17

I recently read a French post on Facebook saying that only 3% of white people owned slaves of African descent in the United States.

Here is a part of a debate in French quoting a twitter post about the initial claim.

Debate

Some more examples:

Jim Goad, Hey Kids... It's Time For Some Dumb Myths And Smart Facts About Slavery!, Vice, November 30, 2005 *

At the peak of black slavery in the South, only 6 percent of Southern whites owned slaves. If you include the white people in the North, it means that only 1.4 percent of white Americans owned black slaves at the HEIGHT of slavery.

Subreddit "AskHistorians" *

One-quarter of all free families in the South (note, this category includes black free families) owned slaves, according to the 1860 U.S. Census. [...] In the Lower South (seceded before Fort Sumter), 36.7 percent of white families owned slaves. In the Upper South (seceded after Fort Sumter), the proportion was 25.3 percent. In the Confederate states as a whole, it was 30.8 percent. In the border states (which did not secede), the percentage of slave ownership was 15.9 percent. *

Another website with another numbers

(unless otherwise noted, all data is as of the 1860 census) Total number of slaves in the Lower South : 2,312,352 (47% of total population). Total number of slaves in the Upper South: 1,208758 (29% of total population). Total number of slaves in the Border States: 432,586 (13% of total population).

* Links provided by tim

What percentage of white people owned black slaves in the US?

Community
  • 1
MoebiusCorzer
  • 273
  • 2
  • 9

2 Answers2

20

There is surprisingly little information about this available. The information that does exist is almost entirely based on different interpretations of the 1860 census (except of course for the claims that do not name any sources).

This seems to be the more relevant claim from Goad:

At the peak of black slavery in the South, only 6 percent of Southern whites owned slaves. If you include the white people in the North, it means that only 1.4 percent of white Americans owned black slaves at the HEIGHT of slavery.

Goad doesn't name any sources, and I was unable to find any for this claim. A more common claim is 1.6%, see for example this meme, which claims to be based on the 1860 census. The only way to get that number from the census data is to calculate the relation between slave-owning households and total population (including slaves), which doesn't make a lot of sense.

You can view an interpretation of the 1860 census data at civil-war.net which notes that 8% of US families owned slaves (note that this already includes the north as well as the south). You can retrieve the original census data yourself, but it's not very readable. You can also access the census data digitally at the University of Virginia.

There are a couple of problems with the 8% number, for example that it includes the northern states which didn't actually own any slaves at that point, or that it includes black slave-owners.

In general, there is also the problem of who you consider a slave-owner. Do you calculate it on a per-family basis? Do you consider the wife of a slave-owner a slave-owner as well? If you calculate the percentage of slave-owners of the entire population, do you include non-free people as possible slave-owners? And of course who do you consider to be white? You could make your own calculations based on the census data, but it seems like original research to me, so I'll stick to what the sources claim.

The most notable interpretation that I could find, and likely the closest to an answer to your question, is from Armstead L. Robinson - a history professor at the University of Virginia - in Bitter Fruits of Bondage: The Demise of Slavery and the Collapse of the Confederacy, 1861-1865:

[...] while slaves made up 40 percent of the total population of the South, only 25 percent of free families, most of them white, owned any slaves at all

A different interpretation by slate.com states:

According to the 1860 census, taken just before the Civil War, more than 32 percent of white families in the soon-to-be Confederate states owned slaves

tim
  • 51,356
  • 19
  • 207
  • 177
  • Very good answer! – Sklivvz Oct 03 '16 at 16:09
  • 1
    Although it is not complete (in the sense that it didn't entirely addressed my question), it is a very interesting answer that make me see these numbers with more clarity. Thank you :) (and thanks to @Sklivvz for his patience and edits in order to make my question more appropriately fit for this site). – MoebiusCorzer Oct 03 '16 at 16:30
  • 1
    Bad conclusion that "if only 6% of people in the south owned slaves only 1.4% total owned slaves". It assumes nobody in the north owned slaves, which is blatantly wrong. – jwenting Oct 04 '16 at 06:47
  • 3
    @jwenting the 1860 census counted 393,975 slave holders and 26,690,781 white people. So 1.5% applies to the whole country. However, in the case of joint ownership only one owner was listed, and some of the listed holders were trustees of a trust or corporation that owned slaves, rather than owners personally, and others are people who didn't own the slaves but hired/rented the slaves. – DavePhD Oct 04 '16 at 08:23
  • 1
    @DavePhD I don't contest the 1.4% but the logic in the statement that derives it. – jwenting Oct 04 '16 at 08:33
14

In the 1856 abolitionist speech before the House of Representatives Modern "democracy", the Ally of Slavery by Mason Weare Tappan it is stated that there are 347,525 slave holders in the United States, out of 6,222,318 white people in the 15 slave-holding states plus DC.

Specific numbers are given for each state and DC.

It is stated following the data table:

From this it will be seen that there is not a slaveholding state in the Union in which the slaveholders constitute one-tenth part of the aggregate free white population, and in some of them not one thirtieth part.

It is also explained that the census figure "347,525" includes people who hire slaves, without actually owning the slaves.

The total white population (including non-slave states and territories) was 19,553,068 in 1850.

The 1850 census was the first census that attempted to specifically list every slave, and for each slave listed a slave holder was to be listed. Data from the 1850 census are tabulated in the 1854 US government publication Statistical view of the United States, compendium of the seventh census

The listed number of slave holders (347,525) corresponds to 1.8% of the white population.

However, the instructions of the 1850 census were:

Where there are several owners to a slave, the name of one only need be entered, or when owned by a corporation or trust estate, the name of the trustee or corporation

and

The person in whose family, or on whose plantation, the slave is found to be employed is to be considered the owner—the principal object being to get the number of slaves, and not that of masters or owners.

