Reflections on World on Fire
Negative reviews of Amy Chua’s Battle Hymn of the Tiger Mother often begin by praising her earlier book, World on Fire: How Exporting Free Market Democracy Breeds Ethnic Hatred and Global Instability. I was inspired to read it – and found the book so hypnotic I finished it in a day.
Chua’s thesis is that there is a major tension between capitalism and democracy – especially in the multi-ethnic developing world. Most less-developed countries have relatively prosperous commerce-oriented minorities; Chua calls them “market-dominant minorities.” The Jews in Russia, overseas Chinese in Southeast Asia, Indians in East Africa, Lebanese in West Africa, and Latin Americans of European descent all qualify. The impoverished majorities of these countries have long resented their market-dominant minorities.
Given these initial conditions, both free-market reforms and democratization have negative side effects.
The negative side effect of free-market reforms: The market-dominant minorities disproportionately benefit, increasing popular resentment. The negative side effect of democratization: Market-dominant minorities disproportionately suffer, because the majority finally gets a chance to legally enforce its resentment. Pushing both reforms on developing countries simultaneously – which Chua claims the U.S. government habitually does – gives the worst of both worlds: Increasing resentment – and the opportunity to politically act upon it. If the stars align badly enough, preaching democratic capitalism gives you Yugoslavia or Rwanda.
Before I critique Chua’s thesis, let me begin with some praise. World on Fire is a great example of what I call “behavioral political economy.” Her model of politics begins with psychologically plausible assumptions like “Unsuccessful majorities resent economically successful minorities,” and “Politicians can win power by satisfying this resentment in ways that make the majority even poorer.” She takes anti-market and especially anti-foreign bias very seriously. Scientifically speaking, this is light-years ahead of hundreds of obtuse rational voter models. Even if Chua were completely wrong, she’d still be pointing social scientists in the right direction.
Still, despite many thought-provoking insights, I see two glaring problems with World on Fire.
1. The book actually inspires me to stand up and defend democracy. Sure, a civilized traditional dictator is better than a genocidal median voter. But empirically, dictatorships have murdered vastly more people than democracies. Communist China, the USSR, and Nazi Germany top the list for the 20th century. The standard runner-ups include Imperial Japan, Cambodia under the Khmer Rouge, Turkey (the Armenian genocide), Pakistan’s attempt to hold onto Bangladesh, and Communist regimes in Vietnam, Poland, and Yugoslavia. While you might argue that a few of these regimes had some democratic elements, none approached free elections based on universal suffrage.
On reflection, dictatorships’ bloodthirstiness is no coincidence. Dictatorships have high variance; if the men at the top happen to want blood, they get it. Democracy, in contrast, systematically produces moderation. That’s the essence of the median voter theorem. Even if almost everyone resents a market-dominant minority to some degree, politicians don’t win by advocating genocide. Politicians win by advocating a policy that half the electorate thinks is too harsh and half thinks is too soft.
In theory, of course, resentment could be so intense that even “moderates” favor mass murder. Chua apparently believes that this is frequently so:
In a frightening number of cases, democratization in the face of a market-dominant minority has led to government-encouraged attempts to “cleanse” the country of the minority altogether. Strategies for doing so include forced emigration, expulsions, and in the worst cases pogroms, extermination, and genocide… Almost always, such policies are passionately supported by an aroused and angry “indigenous” majority, motivated by tremendous feelings of grievance and inferiority.
But how does she know what the majority actually favors? As far as I’ve heard, no country on the verge of genocide has ever run a scientific survey to measure popular support for it. And even if there were such a survey, people in a country poised on mass murder are extremely likely to engage in preference falsification – to feign support for horrors they inwardly oppose. Chua blithely assumes the worst about median voters around the globe based on little more than anecdotes and introspection.
Even worse, there is strong indirect evidence that her anecdotes and introspection are unreliable.
a. Notice how she refers to “government-encouraged” cleansing, not government-enforced cleansing. Why the distinction? Because she doesn’t offer any examples where clearly democratic governments actually engage in mass murder. The worst democracies do, rather, is impotently twiddle their thumbs in the face of ethnic riots. But if the median voter affirmatively wanted genocide, democratically elected governments wouldn’t merely “encourage” private violence; they would establish, fund, and staff killing fields and death camps. Democracies virtually never do this.
b. The mere fact that democratic governments fail to suppress riots hardly shows that the median voter wants riots. A more plausible story is that the median voter opposes riots, but – like the typical American during the Rodney King riots – is too squeamish to back the harsh police measures necessary to crush them.
c. Governments engaged in mass murder almost always conceal it from their own populations. Even the Holocaust wasn’t publicized. The men who ran it operated under orders of strict secrecy. The most plausible explanation is that even the Nazis realized that they were vastly more murderous than the median German.
d. If Chua were right about the savagery of the median voter, the fact that democracies almost never fight each other would be completely mysterious. After all, if the median voter were eager to exterminate the scapegoat within, you’d expect him to occasionally vote to exterminate the scapegoat without as well.
2. World on Fire fails to appreciate how wonderfully free-market reforms seem to work. Brookings’ Chandy and Gertz lay out the basic facts for 2005-2010. GDP in the Third World is exploding:
The economies of the developing world have expanded 50 percent in real
terms, despite the Great Recession. Moreover, growth has been
particularly high in countries with large numbers of poor people. India
and China, of course, but also Bangladesh, Tanzania, Ethiopia, Vietnam,
Uganda, Mozambique and Uzbekistan – nine countries that were
collectively home to nearly two-thirds of the world’s poor in 2005 –
are all experiencing phenomenal economic advances.
Nor is there any reason to believe Chua’s repeated claim that market-dominant minorities scoop up most of the gains. Global poverty is plummeting, too.* Chandy and Gertz:
We estimate that between 2005 and 2010, the total number of poor people around the world fell by nearly half a billion people, from over 1.3 billion in 2005 to under 900 million in 2010. Looking ahead to 2015, extreme poverty could fall to under 600 million people–less than half the number regularly cited in describing the number of poor people in the world today. Poverty reduction of this magnitude is unparalleled in history: never before have so many people been lifted out of poverty over such a brief period of time.
Now you could admit the reality of massive progress, but deny free-market reforms the credit. But this course is not open to Chua. Throughout World on Fire, she freely gives market reforms the credit for enriching market-dominant minorities – and the blame for failing to spread these riches more widely. Now that the developing world has prospered for eight additional years under these very policies, it seems time for her to change her tune. There’s still plenty of room for further progress, but the global move toward democracy and capitalism is working out amazingly well.
* For 2005-2010, Chandy and Gertz admittedly assume constant income distributions. But their results are consistent with poverty research based on micro-level data.