An Economic Portrait of Terrorism

August, 2002

ELEVEN MONTHS AFTER the disastrous attacks on the World Trade Center, the American government and its allies are grappling with how to wage an international "war on terrorism." It would help to know more about how terrorists are created, why they operate the way they do, and how they affect the globe's economic health and political stability. We asked a Stanford political scientist and an economist, each of whom have studied some of the causes and effects of postöWorld War II political violence, to apply their insights to the current situation with terrorism. Their remarks to a Business School audience in March are excerpted here.

PHOTO ILLUSTRATION BY EDWARD GAJDEL

Political science Professor James Fearon finds evidence to dispute some of the commonly suspected triggers of political violence, and he suggests that government responses to terrorist actions are critical to terrorists'success.

The Business School's Romain Wacziarg, assistant professor of economics, highlights how political violence affects a country's economy. While military spending can expand an economy in the short run, he says, "it's hard to find a measure of violence that actually has a positive effect on economic growth." GSB Professor David Brady, a political scientist whose expertise is American politics, moderated the discussion.

JAMES FEARON
Almost all deliberate attacks by organized groups on civilians occur in the context of civil wars. (That isn't a great definition of terrorism, but it's a decent start.) In my research with another Stanford political scientist, David Laitin, we define civil wars by several criteria, but basically as a conflict internal to a state that kills at least 1,000 people. Over the whole postöWorld War II period you see a steady, gradual increase in such conflicts.

There are about 25 such conflicts going on now; about 122 or so since 1945 have reached the 1,000-dead threshold. They have killed÷this is a very rough but probably conservative estimate÷about 16.5 million people, compared to about 3 million for interstate wars in the same period. We don't have exact numbers, but certainly a very large majority of the dead are civilians who were killed in the context of what looked on the local level like terrorist attacks÷by rebel forces or by government forces in reply.

Why so many civilians dead? Overwhelmingly, these wars are fought as guerrilla wars. If the environmental conditions are right, you need very few people to get a significant guerilla war going, in some cases hundreds but typically 1,000 or 2,000 people. Such small numbers of rebels can survive by hiding among the general local population or in rough terrain. A standard rebel tactic, especially at the start of these wars, is to intentionally provoke violent retaliation by government forces, hoping that they will burn, loot, pillage, and kill in an indiscriminate fashion. This helps small rebel groups recruit more people to the movement.

There's a parallel with Al Qaeda. We should think about the 9/11 attacks more as a recruiting effort directed at a Muslim audience than as just an expression of hatred of the United States. As in the early phase of a guerrilla war, the idea was to provoke a big response that the terrorists hope will be highly indiscriminate. The idea was to give credence to the notion of a war against Muslims in which the alleged enemies of Islam÷the U.S. and the western powers÷are aligned with the Muslim governments in the region that Al Qaeda doesn't like.

Why do some states have civil wars and others don't? One very common theory argues that countries at risk for civil war and terrorism are states that have lots of ethnic groups or religions. We don't find much evidence for this hypothesis. We have coded the average number of civil wars in about 160 countries over this whole postwar period by degree of ethnic diversity and by level of per capita income. You don't see any consistent relationship with increasing ethnic diversity when you compare countries that are at roughly the same level of economic development. On the other hand, if you look at the effect of income while controlling for ethnic diversity, you see at every level of diversity a very pronounced effect of higher per capita income. Namely, richer countries have fewer civil wars.

The second story one often hears about the causes of civil war seems overwhelmingly plausible at first glance: The claim is that we find civil wars and terrorism where some segment of society is being oppressed or suffering discrimination. If that were true, one would expect that the greater the degree of democracy in a country, or the greater the degree of observance of civil liberties, the less likelihood of civil war occurring. Interestingly, this doesn't appear to be the case either. Among countries at similar levels of economic development, there is no tendency for the more democratic or civil-libertiesöobserving ones to have fewer civil wars. Once again, the dominant factor is per capita income, which is associated with less civil violence at every level of democracy.

