The Effects of Conflict on the Structure of the Economy

The presence of conflict affects peoples economic incentives. Some sectors of activity flourish, while others suffer. For understanding structural problems in developing countries and designing appropriate post-conflict reconstruction policies, it is essential to understand in what ways conflict affects the structure of the economy. We develop a simple model of conflict and multiple sectors of activity, where conflict efforts, the allocation of factor endowments and the production outputs are endogenous. We predict that for moderately destructive conflicts labor-intensive sectors are most affected by fighting, while for highly destructive conflicts capital-intensive sectors suffer most. In the latter case, under some conditions it is also possible that in the presence of endogenous conflict - an increase in the price of the capital-intensive commodity reduces the output of this same good. The model further predicts that export-sectors and sectors that require inter-temporal investments are particularly exposed to conflict activity. In the empirical part of the paper, we study the impact of various forms of conflict, separately and as an aggregate conflict index constructed with principal component analysis. We present some basic stylized facts about the effect of conflict on the productive structure of the economy. Conflict reduces the share of the manufacturing sector in the GDP, increases the exploitation of some simple natural resources (i.e. forestry) and reduces the production of crops. Using industrial level data for developing countries we study the channels through which conflict affects the manufacturing sector. As expected, we find that industries that are more institutional/transaction intensive are the ones that suffer most in conflictive societies. Laborintensive sectors are also negatively affected by conflict. It is also found that exporting industries and sectors requiring external financing suffer more during conflict. Our results are robust to sensitivity analysis.