Inequality is one of the best predictors of conflict ever found. Except when it isn’t.
Consider homicide in the United States. In 1990 and again in 2010, there was an impressive correlation between income inequality and the homicide rate. Among the states where inequality was high, so was homicide; and where inequality was low, homicide was too. But over those same 20 years, inequality grew while homicides fell — the opposite of what you would expect.
Call it the inequality paradox: The effect of inequality on conflict depends profoundly on the way it’s measured.
You can see the same contradiction in political violence. When measured over whole nations, inequality fails to predict the risk of civil war. Yet when measured between groups within those nations, inequality foretells both its outbreak and duration. Why is inequality so fickle?
The answer has to do with competition. Or, more precisely, who your competition is.
Historically, our competitors were our neighbours. They sought the same marriage partners, the same wealth, the same land. Local competition like this makes it hard for neighbours to co-operate, because one person’s gain is the other’s loss.
With the growth of trade and the free movement of people, however, we have shifted towards more global kinds of competition. Complete strangers, spread all over the world, now compete with each other for many of the same things that were once available only locally. As a result, neighbours have an incentive to work together, beating their competition abroad and sharing the wealth at home.
From a whisper to a shout
The upshot is that competition can amplify inequality. Global competition keeps the signal steady, tying small differences in wealth, resources or power to equally small impacts on conflict.
Conversely, local competition turns up the volume, transforming small differences in inequality into large effects on conflict. The irony is that it hurts more to lose a minor promotion to a colleague than it does to lose vastly greater sums of money to the faceless “one per cent.”
With the help of a colleague, I tested this idea in an experiment just published in the journal Psychological Science. In it, more than 1,200 participants competed in an economic game, and the winners were paid in real money.
Participants played the game in pairs and could choose to be either nice or nasty for points. Nice players share their points whereas nasty players take everything for themselves. When both partners are nasty, however, they get into a “fight” so costly that they wind up losing more points than they gain.
Thus, nasty players do well against nice players, but fare poorly against, and along with, other nasty players.
In the experiment, we varied the number of points at stake, creating inequality. We also imposed a 100-point penalty per fight, which the nasty players split. As shown in the payoff tables below, participants competed for zero points in the “No Inequality” condition, 10 points in the “Low Inequality” condition and 90 points in the “High Inequality” condition.
In the High Inequality case, for example, two nice players would earn 45 points each; a nice player and a nasty player would earn zero and 90 points, respectively; and two nasty players would lose five points each.
Crucially, we also varied the degree of competition between partners. We created local competition by paying participants for scoring more points than their partner, and we created global competition by paying participants for scoring in the top half of a larger group of 200 players.
Hence, a player facing local competition needed to beat her partner to win, whereas a player facing global competition could work with her partner to win against the other participants in the study.
Since competition amplifies inequality, we expected that small stakes can have a big impact on conflict — the number of participants choosing to play nasty. We found exactly this.
Our participants played nice when there was no inequality, and became nasty as inequality grew. But they were most likely to play nasty, get into fights and lose points when competition was local, even when there was only a small amount of inequality between them. Thus, under local competition, a little bit of inequality went a long way.
So inequality causes conflict, but the strength of this relationship depends on who’s competing.
Paradox no more
With this, the inequality paradox disappears. Homicides fell, rather than grew, over the last two decades because wealthy people moved away from their less affluent neighbours, reducing local inequality, and because globalization increased, diluting local competition.
Likewise, civil war reflects inequality between groups, rather than inequality between individuals, precisely because it is competition between groups that causes war. In the end, a measure of inequality is only as useful as the competitors it chooses to count.
And while years of open trade and migration have led to more global levels of competition, recent turns towards protectionism and closed borders may reverse this trend. This will make competition local again, raising the risk of conflict — and its cost.