Chris Blattman

Search
Close this search box.

The lingering effects of the slave trade

We investigate the historical origins of mistrust within Africa. Combining contemporary household survey data with historic data on slave shipments, we show that individuals whose ancestors were heavily raided during the slave trade today exhibit less trust in neighbors, relatives, and their local government.

We confirm that the relationship is causal by using the historic distance from the coast of a respondent’s ancestors as an instrument for the intensity of the slave trade, while controlling for the individual’s current distance from the coast. We undertake a number of falsification tests, all of which suggest that the necessary exclusion restriction is satisfied.

Exploiting variation among individuals who live in locations different from their ancestors, we show that most of the impact of the slave trade works through factors that are internal to the individual, such as cultural norms, beliefs, and values.

That is Nathan Nunn and Leonard Wantchekon in a new NBER working paper (ungated version here). It is the kind of paper that makes me wistful to be a macro-development researcher.

The basic result–that levels of trust trust today decrease with the intensity of the historic slave trade–is convincing. The suggested channel, cultural norms, seems less so. More plausible (to me): the slave trade undermined the development of legal and political institutions, and some of those deleterious effects (and the attendant insecurities) persist to this day.

The authors argue otherwise, but we need to be very careful with Afrobarometer data. A useful exercise for a budding development researcher: take an Afrobarometer survey (or a DHS survey, or a LSMS survey…) and try it out in the field. Fear and data humility will descend upon you in a cascade.

8 Responses

  1. Internal norms are political institutions; the personal is always political. No institution survives a population hostile to it, and populations can generate institutions out of whole cloth, if tradition is strong enough.

    Simply put — put twenty Americans in a room and cause a complicated problem that they need to solve. There will be a committee within an hour and a plan within three.

  2. Interesting conclusion.

    @Joseph Holden

    I heard a fascinating talk last week by Gareth Austin in which he noted that the number of Africans being enslaved AFTER the end of the trans-Atlantic slave trade was actually higher than the numbers during the height of the trans-Atlantic slave trade period. Basically, slaves became cheap on the domestic market, and the demand for slaves on the continent went up. Wonder if there’s any variation on this as to where those slaves originated and how it relates to modern-day levels of trust among ethnic groups.

  3. @ Matt C

    It’s worth having a look at the raw evidence for your question. See here:
    http://www.upload4free.com/download.php?file=1314179702-slaves.jpg

    It seems pretty clear that states where slaves were taken from are coastal – the biggest factors appearing to be population densities and geography – although at a glance, Mali seems to be an exception.

    Although I’ve read some papers positively relating slave numbers to the propensity to enslave already within the African culture. The Trans-Saharan slave trade is ancient, and that would explain Mali’s slave numbers.

  4. Fascinating, I’ve really enjoyed Nathan Nunn’s articles over the past few year.

    I wonder if he has exclusive ownership of the data set on slavery he developed or whether other researchers can use it. Any ideas?

  5. Omair,

    Good point, although I still find the concept that there’s some level of mistrust between peers that’s lasting from over a hundred years ago in societies that have had very high mortality rates that isn’t channeling through some other factor.

    How might have ethnicity played a role in this? Might more ethnically diverse places (where there is a higher likelihood of Africans enslaving Africans) have been targeted more by Arab and European traders. Ethnically diverse regions would have a higher levels of mistrust in general.

  6. @Matt C – When Europeans bought slaves, it usually wasn’t a case of people selling their peers, but rather selling captured people from other areas, who had either been captured during war or occasionally for the purpose of being sold to the Europeans.

    Of course it is worth noting that they didn’t always pay for slaves either; sometimes they just took people captive themselves, in which case I would guess trust between Africans played in even less. However, this could still have an effect on modern trust, though I expect Chris’s explanation to be more accurate in these areas than Nunn and Wantchekon’s.

  7. I’ve commenting without reading the paper: isn’t their a fairly big endogenous problem? Say you are a slave trade, and have a choice of different types of communities/cultures to enslave people from. Aren’t you more likely to choose areas where the people already have less trust in each other? Places that cultivate cultures of mutual trust and support are less likely to sell out their peers when the traders come knocking.

Why We Fight - Book Cover
Subscribe to Blog