Chicago Booth Review Podcast Why Is Housing So Segregated?

Milena Almagro: However, by contrasting our research, what we find is that out of the segregation that we observed, one-third can be explained by these restrictions.

Hal Weitzman: Many American cities are highly racially segregated. Explicit segregation is illegal today, but some patterns of settlement haven't fundamentally changed. So why does housing segregation persist? Welcome to the Chicago Booth Review Podcast, where we bring you groundbreaking academic research in a clear and straightforward way. I'm Hal Weitzman.

Today I'm talking with Chicago Booth's Milena Almagro about her research on racial housing covenants. What do the racist housing rules of the 1940s tell us about housing policy today? Milena Almagro, welcome to the Chicago Booth Review Podcast.

Milena Almagro: Thank you. Thank you for having me.

Hal Weitzman: We're delighted to have you and we're here to talk about housing markets and your research on housing, what drives housing markets. And you talk about two different things that do that. One is household preferences, I guess that's demand, and one is supply side restrictions. So just explain, if it's not too obvious, explain that distinction and why that's important in your work.

Milena Almagro: Let me give you a very clear example and I think the best example is going to be Chicago. So if you look at racial segregation in Chicago, you see that non-Hispanic white people live in the Northeast Lincoln Park, Hispanic people live West, so Little Village, La Villita, African American people live in the South. So one question is if you see that pattern of segregation, what is driving that?

Is it the fact that Hispanic people want to live near other Hispanic people? So that's what we call preferences. Or is it that they can only consider the neighborhoods in the West, meaning that, for example, they cannot afford to live in Lincoln Park, so Lincoln Park is not even in their consideration set. So that's why we call supply side restrictions or the formation of consideration sets.

Hal Weitzman: Just explain that because if people choose to be in a neighborhood that's affordable versus people choosing to be in a neighborhood that's fancy, there's not necessarily a racial component to that.

Milena Almagro: Exactly. But one of the hypotheses that we entertain in urban economics is that people have homophily bias, there is homophily bias, which is a term that means that you want to live near other people of your own ethnicity or race. And that's conditional on income. So you could be high income and you still might want to live near people who look like you, or you can be low income and still exhibit those preferences. And it's very hard to separate those things because race and income are highly correlated. So that's what we are trying to disentangle here.

Hal Weitzman: Got it. Okay. And to do that, you actually went and did some historical research focusing on Minneapolis in the year 1940 and a lot of your research revolves around racial covenants. So first of all, set the scene for us, why Minneapolis 1940 and what are these covenants?

Milena Almagro: So first of all, let me define what racial covenants are. So racial covenants are clauses in property deeds that prevent the sale, occupancy, or rent to households of certain ethnic background or race. So these are private contracts and these are typically written by the developer. And the nice thing about Minneapolis is that there is this project run by the University of Minnesota called Mapping Prejudice that what they've done is they've gone through all the historical archives and digitized the individual covenants at the house level.

Hal Weitzman: So what would the covenant say? How would it restrict people from moving in?

Milena Almagro: Literally with a person of ethnicity blah shall not live, rent, or buy these premises.

Hal Weitzman: Okay, and this was completely legal?

Milena Almagro: This was completely legal until 1948. There's a Supreme Court ruling Shelley v. Kraemer that makes them unenforceable. These were contracts that were enforced by the local police or the local government, yet you see them being written in properties until 1960, but they were not legally enforceable. So that's kind of a question of why, which we don't explore, but it's an open question of why do we see racial covenants even after they were made illegal.

Hal Weitzman: I see. Okay. But the period you're talking about they were legal.

Milena Almagro: They were legal.

Hal Weitzman: And importantly, they're written, as you say, by the developers themselves.

Milena Almagro: Exactly.

Hal Weitzman: They put up the buildings and they decide who can live there. I mean, they're the primary gatekeepers. We don't necessarily think of developers as setting race rules.

Milena Almagro: Yeah. And what is interesting is there's some covenants that are written by the owner. Those are retractive covenants, but there are very few of them. Most of them are, as you said, written by the developers. So the idea is that think of Minneapolis in the 1930s and '40s as a city that is expanding. It's basically becoming urbanized. We're transforming farmland into suburban subdivisions.

