Public and private capital productivity in Russia: a non-parametric investigation

The productivity of public capital has been very popular research topic for US and other OECD countries, while studies using data from transitional countries are almost non-existent. In this paper, we analyze the productivity of public and private capital in Russia with parametric and non-parametric regression methods utilizing a unique regional level panel data from 2003 to 2007. More specifically, we assess public capital's spillover effects, i.e., the productivity of public capital on private output, as well as the productivity of different capital ownership types on total output. We find that public capital has a clear positive effect on private output. However, our estimates and test statistics show that parametric methods are not able to grasp vast non-linearities and heterogeneity present among Russian regions, while the non-parametric approach can capture these important features of the data better. Furthermore, we find that multicollinearity is an important methodological problem which should be accounted for in analysis concerning capital data. Our results also suggest that the impact of public capital in Russia is heterogeneous in the sense that for some regions its contribution to private output is insignificant or even negative while it has a considerable positive role for most regions. Concerning the capital elasticities of total output, we find that public capital is less productive than private capital and roughly as productive as joint private-public capital.