Does trade promote peace? squared: a gravity equation in a rectangular panel world

The purpose of this paper is to revisit the question if trade promotes peace or not? I account for heterogeneity of trade dyads over time in using panel estimation techniques. The world is modeled as a rectangle. I present models focusing on how conflict affects trade, and in another set of models how trade affects conflict. To account for simultaneity I use past values of trade and conflict, as well an instrumental variable approach. My instruments to explain conflict are military expenditures and a military capability index. The instrumental variable to explain trade is annual rainfall. I find in most setups that trade and interstate conflict are reciprocal. Trade indeed promotes peace because of welfare gained from international trade. Past values of conflict or trade have a negative impact in their respective models. Only after accounting for endogeneity in using an instrumental variable approach, the negative relationship becomes insignificant or positive in some setups. I employ a dynamic panel estimator to deal with possible limitations of the instrumental variables.