So think about a developer buys a big plot of land, and all of a sudden you build 100 houses that look identical and you be like the streets. And the developer is writing the properties of those houses. And the developer thinks, "Okay, I add racial covenants." Why? Because back in the day, unfortunately, these preferences about or trying to move away white flight, I don't know if you have heard that expression, those preferences were very, very strong and very, very prevalent.

Thankfully, things are changing over time. So developers thought, "Oh, so if I put a racial covenant and I basically prevent people from certain race or ethnicity to come into my neighborhood, this is going to be perceived as a higher quality neighborhood, so I might be able to charge a higher price." So there could be a monetary motive, financial motive for these covenants. It could also be that the developer is a racist person and thinks that's the right thing to do.

In this paper, we cannot tell those apart. But what is interesting is that these admission decisions were to admit people of certain race or ethnicity are done at the neighborhood level. So basically covenants in this historical setting basically create restrictions at the entire neighborhood level, which is something that you barely see in real life these days because actually it's illegal. Actually to have any form of racial zoning has been illegal since 1917.

Hal Weitzman: Right. So as you say, I mean, we don't know what the motivation was, but the effect of it is racist. I mean, the effect is to limit who can move in based on race. You said that you... Well, first of all, I'm interested in Minneapolis. I mean, you said that it has particular characteristics. It's growing. Is that true of many? I mean, does Minneapolis stand in from many US cities in 1940?

Milena Almagro: So I think many US cities in the North of the US experience a similar path. Why? Because you get a lot of immigration from Europe in the first half of the 20th century and a lot of immigration from Scandinavia, Western European countries. And you have all of the Midwest and North Dakota, South Dakota being slowly populated by these European immigrants that are arriving.

Why Minneapolis? Because the racial covenants, the nice thing about having this data is that we have data on explicit discrimination, and this is very rare. Typically, when you have research on discrimination, you have to come up with a very clever identification strategy or an experiment. However, for this particular setting, we do see that they...

Hal Weitzman: Right. They were quite explicit about it. Yeah.

Milena Almagro: Yes. Absolutely.

Hal Weitzman: And so the data you use is from the project that you mentioned, which actually has... You've collected all these covenants and coded them somehow, I guess.

Milena Almagro: Yeah, they did. They had five people reading each property deed. This was pre-AI.

Hal Weitzman: Right. Laborious work.

Milena Almagro: Yes, yes. And sometimes what is interesting is that the language, they know that what they are doing is not ethical. So the language sometimes is very obscure. Sometimes you read a covenant and a covenant is very clear, and sometimes a covenant is not that clear in how they're expressing those restrictions.

And what gets more interesting is that the language of the covenants is changing over time because the race and ethnicity towards we want to discriminate too is changing. In the 1940s, there's a lot of religious discrimination, a lot of discrimination toward people that are not from the Western world, which includes Southern Italians, Hispanics, not Hispanic from... I mean, obviously Hispanics from...

Hal Weitzman: Roman Catholics, basically.

Milena Almagro: Correct. Yes. Yes.

Hal Weitzman: I see. Okay.

Milena Almagro: So it would be like a minority back in the 1940s.

Hal Weitzman: Amazing. Well, we're going to get into that a little bit, but that's such a fascinating part. This is who's considered to be white or who considers me acceptable, white, or whatever. So just one other thing, because a lot of us, particularly in Chicago, will have heard of redlining. What's the difference between these covenants and redlining?

Milena Almagro: So these are two instruments that can be used to discriminate people, but they work in very different ways. So covenants, as I said, those are private contracts in property deeds. Redlining, it's a grading created across different neighborhoods, A, B, C, D. There is a lot of discussion, whether it was the Home Owners' Loan Corporation, HOLC or the FHA, the institution that was creating this grading, and those grades were used to rate the risk of mortgages.

So disadvantaged neighborhoods, neighborhoods with low income people, and neighborhoods with racial minorities typically were given a lower grade or the lowest grade, which meant that it was harder to get a mortgage in that neighborhood because the rate that you had to pay, the interest rate, was higher just to cover for what was a higher perceived risk.

Therefore, that made homeownership for these minorities harder relative to the non-minority neighborhoods. So these two instruments can be used to discriminate, but they work in very different ways. One is through a financial instrument. The other one is directly preventing who can live where.

Hal Weitzman: Okay, fascinating. And part of your methodology talks about what you call a latent choice set. Tell us what that is and why it's important.