So because only one holder is listed for a given slave, and because slaves usually served a family a whole rather than just the technical owner(s), some people try to attribute a slave to a family as a whole. The problem with this strategy is the census definition of "family" used at the time:

all the inmates of a boarding house or a hotel will constitute but a single family, though there may be among them many husbands with wives and children. Under whatever circumstances, and in whatever numbers, people live together under one roof, and are provided for at a common table, there is a family in the meaning of the law.

This extended to whole military garrisons, jails, asylums etc., each being considered a single "family".

For this reason, use of the census data number of families overestimates the percent of families that owned slaves.

Another way to analyze the situation is slaveholders as a percentage of adult whites (over age 15) as listed in the 1850 census, which (using age statistics on page 54 of the 1854 publication) was 347,525 out of 11,550,353 or 3.0%.

DavePhD
  • 103,432
  • 24
  • 436
  • 464
  • It seems that 1854 is chosen rather randomly. The reason that 1860 is often used is that it is shortly before the civil war. And while there are obviously various ways in which these number can be interpreted, note that the total population includes women and children, both of whom had difficulties owning their own slaves. It still seems fair to say that they may have "used" slaves though (the slaves of their father/husband; which is why it may make more sense to calculate the number of slave-owning families). – tim Oct 03 '16 at 17:13
  • @tim your source says 393,975 slave holders in 1860. http://www.civil-war.net/pages/1860_census.html Don't assume that 1860 is the maximum slaveholding percentage. Importation of slaves was banned since 1808, while whites were still immigrating all-the-while. 1850: 347,525/19,438,451 (1.8%); 1860: 393,975/26,690,781 (1.5%) – DavePhD Oct 03 '16 at 17:29
  • Fair enough, that's a good point about the percentage (it would be interesting to see it for other years as well). I still think that the calculation doesn't make much sense though, as the resulting number doesn't represent the circumstances very well. – tim Oct 03 '16 at 17:37
  • 1
    @tim but my reference shows that abolitionists considered that the correct way to calculate the percentage at the time. It is abolitionist Lewis Tappan who is making the calculations "one-tenth" and "one-thirtieth" in the reference. – DavePhD Oct 03 '16 at 17:54
  • Sure, but they were trying to make a political argument with these numbers. Specifically, their argument was that slave-owners had a disproportionate amount of power. So they would be interested in a calculation that keeps the number low, even if that number doesn't represent who benefited from - or "used" - slaves. An easy way to see the problem with your calculation is to consider a year with a very high birth rate. The calculated percentage of slave-owners would be lower in that year (because of the higher population), even if the number of slave-owners and slaves did not change. – tim Oct 03 '16 at 18:05
  • 2
    @tim So basically you're saying that if a white baby is born she is instantly a slave owner, if one of her family members is a slave owner, and therefore my calculation is wrong because it fails to includes slave owners who have just been born. – DavePhD Oct 03 '16 at 18:31
  • 2
    No, my point is that it makes more sense to calculate slave-owning families. That way, if a white baby is born, it doesn't affect the percentage of slave-owners one way or the other (instead of reducing the percentage as it is the case with your calculation, which makes no sense at all). Alternatively, you could use the adult male population for calculation. – tim Oct 03 '16 at 18:43
  • 1
    @tim I agree with using families generally, but the question says "people". More importantly, the 1860 census definition of "family" is not an actual family. A college counted as one "family". A prison one "family". Servants were not counted as separate families. Basically there are a lot of non-slave-owning people that should be considered as separate families that are included with a slave-owning family or considered one giant non-slave owning family (hotels, colleges, asylums, prisons). see reference below: – DavePhD Oct 03 '16 at 19:06
  • https://books.google.com/books?id=fv7Ulh3zZokC&pg=PA41&dq=1860+census+definition+of+family&hl=en&sa=X&ved=0ahUKEwiz2IObp7_PAhXKKiYKHSpXAC0Q6AEIKjAA#v=onepage&q=1860%20census%20definition%20of%20family&f=false – DavePhD Oct 03 '16 at 19:06
  • Women and children were chattel in those days; they were the property of the husband/father. When a women married her dowry (including any slaves) passed to her husband. So married women and little children could not own slaves as such - they didn't own anything. – RedSonja Oct 04 '16 at 07:09
  • 4
    @RedSonja Women definitely owned slaves. See this census listing of slave owners: Elizabeth, Mary, Jane, Susan; many women owners each owning multiple slaves http://boonecountyky.us/slave-schedules-1850--a-b.htm#A1850 – DavePhD Oct 04 '16 at 07:44
  • @DavePhD I did say married women. Single (adult) women or widows could, of course. The point is, you can't count all family members as slave owners because not just the slaves but also the wives and children were the property of the husband/father. But this would be different in every state and every decade, I imagine. – RedSonja Oct 04 '16 at 08:47
  • 7
    @RedSonja Martha Washington separately owned slaves. Martha wasn't George's property, but instead was capable of owning slaves. George willed that his be freed, and they were, but Martha never freed hers. http://www.mountvernon.org/george-washington/martha-washington/martha-washington-slavery/ But I agree with you main point that you can't lump all family members together. – DavePhD Oct 04 '16 at 09:11
  • 4
    @RedSonja Married women started getting property rights by the early to mid 19th Century in the US. By the 1860s, married women could have legally owned property apart from their husbands in many states. See [this](https://en.wikipedia.org/wiki/Married_Women%27s_Property_Acts_in_the_United_States) – KAI Oct 04 '16 at 17:42