You have a better chance of achieving “the American dream” in Canada than in America

Raj Chetty is a Harvard economist who has been called “the most influential economist alive today.” He’s considered by his peers to be a shoo-in for the Nobel Prize. He’s transforming how economics is taught. And he specializes in bringing massive amounts of data to bear on the question of social mobility in America: which communities have it, how they got it, and what we can learn from them.

You can listen to our full conversation by subscribing to The Ezra Klein Show wherever you get your podcasts, or streaming it below. A partial transcript, edited for length and clarity, follows.

Let’s start with the Creating Moves to Opportunity project (CMO). Tell me a bit about the experiment and what it found.

CMO is a study we just released that we ran in Seattle and King County. The basic idea of the project stems from earlier work we’ve done showing that the neighborhood where kids grow up has a profound effect on their chances of upward mobility. The starting point for this project is recognizing that many low-income families in America live in neighborhoods that have historically low rates of upward mobility.

We came to the data asking, why are people living in these sorts of neighborhoods? Is it because they have other reasons for wanting to live there, like family who live close by or a nearby job? Or are there barriers that are making it hard for low-income families to move to higher-opportunity neighborhoods, such as lacking information, lacking support in the housing search process, or landlords who don’t want to rent housing to them?

Before this experiment, it was generally assumed to be the former because a lot of people end up in these high-poverty, low-mobility neighborhoods even when given rental subsidies, right?

That’s exactly right. For instance, there are about 2.3 million families in the US that receive housing vouchers worth around $1,500 a month in terms of rent in the Seattle area. As you noted, most people take that money and still end up renting an apartment in a pretty high-poverty, low-opportunity area. That’s led some to think that maybe this is just a preference that folks have for good reason.

So, coming back to CMO, we ran a randomized trial where we took about 500 families who are applying for housing vouchers through the standard housing voucher program in Seattle and King County. For half of those families we gave them additional services in the form of “search assistance.” This assistance is a counselor who spends time with them, helps them find units that might work for their families in high-opportunity areas, helps contact landlords in order to identify suitable units, and also provides some financial assistance averaging about $1,000 per family to pay deposits or initial bills.

What we found was a remarkable change in terms of where families ended up choosing to move. Of the families that did not receive these services, only something like 18 percent moved to high-opportunity neighborhoods. In contrast, in the treatment group, that number jumps to 55 or 60 percent. So the majority of families are moving to these more mixed-income, higher-opportunity areas.

I really found this striking. We’re talking here about a roughly 40 percentage point change in one of the biggest decisions people make in their lives — all based on a pretty modest intervention in the search process.

That’s exactly right. These are some of the largest effects I’ve ever seen in social science, and you’re right to note that this is not a very big intervention in the grand scheme of things.

One way I like to benchmark this is relative to the cost of the housing vouchers we are already providing, which are about $1,500 a month for seven or eight years. The cost of our intervention is about $2,500 a family, which is about a 1.5 percent increase in the total cost of the program. So it’s a small expense relative to the overall amount of money we’re already spending that ends up having a huge effect in terms of where families end up living.

I’m curious what your team learned talking to the people in the study. What did it reveal to them running this about the obstructions people actually face?

We are finding that it’s not the traditional factors that economists would focus on, such as the costs of moving to a different place or getting [the right] information. Lots of families had the sense that these would be better neighborhoods for their kids and having additional money to move wasn’t a huge factor. What was more important were a few things that are easy to take for granted.

One is having the psychological support of someone acting as an advocate for the families. Someone who would be willing to help them at every step of the way: to help them fill out application forms, to work with them if they receive questions about their credit history, and so on.

Another very important aspect is the landlord-facing side. Many of us take for granted that if we want to rent a particular unit, we can walk in the door, fill out the forms, and we’re going to have a shot at getting that apartment. But many of the tenants we spoke with had submitted 30 applications, each with application fees, to different apartments and every single one was denied. They end up getting discouraged and settled for a unit in one of these higher-poverty areas. What we’ve done in this program is work to identify a set of landlords who want to rent their apartments to voucher holders, and work directly with them on things like creating a damage mitigation fund.

