Over the last years, a large body of literature has investigated the ability of member countries (or regions) of a monetary union to share part of their idiosyncratic risks with the rest of the group. The question is particularly relevant, as country specific shocks of different nature might put pressure on the fixed exchange rate, undermining the sustainability of the union. On the one hand, standard Optimal Currency Area (OCA) theories look at the costs and benefits that countries face when they decide to join a monetary union. The risk sharing approach, on the other hand, looks at group of countries in a monetary union as part of an insurance pool. In this sense, some degree of heterogeneity in the economic structures becomes an asset, as different countries will be exposed to different shocks and this would increase the possibilities of risk mutualisation.
While both approaches look at financial markets only as channel of risk absorption, a growing literature is looking at financial shocks as a driver for real economic variables. What I argue is that the available methodology, while having indisputable merits, is not able to properly identify the different shocks to which economies are exposed and the spillovers of disturbances across countries. Moreover, focusing on exogenous shocks to GDP, it fails to capture the disturbances originated in the financial sector. To fill this gap, my aim is to set a link between the recent literature on financial shocks and the one on risk sharing in monetary unions. What I propose is an extension of the standard method, which is able to properly identify financial shocks and how the different absorption mechanisms react to them.
Several important methodological issues are still present in this setup and need to be tackled. First, shocks to GDP are assumed to be exogenous and to be the main trigger behind net factor income flows, international transfers and saving/investment behaviors. In this static setup, where reverse causality and feedback effects are ruled out by assumption, endogeneity issues are likely to arise. Second, the identification strategy does not allow to capture the nature of the shocks and their interaction with the channels of absorption. Indeed, it would be crucial to understand whether different channels of absorption are able to deal with some, but not other disturbances. Third, financial linkages among different countries are only considered as a mechanism of shock absorption but not as a possible source of variability of real macroeconomic variables. Indeed, the amount of risk sharing in one country is measured as the fraction of idiosyncratic shocks to GDP that does not translate into final consumption. However, take the case where a financial shock shrinks the supply of credit, with a reduction in firm investment, increase of unemployment and private consumption. Or alternatively, suppose a financial shock has an adverse effect on households¿ expectations and thus on consumption. In both cases, the shock would have a direct effect on consumption and a regression of consumption on GDP, as the one proposed by ASY, would be miss-specified.
Asdrubali and Kim (2005) extended the basic framework using a VAR model estimated with panel data. The dynamic framework they propose treats each variable as endogenous and allows dynamic feedback effects and interactions among all the variables in the model. However, using the same variables of the basic framework, they are only able to simulate exogenous shocks to GDP. This partially solve the first, but not all the other issues outlined. To fill these gaps, a new framework needs to be developed, one that allows to identify the nature of the different shocks, how they spread from one country to another and how they are transmitted to final consumption or absorbed by the different channels.