I went to grad school at the University of Michigan, starting in 1988 and finishing up in 1993. For someone who studies international security issues and international war, it was a busy time with the Cold War ending and the war between Iraq and the US breaking out in 1991. My first published work involved developing econometric models of war outcomes and durations as a means of testing some arguments about which types of military strategies were most effective.

“...our papers on war durations generally, and specifically, have proven regrettably accurate in forecasting the length of the current wars in Iraq and Afghanistan.”

I served in the US Army Special Forces in the 1980s. When I went back to college in 1986 I was confronted by a gap between the nuclear weapons policy I learned about in the service, which was focused on tactical nuclear weapons and the policies I learned about in college, which focused on strategic nuclear weapons. It seemed to me that many academics who worked on security issues were focusing on the wrong questions and missing the aspects of military competition that were particularly destabilizing.

EUGene is a software package that a co-author and I developed to help smooth the data-management aspects of doing econometric work on security issues. Many graduate students shy away from doing relatively more sophisticated work because of the time-consuming aspect of data management. We hoped to be able to simplify this process, of developing and querying complex data sets with an eye towards shaping the research agendas of a future generation of people working in this area.

The software makes simple a variety of tasks necessary in order to merge and query dozens of what are commonly disparate data sets. Basically, with the software, a user can create very complex SQL queries simply by pointing and clicking on a series of radio buttons and tabs. The software then outputs a complete data set into a variety of formats for commonly used statistics packages. It's updated about once a year, and is widely used both in scholarly work and as a teaching tool.

We were interested in comparing the effects on war durations of material military and economic factors such as troop strength, GNP, and population to more political factors such as a state's relative degree of democratization or level of political repression. We found that while the material factors matter most, the wars that democracies have fought in are significantly shorter than we would expect otherwise.

This finding led to a group of papers and a book investigating the link between states' political institutions and wartime behavior. The consistent finding is that democratic states fight in shorter wars, and win more often than previous scholars had expected. The most likely reason appears to be that democratic states tend to shy away from particularly tough fights, unless they are attacked first.

One of the current conjectures about the fundamental causes of war is that disagreement about the nature of war makes it much more difficult for potential adversaries to negotiate a peaceful solution to their disputes. The common way to model disagreement between rational actors is to assume that it is private information, or secrets, that drives the disagreement. We argue that while this is a possible explanation, it is not the only one, nor likely the most prevalent source of disagreement between potential adversaries. We argue that more commonly, adversaries have pretty good knowledge about each other’s capabilities. What drives their disagreement is that they have different theories about how those capabilities will interact to influence the nature (cost, duration, outcome, etc.) of any potential future war between them.

In this paper, war then serves as a theory-testing mechanism rather than a process of revealing previously private information. Some of the implications of the model are that longer, less intense wars may lead to longer periods of peace afterwards; that brief, very intense wars are likely to be followed by very unstable relations between the warring parties; and that third-party mediators will have a difficult time substantively affecting the states' negotiations whether the mediators are biased towards one side or not. Our approach is controversial because we relax or modify a couple of the foundational assumptions in game theory, namely how it is that rational actors disagree.

Well, our papers on war durations generally, and specifically, have proven regrettably accurate in forecasting the length of the current wars in Iraq and Afghanistan. Our 1998 paper on the declining advantages of democracies is also holding up well in light of current events in Iraq. Our paper on bargaining and war helps provide some insights into why the 1992 settlement in Iraq between the Iraqis and the UN proved notoriously unstable.

No secrets here. We're working on a variety of things. In one paper we are estimating some Bayesian models to try and better understand the targeting policies of Islamic terrorist groups. We're looking to see whether it is countries' freedoms, policies, or degree of integration into the international economic system that best predicts the location of terrorist attacks.

In another set of papers we are doing some GIS (geographic information systems) work on the Rwandan genocide to try and determine the relative culpability of the various groups and parties involved in the civil war that took place during that period. We're comparing daily data on violent attacks at the political sector level to troop locations and both contemporary reporting as well as retrospective survey data on the nature of violence in various locations.

We're also working with an Indian NGO (non-government organization), the Navsarjan Trust, to execute a village and household census in Gujarat, India to try and track the evolution of discrimination against members of the untouchable castes.