What is the impact of fiscal decentralization on the tax decisions of Russian regional authorities? Russian federalism means, in practice, that central and regional officials co-occupy the same tax base. Such shared tax bases create a common pool problem, with the tax levies of each level of government inducing responses by private producers that will impact the tax base of both. We present a study of the determinants of tax effort in Russia's regions, asking whether a rising federal tax rate is associated with a "crowding out" or a "crowding in" of regional tax effort.
We first report results of estimating the tax effort model on a panel data set of Russia's regions, using pooled OLS, and then fixed effects, to deal with unobserved effects. These results consistently show that a rising federal tax share is associated with increasing tax effort by regional officials. However, with large geographical regions, the random sampling assumption in the cross section dimension of our panel is flawed. We test for spatial autocorrelation and augment the previous economic model with spatial statistical methods. By combining economic and GIS data, we are able to add an intrinsic autoregressive error term (Besag and Kooperberg, 1995) to our economic model and to estimate model coefficients via Bayesian methods and MCMC. The economic portion of the model identifies variables impacting tax effort, while the spatial component allows us to identify and account for regional heterogeneity and interaction among regions linked by proximity, trade links, and resource flows. Identification of spatial commonalities is an interesting outcome of the methodology, which suggests future research questions
