International Real Business Cycles

Backus, Kehoe, and Kydland build a two-country extension of Kydland and Prescott’s (1982) closed-economy real-business-cycle model – with complete international asset markets, country-specific technology shocks, and costless trade in goods and capital – and ask whether it can account for international as well as domestic business-cycle comovements. The model reproduces several closed-economy facts reasonably well (output volatility is 91 percent of the U.S. figure; consumption is smoother than output, 0.40 versus the data’s 0.49), but it badly misses on the international dimension: investment and the trade balance are far more volatile than in any of the twelve countries examined, and – the paper’s “most robust” finding – the model generates a cross-country consumption correlation (0.88) that greatly exceeds its output correlation (which is actually negative, -0.18), the exact opposite of every developed-country pair in the data, where output correlations consistently exceed consumption correlations. A battery of robustness experiments – raising risk aversion, adding a distributed lag on leisure, shortening time-to-build, adding a transport cost on trade, and even eliminating all trade in goods and assets (autarky) – narrows some gaps (notably investment/trade-balance volatility and the sign of the output correlation) but leaves the consumption/output correlation anomaly essentially intact, including in autarky, where no risk-sharing mechanism exists at all. The authors call this their “most robust” finding and leave open whether extensions such as non-separable preferences or taste shocks, pursued in contemporaneous work the paper cites, can resolve it.

F44 | Macro Paper Warehouse

International Real Business Cycles