East-West Conference 2003

Es gilt das gesprochene Wort.

It is well known that developed financial markets allow investors to efficiently pool idiosyncratic risk. Agents can protect themselves against stochastic fluctuations in their incomes through trade in assets with appropriate payoff structures. Open and integrated financial markets allow to choose from a larger set of assets and to pool risks across borders. 

However, the usual finding in the literature is that international risk sharing is rather limited. Backus, Kehoe, and Kydland (1992) demonstrate that cross country consumption correlations are too low to be consistent with a model characterized by complete markets and perfect capital mobility. In addition, French and Poterba (1991) report a large home bias in equity holdings and consequently only a small degree of international diversification. 

Moreover, various authors have empirically tested for risk sharing using consumption data and find that the implications of complete market models are largely rejected. In particular, a common result is that the cross country correlations of output growth rates are higher than those of consumption growth rates which indicates that the opportunities for international risk sharing are not fully exploited. Moreover, consumption is usually found to react to country specific shocks, which is inconsistent with perfect risk sharing. 

However, the ongoing process of globalization and financial market integration has increased the amount of international financial transactions. Tesar and Werner (1998) present some evidence that the home bias, although still substantial, has somewhat declined over time. Thus, one might expect that risk sharing has improved over time. 

Europe appears to be particularly interesting in this context, since the creation of the EU and the EMU were to a great extent motivated by the idea to promote the integration of financial and goods markets. Blanchard and Giavazzi (2002) for instance report that the famous Feldstein and Horioka (1980) puzzle has basically disappeared in Europe over the last years, suggesting that capital mobility has indeed increased. 

The purpose of this paper is to test whether the continuing process of European integration coincides with an increase the extent of risk sharing among European countries. In particular, we search for breaks in risk sharing relationships. Our analysis is closely related to Obstfeld (1994), who explores risk sharing patterns among the G-7 countries and finds some evidence that risk sharing has increased after the end of the BrettonWoods era in 1972. We run regressions similar to those in Obstfeld (1994) with data from the current member countries of the EU and perform the tests for stability developed in Bai and Perron (1998b). Thus, the innovation of this paper is to apply a formal test to detect breaks in risk sharing relationships instead of assuming stability or comparing sub samples chosen on an ad hoc basis. 

Of course, we are not the first to analyze consumption risk sharing in Europe. Various authors have explored to what extent consumption risks are pooled between European countries and regions. A general conclusion of this literature is that risk sharing among EU countries is limited and is exploited to a smaller extent than in the US. 

Our main result is that risk sharing does not appear to have improved among European countries. For a few countries we even find indications for declining risk sharing over time which suggests that these countries have become less financially integrated with the rest of the EU.