How Slavery Changed the US South

Avidit Acharya, Matthew Blackwell and Maya Sen find in a new paper that if it weren’t for the legacy of slavery, white Southerners today would be politically indistinguishable from Northerners.

Drawing on a sample of more than 39,000 southern whites, we show that whites who currently live in counties that had high concentrations of slaves in 1860 are on average more conservative and express colder feelings towards African Americans than whites who live elsewhere in the South. That is, the larger the number of slaves in his or her county of residence in 1860, the greater the probability that a white Southerner today will identify as a Republican, express opposition to race-coded policies such as affirmative action, and express greater racial resentment towards African Americans. We show that these differences are robust to a variety of factors, including geography and mid-19th century economic conditions and political attitudes. We also show that our results strengthen when we instrument for the prevalence of slavery using local measures of the agricultural suitability to grow cotton. In fact, our findings indicate that in the counterfactual world where the South had no slaves in 1860, the political views of white Southerners today would be indistinguishable from those of similarly situated white Northerners.

12 Responses to How Slavery Changed the US South

  1. jonathan September 7, 2013 at 1:58 pm #

    This reminds of the actuality of the composition of the Confederate armies. We have solid information mostly about Lee’s Army of N. Virginia. Contrary to the public image of a widespread public enlistment, the army was composed significantly of people with direct ownership interests in slaves.

    Here are some facts. Most of the men who fought for the army over the course of the war joined in 1861 or 62. Of the enlisted men who joined in 1861, 36% lived in families which owned slaves and 44% lived in households (which includes non-family) that owned slaves. Over half the officers directly owned slaves. The slave owning percentages went up in 1862, with about 48% of enlisted men living in slave-owning households.

    In other words, the popular image of the Southern military is skewed. The reality is it was dominated by slave owners who had a direct financial interest in defending the right to own humans as chattels.

    It’s interesting to note the rhetoric of that day is similar to today when talking about health care! Here is one of my favorite quotes:

    “It is a war of defense against wicked and cruel aggression; a war of civilization against a ruthless barbarism, which would dishonor the dark ages; a war of religion against blind and bloody fanaticism. It is war for your homes and your firesides – for your wives and children – for the land which the Lord has given us for a heritage. It is a war for the maintenance of the broadest principle for which a free people can contend – for the right of self-government.”

    That was Rev. Dr. Benjamin Moore Palmer speaking at City Hall in New Orleans in May, 1861.

    And this quote is nearly as cool. John Dooley, whose father fought in the Revolution, wrote about anyone fighting for the Federals would “consistently turn his back on his principles and for the hire of a few pitiful dollars do all in his power to crush a brave people asserting their right to self-government.” He wondered how someone could “engage in the cause of tyranny, fighting against honesty, Justice, and right.”

    That to me sounds like the words used to attack ObamaCare. In other words, it isn’t just the attitudes but the rhetoric of slave owners that carry through to today.

  2. jonathan September 7, 2013 at 2:00 pm #

    I should note that in contrast to the nearly half of enlisted men and over half – approaching 2/3 – of all officers being slave owners, fewer than 1 in 10 Southerners lived in slave owning households. Slave owners were dramatically over-represented in the army and one can argue the war was the imposition of their views on people who would otherwise rather have not fought.

  3. Avidit Acharya September 7, 2013 at 2:04 pm #

    Hi Henry,

    Thanks for your post! My first name is spelled Avidit, not Avidat. :)


  4. Henry Farrell September 7, 2013 at 2:11 pm #

    Apologies – fixed! Thanks for pointing out.

  5. Andrew Gelman September 7, 2013 at 3:09 pm #

    Not to be picky, but I’d omit the “if it weren’t for the legacy of slavery” bit, which seems meaningless to me. It seems strong enough to just state was was found, “that whites who currently live in counties that had high concentrations of slaves in 1860 are on average more conservative and express colder feelings towards African Americans than whites who live elsewhere in the South.”

  6. Avidit Acharya September 7, 2013 at 5:43 pm #

    Actually, just to clarify: Beyond the baseline analysis, we did a separate counterfactual comparison between southern counties with few slaves, and northern counties that were similarly matched on a variety of attributes, and we see no differences in political attitudes. This is the basis for our claim about Southerners vs. Northerners.

  7. julie September 9, 2013 at 11:06 am #

    Did you run a separate analysis where you included as many counties as possible (northern and southern) and then included those attributes as control variables?

  8. Corrine McConnaughy September 9, 2013 at 12:19 pm #

    Let me begin by saying – I think this is very interesting work, and is importantly pushing behavior research to deal with more than snapshots in time. And yet, I’m with Andrew on laying off the legacy of slavery claim somewhat. In particular, let’s consider this last statement you make here: “In fact, our findings indicate that in the counterfactual world where the South had no slaves in 1860, the political views of white Southerners today would be indistinguishable from those of similarly situated white Northerners.” What is the counterfactual world where the South had no slaves? Does it include mass numbers of blacks imported as indentured servants? Or still imported as slaves but freed much earlier in our history and then worked longer than they did in our actual history as sharecroppers? Freed more slowly that the “sudden shock” you claim as a key mechanism in the paper (weren’t there places with greater concentrations of freed blacks? what would that tell you?)? Is it where some other group was imported to the American South to drive the particular version of elitist agrarian society the South displayed? I could, of course, go on, because the historical couterfactual you’re claiming is simply unclear – and many would argue that the political economic incentives of the South would have lead those elites to some way to use the apparatus of the state to produce incredibly cheap labor to make the agrarian economy best serve the self interests of the Southern agrarian aristocracy. In fact, your instrument of “King Cotton” is consistent with the argument that the economic incentives of these places were key to their history – and that “absent slavery” as we know it, the likelihood of some other form of oppression for the sake of cheap labor for labor intensive mass crop production was incredibly high.