Milena Almagro: Yes. So latent choice set is a fancy term for understanding what you can consider in your choice set and what you cannot consider in your choice set, even though I don't observe your choice set. So let's go back to the example of the Hispanics in Chicago. Maybe some people cannot afford to pay, I don't know, a two million condo in Lincoln Park because they just don't have the financial means to do that.

And we have, unfortunately, borrowing constraints and there are constraints to credit. Even though latent choice set would be, what is the consideration set of these Hispanic households? What are they considering in their location choices, in their residential choices? The methodology that we have... We haven't invented it. There is a lot of people in marketing, an industrial organization trying to think about what consumers consider.

We are importing those techniques and putting them in an urban economics context with the idea that even though we don't observe what African Americans versus Hispanics might consider, with the right variation, we can actually elicit and estimate different consideration sets for minority versus non-minority households. And that's what we do in the paper. And once you have the consideration set and I know you're choosing from A, B, C, and D, I can actually elicit the preferences that you have for living in A, B, C, and D given the characteristics of those neighborhoods.

Hal Weitzman: And that tells us what about demand?

Milena Almagro: Exactly. So once I have the consideration set, I can construct demand, but that's why we call the consideration set supply side restrictions. In this context, it's because it's the developers putting the racial covenants and the developers are steering what you can consider and what you cannot. The demand part is given what you can choose, tell me what's your favorite place and I'll tell you your preferences for living near people of your own ethnicity or how sensitive you are to housing prices or how much you dislike commuting to the city center. But that's conditional of the...

Hal Weitzman: So these are the household preferences we talked about at the beginning.

Milena Almagro: Correct. But in order to estimate that in the right way, you need to first have a knowledge of what they are considering. Because if you assume that they can consider everything at once, then you start getting the wrong estimates for these preferences.

Hal Weitzman: Carry the Two is the show that pulls back the curtain to reveal the mathematical and statistical gears that turn the world. Co-hosts Sadie Witkowski and Ian Martin bring unique perspectives from the field of mathematics and statistics to convey how mathematical research drives the world around us. With each episode tackling a different topic, subscribe to Carry the Two, part of the award-winning University of Chicago Podcast Network.

Milena Almagro, in the first half we talked about your research on housing covenants and fascinating that you mentioned how certain groups that we now consider to be white weren't always considered to be white or weren't considered to be desirable maybe is a more accurate way of putting it. And in your research, you actually cite... I found this fascinating.

There's a real estate economist from the 1930s, Homer Hoyt, it sounds like a made up character from The Simpsons, who said he ranked ethnicities from most favorable to least favorable. So just explain that and how it's relevant for your work.

Milena Almagro: So Homer Hoyt, he's an urban planner of the 1930s and he writes this book that had a lot of influence. And basically the desirability of different ethnicities is actually used for developers, like everybody in the real estate world, of how we should classify neighborhoods in terms of like, should we give a mortgage to this person who wants to live in this neighborhood or is this neighborhood going to go in the right trend?

Hal Weitzman: But he was ranking people based on their ethnicity.

Milena Almagro: Absolutely. And this is how sad things were 100 years ago and thankfully we're changing. But the idea is that you have a rank not by race only, by ethnicity. You have the Northern Europeans and people from the Western World as the most favorable. Then you have Northern Europeans. And then you go down and then you have Southern Italians, Moorish people, Black people.

They also used terms that we shouldn't use in this podcast. I'm pretty sure people can think of those terms. So you have this rank of ethnicities and what we do in the covenants paper is that we merged the data on covenants to the 1940 Decennial Census, which is a census with exceptionally good data quality. And for every person, we know the country of origin so we can actually construct the ethnicity. And we see what's the effect of covenants on different ethnicities.

And it turns out that our empirical analysis reveals a very similar ranking to Homer Hoyt's ranking in terms of covenants were hurting people more or were creating stronger restrictions for non-white households for white, but like non-Western white back in the day, which is, as I said, Southern Italian, people from Spain, Greek, et cetera, et cetera. And then you start moving up and it pretty much aligns with his classification, which is something that...

Hal Weitzman: This was really the reference book for developers it sounds like.

Milena Almagro: So it was a reference book. We learned about it after we started doing this research, and we kind of use it as a way of externally validate that what we're getting from the data actually makes sense through the lens of the 1930s, 1940s views of what... I don't know.

