How Space Channels Wage Convergence: The Case of Russian Cities

Existing empirical work on the growth of Russian regions mostly covers a short time period and considers only regional-level data, while citylevel spatial data of the postreform era remain largely ignored. Using citylevel geocoded data covering 997 cities and towns from 1996 until 2013, I nd sigma- and betaconvergence across Russian cities in wages. City wages during the period under consideration display signicant and positive spatial autocorrelation. Spatial Durbin models of the Barro regression are estimated using Markov chain Monte Carlo methodology. Estimates of the spatial models for dierent weight matrices indicate that the city wage growth is signicantly aected by wage growth rates in neighboring cities, after conditioning on initial wages.