@ARTICLE{26543120_228586666_2018, author = {Daria Averina and Taisiia Gorshkova and Elena Sinelnikova-Muryleva}, keywords = {, Phillips Curve, inflation, regions, inequality, clusterizationGMM}, title = {Phillips Curve Estimation on Regional Data}, journal = {HSE Economic Journal }, year = {2018}, volume = {22}, number = {4}, pages = {609-630}, url = {https://ej.hse.ru/en/2018-22-4/228586666.html}, publisher = {}, abstract = {This article is devoted to estimating the Phillips curve on the regional panel data for Russia, that allows to understand the degree of regional heterogeneity of the main macroeconomic indicators.Using the generalized method of moments, the hybrid Phillips curve and its partial cases (adaptive expectations-augmented and neokeynesian) are estimated in this paper. The results show that among three model’s modifications the hybrid model with the output gap and forward-and back-looking expectations is the most fitting to real data. Thus, we can assume that inflation expectations in Russia include both adaptive and rational components.At the same time, the Phillips curve with the unemployment rate incorrectly describes the regional panel data. We associate this result with the specifics of regional labor markets and divide the regional sample into clusters by the indicators of income and unemployment, which gives the formation of three clusters: 1) with high income per capita, high level of the economically active population and low unemployment, 2) with low income per capita, low level of the active population and high unemployment, 3) with middle values of each indicator.The results of the model estimation show that the Phillips curve does not describe the data on the sample of «rich» regions, which is explained by the natural and climatic conditions in these regions and/or by specific state premiums. At the same time, inflation and unemployment in regions with low and/or moderate incomes are adequately described by the Phillips curve. The obtained results allow to understand more deeply the mechanisms of pricing in the Russian regions, the specific features of labor markets and the formation of macro variables at the regional level.}, annote = {This article is devoted to estimating the Phillips curve on the regional panel data for Russia, that allows to understand the degree of regional heterogeneity of the main macroeconomic indicators.Using the generalized method of moments, the hybrid Phillips curve and its partial cases (adaptive expectations-augmented and neokeynesian) are estimated in this paper. The results show that among three model’s modifications the hybrid model with the output gap and forward-and back-looking expectations is the most fitting to real data. Thus, we can assume that inflation expectations in Russia include both adaptive and rational components.At the same time, the Phillips curve with the unemployment rate incorrectly describes the regional panel data. We associate this result with the specifics of regional labor markets and divide the regional sample into clusters by the indicators of income and unemployment, which gives the formation of three clusters: 1) with high income per capita, high level of the economically active population and low unemployment, 2) with low income per capita, low level of the active population and high unemployment, 3) with middle values of each indicator.The results of the model estimation show that the Phillips curve does not describe the data on the sample of «rich» regions, which is explained by the natural and climatic conditions in these regions and/or by specific state premiums. At the same time, inflation and unemployment in regions with low and/or moderate incomes are adequately described by the Phillips curve. The obtained results allow to understand more deeply the mechanisms of pricing in the Russian regions, the specific features of labor markets and the formation of macro variables at the regional level.} }