On average, estimates of trade elasticities are smaller in aggregate data than at sector level. This is an artifact of aggregation. Estimations performed on aggregate data constrain sector elasticities to homogeneity, which creates a heterogeneity bias. The paper shows such a bias exists in two prominent approaches used to estimate elasticities, which has meaningful consequences for the calibration of the trade elasticity in one-sector, aggregative models. With elasticities calibrated to aggregate data, macroeconomic models can have predictions at odds with the implications of their multi-sector counterparts. They do not when elasticities are calibrated using a weighted average of sector elasticities.
展开▼