To take a slight detour here, I’m living in the Bay Area now, which is a place with a lot of opportunity in that there are a lot of jobs, social services, and social capital. But zoning laws, among other factors, have created an extraordinarily difficult path to living here. Your research seems to show that this is a much more unjust outcome than many folks here with Bernie Sanders stickers on their car recognize. If you’re denying people access to these neighborhoods, then you’re really denying them access to an engine of economic mobility.

I’m glad you brought that up. When we prevent dense development in high upward-mobility areas, we are denying many kids that shot at the American dream.

Because you’ve done such granular work on which neighborhoods throughout the US provide opportunity, I’m curious if you’ve developed a view on whether political ideology plays a role here. Does living in a more blue or red area have a significant impact on opportunity?

When you look regionally, you tend to find that the highest levels of upward mobility are in the Great Plains and some parts of the coast, while you have far lower levels in much of the Southeast. If you were to just take that data you will find essentially a zero correlation with Republican and Democratic voters.

If you zoom into local areas, you will find that a lot of the variation in upward mobility is actually coming from neighborhoods that are just a few miles apart from each other — often within the same city. There are parts of the Bay Area, for example, like Redwood City, that are relatively affordable with quite high rates of upward mobility. But just a few miles down the road in East Palo Alto, you find much lower levels of upward mobility.

You see this extremely local variation in all cities across America in a way that doesn’t obviously line up with political factors. The roots of these differences in mobility are complex — they don’t break down along the lines of certain government programs or investments.

In that case, to what extent do policy preferences really matter for increasing mobility? Someone on the left who wants to increase upward mobility will advocate for things like universal health care and pre-K while someone on the right will advocate for things like low taxes and corporate incentives. But the fact that your research shows almost zero correlation between social mobility and political ideology, all else being equal, could mean that all this political fighting is less important in terms of mobility than people would think.

I actually don’t take that view because I think when we look at specific policies like the affordable housing and education policies, we see very clear effects of policy changes on long-term outcomes and rates of upward mobility.

I think what’s going on is that there are various factors at play that don’t go in a systematic direction on aggregate. So there are some places that seem to be getting things right more in the dimension of creating integrated neighborhoods or better schools. There are other places with higher levels of social capital or connectedness. So, in total, there is not a single policy platform that systematically leads to higher levels of mobility.

You’re saying not that policy matters less than you think, but that the packages of policies that have been incorporated into party agendas haven’t necessarily been increasing mobility. So Republicans and Democrats both have some good ideas and have some bad ideas, and, on net, neither one of them has managed to build a platform that is way better than the other for mobility. Is that right?

I think that’s part of it. I also think we need to remember that there are lots of differences across places that exist prior to the implementation of these policies that affect mobility. For example, Democrats tend to win more urban areas, which tend to have higher levels of concentrated poverty to begin with. There are plenty of offsetting factors like that.

It’s not like we’re working in a controlled vacuum where we can experiment with a Democratic package of policies and a Republican package of policies. Those policies are responding to the conditions on the ground, making it difficult to piece apart the causal effect of the policy packages from whatever else is going on.

So my read is, on net, you can’t read too much into how these things correlate with broad policy platforms. You have to dig into the specifics of each policy in order to understand how policies impact upward mobility.

Your team recently published a paper that compared 133 different policies in terms of their “return on investment” defined as the return to society divided by the cost to the government. Do you want to run through some of the big-picture takeaways from that?

Yeah. This was an impressive study conducted by my colleagues Nathan Hendron and Ben Sprung-Keyser at Harvard. What they did is take a set of estimates that previous researchers had constructed on the impacts of a broad range of policies ranging from health care for kids to tax cuts for adults to investments in education. For each of these policies, they constructed a bang-for-the-buck measure: For every dollar I spend from the government’s perspective, how much do I end up improving people’s incomes?

The main punchline that they arrived at is that programs that invest in kids tend to have very high rates of return. Sometimes, they even have infinite rates of return, meaning that they completely pay for themselves. So we actually find that programs like helping families move to better neighborhoods or certain types of early education programs do not cost the government anything on net. In fact, they pay the government back because the higher earnings that kids end up achieving later in their lives translate into additional tax revenue such that they more than offset the costs of the programs.

If the mechanism by which these things pay for themselves is spillover tax revenue, then the actual benefit to the lives of the people being affected must be much larger. Tax revenue is going to be only a fraction of actual income generation and doesn’t come anywhere close to fully measuring the benefits of things like staying out of jail or having a job that you like better or avoiding teen pregnancy.

