Suspiciously vague graph purporting to show “percentage of slaves or serfs in the world”

Phillip Middleton sent this along, it’s from Peter Diamandis, who is best known for his X Prize, the “global leader in the creation of incentivized prize competitions.” Diamandis wrote:

Phillip Middleton,

Is technology making you work harder? Or giving you more time off?

Seriously, it feels like it’s enabling me to work around the clock! Heck, I’m writing this email at 37,000 feet on a Virgin America flight from DC to LA at 11 p.m. ET.

So that being said, I want to share the actual DATA with you about Work vs. Leisure. . . .

It’s easy to forget that for centuries — for millenia — the “workforce” was ALL of us.

A few people lived in luxury, but the vast majority were slaves and serfs who did the work. In 1750, 75 percent of people on the planet worked to support the top 25 percent.

Let’s look at the numbers. It’s extraordinary how this has changed over time.


You’ll notice that by 2000, the global percentage of slaves and serfs in the world is down to 10 percent. As artificial intelligence and robotics come online, this number is going to drop down to zero.

Hey, if only artificial intelligence and robotics had existed in 1863, then Lincoln could’ve freed the—whaaaaa? What’s with that graph, anyway? Let’s look at the data, indeed. That curve looks suspiciously smooth!

Where did “the numbers” come from? The source says “Simon, pp. 171-177” but that’s not quite enough information. Luckily, we make rapid progress via Google. A search on “percentage of slaves or serfs in the world” takes us to this 2001 book by Stephen Moore and Julian Simon and the following quote:

A larger percentage of the world’s inhabitants are freer than ever before in history. Economic historian Stanley Engerman has noted that as recently as the late 18th century, “The bulk of mankind, over 95 percent, were miserable slaves or [sic] despotic tyrants.” . . . The figure shows the decline of slavery from 1750 through the end of the 20th century.

Screen Shot 2014-07-16 at 4.00.18 PM

This one’s kinda weird because they put 1917 exactly halfway between 1750 and 2000, which isn’t quite right. It’s almost like they just drew a curve freehand through some made-up numbers! Also a bit odd is that Moore and Simon’s curve is not consistent with their own citation: in their text, they say the proportion of slaves in the late 18th century was 95%, but in the graph it’s around 70%.

The next step, I suppose, is to track down “Simon, pp. 171-77; and authors’ calculations.” But I’m getting tired. Maybe someone else could follow this up for me?

In summary, the graph looks bogus to me. Some of these tech zillionaires seem to have no B.S. filter at all! Perhaps to be successful in that area it helps to be a bit credulous?

P.S. From comments below it seems clear that this graph has been created from a few nonexistent data points. It’s pretty horrible that Diamandis labeled this as “actual DATA.” I guess that’s just further confirmation that when people shout in ALL CAPS, they don’t know what they’re talking about!

28 thoughts on “Suspiciously vague graph purporting to show “percentage of slaves or serfs in the world”

  1. Stanley Engerman is a real expert on slavery but he’s mostly known for work that describes the actual economics of slave production in the US. His point, btw, has often been that slavery was more profitable than history likes to acknowledge. I don’t recognize the attributed quote. It doesn’t make sense to say 95% of the world’s population was anything other than alive at any given time.

    If you’re interested in the changes in how time has been spent, a very short but good read is Stanley Lebergott’s Pursuing Happiness: American Consumers in the Twentieth Century. He shows how we shifted more and more time to leisure from the drudgeries of trying to stay alive and keep clean.

  2. Unfortunately, “Simon, pp. 171-77; and authors’ calculations.” doesn’t show up on the link to his book on that page. I found it on pp. 255. It does also appear on as one of the “31 charts [which] will destroy your faith in humanity” by Brad Plumer. Also look at #7, Plumer’s take on “Life Expectancy in the 20th Century” which clearly shows that “More 77-year-olds are dying than ever before.” While you are at it, look at Plumer’s #20 to see that “Electrification rates have stagnated [at around 100%] since the 1960s.”

