Economic crisis, unemployment and health: conceptual and methodological issues in the modelling of national contexts

摘要

The study by Laanani et al is a welcome contribution to the growing body of empirical literature interrogating the relationship between the recent economic crisis, unemployment and suicide rates. We nevertheless find that their analysis suffers from three major conceptual and methodological shortcomings. First, despite making an important distinction between the direct and indirect effects of the recent economic crisis on suicide rates, we believe that the relationship between these effects is inappropriately specified in the analysis. According to the authors, the impact of the crisis on suicide rates is explained by an individual-level unemployment effect and a contextual-level crisis effect. To the extent that they are modelled in a distinctly additive fashion, both of these effects are thought to operate independently from the other. The authors distinguish between the two effects in order to parse out any potential bias caused by contextual features associated with the crisis, including an increase in perceived levels of job insecurity and political decisions such as budgetary cutbacks. They control for such contextual features with the intention of isolating the independent effect of unemployment on suicide rates. However, the authors neglect to consider the ways in which the latter effect itself varies as a function of the contextual factors for which they are attempting to provide a control. Ample evidence exists to suggest that unemployment does not transmit automatically from a decline in economic output. On the contrary, unemployment shocks are fundamentally shaped by already-existing labour market institutions -including those that operate as objective and subjective sources of job security-and the political (ie, budgetary) decisions governments make in the process of interpreting and resolving economic crises. In other words, unemployment is endogenous to, rather than independent from, these contextual features of economic crises. Unemployment must therefore be operationalised and interpreted as such if its implications for suicidality are to be appropriately captured by our models.