Spatial effects in a common trend model of US city-level CPI

• We study relative movement prices in 17 US cities.
• We allow for spatially correlated shocks.
• The parameters of the model are estimated using a generalized method of moments (GMM) method.
• We find a slow rate of convergence of the price levels and strong evidence of spatial effects.

This paper studies relative movements in price indices of 17 US cities. We employ an unobserved common trend model where the trend can be stochastic or deterministic with possible breaks or other nonlinearities. To accommodate the spatial nature of the data we allow for spatially correlated short-run shocks. In this way, the speed of convergence to the equilibrium implied by the law of one price is estimated taking into account the effect of distances across cities. The parameters of the model are estimated using a generalized method of moments (GMM) method which incorporates moment conditions corresponding to a generalized least squares-like within estimator of regression parameters. We find a slow rate of convergence of the price levels and strong evidence of spatial effects.