Who fights with whom and why? These questions are as old as war itself.
Montesquieu and Kant posited that countries trading with each other should be less likely to enter into militarised conflicts because they would not want to forego the gains from commerce. In a recent paper, Martin, Mayer and Thoenig (2007, 2008) provide evidence in favour of this view.1 Scores of scholars have also investigated the other “liberal peace” hypothesis – that democratic institutions should reduce the risk of conflict. This view played a prominent role in international policy debates during the Clinton and Bush years and continues to inform the current debate about “ballots and bullets” – most recently, for example, with respect to political developments in Iran.
Several critics have questioned whether trade and democracy have direct pacifying effects and argued that armed conflict has deeper roots, stemming from cultural, ethnic, and religious differences. If this “primordialist” view were correct, a lower risk of conflict would be associated with a higher degree of cultural, ethnic, and historical relatedness. According to this view, it would not matter much whether Iran adopts more democratic institutions or establishes closer economic links with other countries – its likelihood to engage in international conflict would mostly reflect deep long-term differences with its potential enemies. A clear expression of this view was given by Samuel Huntington (1996), whose widely cited book put forth the hypothesis that deeply-rooted differences across culturally and ethnically distant populations should be conducive to wars – salient cultural (particularly religious) cleavages would likely lead to a “clash of civilisations.”
New empirical research on the question
In our recent research (Spolaore and Wacziarg, 2009a), we introduce a novel way to estimate the direct effect of long-term relatedness on the risk of international conflict. Surprisingly, we find not only that the primordialist view is incorrect, but that the opposite is true – the risk of conflict is greater among more closely related populations. This effect is statistically significant, large in magnitude, and robust to controlling for all the usual determinants of interstate conflict considered in the previous literature, including trade and democracy. Most importantly, the effect of relatedness on conflict is robust to controlling for a very wide set of measures of geographic distance and isolation, implying that the relationship we uncover is not due to the fact that many conflicts occur among countries that are geographically proximate.
Our main measure of long-term relatedness is genetic distance – a measure that we also used in our recent work on the diffusion of economic development (Spolaore and Wacziarg, 2009b). Genetic distance between populations is based on differences in neutral genetic characteristics (or alleles) that mutate randomly and are not subject to direct natural selection. Genetic distance has the attractive property of being linearly related to the time elapsed since two populations shared a common ancestor population (that is, since they were the same population). Much like the number of generations separating two individuals from a common ancestor captures whether they are siblings, first cousins, second cousins, etc., so genetic distance describes the degree of relatedness between populations. It is a useful summary statistic of inter-population differences in a host of traits that are transmitted with variation across generations, from parents to children, including cultural traits (preferences, norms, values, etc) that are in fact closely correlated with genetic distance.
Why are more similar populations more likely to go to war?
We develop a simple theory that links the probability of war to the emergence and salience of common issues between populations. Populations that are more closely related and have shared a more recent common history are more likely to have developed a greater set of common issues about which conflict might erupt. We call this the “common-issues effect” and argue that it accounts for the bulk of the positive effect of relatedness on conflict. The effect is especially important when historically related populations care about the same rival good – for example, a land or city of common religious and cultural relevance – and each wants exclusive use and control of that good.
The common-issues effect may be either amplified or countervailed by additional effects. In our paper, we model two other effects – the extent of disagreement over non-rival issues and the ability to solve coordination failures. For instance, if more closely related populations tend to agree more with each other over non-rival international issues (say, international repression of piracy or terrorism), the common-issues effect is reduced. In contrast, if kin populations also disagree more about sets of non-rival issues, the common-issues effect is amplified. The net direction of this “disagreement effect” is an empirical question. Empirically, we looked at “who votes with whom” at the United Nations assembly. Surprisingly, we found that the correlation of members’ votes is positively related with genetic distance, controlling for all sorts of other variables (including differences in income per capita). In other words, relatedness is also associated with greater disagreement over international diplomacy issues, which tends to reinforce the common-issues effect.
Our findings have several implications; here we highlight three.
- First, it is not helpful to think of international conflict as arising primarily from primordial differences between culturally distant populations, leading to clashes of civilisations. In fact, the opposite is true – issues of war and peace are mostly family matters.
- Second, in the course of establishing new results, we confirm old ones; the pacifying effects of democracy and bilateral trade are robust to controlling for long-term relatedness.
- Third, our results may shed light on the sources of conflict within states as well. In their seminal study of ethnic diversity and civil conflict, Fearon and Laitin (2003) showed that, surprisingly, ethnic cleavages have little predictive power as determinants of the onset of civil wars, controlling for other determinants. This is entirely consistent with the idea that armed conflict is far more likely among kin.
1 Their evidence comes with a twist: extensive multilateral trade links reduce the value of any given bilateral link among a pair, making conflict among the countries in that pair more likely. Yet it is still the case that bilateral trade intensity reduces the likelihood of international conflict, controlling for a host of conflict determinants.
Fearon, James and David Laitin (2003), “Ethnicity, Insurgency and Civil War,” American Political Science Review, 97(1), February, pp. 75-90.
Huntington, Samuel (1996), The Clash of Civilizations and the Remaking of World Order. New York: Simon and Shuster.
Martin, Philippe, Thierry Mayer and Matthias Thoenig (2007), “Does globalisation pacify international relations?” VoxEU.org, 4 July.
Martin, Philippe, Thierry Mayer and Matthias Thoenig (2008), “Make Trade, Not War?” Review of Economic Studies, 75, pp. 865-900.
Spolaore, Enrico and Romain Wacziarg (2009a), “War and Relatedness,” NBER Working Paper 15095, June.
Spolaore, Enrico and Romain Wacziarg (2009b), “The Diffusion of Development,” Quarterly Journal of Economics, 124(2), May, pp. 469-529.