That’s exactly right. Tax rates are around 15 or 20 percent of total income for low-income families. So the total income impact of the program would be five times as large as the impact on tax revenue — not to mention all the other benefits that you described. So some of these early programs — like moving to a higher-opportunity neighborhood — can have quite substantial impacts, often hundreds of thousands of dollars of higher income over a lifetime.

You see this for lots of programs that target kids. And it’s not just programs that target the youngest kids, but also programs that target kids in their teenage years or certain types of college access programs. Many of these have incredibly high rates of return or pay for themselves.

Programs targeted purely at redistribution for adults do not typically [pay for themselves]. A good example is a disability insurance — that’s not going to be a program that pays for itself, but that was never the goal to begin with. [The goal] is to provide support for people who really need help.

The lesson of this analysis is that we should look at programs that invest in kids differently. We shouldn’t look at them as costing society. If designed well, many of them could actually save taxpayers money while increasing mobility and reducing inequality.

I want to be careful with these kinds of analyses because, to use a slightly different example, providing end-of-life care through Medicare does not provide some large social return in terms of actual money coming back into the government. That’s not how end-of-life care works — it doesn’t bring in more tax revenue. You do it because you’re a decent society.

But it does seem to be that the implication of this research is that we are way underinvested in children. What we could be doing in terms of investing in a better society tomorrow by investing in children seems to get down-weighted given the narrow bandwidth of our political system.

Your example of end-of-life care is great. But people naturally have different weights on the importance of having a caring society. Some people think that inequality is a big concern and we should try to share the rewards of economic growth more equally. Other people have different views — that if you earned a lot of money you should get to keep that. I think you can reasonably have a debate about that. I think what much of our work shows is, however, that there are certain programs where there really shouldn’t be a debate because they pay for themselves and reduce inequality in the process. Those tend to be programs targeted at kids.

Your team’s study looks at experiments like the Perry preschool experiment, which was a very high-touch, high-quality integrated early preschool intervention that worked great. If you could give everybody access to the Perry preschool services, that would be wonderful. But the problem is that it’s hard to scale.

The research that ends up being the strongest is often the best funded and best conducted with the most committed backers, making it the hardest to scale up. So how reliable are these research experiments to take big-picture, societal-level conclusions from?

We’re increasingly learning that the way you implement things really matters. The details really matter. It’s not just about the dollars you allocate to a program, but who was implementing the program. In the context of education, who are the teachers in the classroom, who’s actually providing the service, and how motivated are they? Are they really invested in seeing success? I think one of the challenges as we expand these policies is not the conventional debate about how many dollars do we allocate to bucket X or bucket Y, but rather how we maintain fidelity to the original goals of the program as we achieve scale.

For things like neighborhood effects, we have small-scale experiments but also very large-scale evidence where we’re looking at millions of families that move across places and find very similar patterns. So I don’t think the problem is that a given study’s results don’t generalize. I think the key is to figure out how we maintain quality as we scale.

Given the breadth of your research, how would you tell the story of American mobility over the past 50 or 60 years?

America has been perceived as a land of opportunity for a long time. My own parents, like so many others, came here as immigrants searching for opportunity. But we found that America is no longer as much of a land of opportunity as it once was. One cornerstone of [the American dream] is the idea that through their hard work, any kids should be able to go on to have a higher standard of living than their parents. That was basically true for kids born in the middle of the last century. Something like 90 percent of kids born at that time went on to earn more than their parents did.

Over time, though, you see a really dramatic fading of the American dream. For kids who are entering the labor market today, it’s basically a 50/50 shot now as to whether they will do better than their parents. I think that’s what drives a lot of the frustration that people in America are expressing.

The critique you hear on this from conservative scholars like Scott Winship is that mobility has not gone down all that much — that opportunity is still alive and well. He makes two arguments. One is that because we have smaller families now, a given amount of income needs to be spread over fewer people. And when you take that into account, most people are better off than their parents were at the same age. Two is that we aren’t counting transfers — like health care benefits and pension benefits — correctly. He argues that when you take into account these transfers in the system, the mobility picture changes. What do you think of these arguments?