There are several other factors that we have found to matter, even controlling for income differences. More mountainous countries with rough terrain where guerrillas can hide are more likely to have civil wars. Political instability matters a lot, as it creates weakness at the center of the state that provides an opportunity for rural guerrillas. Newly independent states are more likely to have civil wars for similar reasons. External, third-party financing for guerrilla groups and secure base areas over a border also appear to help. But the single most powerful correlate, and the one most robust across regions of the world, is per capita income.

One reason is that in a very poor country, poverty makes it easier to recruit fighters for a rebel band because the alternatives are worse. We think, however, that there may be a more important reason. We suspect that per capita income is really standing in for state capabilities and coherence: the degree of administrative, military, and policing competence of states.

In weak states÷Indonesia or Sierra Leone, for example÷that don't have the capability to pursue intelligent and effective counterinsurgency, the state takes the bait of the rebels. The military leader says to some military people, "Go out and deal with it." The military then "pays itself" by looting, pillaging, raping, or even by starting businesses as they did in East Timor. Of course, these activities tend to exacerbate and keep the civil war going. You see this pattern in a broad range of cases from Colombia to Indonesia. Bad counterinsurgency tends to drive people off the land, giving them fewer alternatives except becoming rebels, and it also angers them, understandably, and drives people into rebel groups that way.

Sept. 11 is not exactly analogous to this scenario of how civil wars get going. For one thing, Al Qaeda has international, interstate dimensions that have no exact parallel in the typical civil war scenario. Let me develop parallels and contrasts by asking: What do economic factors have to do with Sept. 11?

On first glance, it would seem not very much. Some of the attackers were educated and relatively well off for where they were coming from. The things that seem more relevant than money are a sense of rage, religious fanaticism, hatred of the United States, hatred of Israel. I certainly wouldn't discount those factors.

On second glance, there are indirect and important ways that economics did and do matter here. The key word is "oil." Twenty- eight percent of the world's supply comes from the Gulf region, almost half of that from Saudi Arabia. This is incredibly important to the world economy and an easy source of unbelievable profits.

The Gulf monarchies are not really like other states. It's better to think of them less as states than as private firms that are very narrowly held by ruling elites÷royal families in most cases. They are thus particularly attractive as subjects of hostile takeover bids in the form of military attacks from their neighbors. That's one way to think about what happened in 1991 with Iraq and Kuwait.

How do they keep themselves in business? Way back when, they cut an incredible deal with the United States. We protect massive profits for these very narrow, dictatorial elites in return for stable, smooth access to oil.

What does this have to do with 9/11? Here is an anecdote that helps make the connection: In 1990, with the invasion of Kuwait by Saddam Hussein, Osama bin Laden went with a few others to see some of the top people in the Saudi royal family. Reportedly, he said something like, "Look, we're just back from Afghanistan, where we defeated this great power, the Soviets. We raised an army of Muslims from all around the Middle East who fought with the Mujahideen in Afghanistan in the eighties, and could use this army to kick Saddam out of Kuwait." The Saudi monarchs/royal family said, "Forget it," and they invited in the United States. Bin Laden was said to be devastated.

Why was the Saudi monarch not so keen on raising a more indigenous force? One major factor is that he knows (or could intuit) a powerful historical regularity, which is that if you demand that local people fight a war on your behalf, you will end up having to give them more political rights, and that's the last thing he wanted to do.

By this kind of account, economic factors are at the root of the 9/11 attacks. The attacks are part of a strategy to produce a strong Western reaction that will look like a clash of civilizations that, in turn, will help bin Laden and people like him mobilize support for taking power from these narrow, authoritarian regimes in the Middle East.