  3. Andrew is encouraging others to “follow this up for me” so let me suggest that readers go to to see Julian Simon’s (right-wing) philosophy as illustrated by his wager with (left wing) Paul Ehrlich. Ehrlich believed that the population explosion would produce huge increases in the cost of raw materials while Simon countered that the commodity prices would decrease in the 10 year period because “in a free society, solutions are eventually found. And in the long run the new developments leave us better off than if the problems had not arisen. That is, prices eventually become lower than before the increased scarcity occurred.”
    Wikipedia says “Ehrlich lost the bet, as all five commodities that were bet on declined in price from 1980 through 1990, the wager period…Paul Ehrlich mailed Julian Simon a check for $576.07 to settle the wager in Simon’s favor.” Wikipedia however, also says “Ehrlich would likely have won if the bet had been for a different ten-year period.” Moreover, a longer period and more natural resources, Simon would have lost “by a lot.”

  4. He didn’t write slaves; he wrote slaves and tyrants. Presumably, he’s counting indentured servants as slaves. So if 20% of the people in 1751 owned someone else then that’s 95%.

  5. I’ve tracked down the pp. 171-177 reference in the volume edited by Simon. The article is written by Stanley L. Engerman. I can’t find anything in the Engerman article that would explain the two graphs. The Moore and Simon book states that “Economic historian Stanley Engerman has noted that as recently as the late 18th century, “The bulk of mankind, over 95 percent, were miserable slaves or [sic] despotic tyrants.” Actually, it’s Arthur Young, in 1772, who stated this. Engerman just quotes Arthur Young. Such looseness with the facts is worrisome. The graphs are likely wishful thinking. The good news: This could be a fun example in an introductory Statistics class!

  6. “Simon, pp. 171-77”, I think, refers to Julian Simon (ed.) “The State of Humanity” (1996). The quote about “slaves or despotic tyrants” is a misquote, misattributed (in “It’s Getting Better All the Time”) to Engerman. Both Engerman and Simon (1996) point to the actual source, which is Arthur Young, “Political Essays” (first printed 1772). The actual quote is “What a melancholy reflection is it to think that more than nine-tenths of the species should be miserable slaves of despotic tyrants!” (p. 19).

    Simon seems also to have taken the 95% number from Young, as he provides an estimate of the number of people in the world living under “arbitrary governments” in 1772, which he comes up with by adding up the free peoples of Britain, Holland, Switzerland, Geneva, Sweden, British America, plus “say, a fifteenth of Germany” and “suppose” 10,000,000 “Free Indians” (giving a total of 33.5 million), and subtracting that from his estimate of the world population (775.3 million).

    As for source of the graph, there is a table on page 174 of Simon (1996), but it only provides the percentage of the population in slavery in certain parts of the Americas in 1750 and 1830. The best clue as to the possible source of the data behind the graph seems a book Simon references as the source of many of his estimates: Appendix C of Orlando Patterson, “Slavery and Social Death” (1982).

  7. This looks like an Excel chart. A “line” chart in Excel will distribute the data points uniformly along the x axis, then add a smoothed line.

    I tried the following data:

    1750 75%
    1850 55%
    1917 36%
    1944 50%
    2000 10%

    And wound up with this:

    At a minimum, it’s a very poor graphic, and as you all have already ascertained, the data are pretty suspect as well.

  8. For a site whose title celebrates social science, I think you forgot some key population and social behavior dynamics. Modern day slaves include MANY couples and individuals in committed or recreational relationships that celebrate Dominance/submission, BDSM, erotic slavery and practical slavery in general. I would wager the graph is actually upward sloping when you properly account for the larger population of people who are predisposed to be slaves and act on those tendencies. I also think the graph masks some import trends in emerging countries engaged in the earliest stages of nation building. These countries include many in the Middle East and Southeast Asia where growth rates are high thanks to economic slavery or serfdom in many forms. Bottom line: I dont think the graph is even remotely accurate globally or characteristic of rising trends in many countries.

  9. Pingback: Not the Worst. Graph. Ever. - The Washington Post

  10. Pingback: When is statistics manipulation evil? - The Washington Post

  11. Pingback: Taking Data Journalism Seriously - Statistical Modeling, Causal Inference, and Social Science

  12. Pingback: The uncanny valley of Malcolm Gladwell « Statistical Modeling, Causal Inference, and Social Science

Leave a Reply

Your email address will not be published. Required fields are marked *