The approach has merit: it is useful for us to look for patterns in how things work, even if differences between experiences outweigh similarities. However, it can also bias the choice of the factors we analyse towards those with the most uniform properties, suitable for this kind of theorisation (the obvious counterpoint, that all research methods have biases, is true: this is why it is important to draw on research from multiple disciplines). I saw an example of this in a very interesting book Iâ€™m reading at the moment: When Things Fell Apart, by Robert H. Bates, a political scientist.
Bates is looking at state failure and conflict in late-20th Century Africa, noting an increase in incidence of civil war and predatory state behaviour. He puts forward a basic resource-based theory: essentially, state predation emerges when the discounted returns from predation on the society and seizing or stealing resource are higher than the discounted returns from taxation revenue. The theory predicts that as the time horizon of the political leaders reduces, so their tendency to predate on society should increase; similarly, if possible tax revenues fall, they will again tend towards becoming predatory.
Obviously, Bates would recognise that this is not the whole story, but his argument is that this is the basic starting point for understanding civil war. His approach is useful: resources are crucial to understand in conflict, though their role can be different in different circumstances. Yet the method can obscure understanding of other issues. To give a minor example, he writes:
I argue that ethnic diversity does not cause violence; rather, ethnicity and violence are joint products of state failure.
This kind of statement bothers me; it betrays far too rigid a conception of the world. Ethnicity is not â€˜causedâ€™ by state failure. Ethnic identities exist everywhere. Hong Kong has a dominant ethnic identity of Han Chinese, and many minority ethnic identities. I would be extremely dubious of anyone who claimed Hong Kong was a failed state. Further, in the African context, we could argue that the statement is turned upside down. It may well be far more useful to say â€˜unified identity is the product of state successâ€™. The history of African state-formation suggests that examples like Tanzania where national identity supersedes ethnic identity in many contexts are rare, because pre-existing ethnic identities were welded into states, and these identities have continued to evolve over time. On this particular point, this is just nitpicking. His analysis of ethnicity does not undermine his central argument. But I would argue that ethnicity (and identity more generally) is one of the concepts that model-based analyses struggle with, which is not to say that no models using ethnicity are worthwhile. Some other concepts Iâ€™d group in this â€˜troublesomeâ€™ category for economics are leadership, ideology and motivation.
It may seem unusual to claim that motivation is difficult to model. Bates for example, implies that actors are pursuing self- or state-enrichment in his model. Itâ€™s important, however, to distinguish between motivation and incentives. Economics is exceptionally good at thinking about and modelling incentives. We know, for example, that all actors interact with each other and their environments responding to a range of different incentives, economics is probably the best tool out there for trying to understand how they operate and change behaviours. Incentives are essentially underlying relationships that an actor pursues.
Motivation is more complex and less modelable; it may change more rapidly and can reflect quite a few influences and incentives. While incentives of a state may be self-enrichment, its motivation in becoming predatory may be more complex. It may be seeking to self-enrich by destroying a particular section of society which poses a threat; alternatively it may seek to increase its power base by playing sections against each other. Motivations can also be more prosaic: you may simply wish to destroy enemies. Some historians have Stalin as an essentially paranoid-schizophrenic who would have a party member purged if he didnâ€™t use a coaster. Such motivation can be just as important as incentives, though much less structured.
Ideology is another crucial aspect of how people think and behave that is particularly difficult to model. Ideology does not always provide the basic motivation or cause for conflict or action but may have important consequences for the form in which action is taken, and the way in which it proceeds. Take Tanzania, for example. Nyerere forged an ideological basis for the changes he sought by welding a mixture of Chinese Communism and his vision of African Socialism into ujamaa. He was a developmental leader in that his aim was the enrichment of his country; his ideology was the most important factor shaping how he sought to achieve it: pan-Africanism; promotion of Swahili and ujamaa. It also did much to shape the way opposition formed to him: villages refusing to collectivise and seeking market-based solutions to their constraint. They were met with a violence that reflect the motivating ideology behind it: individual farms had their crops and homes burnt; â€˜capitalistsâ€™ in the cities were forced to move to the shamba to farm.
Whatâ€™s more ideology can at some point supersede the importance of the original motivation or incentives behind a conflict. For example, the religious and ethnic aspect of violence between Israel and Palestine has developed a life of its own. Not only does it help determine how violence occurs, it also has begun to be part of the central problem. Where once land was sought because it was land, it is now land is invested with an identity by both sides of this conflict; its identity is now part of the conflict, living a separate life from the resource they are fighting over. Ideology is an exceptionally difficult idea to generalise based on precisely because itâ€™s a high-level abstraction: in practice it encompasses so many different movements.
Leadership, too, matters. This is something that economics can deal with much better than it does, but sometimes seems to underestimate. In my MSc thesis, I looked at the collective action problem of unionisation in Sri Lanka; what I found was that the literature on collective action was very useful in suggesting what barriers might be observed, and how the game theoretic approach provides us with a useful framework for thinking about how these can be overcome from an essentially liberal viewpoint or an essentially coercive viewpoint. The dynamic aspect that is much less well covered is how leadership and in particular, charismatic leadership can change the ways in which people perceive the players (the structure of the game) or their payoffs. In Sri Lanka, for example, estate Tamils, a particularly discriminated-against group, achieve political importance and large gains to their welfare through a union dominated by its leader, a central figure in Sri Lankan public life at the time. After his death, fortunes declined. It is virtually impossible to â€˜proveâ€™ causation here, but evidence of sequencing, contemporary sources and our knowledge of the lack of coercive punishments for defection in this union all point towards the conclusion that leadership was crucial.
What makes many of these things difficult to model is that they are so contingent: on time, place, luck. They may have very little effect or may play a crucial role in a countriesâ€™ history; they may affect only forms of action in some places, and may change its direction altogether in others.
To say, then, that economics or political science does not model them well is not particularly a criticism of either. Their best use is looking at more structural relationships. However, it is a hint that we should be considering how a multidisciplinary approach will improve our work further. Iâ€™ve discussed this with friends, who have raised objections: one suspects that specialisation will yield greater returns in academia, and another points out that it is simply extremely hard to be good at two different disciplines, when there is so much to read and know in any one. I think these are both valid points, though specialisation is also to blame for the unwillingness of many scholars to engage with criticism from outside their own discipline.
Ultimately, a multi-disciplinary approach can come within individual scholars or from broader use of a range of source by the professionals and policy makers who ultimately actually make change. It might be just my experience, but I see little evidence of either making much impact.