Military expenditure in post-conflict societies

Abstract

During civil war, government military spending rises sharply as a share of GDP. In previous work we have modelled military spending and find that on average a civil war raises it by about 1.8 percentage points (Collier and Hoeffler, 2004). This figure of course understates total military spending during conflict since it omits rebel expenditures. With the reversion to peace, the society therefore has the potential for a substantial ‘peace dividend’. Indeed, taking into account rebel spending, the scope for a peace dividend from reduced military spending is likely usually to be even more substantial once a civil war has ended than after an international war. However, peace after a civil war is often fragile. Using our model of the risk of civil war (Collier et al., 2003), we estimate that during the first five years post-conflict there is a 44 per cent risk of reversion to conflict. Any sensible government facing such high risks of civil war would tend to be cautious about rapid reductions in military spending. Further, post-conflict societies inherit at least one powerful military lobby, and quite possibly two if the rebel group is brought into government. Controlling for other characteristics, military dictatorships spend two percentage points of GDP on the military in excess of what fully democratic countries choose to spend, suggesting that where the military is powerful it lobbies for its own interest. Hence, the case for military spending based on prudent deterrence is likely to be reinforced by the atypical power of the military lobby.