We use new data to examine the effects of giant oilfield discoveries around the world since 1946. On average, these discoveries increase per capita oil production and oil exports by up to 50 percent. But these giant oilfield discoveries also have a dark side: they increase the incidence of internal armed conflict by about 5-8 percentage points. This increased incidence of conflict due to giant oilfield discoveries is especially high for countries that had already experienced armed conflicts or coups in the decade prior to discovery.
Work in progress:
Proxy Warriors: A Theory of Military Assistance
Observing the recent conflicts in the Middle East, and, in particular, the issue of the spread of military resources well beyond government’ control, one main contributing feature seems to be that governments sponsor weapons to non-state agents (militias) to fight the state’s enemies, but after the war ends these weapons cannot be called back. I develop a theory to study government's optimal strategy in sponsoring weapons to militias. Although giving militias more weapons may win the war more quickly, it puts more weapons into the hands of militias afterwards. Militias learn that they can benefit from holding more weapons when hostilities end. This incentivizes militias to save the weapons from loss. Given this governmental dilemma, we analyze in a repeated game two possible government strategies: a supply fixed for every period vs. intermittent supply – and extend them to other scenarios.
Lightning Wires: The Travel of Information and Market Integration (with Pei Gao)
How much does information friction distort market integration? How far can integration translate into regional insurance against extreme weather conditions in a pre-modern society? This paper uses the introduction of the telegraph network in late 19th century China to look at the effect of the reduction in information friction on market integration. Using a newly collected archival dataset on early telegraph network, we first examine the effect of this reduction on price movements and dispersion of crops across regions. We then turn to look at whether the reduced information friction affects insurance across regions in extreme weather conditions.