HKS Professor Maya Sen, co-author of the new book “Deep Roots: How Slavery Still Shapes Southern Politics,” explains how she and her colleagues were able to pinpoint the extent to which slavery continues to affect political beliefs to this day.

Featuring Maya Sen
May 21, 2018
34 minutes and 27 seconds

In the aftermath of most major election cycles, researchers and academics spend considerable time and effort divining how events leading up to election day might have swayed voters towards one candidate or another. But while these analyses often focus on factors such as economics or the media, they too often overlook what might be the most critical factor: history.

That’s one of the conclusions reached by HKS Professor Maya Sen, Harvard Professor Matthew Blackwell, and Stanford Professor Avidit Acharya in their new book Deep Roots: How Slavery Still Shapes Southern Politics.

To illustrate the point, the three co-authors took a close look at states in the American South where slavery was once legal, breaking them apart at the county level. They found that counties across the south where slavery was most prevalent in 1860 were distinctly more politically conservative today than counties where there were comparatively fewer slaves.

In this episode of PolicyCast, Professor Sen joins us to dig into her findings, describing the difference between political attitudes and behaviors (with particular attention to the impact of the Civil Rights movement,) and explaining how attitudes persist over time thanks to “Behavioral Path Dependence,” a process by which families and communities establish and maintain attitudes over generations.

Hosted by

Matt Cadwallader

This episode is available on Apple Podcasts, Spotify, and wherever you get your podcasts.


Note: This transcript was automatically generated and only lightly edited.

Matt: So this was a fascinating book, I really enjoyed going through it. What it was that made you look into this in the first place?

Prof. Sen: Yeah so there are a couple of different paths that lead us to this book, but actually the one that was the most straightforward, was that we’re at a time when a lot of scholarship is being done on the long term impacts of institutions, and that’s mostly been done outside of the US. So there’s a lot of really, really excellent research out there kind of showing institutions that kind of took place or maybe collapsed earlier in human history, still have a lingering impact on us today.

So I’ll give you a couple of examples. So there’s a really, really very interesting paper that two economists Nathan Nunn and Leonard Wantchekon wrote. Where they show that people in Africa who live in places that were exploited by the slave trade 400 years ago, those people in Africa today in those parts of Africa are today more distrustful of strangers. So 400 years after the slave trade collapsed, you can still detect kind of this lingering persistence of it.

There’s another example from a paper by Melissa Dell who’s an economists on the Econ department here. Who shows that places in Colombia that had kind of a reliance on 18th century form of labor coercion similar to slavery. Those parts of Colombia today have actually poor outcomes for kids. So there’s these long term impacts of that kind of earlier system of labor coercion. And they’re just these examples that we’ve noted kind of all over the world.

I study American politics, and to me the relevant question coming out of that research was, we see that these institutions had impacts, downstream consequences in other parts of the world, it surely must be the case that the United States is also a product of earlier institutions, institutions earlier in our history? And my collaborators and I got to talking like what is the institution that really came to shape the early United States? And immediately we started thinking about slavery, and the long term impacts that slavery could have had on our politics and our political attitudes.

We were also inspired by other people who’ve worked on the lasting legacies of slavery. There are just numerous historians obviously who’ve worked on this topic, but also in terms of social scientists and people who work with quantitative data, there have been … There’s one sociologist who documented that slavery appears to have had long term consequences on income inequality in the south, which is a finding that’s been supported by economists such as Nathan Nunn.

So we got to thinking, well what would it mean for politics? And that’s where we got to thinking about the project and the long term impact that slavery could have had on political attitudes.

Matt: When you’re looking at these things, we’re talking about centuries separated events. How can you be sure that this particular catalyzing event, I believe you called them critical junctures. How can you be sure that they are the cause of what you find down the road, 150 or however many years later?

Prof. Sen: I would say it’s actually not that long ago. We’ve had two kind of responses, some people say exactly as you did. How could something that happened so long ago possibly impact me today? The counter to that which people have pointed out to us is 150 years is basically the lifetimes of two 75 year olds put together. We all know, most of us know someone who’s 75 years old and chances are that that person that we know probably knew someone who is also older when they were younger right?

