Comments and Discussion

Michael Kremer

The papers by Nicholas Sambanis and Carol Graham not only are very interesting but also raise some very important questions, and it makes sense that the two were paired together in the session. Sambanis discusses the quite complicated relationship between economic status and political violence. Many high-income countries are stable democracies, but the relationship between income and stability seems much more complicated among lower-income countries. For example, although civil war seems to decline as income rises, that is not necessarily the case for terrorism. Obviously, the September 11 hijackers were not poor or uneducated.

Carol Graham looks at the relationship between income mobility and satisfaction, or happiness, within countries. She finds some expected relationships but also discovers a large group of frustrated achievers who, though having apparently risen in socioeconomic status, express unhappiness and sometimes discontent with democracy.

I will discuss Sambanis's work first. I have mixed views about the idea that state capacity is critical to avoiding violence. On the one hand, it has a lot of intuitive appeal. For example, the Kenyan state functions much more effectively than Zaire did. Soldiers follow orders in Kenya and put down rebellion when instructed to do so. This presumably has something to do with Kenya's relative stability. However, that type of explanation risks being circular: How does one know that the state capacity is high? Because the state maintains law and order.

The idea of looking at regional inequality or inequality between other salient groups is a very good one. What matters for violence may not be the variance of income in society but rather the inequality between different ethnic groups or religious groups or geographic areas.

The finding that terrorists are highly educated is not so surprising. That may be partly the production function for terrorism. It is amazing how many terrorist plots fail, for example, in Israel. Perhaps terrorists need to be educated to be effective. Another good point made by the author is that low-level violence often precedes war and may have economic consequences, making it very difficult to disentangle the relationship between violence and economics.

There are a couple of points I would like to add. A lot of the discussion approaches the question of violence by asking when the disadvantaged will rebel. However, it would be informative to look more carefully at the behavior of the powerful. For example, just as the Romans did not find it worthwhile to pacify northern Britain, the Ugandan and Kenyan governments may not have big incentives to pacify the remotest regions of the country where people are nomadic. It is expensive and the rewards may be minimal. To take another example, the U.S. government does not have much incentive to pacify Liberia or Haiti.

In terms of an ultimate policy conclusion, Sambanis is skeptical of the notion that putting a lot of money into poor countries will help reduce the violence there. I agree with the author that pouring in money to redress grievances is not a panacea. However, the alternative strategy of military intervention to control violence does not seem to be all that effective either, as currently demonstrated in Iraq. Given that the impact of economic aid is limited and long term, it may still be a more cost-effective way of buying security than invading a country or putting air marshals on planes.

It may be naïve to think that if we just pour money into violence-plagued regions, such as the Middle East, then the people there will no longer have grievances. However, money can be used to get leaders to act in ways that serve U.S. security interests. Aid can be provided to leaders who make the kinds of changes that reduce the risk of violence, and it can be provided to bolster democratic leaders against violent opponents.

I liked the paper by Graham. Of course, it is obviously difficult to rule out the possibility that frustrated achievers were...