    Also, there’s this bit from the paper, your claim that you have “offered a theory in which slavery affects racial attitudes historically, and these attitudes are passed down through generations to shape contemporary attitudes.” Did I miss the evidence on migration patterns? I’m wondering where the confidence that this is inherited social psychology is coming from absent that sort of evidence? Is it not possible that the social structures of these places were created inherently differently — including sharper incentives for separate social institutions, including schools, churches, etc. — such that the historical sociology of these places – maybe instead of, or maybe in addition to individual- or familial-level psychology – is responsible for the enduring difference? What happens, in other words, when we implant new whites into these places? Or is it that these places uniquely isolated themselves from instreams of new populations?

    Again, very interesting work. It just seems, despite the careful work done so far, there are still big steps to take between the specific conclusions you draw and what your evidence says so far.

    • Matt Blackwell September 9, 2013 at 5:04 pm #


      Many thanks for your comments. These are issues that we have discussed with each other and we acknowledge they are important. On population sorting, Section 5.5 of our paper presents various pieces of evidence that rule out the idea that contemporary cross-county migration is driving the results. However, as we discuss in the paper, we accept that historical sorting may be an important channel by which the effects of slavery persist. On institutional mechanisms of transmission, we are beginning to do follow up research. Our claim was only that parent-children socialization is one plausible way in which the effects of slavery might persist, especially given over 40 years of research in political behavior (see, e.g., Campbell et al 1980, Jennings and Niemi 1968, 1981, Glass, Bengston and Dunham 1986, and Jennings Stoker and Bowers 2009, among many others). But you are right that we do not have direct evidence for this in the context of slavery.

      On the counterfactual question, in Section 4.3 we compared Southern counties with few slaves (< 3% of the population) to Northern counties that were similar, using coarsened exact matching. We found no differences in political attitudes. We conjecture from this that if the economic incentives of the postbellum South had not been in place, the high-slave areas of the South would be similar to the low-slave areas, which appear very similar to the North. This assumes that slavery’s effects are “local” and homogenous, and these may or may not be reasonable assumptions. It is, of course, a fictional counterfactual. However, you are right that we should be clearer that this is the gedanken-experiment we are conducting, and it requires some additional assumptions, which we think are reasonable.

      Glad to know that you found our paper interesting!

      Avi Acharya, Matt Blackwell, and Maya Sen

  9. Robert Savage September 9, 2013 at 6:24 pm #

    What is/was the impact of the republican’s “southern strategy” and the ongoing thinly-veiled racism employed by the republican party?

  10. Winnie Smith September 11, 2013 at 9:07 pm #

    I grew up on the Piedmont in North Carolina (raised by a chief of police), and most of the tales I heard concerned the carpetbaggers. The feeling was that “we” would have done “fine”, even after losing the war, if so much of what was left hadn’t been stolen in the following decade by the Yankee carpetbaggers; especially “our” homes and farmlands. Yankees are still considered basically untrustworthy and crass.

    The most unfortunate residual from slavery itself is an enduring attitude that blacks are incapable of caring for themselves and are inherently lazy due to their genetics. As everywhere in this country, the races were separated and so never able to learn differently: much like the rich today who reflect similar attitudes toward the poor.

    I think the anger toward Obama is partly from being proven mistaken in that belief and thus unjustified for racist assumptions. (I also think that’s why the rich are now privatizing schools: poor sons doing as well and sometimes better than rich sons is unacceptable.

  11. Harry Watson September 12, 2013 at 10:08 am #

    In regard to southern migration, the direction of movement from about 1800 to 1950 was only one way–out. The slave economy and the combination of share-cropping and low-wage manufacturing that followed it simply did not generate the high-paying jobs that would attract newcomers. As a result, almost all foreign immigrants avoided the South and most southern migrants left the region. As recently as 1950, the states with the highest proportion of residents who were born in that state were almost all in the South, and most white and black southerners were native-born Protestants descended from those who arrived in the 17th and 18th centuries.

    What was true for the South as a whole was almost certainly more true for the high-slave, plantation counties within the South. Of all southern southern locales, these were about the least likely to attract outside migrants (the others being the impoverished mountain regions). The economies of the black-belt plantation counties relied on exploited black labor (first enslaved, later “free”), kept under control by the ever-present threat of white violence. To be credible, this threat had to be maintained by consistent transmission of racism from one white generation to the next, both by parents and by broader community institutions like church, school, law enforcement, politics, and so on.

    Even today, when much greater migration has clouded the picture, newcomers are likely to take on the cultural “coloration” of the communities they move into, and those who move to conservative communities are probably more conservative to begin with than the average in their communities of origin. Migrating people look for other people similar to themselves, and communities are becoming more politically homogeneous all over the country. Given all that, it shouldn’t be surprising that “black belt” whites remain more deeply imbued with conservative values than other white southerners or Americans in general.