We assessed those arguments in a paper that we published in Science a couple of years ago when we first reported these results. We have a version of the analysis where we adjust for household size and it makes a small difference. Instead of going from 90 percent to 50 percent, the chance of making more than your parents goes from 90 percent to 60 percent. The point here is not literally whether the number’s 50 percent or not, it’s whether it’s declined relative to the past. I’ve spoken with Scott about this and I don’t think there’s any disagreement here. There’s some debate about the magnitude of the decline, but in my view that’s not the core issue here. The point is there’s been a big change and we should try to understand what the drivers of mobility are.

I actually think you’re giving that argument too much credit. I think it’s weird to say that you should adjust for a smaller family size given that a lot of people would like to have bigger families but don’t feel they have the income to do so. If I was not able to have more children because I didn’t feel I had the money to support them, then maybe I have more money, but it would be a terrible hit to the kind of life I want to live. So that argument always strikes me as very strange.

I totally agree with you, but even if somebody has a different view on that question, you’d still come to the conclusion that there’s been a big fading of the American dream.

It’s a similar answer for transfers. We have a version where we add in transfers and account use different price indices. The qualitative patterns change a bit, but the question is, what are you trying to measure? What’s the right way to think about the American dream? Is it that you are doing better than your parents because you’re getting more transfers from the government? Or is it that you actually have the opportunity to earn more than your parents did?

I think for many people the dream is about having the opportunities to have the type of job that yields a certain level of earnings. It’s the dignity associated with that. It’s the freedom and opportunities associated with that. So while we focus on incomes, I think we should recognize that the idea of the American dream is actually much broader than that.

What are the non-policy factors that you have found are most predictive of whether or not a place is going to have high opportunity?

I’m not sure I would divide things neatly into policy and non-policy factors because I think a lot of things that might appear to be non-policy factors are indirectly shaped through policy. So I’ll just list what I see as the strongest predictive factors.

High-opportunity areas tend to have less concentrated poverty and tend to be more mixed-income integrated areas. They also tend to be places with more two-parent families. This is actually one of the strongest patterns in the data. Places with more stable family structures, lower divorce rates, and higher marriage rates tend to have higher levels of upward mobility. Initially, the explanation that comes to people’s minds is that it’s better for a kid to be raised in a two-parent family than a one-parent family. But that is actually not what’s driving this pattern. We find that even for kids growing up in two-parent families themselves, they are less likely to climb the income ladder if they live in an area with more single parents. Opportunity is not a direct effect of whether your own parents are married. Instead, there is some community-level factor that’s being picked up by these measures of family stability.

Another factor is social capital. The simplest way to think about social capital is the old saying that “it takes a village to really raise a child.” Salt Lake City with the Mormon Church is thought to be a place with a lot of social capital. Correspondingly, Salt Lake City has very high rates of upward mobility. Generally, places where somebody else will help you out if you’ve fallen on hard times tend to have high levels of upward mobility. Is that a policy factor or not? Obviously, we don’t have policies that directly try to increase social capital, but it’s possible that the way we set up our cities or the way we structure our schools and so forth ends up having indirect effects on it.

Something that your research seems to show again and again is that it is very important whether or not you are grounded in a particular culture. You have interesting research showing that if you grow up in an area with more inventors, then you are more likely to be an inventor, and also far more likely to produce the same kinds of inventions, like medical devices, as others in your area. The prevailing norms in your direct community are incredibly important, even if they’re totally separate from what’s happening in your household.

That’s exactly right. One of the strongest patterns that emerges in all of these studies is that what you’re exposed to as a child — in terms of career pathways, crime, marriage, etc. — impacts how you grow up.

In the inventors study, for example, we find that increases in inventors occur in a gender-specific manner. If women grow up in an area where there are more female inventors in a particular field, they are more likely to become mentors themselves in that field. But if there were more male inventors in that area, it has no impact at all on women.

These patterns are very specific and they reflect the fact that children absorb what’s in their family and what’s in their surroundings — all of which might ultimately be shaped by government policy. That really seems like the critical factor leading to different trajectories in different places.

That paper, along with other work you’ve done, has led me to view representation as much more important. There are many who hold the view that increasing diversity for the sake of representation is ridiculous. But the work you’ve done shows that people who dismiss its importance in shaping children’s life outcomes are really missing the boat. Not everything is about material benefits. Not everything is about direct chances. Much of what matters is a person’s horizon of possibility.