Matt: Sure.

Prof. Sen: So this is not that long ago, it’s about two generations ago that this happened. When you think about slavery’s legacy moving forward as well, emancipation happened about 150 years ago, but following that we had reconstruction. We had Jim Crow which essentially didn’t collapse until the 1950s and 1960s. We all know people who were alive in the 1950s and the 1960s. So this history is not really that ancient in the grand scheme of human development.

In fact when you think about it, racial violence continues to this day but lynchings really continued into the ’30s and ’40s. So people who are alive and voting today, may have been participants in some of the racial violence of Jim Crow. So we don’t think it’s too far removed actually from where we are today. That’s sort of a separate question of like how do we know whether this caused it. The truth is that slavery was such a defining institution in early American history that had an impact on nearly every facet of American life.

So it had a direct impact on the Southern economy. The Southern economy had a direct impact on the national economy. It’s structured the way that politics developed in the south following from its collapse through reconstruction and through the early part of the 20th century. In terms of demographics it profoundly changed what the United States look like. At the time of slavery’s collapse blacks made up about 30% of the entire southern population, about 10% of the American population overall.

African-Americans to this day are about 8 to 10% of the United States population. A lot of them are clustered in parts of the South that once were historically reliant on slavery. So this has had a economic, political, and demographic impact. So we think that may be one way to think about how do we know this has had an impact is actually think holistically about all the ways in which slavery has actually impacted American politics and culture.

This is all a long winded way of saying it’s obviously impacted the way that we think about politics today. What we do in the book is we very crisply trace it. And we do so looking at geographic units. So we look at counties that were very reliant on slavery, and what we show in the book is that those counties are places where whites today are more conservative. We look at a host of other possible explanations. So we spent I think three or four chapters kind of ruling out other explanations, and we see no evidence for these other explanations.

In fact, all of the analyses that we look at essentially point to and keep pointing to slavery as the driver of these kind of regional differences.

Matt: Well what’s fascinating is how you’re able to break up, this isn’t about the north versus the south, or having racism more prevalent in the South. It’s about within southern states where there was … I’m sorry, slavery was legal. Even between counties you were seeing a difference that correlated to what was true … Have the number of slaves in a particular county in 1860. That seems extraordinarily specific even given the relatively recent past that 150 years is.

Prof. Sen: Yeah that’s right. I think, just a couple of points on that. I think one thing that’s hopefully impactful about our work is that I think a common misunderstanding is that there’s not much variation in the south in terms of political attitudes. And I think this is something that a lot of people are guilty of assuming. Scholars, political observers, people in the north are guilty of assuming this. But I think what we find when we look at the data is that there’s actually a lot of regional variation in southern attitudes.

Some places are actually fairly liberal on race and race related issues, some places are quite conservative on race and race related issues. In the book we juxtapose a lot of these different cities. So in the introductory chapter we look at Asheville North Carolina, which historically has been relatively progressive by Southern standards. And we juxtapose that with places in the deep black belt south. So the city that we look at really closely is a place called Greenwood Mississippi, which is in the Mississippi Delta.

So places there whites are very conservative. And given that the South is the most conservative region of the United States, these are individuals who are probably amongst the most conservative in the entire country. This is just kind of on a aside point which is that there’s tremendous variation in regional attitudes. Now what we do in the book is we help explain some of that variation by looking at the political economy, and the history of these places, and tying that back to the institution of slavery 150 years ago.

Matt: I think one thing that some people listening to this might hear is a conflation of these negative racial attitudes, prejudice with conservatism.

Prof. Sen: Yeah.

Matt: Did you consider that, and how did you think about that?

Prof. Sen: Yeah so it’s an excellent point, and we thought about it in a bunch of different ways. So It’s true that in the South generally true, that race tends to protect partisanship sort of most crudely. So the South was sort of solidly controlled by a Democratic Party for a large swath of the 20th century moving into the 1960s, when the party sort of realigned and the Republican Party sort of then became the racially conservative party. So it’s true that the party represented racial attitudes particularly so in the south. Today we understand the Republican Party sort of mean the party that’s most conservative on race and race related issues.

