City Size Distribution Dynamics in Transition Economies: A Cross-Country Investigation

Project type: CERGE/GDN IRC, 2009-2010

Project team:

• Ciprian Necula, Bucharest Academy of Economic Studies - coordinator
• Uladzimir Valetka, Belarus State Technological University, CASE
• Marat Ibragimov, Tashkent State University of Economics
• Gabriel Bobeica, Bucharest Academy of Economic Studies
• Alina Nicoleta Radu, Bucharest Academy of Economic Studies
• Kamila Mukhamedkhanova, Center for Economic Research
• Aliaksandr Radyna, Belarus State University

Abstract

The empirical literature in urban economics have come to an agreement that, in the long run, the growth rate of a city is independent of its initial size (Gibrat`s law) and that the city size distribution could be closely approximated by a Pareto distribution with an exponent that varies among countries. It is of great interest to determine the factors that influence the value of the exponent, for such a relationship may point to policy-related issues. The general objective of the present Project is to study the dynamics of the city size distribution in CEE and CIS transition economies, and identify the determinants of the variation of this distribution in time and across countries. We will build a comprehensive unified database for CEE and CIS countries concerning city dynamics. We will test the Gibrat`s law employing Pesaran panel unit root test that takes into account the presence of cross-sectional dependence and Nadaraya-Watson non-parametrical kernel regression. The main novelty of the Project resides in constructing a consensus estimate of the Pareto exponent of the city distribution using various econometric methods. We also test for non-Pareto behavior of the distribution of when all the cities in a country are considered, using the Weber-Fechner law, the logarithmic hierarchy model, and the log-normal distribution. Not only we consider various distributions, but also study the �within distribution� dynamics by analyzing the individual cities relative positions and movement speeds in the overall distribution using a Markov chains methodology. In order to explain the differences in the city distributions and obtain valid statistical inference, we will estimate, using Driscoll-Kraay robust standard errors, a panel data fixed effects model to control for unobserved country specific determinants.

Completed paper

Home
News
AEFR - Working Paper Series
Research Projects
Publications
Members

..............................................................