I completely agree with that. I think people’s aspirations are shaped by what they see in others around them who they can relate to. One place I might push back is that we don’t know for sure yet whether it’s adequate to see people on TV or in the media. My sense is the deeper interactions — such as interacting with somebody in your community — has the strongest influence on what you choose to do. That’s obviously also much harder to change, though.

If I were a policymaker, one of the big takeaways that I would have from your work is that if I care about social mobility, then two of my overriding goals should be promoting economic and racial integration and developing social capital. I’m curious if you would add anything there. Are there other macro takeaways that you think policymakers should take from what you found?

I think the other policy takeaways are in the domain of education. Trying to improve the quality of elementary education by retaining and recruiting the most effective teachers has been shown to have extremely large long-term effects.

Likewise, in the higher education system, increasing access to institutions that provide pathways to upward mobility is an extremely important goal. Right now, we have extremely few low-income students at institutions like Harvard and Stanford, which are gateways to elite positions in the country. There’s room for policy to do that, such as performance-based funding. Rather than anchoring the amount of funding that colleges get to other metrics, you could reward colleges based on how many kids come from low-income families and end up moving into a higher place in the income distribution. I think that kind of thing would move things in the right direction.

In a study you released a number a years ago, you showed that having a good teacher early in your life has a huge impact on lifetime earnings.

That’s right. We see quite substantial increases in lifetime earnings on the order of hundreds of thousands of dollars per student.

That is remarkable. So it seems to me that if we are building a society based off Raj Chetty recommendations, the base pay for teachers should just be $100,000. Considering the incredible value that teachers generate for society, that should be a very high-status profession. The idea that the base pay of management consultants is super high but the base pay of teachers is relatively modest seems crazy to me.

I agree but would also say that we should have a culture of trying to recruit and retain the top talent in teaching as we do in these other highly paid professions. At present in our teaching workforce, we don’t necessarily have the evaluation systems and the recruitment systems to reward star teachers, making teaching unlike any other sector of the American capitalistic economy. So I think providing that kind of flexibility could be extremely valuable.

I want to talk about some recent work you and your team did on life expectancy. You found that the richest American men live 15 years longer than the poorest men, and the richest American women live 10 years longer than the poorest women. And in recent years, we’ve continued to see these gaps widen. That’s a really profound kind of inequality. It’s one thing to say we have economic inequality; it’s another thing to say that we have life expectancy inequality on the order of 15 to 20 percent of people’s lifetimes. I would like to hear a little bit more about your findings here. What do you think are the probable contributors?

There are stunningly large differences in life expectancy between the rich and poor in the US. The benchmark I like to use is the CDC estimates that if we were to eliminate cancer as a cause of death, we would increase average life expectancy in America by about three years. So think about the 15-year gap in that perspective — it’s five times as large. This 15-year loss of life for big groups of our population should be considered a public health issue, and we should devote an enormous amount of resources to try to understand and identify the cure, as we do for cancers.

So how do we go about doing that? I break this into two classes of factors. One is the set of factors that are related to economic inequality — more segregation, have weaker schools, less social capital, and so forth. But then there’s a second type of what you might think of as pure health inequality. If you take people at a given level of income, say, earning $25,000 or $30,000 a year, living in New York City versus Detroit, you see very different life expectancy. In particular, the person in New York lives about three or four years longer than the person in Detroit with comparable income.

What’s driving that variation across places across different parts of America? What we find is that the drivers are health behaviors, not health care access. The fact that some people have access to health insurance and better health care doesn’t correlate much with these differences. What does correlate is smoking, obesity, rates of exercise, and so forth. If you look at maps of smoking rates in America for low-income individuals, they look almost identical to these differences in life expectancy for low-income Americans. So I think it again comes back to issues of changing social norms and changing behavior. How do we change health behaviors in a way that would narrow some of these gaps, either through preventive care or through education? I think that’s an important area to focus on.

I’m somebody who spends a lot of time arguing with people about the best way to design health insurance systems, and in the back of my mind, I worry that the true answer to the whole question is instead of spending money on health insurance — which, to be clear, I think you should do — we would have a bigger effect on health outcomes simply by tripling cigarette taxes everywhere in the country.