And so we would not be surprised to see the attitudes on race than correlate specifically with party affiliation. Now just to separate that aside, we’re particularly interested in how slavery impacts political attitudes. And we’re also interested as part of political attitudes, we think attitudes on race and policies that surround race are really important. Not only do those impact the party platforms, but they also shed light on people’s racial views. One of the things that we tried to do in the book is we look at how slavery predicts whites attitudes today, not just in terms of their party affiliation but also in terms of attitudes that were specific to race and race related policies.

We specifically tried to disaggregate your exact question. For example we looked at, on the party side we looked at the share of the county that voted for Democratic candidates or who self identified as a Democrat or a Republican right doesn’t really matter there are two sides. But we also looked at white support for policies that could potentially help African-Americans such as affirmative action. And we found that slavery in 1860 negatively predicts white support for affirmative action.

That is if you’re a white person today who lives in a county that was heavily reliant on slavery in 1860 you’re much less likely to support affirmative action than if you were a white person who lived elsewhere. Now the same is true for other measures of race related attitudes. For example we looked at what political scientists would call racial resentment, which is this idea that you’re not asking a person to express racist views, but you’re asking them about questions that indicate a resentful view toward minority groups.

The classic example is something along the following. I’m sort of paraphrasing the question that we looked at but it’s a question that says, “Italians, the Irish and Jews have been able to work their way out of the lower class, African-American should be able to do the same without government help. So agreement with the question would indicate that you believe that African-Americans should pull themselves up by their bootstraps, should not be entitled or expected to receive government assistance in doing so.

Scholars have designated such responses indicating a higher level of racial resentment. Now we can sort of quibble with whether that’s actually capturing this kind of racial resentment idea. But we did find a very strong correlation between slavery 1860 and whites today being more “racially resentful.” Now to further buttress that we actually look specifically at whites attitudes toward African-Americans, asked by this … We would call this a thermometer score question. The question is, how would you rate this group on a zero to 100 scale based on how warmly you feel toward them?

And the surveys tend to ask this question of all kinds of groups. Democrats, Republicans, whites, blacks, Latinos et cetera. And the places that were heavily reliant on the institution of slavery 1860, those are places where whites were the coolest toward African-Americans as a group. So there is a important pattern here which is that these patterns that we’re seeing do tend to be race related. They’re not just ideological or are party oriented. They do have some racial component to it that reflects we believe racial views.

Matt: So how is it that these kinds of views persist? Is it just because you teach your kids a certain way of treating other people and that passes down through the generations? Are there institutional elements to this?

Prof. Sen: So we think they’re both. And I think all of us can come up with examples and instances where our parents taught us important lessons about political and moral values. And we think that roughly that similar things are in play when we’re talking about racial values as well. So I’ll give you an example that’s somewhat lighthearted, but I think makes the point. Let’s say that I grew up in a household where my parents were huge Chicago White Sox fans, I’m making this up.

But let’s say that that’s like a really strong tradition in my family. I grew up going to Chicago White Sox games with my parents or my brothers or my sisters. I got Christmas presents with Chicago White Sox gear on them, and moving forward with my family I take my son to Chicago White Sox games, and he’s grown up to love the Chicago White Sox as well. That’s a really goofy example of something that we would call intergenerational socialization.

Which is that my parents taught me to kind of appreciate the sports team and I grew up to be a fan of the sports team, and in turn I’m teaching my own children that they should be fans of the Chicago White Sox too. And that’s … It’s a goofy example but we can think of other examples such as religion. Religion is something that’s passed down from great grandparents to grandparents to parents to children.

Even things like language passed down from parents to children. Accents, cultural preferences, customs. These attitudes I think there are a lot of examples in which these are passed down over time. We think that political attitudes are as well. And the research actually is really consistent with that. So there’s not our work, but the work of other scholars has kind of indicated that a young person’s political attitudes and even partisan preferences really strongly correlates and tracks with that of their parents. And so it’s very sticky.

