In the last decade an increasing interest has been devoted to concepts such as social interactions, peer effects, social norms and band-wagon effects. The theoretical underpinnings of these concepts are relatively clear, and boil down to the workings of some kind of market failure such that in equilibrium the dependence of individual behaviours on the aggregate behaviour of some reference group is not considered by agents when making their decisions. The empirical counterpart is much weaker because of serious identification problems due to the fact that average behaviour, which supposedly feeds back into individual outcomes if such mechanisms are at work, carries no additional information on top of that already embodied into observed individual outcomes. The rewards for a reliable estimation of such effects are high, especially when it comes to policy design, due to their increasing importance as a major ingredient in several fields of economic theory such as economic growth, spatial agglomeration, inequality, technological choices and so on. We exploit the 1980 earthquake in southern Italy and the subsequent relief from mandatory military service granted to all males in the regions hit by the seism to estimate the strength of endogenous social interactions in schooling achievements. Preliminary results point to a significant and strong effect of interactions not mediated by markets. |