It depends what your goals are. I think providing health insurance obviously can have tremendous value even if it’s not directly changing overall life expectancy or health outcomes all that much. It tremendously reduces financial risk and financial burden.

But I do think we need to have that long-term view that health builds up over time. Many of our health behavior patterns go back to our childhood years. So focusing on health education for kids may have a higher payoff in the long run.

One way you could interpret the conversation is that the life expectancy gap is based largely on individual choice. Health behaviors are the driver here, and health behaviors — like smoking, eating healthy, exercising, etc. — are largely the product of individual choice. So what can we do?

But what you just pointed out is that these behaviors are at least partially cultural. People act differently in different places and eras. I grew up in Southern California, and I think that is reflected in my attitudes toward health and diets and exercise today.

In that case, things like economic integration policies really matter because changing locations changes what children absorb. To the extent you don’t do anything about pockets of unbroken poverty, you are setting children in those communities up for failure. Whereas, developing policies for integration creates more cultural mixing which can have a big effect in the long run.

I think that’s right. My view is that behavior and individual choice matters a great deal, but choices occur in a framework that we create. The set of norms that prevail in a given area are a function of the types of institutions and policies that we have.

So when people talk about culture being totally outside the domain of what government can do, that’s incorrect; those cultures are shaped by the policies that we set up. A good example of that is Richard Rothstein’s The Color of Law, which shows how government policies led to the racial segregation that we’re seeing in America today. That is an illustration of how we’ve created certain structures through government policy that lead to the development of a certain set of cultures and norms that in turn affect behavior. I think that is the productive way to look at these issues.

So if you were designing policy here, what are the two or three things you would do to close the life expectancy gap?

The first set of things I would do is come back to what we’ve talked about before: trying to reduce economic inequality and create a greater social mobility through school policy and integration. In the long run, those are factors make a huge difference.

Second, I would try to directly change health behaviors. This is outside the scope of work I’ve done myself but includes public health campaigns to reduce things like smoking. I think that can be incredibly effective. Also, we’ve effectively subsidized many unhealthy foods in the United States like high-fructose corn syrup. There’s no reason to be making those products cheaper.

There have been a couple of great profiles written about you recently. One of them was by my colleague Dylan Matthews focusing on the way you’re teaching a major course at Harvard. There’s another one in the Atlantic that caught my eye because it shows how maps that you’ve created look a lot like maps of slavery. They put up one of your maps side by side with an antebellum map showing places that had the highest concentration of slaves.

Then, as I was looking through more of your papers to prepare for this interview, I was noticing those maps again; I just couldn’t stop seeing the legacy of slavery and racism on those maps. I’m curious what you think in terms of our historical legacies around race and the role they play in a lot of the questions you’re studying.

That’s a clear fact in the data: Places that had slave plantations have much lower levels of upward mobility today. There are a few different factors that drive this. First, in the United States today there are extremely large racial disparities in upward mobility, particularly between black and white men (with black and white women we actually see more similar rates of mobility). A black man born to a low-income family has much lower odds of climbing the income ladder than a white man born to a family of comparable income. Conversely, black men born to high-income families have a much higher chance of falling down the income ladder in the next generation relative to white men.

The way I like to think about this is that the American dream is like climbing a ladder for white Americans, whereas it’s more like being on a treadmill for black Americans. Even after you climb up in one generation — if you’re living in an affluent neighborhood and going to a good school and so forth — you still have a great chance of falling down in the next generation. It’s that treadmill feature that leads to the persistence of racial disparities in the United States generation after generation. We really have to fix that in order to have a sustained impact on racial disparities.

I think one of your most striking findings is something you just referenced: the huge difference between intergenerational opportunity for black men and black women. I think this disparity is quite damaging to explanations for the racial wealth gap that attend to culture or genetics. Given that we know there are very different stereotypes and societal treatments that attend to black men and black women, it really seems to suggest that how society is seeing people is what matters, not the genes or culture of the people themselves.

I totally agree. If the explanation was genetic, there is no way you would see gender differences. In fact, if you look at earlier arguments looking at test score differences, you actually see no differences in terms of test score outcomes between black and white women. Yet you see tremendously different outcomes in the labor market where black women have college attendance rates comparable to white women and higher than white men, controlling for their parents’ income. Whereas black men have incredibly low employment rates, particularly if they grew up in low-income families. And I think this reflects environmental factors that are quite prevalent across the United States.