So if I’m someone who grew up in a sort of a central left household, I will very likely be a center left person myself. If I grew up in a conservative Christian right household, I will be much more likely to vote that way and then raise my children that way. These preferences are pretty sticky, it takes something … Other research has shown this, but it takes something kind of like moving cross country or marrying someone with a different political viewpoint to really shake people out of these persistent patterns.

Now that’s … socialization that happens at the family level, but obviously institutions are huge component of this as well. So things like churches and schools, community organizations, even informal social organizations are really powerful and carrying on these norms. We have lots of examples of this in the south where children were really raised to hew to this, what sociologist have called this racial adequate. And here in the book we really build on really terrific work from sociologists and historians who’ve done a lot of for kind of building up this narrative that white children from a very early age knew their place in the broader southern society. And they were taught how to treat people who were black or minorities in a way that reaffirmed their “superior status.”

One concrete example of us that I’ll share with you is we read many accounts and we actually saw pictures where children would actually … And this in the ’20s, ’30s actually even through the early 1940s, would actually be participants in racial violence and in lynchings. So in several communities there are examples of this where children would get a day off of school to actually come see someone be lynched, which is really provocative and kind of shocking to our minds today, Again this is not that long ago, we all know people who had parents who were alive at this time period. Maybe if they themselves were not alive at this time period.

There’s actually a picture that we show when we present some of this work of a lynching that happened in 1935 in Fort Lauderdale. And 1935 is not that long ago. Where there’s a picture of someone who’s been lynched and below the body there are actually several little girls kind of sitting there watching the whole thing and they’re actually smiling. And we think this is a … The whole socialization component of children happen through families, institutions but it also happened at a community level, informal community level.

And so seeing these stories I think really illustrated for us why we see such a strong relationship between slavery and racial attitudes’s because of these stories which are so hard for us to understand and listen to.

Matt: Right. So one thing that you’re talking about here is attitudes, but one thing that you looked at in the book was also behavior. And once you got to past slavery and up to civil rights, you have this whole generation of Jim Crow. But after the Civil Rights Act obviously the federal government changed a lot about behavior and there’s a difference in that. Can you talk about that a little bit?

Prof. Sen: Yeah that’s right. So this is an excellent point. So this is actually a way in which we depart from previous scholarship that has looked at this persistence of institutions on contemporary outcomes. A lot of Economists who’ve looked at kind of the long term persistence of institutions have looked at what we would call economic indicators. So like inequality, earnings, employment educational outcomes, these are what we would call sort of more yeah, economic indicators.

Matt: Sure.

Prof. Sen: Which are incredibly important. There’s no question about that. But these are things that can be targeted through very effective legislation. And two examples of very effective legislation in the 1960s where the Civil Rights Act of 1964 and then the Voting Rights Act of 1965. So we do a fair amount of work in the book trying to figure out kind of what the effects like … Not the rigid causal effects but more colloquially what the effects of those pieces of legislation were in terms of attenuating slavery’s effects.

And here I’ll give you, I’ll kind of explain to you what we found. So let’s take the Civil Rights Act, So the Civil Rights Act addressed educational inequalities. So it addressed segregation in education, and it addressed segregation targeted against African-Americans by private businesses. So it addressed these things. I mean it did a lot of other things but I’ll focus on those two things. So before the Civil Rights Act was enacted, we saw that slavery really predicted educational inequalities.

So blacks who lived in former slave holding areas, they had big gaps between them and white children. Whereas blacks and whites who lived outside of slave holding areas we didn’t see these huge gaps. After the Civil Rights Act was passed we saw these differences narrow quite a bit. They don’t fully go away, so people who live in former slave holding areas still see worsens inequalities. But that gap between former slave holding and non-slave holding areas really goes down quite a bit.

