Bases, Bullets, and Ballots: U.S. Military Aid and Conflict in Colombia

January 5, 2010

My guest this week is Oeindrila Dube, a postdoctoral fellow here at the Center for Global Development and an assistant professor of politics and economics at New York University. She is the author, along with Suresh Naidu, of a new paper that examines the relationships between U.S. military aid to Colombia and paramilitary violence and electoral participation in that country. Her paper reaches the unsettling conclusion that U.S. military assistance dollars may in fact be responsible for raising the levels of political violence.

At the heart of Oeindrila’s paper is an innovative approach that uses detailed data on paramilitary attacks and assassinations (available from 1988 to 2005) to establish quantitative evidence for a phenomenon that has long been suspected. “For decades,” says Oeindrila, “many NGOs have anecdotally been describing links between paramilitary groups and the government military that has this implication that … resources going into the country … might be diverted to these groups. The nice thing is we are able to show that quantitatively.”

Oeindrila’s data show that municipalities in Colombia that house military bases show statistically significant increases in paramilitary violence following stepped-up U.S. military assistance. She explains that the extremely localized effects provide strong evidence that government forces are not only providing arms and ammunition to paramilitaries but have also conducted joint operations with them.

Her quantitative approach also allowed her to answer a range of other important questions. She finds that while an increase in aid correlates with increased paramilitary activity, it has no impact on guerrilla attacks or on counternarcotics operations, calling into question whether military aid to Colombia is an effective means to either securing the country or decreasing narcotics trafficking.

Oeindrila’s findings carry important lessons for policymakers studying any country where relationships between the government and paramilitary groups factor into the equation. In the last several minutes of the podcast, we examine several of these cases, including Mexico, Iraq, and Afghanistan. “We have to consider the links between armed non-state actors and the state anytime we start disbursing money,” Oeindrila tells me. “Otherwise, our military aid is going to end up financing groups that we are … trying to counter.”

Please do listen to the interview and read Oeindrila’s paper here. Have something to add to our discussion? Ideas for future interviews? Post a comment below. If you use iTunes, you can subscribe to get new episodes delivered straight to your computer every week.

Podcast: Play in new window | Download

You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.

1. Dr. Paul Rigterink Says:
   January 8th, 2010 at 5:35 pm

   I lived in Colombia from 1964-1966 and recently started a project there as described at my website at http://home.comcast.net/~prigter/site/. I have no doubt that there is collusion between the military and illegal armed groups. However, I don’t believe that the primary reason that there are problems in Colombia is due to US foreign assistance directly enabling illegal groups to perpetuate political violence and undermine democratic institutions, such as electoral participation. You are underestimating the effects of the guerillas that are sent from Cuba and Venezuela as well as the effects of massive amounts of money coming from the illegal drug trade. In addition you are underestimating the effects caused by Colombia’s history of violence. Violence in Colombia has been an on-going problem for more that 60 years (your study focused on only the period between 1989 and 2005 and ignored the Violencia period in Colombia, for example) I know some men who were told to become members of the FARC or die. I have known many more people who died of disease and local crime. I also have had to avoid guerillas sent from Cuba who were assigned to cause problems in the region where I work. I believe your study was too narrow in focus and did not come to the right conclusion, unfortunetly.

2. Oeindrila Dube Says:
   January 9th, 2010 at 10:35 pm

   Dr. Rigterink,

   Thank you for your comments. We were also intrigued by the proposition that resources from Venezuela affect FARC activity, for two reasons. First, we wanted to see if this effect was detectible. We decided to use the value of oil as a proxy for potential Venezuelan resources, since oil is a major revenue source for the Venezuelan government. We tested whether changes in the price of oil affected FARC violence more in Colombian regions closer to the Venezuelan border. We found that it did not. One caveat here is that the value of oil may not be a sufficiently precise measure of how resources get transferred, but our non-finding does provide some suggestive evidence. Second, we also wanted to ensure that controlling for this Venezuela effect does not change our estimates of how U.S. military aid affects paramilitary violence. In fact, we wanted to see if controlling for any year to year change, not just oil prices, affects our results, so we controlled for the region’s proximity to the Venezuelan border interacted with year indicator variables. It’s worth noting that statistically speaking, the point you raise should only bias our estimated effect if changes in Venezuelan funding also increased paramilitary attacks, and if there were a positive correlation between regions with bases and distance from the border, which is not the case since bases are spread out throughout Colombia. This is in part why we find that controlling for the Venezuela border effect did not change our estimates.

   Now, there are other shocks that guerilla violence does respond to. For example, in other work with Juan Vargas I have shown that a fall in the prices of agricultural commodities such as coffee does in fact lead to more violence perpetuated by both guerillas and paramilitaries in regions dependent on these commodities. (See http://politics.as.nyu.edu/docs/IO/13746/Commodity_shocks_civil_conflict.pdf). But that’s exactly why its meaningful that changes in U.S. military aid have this asymmetric effect on paramilitary violence in particular, which provides evidence for the idea that aid leaks to the group that colludes with the military.

   Finally, your comment suggests that we may be underestimating the effects of the drug trade. Let’s be explicit about how drug crops could lead to a bias in our results: if there is a positive relationship between regions that grow coca and regions with military bases, then changes in military aid may increase paramilitary violence due to the presence of coca rather than the presence of the base. If we did not control for coca, then we would be overestimating the military base effect. However, we are able to account for coca in a number of ways in the paper, since we have precise measures of how much coca is cultivated in each municipality based on data from UN satellite imagery. We include the interaction of this coca variable with military aid as a control variable. Also, we eliminate every municipality recorded as ever producing coca from the sample, and still find the military aid shock leads to increased paramilitary attacks in both cases. This helps ensure that drug related factors cannot account for our findings.

3. AdamMichnik Says:
   January 10th, 2010 at 7:33 am

   Pěkný článek a i celkově máte hodně zajímavé stránky. Jen tak dál. Dobře se to čte a je to inteligentní – to se často na internetu nevidí.

4. Dr. Paul Rigterink Says:
   January 10th, 2010 at 12:02 pm

   You may want to ask the Colombian Ministry of Defense about the security situation in Cólombia. When I inquired about whether it is considered safe for an American citizen to visit Cordoba, they sent to me a Microsoft Excel file which lists the deaths and injuries in that region from Police data as well as the cause. The Ministry of Defense can be contacted directly at http://www.mindefensa.gov.co/. The Minisry of Defense put more troops in regions where they can be most effective based on the data that the police keep. I think they are doing an outstanding job.

   In case you are interested, many of the top Government officials in Cordoba are women. They are doing an outstanding job also. Many of the current terrorists problems are near Valencia, Cordoba where the terrorists at one time felt they had a safe haven inadvertently created by European NGOs. In 1966, the Cuban guerillas lived near Ayapel and near Los Cordobas (I worked for the Peace Corps in Buenavista and Los Cordobas). Most of the municipios in this region had been burned during the Violencia period. Coroba was considered a very unsafe region by Colombians although the Peace Corps did not know this at the time. When my Peace Corps friend visited Buenavista last year, she was told that one of the leaders of this community had become a senior FARC leader out of frustration and that many of our friends had been killed by FARC leaders or forced to become members of the FARC many years ago.

We value frank and constructive exchanges and encourage you to use your real name in your comments.