RDP 2013-08: International Business Cycles with Complete Markets 5. Conclusion

We show that a complete markets model driven by productivity shocks alone can help to resolve two major puzzles in the international business cycle literature: the ‘international co-movement puzzle’ and the ‘quantity anomaly’. Instead of restricting international asset markets or introducing new disturbances, we deviate from standard Cobb-Douglas preferences. Our model incorporates a time non-separable version of GHH preferences consistent with a small (but non-zero) wealth effect on labour supply.

The model accounts for positive cross-country co-movement of investment and employment, and delivers a plausible international consumption correlation. In contrast to incomplete market models, ours is robust both to the persistence of shocks and the degree of cross-country spillovers. However, it fails to reverse the inconsistency of the relative magnitude of consumption and output correlations with the data, and predicts that the ratio of net exports to output is procyclical, whereas the ratio is countercyclical in the data.

A limitation of our one-good model is that it abstracts from variations in relative prices. Terms of trade movements provide an extra channel in the international transmission of productivity shocks. Incorporating this channel might improve the model's fit with respect to quantity aggregates. Furthermore, a two-good extension of our model might shed some light on price anomalies, such as the ‘Backus-Smith puzzle’. This puzzle highlights the high positive correlation of the real exchange rate with consumption in one country relative to another observed in models featuring complete markets, compared to a correlation of close to zero in the data (see Chari, Kehoe and McGrattan (2002)). We tend to agree with Huang and Liu (2007, p 1288) who conjecture that ‘non-time-separable preferences … may help break the tight link between the real exchange rate and relative consumption’. We leave the testing of this idea for future research.