ROMAIN WACZIARG
I will talk about the economic consequences of terrorism. We have measures of violence or instability from which we can perhaps infer the effects on a country's long-term economic growth. I'll focus on a period of 30 years, 1960ö1990, and control for other growth determinants, because it may be that countries that have a lot of violence also have other factors, such as a low savings rate or low human capital, that make them grow slowly.

One measure is war casualties. The relationship between war casualties and growth is negative with about 99 percent statistical confidence. I'm only correlating unexplained growth with the unexplained portion of war casualties. While wars, per se, seem to have expansionary effects on growth because of increased military spending, war casualties÷and especially severe war casualties÷have a negative effect.

There is also a big literature in economics and political science about the effects on growth of revolutions and coups. Even if you control for a variety of things, there's a negative correlation between how many revolutions and coups you have had and your growth rate. However, if you look at the number of government changes in a country, you don't find a negative relationship. That tells you that the dataset on revolutions and coups has to do with the violent or unpredictable nature of those changes.

This political instability and violent change in government affects growth mainly through investment rates. That's a sign, perhaps, that increased uncertainty brought about by unpredictable government change is what explains this negative relationship with growth.

I also looked at the relationship between number of guerrilla movements within a country and economic growth. If you control for national income, you don't see much of a relationship. My view on this is that there's something about whether or not the movements induce severe violence. If you think of France, our guerrilla movements are pretty tame. (I'm saying "our" because I'm French.)

Political assassinations have a negative correlation with growth but with only 75 percent statistical confidence. Perhaps this measure captures the degree of arbitrariness or uncertainty that entrepreneurs face in those economies.

Another surprise to me was that riots don't predict growth very well. In fact, statistically it's indistinguishable from zero.

There are bottom lines here: Because war involves expansionary expenditures on defense, wars are actually good for economic growth. But if you have sustained war casualties, that's bound to impact your growth negatively. There's negative effect of violent changes in government, but not of the number of peaceful changes, an indication that instability of this form has an effect on growth if it's unpredictable.

There's mixed evidence on the internal strife measures. Some display negative effect, some a zero effect. Very severe internal strife probably depresses growth, but most of what we observe isn't severe enough to be a primary determinant.

The last bottom line is that it's hard to find a measure of violence that actually has a positive effect on growth. If anything, it has a negative effect or zero.

An Economic Portrait of Terrorism

Organizational Ecology

Wanted: A Company to Call Their Own

Living in Limbo with “Don’t Ask, Don‘t Tell”

It's the Cash Flow, Stupid!

QUESTIONS FROM THE AUDIENCE

What would your policy recommendation be for postö9/11?

ROMAIN WACZIARG
We have some good lessons on what improves the economic growth of countries. Trade openness is one, having good institutions, good governance, policies that don't distort the economy, exchange rate regimes that aren't completely out of whack with reality. Conditioning the aid that rich countries give to poor countries on getting your act together on those fronts, more than is done now, would be, in my view, the long-term response.

DAVID BRADY
That sounds good, but across the world that's precisely what countries object to. When you read Robert Bates' work on Africa, it's perfectly clear that throwing money there is not going to be very useful unless they set up markets, but we don't know how to get people to accept that. It was much easier after the war in Europe and Japan because there was a tradition of a capitalist economy that could be built back up. Building a nation and the concomitant economic development that characterize modern nation-states is not so easily done.

JAMES FEARON
Some people have talked about a Marshall Plan for terrorism based on the notion that poverty is causing this. I don't think that's true. This isn't quite correct but it's probably not far from it: People who are interested in committing terrorist acts may be randomly distributed with respect to income. Rich countries are able to keep the threat at the level of the Red Brigades in Italy, but in poor countries you get this escalation process that produces the violence. The big question in Afghanistan is, How do you reconstruct order in a country that's been divided and destroyed by horrible civil wars for more than 20 years? Some kind of international intervention is necessary to make it possible to rebuild state institutions. It's a costly job that few countries are willing to pay for, and even if they were, it's not clear that we have great ideas about how to do it.