Essays on sticky price models of the international business cycle.

Abstract

The first chapter of the dissertation connects two empirical investigations on the issue of which assumption is more appropriate for international business cycle research: sticky export prices or sticky import prices. First, vector-autoregressions demonstrate that in the United States and Canada, import prices tend to react faster to money and exchange rate shocks than the general price level and export prices. The second investigation of the first chapter concerns an important implication of export price stickiness. In an economy with a high level of import penetration, domestic prices should be more responsive to shocks than in a closed economy, due to volatile import prices. Domestic producers find it costly to keep their prices fixed. In the case of fixed exchange rates, on the other hand, import prices are independent of domestic nominal shocks, which makes domestic prices more rigid. Hence, the short-run Phillips curve should be steeper in an open economy with flexible exchange rates, and flatter with fixed rates. A panel version of Ball, Mankiw and Romer (1988) empirical procedure gives support for this implication. This evidence indirectly suggests presence of export price stickiness. The second chapter presents a sticky price dynamic general equilibrium model of a small open economy, which demonstrates several powerful results. The presence of physical capital in the model helps achieve a current account deficit in response to a money shock, which is unusual in the literature, but has empirical support. The trade balance is then found to be countercyclical in the simulated economy, in which the business cycle is driven by either monetary or persistent technology shocks. The nontradable sector in the model helps achieve persistent deviations from purchasing power parity with a high degree of real price rigidity. Likewise, the nontradable sector helps generate overshooting of the exchange rate, which depends on non-separability of labor and consumption, as opposed to the more traditional liquidity effect based overshooting. However, the volatility of nominal or real exchange rates observed in the data can only be reproduced after the introduction of the local currency pricing assumption. Lastly, a virtue of the proposed model is its analytical tractability, which helps develop firm intuition about the model before using numerical solutions.