Consumption Risk-Sharing and the Real Exchange Rate: Why does the Nominal Exchange Rate Make Such a Difference?

A basic prediction of effcient risk-sharing is that relative consumption growth rates across countries or regions should be positively related to real exchange rate growth rates across the same areas. We investigate this hypothesis, employing a newly constructed multi-country and multi-regional data set. Within countries, we find signifcant evidence for risk sharing: episodes of high relative regional consumption growth are associated with regional real exchange rate depreciation. Across countries however, the association is reversed: relative consumption and real exchange rates are negatively correlated. We identify this failure of risk sharing as a border effect. We find that the border effect is substantially (but not fully) accounted for by nominal exchange rate variability. We then ask whether standard open economy macro models can explain these features of the data. We argue that they cannot. To explain the role of the nominal exchange rate in deviations from cross country consumption risk sharing, it is necessary to combine multiple sources of shocks, ex-ante price setting, and incomplete financial markets.