Landmines, poverty and recovery : instrumental variables evidence from Mozambique

The International Campaign to Ban Landmines production and use estimates that there are more than 80 billion landmines in the ground in more than 80 countries. Despite the scale of the problem and large investments by OECD countries to clear mines in low income countries, the economic consequences of landmine contamination have been so far unexamined by economists working on the economics of wars, perhaps due to the lack of data thus far. This paper exploits a unique dataset on landmine contamination intensity covering 126 Mozambican districts. Because landmines (unlike other weapons) are used as a weapon of choice to protect territories, the empirical strategy uses an indicator of distance to strategic borders as an instrumental variable to correct for selection in landmine placement. Instrumental variables estimates indicate a large effect of landmine contamination on poverty and consumption several years after the ceasefire. Hence, despite the very high cost to clear a mine a conservative costbenefit evaluation of the national demining program indicates that the program generates a large positive return.