Hal Weitzman: And it correlates pretty strongly, you're saying?

Milena Almagro: Yes. Yes.

Hal Weitzman: Okay. Fascinating. Now, you said that Minneapolis was interesting because it was expanding into these sort of greenfield sites as one day it was a field, next day it was a new neighborhood with a hundred houses exactly the same as what you said. And you use what you call an Urbanization Frontier Index to track how as the city expanded, these covenants were put in place. Tell us about that.

Milena Almagro: So the reason why we construct this Urbanization Frontier Index is because the placement of covenants is not random. First of all, it's a decision made by the developer that can correlate with the underlying characteristics of neighborhoods or of the location. So my co-author Aradhya Sood in a previous paper documents that covenants are more prevalent in middle income neighborhoods. Why? Because even though people might want to discriminate in high income tracks, but those places are very expensive.

Hal Weitzman: They don't need to.

Milena Almagro: Exactly. Exactly.

Hal Weitzman: So this is where the races might mix will be in the middle income elements.

Milena Almagro: Exactly. Exactly. So we have evidence that racial covenants, the placement of those are not random across the space. So we construct this index to try to purge out this endogeneity, that actually to estimate the casual effect of covenants on the presence of different races in that neighborhood or ethnicities if I have to be more precise. The idea is that we combine three facts.

First of all, is the urban expansion of Minneapolis that we already discussed. Second is that if you see the urban expansion of a city, it happens in a radio gradual way. You don't see a house popping up there and the other house popping up there. It kind of follows a pattern, like a concentric spatial pattern. And the third fact is that during this urban expansion is when racial covenants are the most or is a very popular instrument in the US.

So we're combining the fact that covenants are gaining traction with the fact that Minneapolis is expanding. So what does this tell us? It tells us that if we look at sites that are candidates to be developed next, which is the Urban Frontier Index, because developments in the 1930, 1940 are more likely to have covenants just because of the popular trends of covenants themselves.

Then if we put these two together, we say, "Ah, developments that were built around the frontier during these decades are more likely to have covenants." And the idea is that we know that you're going to be along the frontier, but exactly which plot of land gets available is random because we're converting farmland into suburban additions. And this conversion usually happens because the farmer dies or the farmer enters into a financial crisis.

And there's a lot of documentation of that, that the way that new land was bought and being urbanized didn't follow a precise pattern. It was more of coming from these random shocks.

Hal Weitzman: So we started by talking about what's driving housing markets. Is it demand basically? Is it household preferences, or is it these restrictions? Do we have a definitive answer to that in the case of Minneapolis?

Milena Almagro: Not yet. We're still working on it, but some preliminary numbers that I want to give you is that a lot of working urban economics have assumed that everybody can choose from everywhere. And they get very strong homophily bias preferences because you need to explain these segregation patterns that we see in the data, the US is super segregated. So you get this...

Hal Weitzman: That would say that the Hispanics want to be with other Hispanics because they just want to be with them.

Milena Almagro: Exactly. However, by contrasting our research, what we find is that out of the segregation that we observe, one-third can be explained by these restrictions. So legal restrictions, these legal contracts were explained one-third of the way, and two-thirds are these preferences for living near your co-patrons.

So the punchline is that removing restrictions can get you halfway there. But if we want to get rid of segregation because we think segregation is bad for whatever policy we think could be bad for, then there's a part of preferences that is still going to drive segregation and that's something that you cannot tackle with policy. So policy at least in this context...

Hal Weitzman: It looks like self-selection, whatever, right? People-

Milena Almagro: Exactly. Exactly.

Hal Weitzman: ... moving close to other ethnicities. As you say, we're in Chicago, which has every town, Greek town, even Japan town. There's all these little areas. So that would suggest that some of that at least is because of people want to be with other people who share their ethnicity. So you actually ran this counterfactual simulation, see what would happen if you didn't have these covenants. What did you find there?

Milena Almagro: So precisely that the similarity index, which is an index that you use to measure segregation, falls a little bit from 0.49 to 0.42. I'm just giving numbers that are hard to interpret, but for people who have studied segregation, it's a city that is quite segregated. If you remove covenants, segregation falls by a bit, but it doesn't fall by a lot precisely because these preferences are still in place. And it's only when you remove the preference...