And to us this suggests that the Civil Rights Act was really quite effective in terms of addressing educational inequalities. Like these persistent, really historically persistent educational inequalities, it was very effective. We see a similar pattern actually with regards to the Voting Rights Act of 1965.

So before the Voting Rights Act of 1965 you see huge, huge gaps in voter turnout by race. And you see them the highest in states in the deep south, Alabama, Mississippi, South Carolina, Georgia, Louisiana, the huge, huge gaps. So I think it’s something like in Mississippi I think it was something like between 6 and 8% of African-Americans were actually registered to vote. Right, it’s really astounding. After the Voting Rights Act is passed in 1965, you see massive voter enfranchizment of African-American.

Voting Rights Act here doesn’t get exclusive credit because you had also massive voter registration drives that happened as a result of black activism, but it’s a very effective piece of legislation. And in fact skipping ahead to the modern day period, it’s so effective that part of it is struck down in 2013 the Supreme Court because it’s so effective. Our findings are completely consistent with that narrative, that these two pieces of legislation were very effective at addressing these sorts of outcomes.

So educational inequalities, income inequalities, employment, voting turnout. What these two pieces of legislation could not do, and what other pieces of legislation have not been able to do is change people’s attitudes, and that’s where the difference between behavior and attitudes becomes quite important, in terms of policy interventions. Ultimately our story is … It’s a little pessimistic I’m not going to like. We’re still trying to figure out kind of what the policy intervention takeaways are.

We think that the government it has a huge role to play in attenuating in ameliorating in addressing these behavioral outcomes. So things like employment, education, healthcare, voting like voting turnout, legislation is effective for those things. We are not as optimistic that legislation can actually change people’s hearts and minds, that it can change attitudes. A lot has changed between now, between 2018 and 1960s in terms of people’s attitudes. Things that were acceptable to do and to say in 1960 are not acceptable to do and to say today.

We can talk more about how the current political climate maybe has changed that statement. But it’s true that we’ve come very, very far. We think that might be more likely due to kind of the Civil Rights movement, perhaps the spillovers from these pieces of legislation. But the fact is that you can still see regional differences in people’s attitudes persist even after these interventions suggest to us that changing hearts and minds is incredibly difficult. And much more difficult than changing these economic indicators and other kinds of outcomes.

Matt: Right, you’ve mentioned a number of other situations globally that mirror this type of research. Are there examples out there where you see a change? Where there has been this critical juncture. Something as big as slavery or maybe even smaller than that, that has changed attitudes and then there was another shift in the future?

Prof. Sen: Not to knowledge no. But there is some … and we discussed this a little bit in the book. The attitudes that you get from your parents and your grandparents, I would say this is going to sound kind of technical but I think you know what I mean, the transmission is not perfect. I am not 100% of my parents attitudes. So there’s a little bit that’s lost with each generation. And it’s true that the attitudes, that the baby boomers had is really different from that attitudes that millennials have across a variety of different things.

Transmission’s not perfect which is something we just can’t measure how quickly attitudes dissipate, in the book we sort of call it a half life. So we think there’s a half life. We think eventually over time there’ll be sort of attenuation in these regional differences, we just don’t see it now yeah. I’ll give you one particularly depressing bit of research, which is that there’s this paper by [Voint Lener 00:26:32] and [inaudible 00:26:32], they’re two economist. And they look at anti-Jewish programs around the time of the Black Death which is 15th century. And they actually find that places that had these anti-Jewish programs actually today are places in Eastern Europe that have higher rates of anti-Semitism.

So that’s a persistence of about five or 600. So given how long slavery was sort of in effect. Given how much it permeated the natural culture and national politics. I think we’re looking at a bit longer.

Matt: I can understand that. You’ve termed this whole idea. This kind of research realm. I don’t know if it existed before but you use the term behavioral path dependents. For those who don’t know about path dependency as it already existed, could you just define that?

Prof. Sen: Sure so we look just the phrase path dependence kind of refers to this body of literature and it’s fairly robust. That in politics and economics suggest that institutions create corresponding behaviors that then reinforce the institutions moving forward. So institutions have a way … Let me put it more simply. Institutions have a way of keeping on, they sort of keep going even after the incentives that led to their creation have sort of gone away.

