The Scioeconomic Determinants of Suicide in the OECD: Can anything be learned from cross-country regressions on suicide rates?

In this paper we first review the economic and econometric literature on the social and economic determinants of suicide rates. We find that there is a large amount of variation in the results across different studies. We go on to specify models of male and female suicide rates using a panel of 13 OECD countries. Our estimation strategy is, first, to estimate a simple OLS model and then to loosen a series of restrictions implied by standard OLS and by our chosen sample. We find that every single loosening in any possible direction of the restrictions implied by the base model changes coefficients, often dramatically. We go on to find that time series estimation methods do not improve the consistency of the results and conclude that there is probably little to be learned about the determinants of suicide rates using standard panel estimation techniques on international aggregate data.