Hal Weitzman: But you're talking there about the demand side?

Milena Almagro: Correct. The preferences for living with the homophily bias. I don't like to use the term homophily bias. That's the term that people use in urban economics. I like to say more preference for co-patrons because I don't think it's a bad... I don't know. It's a controversial...

Hal Weitzman: You don't like the term bias when it comes to race.

Milena Almagro: Yeah. Yeah. Yeah.

Hal Weitzman: Well, it's sort of counterintuitive here. Actually, so going back to the real data, even in the neighborhoods without these explicit covenants, minority households was still less likely to be admitted. So there's implicit discrimination.

Milena Almagro: Absolutely. Yeah. So we not only find discrimination coming from covenants, we also find that every neighborhood inside Minneapolis, a minority household is eight percentage points less likely to be admitted into that neighborhood just by baseline.

Hal Weitzman: What do you mean by admitted?

Milena Almagro: Like that neighborhood is in their choice set, in their consideration set. And we say admitted because the developer is making the admission rule. So that's kind of a language that we're probably misusing in the paper. But basically every neighborhood, the baseline probability of considering that neighborhood is eight percentage points less likely for minority versus non-minority, which tells you that even without covenants, even without the instruments of covenants, or even without financial constraints, minorities are facing a harder time in being able to consider all neighborhoods.

Is this implicit discrimination? Is this something else? We don't know. We cannot answer that with our research, but at least we get evidence that it seems that there is a gap even before covenants and financial restrictions.

Hal Weitzman: Okay. Now, some people be listening to this, Milena, and thinking this isn't history, this is just today. What you're describing as a more formal version of what goes on today. What we used to call covenants, we now call zoning.

Milena Almagro: Exactly. Exactly.

Hal Weitzman: I mean, talk about that. What does your research tell us about housing policy today?

Milena Almagro: So let me go back to a very clear example because I think that's the best way of crystallizing ideas. Think about minimum lot sizes, which is a zoning policy that is prevalent in the US these days. And even though this policy is not about race or about income directly, but it has indirect consequences on who can live where. Why? Because you have these minimum lot sizes that, of course, are going to push prices of neighborhoods up because you have to buy a bigger plot of land.

You have to build presumably a bigger house that could also be part of the zoning regulation. Therefore, that's going to leave out everybody who cannot afford to live in that neighborhood. And even though it's not explicit about race, we know race and ethnicity is highly correlated with income, unfortunately, even until today. I hope that changes in the next decades.

But what we are trying to say is that, okay, you cannot assume that low income people can choose to live in these areas with huge minimum lot sizes because of regulation because these neighborhoods are not even in their consideration set. If you assume that, you're going to get bias estimates for demand, like my preferences for living with co-patrons or my sensitivity to prices.

So let me give you a very clear example. Let's say that these restrictions are in places that are very, very expensive and those places for a low income household are not in their consideration set because I cannot pay the down payment. But I wrongly assume that you can choose from there, yet I never see you choosing from that super expensive neighborhood.

What is it that I'm going to infer? That you're actually more sensitive to prices because you restrict your choices to cheaper places. And it's not about the sensitivity to prices, it's about I couldn't get the down payment for this mansion in this huge minimum lot size area. So that's why even though this is for Minneapolis, a historical context, there are many things, many regulations that create these restrictions that...

Hal Weitzman: Without being explicit about race.

Milena Almagro: Yes. And that's why another lesson that we want to bring forward with this paper that it's not only also about preferences, but also these are going to have equilibrium consequences. We were talking about segregation in our counterfactual simulations. These restrictions also have equilibrium consequences that affect housing affordability. This is like the discussion that we want to put forward with this paper.

Hal Weitzman: Well, thank you for coming here to start that discussion. We'll continue that discussion here on the Chicago Booth Review Podcast. Milena Almagro, thank you very much.

Milena Almagro: Thank you.

Hal Weitzman: That's it for this episode of the Chicago Booth Review Podcast, part of the University of Chicago Podcast Network. For more research, analysis, and insights, visit our website at chicagobooth.edu/review. When you're there, sign up for our weekly newsletter so you never miss the latest in business-focused academic research. This episode was produced by Josh Stunkel. If you enjoyed it, please subscribe and please do leave us a five-star review. Until next time, I'm Hal Weitzman. Thanks for listening.