The other thing those maps reveal to me is the absurdity of what I like to call the “zero-hour myth” — the myth that at some moment we hit a zero hour where previously America was unequal and discriminatory, but now that we’ve passed this one law or had this one court decision, there is equality from here on out. Therefore, anything that is redressing past historical injustice is interpreted as a present discrimination. But the similarities between your maps and maps of slavery are a reminder of the way that our past injustices and discriminations and conflicts affect things today.

That’s right, but an example of how they affect things in non-obvious ways is that when you look at who has very low rates of upward mobility in areas with a high rate of slavery in the past, it’s not black Americans — it’s actually white Americans in those areas who have the lowest rates of upward mobility. If you look at a map among white Americans, the lowest rates of upward mobility for whites are in highly segregated areas with a history of slavery, racial discrimination, and so forth. I think what that shows us is that the legacy of some of these institutions from the past created more segregation by income and race, which reduces mobility.

That actually isn’t what I had realized was happening in those maps. But something you just made me think about is something I heard listening to Nikole Hannah-Jones recently. She makes the point that because of the way Brown v. Board was enforced the South has the least-segregated school districts in America. So I wonder if one reason that African Americans are doing better than one might expect is because you actually have more integration and, for a period of time, there was a lot more attention on trying to right some of the historical wrong.

I think that’s right. Also, traditionally in the South, there has been less funding for public schools than in the rest of the country, and less public investment in general. I think that could be part of why you see lower rates of upward mobility for whites.

So you do not see a difference in African American mobility in those areas?

Imagine that you’re looking at two separate maps: a map of upward mobility for African Americans based on where they grew up, and a map for white Americans. If you look at the first map for black Americans, you would actually find slightly higher rates of upward mobility in much of the Deep South. But for whites, there are much lower rates of upward mobility throughout the South than you see in the rest of the country.

I should add one important caveat here, which is when you look at black versus white men uniformly across the country, you see lower rates of upward mobility for black men relative to white men. There’s a big racial difference throughout America. However, black men often have higher mobility rates in the South than other places.

I want to ask you about another kind of mobility. You’ve spoken about the fact that your family was able to immigrate here from India when you were young and that’s the reason you’ve had the particular life you’ve had. Borders are one way we keep mobility extremely artificially constrained. So I’m curious how you think about the opportunities that exist by tightening or loosening border control and immigration laws.

You’re right that immigration played a profound effect in my own life. And when you look at the data, you see that immigrants who come to the United States have the highest levels of upward mobility — much higher levels of upward mobility than children born within the United States. I think one way to think about what’s going on there is many immigrant parents come to the United States in search of opportunity. They themselves may not be able to climb the income ladder much in their own generation, but their kids often end up doing extremely well, even if the parents are quite low income. So from a global welfare perspective, I think giving more people access to those types of opportunities can be extremely valuable.

The best evidence that we have on the impacts of immigration on natives from scholars like David Card and others is that those effects aren’t all that significant. In some studies they actually look somewhat positive. High-skilled immigrants in particular increase productivity and lead to more innovation. So from a social welfare point of view, there’s certainly something to be said for having high rates of immigration and allowing people to flow across borders.

I have one more question on the international scene. I saw some research a number of years ago arguing that mobility is higher in much of western Europe and Canada than it is in America. I know this is not primarily where your research is focused, but I’m curious if you have a global take on how we compare with other peer nations in terms of mobility.

If you were an immigrant choosing where to go and have the best chances of climbing the income ladder, then statistically you’d have a better shot of achieving “the American dream” if you’re growing up in Canada or in many Scandinavian countries than the United States. That’s just a fact. If you’re in a lower-middle-income family, you’re more likely to climb the income ladder in those countries.

A more nuanced take on that, which I think is pretty interesting, is that the US has a lot more variation across places than other countries do. In Sweden, it matters much less which specific neighborhood you grow up in; in the US, it matters a ton. There are parts of the US that have even higher rates of upward mobility than Sweden on average, but there are also parts of the US that look much, much worse than any country for which we currently have data. So the US has this incredibly varied mix where there’s some places that are truly lands of opportunity and other places that are lands of persistent poverty.