons why you might find something different when kids are in primary school versus later on that might look like fadeout. But that’s a stretch.

Currie: I think middle school is a difficult period in general because the kids are hitting puberty at different ages, and they’re having growth spurts at different ages. They’re having all sorts of social things going on, and so a lot of different kinds of interventions may not find a very uniform effect at that age.

Boushey: It’s fascinating, though, because there’s been a number of different policies that when scholars have gone back and looked at them we come to different answers now. It does seem that both the access to longer-term data or administrative data that looks over these long periods of time or new methods or computing power makes it easier for more scholars to evaluate a certain question. It’s quite powerful just how much advice we can give policymakers now.

Boushey: Can you identify what you think are some of the most important policy interventions to boost people’s human capital? If you were a policymaker with a magic wand, what would be your top three investments based on your research?

Currie: Based not only on my research but on what I know of other’s research, healthcare is really important and healthcare for kids, anyway, is relatively cheap. So, just preserving the access that we have gained for young children so that all children are covered for basic services is really important. We saw declines last year in the number of children covered by Medicaid for the first time in almost 30 years, and that is deeply disturbing.

Similarly, all pregnant women should have healthcare coverage. People shouldn’t have to worry about where they’re going to get maternity care. They shouldn’t have to worry about what happens if their child needs some kind of intensive care at birth. And I think we have a lot of evidence now that care at that stage is critical for future outcomes.

People aren’t aware of all of the research showing that poor health is one of the most important reasons why people don’t participate in the labor market, and why they fall out of the social fabric. Poor health can even prevent people from getting married or finding stable partners. We now have a good deal of excellent social science research tracking the cohorts of children who gained access to healthcare when Medicaid was expanded in the late 1980s and early 1990s. Young adults who were covered from birth have more education, are more likely to work, pay higher taxes, and are in better health. So, we know that public health insurance for poor women and children can lead to positive effects on all these long-term outcomes.

The gains that we’ve made in terms of early childhood education are important, not just for teaching kids colors and shapes and so on before they get to kindergarten but because a good childcare is a safe space for a child that helps their parents to be able to work. Early childhood education programs provide an important support for families and reduce parental stress, which also is good for the child.

Boushey: Those are super important. Both are big-ticket but cost-effective interventions. And both kind of focus on the whole child rather than the nature/nurture distinction by focusing on the physical child and the child’s brain as the child develops.

Boushey: The Washington Center for Equitable Growth, as you know, is a grantmaking institution, and you sit on our Steering Committee and help us with this. Based on where you see the research, what do you think are the most important questions we should be asking and pushing other scholars to ask in this area?

Currie: If someone were to ask, ‘I have $100 million, so what’s the policy that will give me the biggest bang for the buck in terms of improving child outcomes?’ I don’t think anybody could answer that question. That’s a problem for policymakers. It may be that the best answer is a package of programs rather than a single program, and that is something that we as academics have not given a lot of thought to.

We don’t know very much about which kinds of programs are substitutes and which are complements, or about how substitutable interventions are over a child’s life. If this child didn’t get an early childhood education, can we make up for it by putting him in some kind of afterschool program now?

The answer to that particular question is probably no, but these kinds of trade-offs between programs and knowledge about which packages of programs would work well together is important but is sorely lacking.

Boushey: That’s an interesting series of policy questions that connects the dots between what we know about program evaluation and the effects and the intersection. That’s a great place to leave it. Thank you again so much.

Currie: Thank you.