We think that that’s part of what’s going on in the south is kind of due to institutions and institutional persistence, but this term that we coined behavioral path dependence is specifically referring to not persistence in institutions but persistence and behaviors. And as a way of explaining that we point to two possible pathways of this persistence. One is intergenerational which is the process that I described earlier which is that all my ancestors have been Chicago White Sox fan and I’m a Chicago White Sox fans. My Chicago White Sox fans. And my kids are going to be Chicago White Sox fans.

So everything that I know I …A lot of What I know and the person who I am is a reflection of my grandparents and parents. But the other part of it is this institutional part of it which is that institutions create attitudes or further attitudes and those persist also right? So we see this literature … Let me rephrase that, see this concept really illustrated very well in some of that literature that we talked about.

So the literature on Africans distrust of strangers in places that were exploited by the historical slave trade. Or persistence in anti Jewish or antisemitic attitudes in Eastern Europe, in places that were the sites of programs during the Black Death. So those are examples of behavioral path dependents, persistence and attitudes.

Matt: Obviously something like the Black Death which killed a huge chunk of the human population, and something like slavery which as you said before affected really everything about early American life. Those are enormous institutions or enormous I guess catalyzing events. Can this type of thing be traced if it were say smaller? The creation of the Chicago White Sox team was a decidedly smaller event.

Prof. Sen: Yeah, I think you probably can. You can certainly trace persistence more broadly. So let me give you a couple of examples actually drawing on Boston. So if you were to walk around downtown Boston you’d see schools that … You’d see schools. You’d see streets that are named after buildings that were once there. So for example you’d see School Street, you would see Milk Street, you wold see water Street I.

These are all places that … These where streets that are named after things that no longer exist. So School Street is the oldest school. Water Street is where the water actually came up to the street. The water now I think is a full half mile from where Water Street actually is, maybe a quarter mile. North Straight I assume there was a milk under in the middle of the street.

Matt: That’s where the cows were.

Prof. Sen: Yeah. Benjamin Franklin I think was born on Milk Street.

Matt: He was?

Prof. Sen: Yeah But these are like … And I think that these are all places and institutions that shaped early Boston, whose presence is they’re no longer there, but whose presence still shapes the life of the city in some way . And you an even see it in kind of the way that Boston is laid out. It sort of mimics this colonial era verb. That you know makes in maddening for anyone who lives here. But [inaudible 00:31:33] because it still retains that kind of, that historical charm. There are lots of examples like that, and those are actually fairly small instances compared to these massive important institutions that really defined historical moments in history.

So I think once you start looking at it from the perspective of persistence you kind of see it everywhere.

Matt: Do you think this is something that political scientists in general should kind of see as an opportunity to help explain more contemporary politics?

Prof. Sen: I do, and this is a point that we make when we present the work to scholars. So a lot of scholarship on American politics really tries to understand the moment, and it does so by looking at things that are happening in the contemporary period. So looking at the election in 2016, many papers are now being written about the effect of the Comey Letter. Or the effect of the Hollywood, with the Hollywood access tape.

Matt: Oh yes, yeah of course.

Prof. Sen: Or the collapse of unions and the years leading up to election and things like this. We want scholars to really take a much more long term view. So the forces that we’re talking about in the book date back two, 300 years and we can still detect them. So we can detect these forces in modern day elections. These forces are much more significant than whether a flier will convince you to change your vote, or whether the Comey letter changed the votes of 0.001% of people living in Wisconsin.

The forces that we’re talking about here are much more deep rooted. They’re very sticky, they’re very hard to move and yet they explain contemporary politics reasonably well. And so we want to encourage scholars of politics to really think more about the historical forces at play in terms of contemporary public opinion. It would be nice too if we could get observers of politics to engage with these ideas as well. Because I think “puzzles” like the Trump election would not be as puzzling if we thought about them in a longer historical context.