Estimating Gravity from the Short to the Long Run: A Simple Solution to the 'International Elasticity Puzzle'

We propose a simple and flexible reduced-form econometric approach to estimate gravity models in the short and the long run. The theoretical lens for interpreting our methods amends the canonical Lucas-Prescott adjustment formulation to allow for time-interval-varying depreciation-cum-adjustment. A time-varying trade elasticity in the structural gravity model is implied. Our methods explain the 'international elasticity puzzle,' the discrepancy between trade elasticity estimates from the trade literature and the international real business cycle literature. The same theory-motivated estimating equation applied to the same data generates a distribution of trade elasticity estimates that vary from 0.4 in the short run to 4.8 in the long run. The results offer support for some existing theories of dynamic adjustment in trade costs and imply that the long-run equilibrium in our sample is reached in about 16 to 17 years.