Printer Friendly

Cities in transition.


The world's largest cities are major drivers of economic growth. They attract the brightest minds and the most productive firms. The largest cities in transition countries lag behind. This paper investigates factors that slow down growth of the prime cities in transition countries. In particular, it aims at explaining low population mobility. For a long time, central planners had been restricting internal migration to and industrial production growth in the prime cities, while encouraging development of the secondary cities. De jure, restrictions were softened with the fall of the communism. De facto, economic and political factors that prevent people from moving are still in place, as internal migration in the region remains at a very low level. As a result, convergence of the urban systems in transition countries from the centrally planned equilibrium to the market-based one is slow.

The early stages of transition drew substantial attention of researchers in urban economics. (1) However, the literature has relatively little to say about more recent developments that have occurred during the last decade. To access what factors contribute to low internal migration during the last decade, I look at the Life in Transition Survey (LITS) II that describes, among other things, housing conditions and satisfaction with local governments for almost 39,000 households from 34 countries in 2010. (2) I identify that the housing market inefficiencies played the major role in explaining low internal mobility. Two factors are found particularly important. First, there is a strong and robust positive correlation between population mobility and availability of rented housing. Second, high mobility is associated with the share of households with mortgages in the urban area. The effects are economically important. A standard deviation increase in the share of rented housing in transition countries is associated with a 50% closing the gap in mobility between Western Europe and transition countries, while a standard deviation increase in the share of mortgaged dwellings is associated with a 25% closing the gap in mobility.

Absence of properly functioning land markets during the socialism further exacerbates the problem of low mobility and lead to spatial distortions in the distribution of economic activities within urban areas (Bertaud and Renaud, 1997). In the market-based economy with efficiently functioning land markets, land prices act as signals that help to recycle inefficient use of the valuable land within city limits and replace it with more economically sound activities. Absence of such signals in centrally planned economies led to economically inefficient allocation of economic activities within socialist cities. Since the transition to the market economy has started, land markets have developed in all transition countries, but the institutions regulating land markets are still underdeveloped and lack the legislative base. Quick reformers in Eastern Europe that have created those institutions and developed the legislative base early on were more successful in urban development, while slow reformers are still lagging behind, having difficulties to revitalize their urban areas.

Under-sized prime cities in the region relative to the prime cities in other parts of the world are a challenge for the policymakers. Larger cities are more attractive to skilled workers and investors due to positive externalities brought in play by the agglomeration forces. Therefore the findings presented in this paper lead to a conclusion that the urban development policies in transition countries should be channelled towards increasing internal mobility of population by means of reducing inefficiencies in the housing market and developing market-based institutions that regulate land use within city limits.

The rest of the paper is structured as follows. The next section focuses on urbanization and distribution of city sizes in the region. The subsequent section discusses factors that contribute to low mobility and distort the spatial equilibrium. The penultimate section draws policy implications of underurbanization. The last section concludes.


Economic polices of the central planners during socialism were not favourable to development of large cities. According to Ofer (1976), in the 1960s socialist countries were under-urbanized by about 9 percentage points relative to the OECD countries at the similar level of development. Ofer suggests that the under-urbanization was a result of the development strategy to maximize the rate of capital accumulation by suppressing rural-to-urban migration and diverting resources from the prime cities to the secondary cities. The replication of Ofer's methodology for 2008 demonstrates that currently transition countries have reached levels of urbanization similar to the OECD countries as Table 1 indicates. (3)


To test formally whether transition countries systematically differ in terms of urbanization from other countries, I estimate a linear regression model with the urbanization rate regressed on the set of variables specified in Davis and Henderson (2003):

[urbanization.sub.i] = [alpha] + [transition.sub.i][gamma] + [X.sub.i][beta] + [[epsilon].sub.i] (1)

where urbanization is the share of the urban population in country i; transitioni takes the value of one if country i is a transition country and zero otherwise; X is the vector of controls that includes log real GDP per capita, log population, openness to trade, landlocked dummy, log land area, and voice and accountability indicator that captures the average level of democracy in country i in 1996-2008; [[epsilon].sub.i] is the error term.

First, the regression is estimated on a cross-section of data for all countries in 1991, 1998, and 2008. The results are presented in columns (1)-(3) of Table 2. Conditional on the level of economic development, population size, openness to trade, and geographic characteristics, transition countries do not systematically differ in rates of urbanization from the other countries in the sample. The coefficient of the log of GDP per capita is positive and significant in all regressions that confirms the finding from the literature of strong positive association between the level of economic development and urbanization. (4) Column (4) presents the random effect estimation of (1) on a panel of all countries in 1991-2008 with no significant effect on the estimate of the transition coefficient.

According to Davis and Henderson (2003), urbanization comes along with a shift from agriculture to industry and modern services. Column (5) controls for the shift by including the share of agriculture to GDP and ratio of services to industrial production for a cross-section of countries in 2008, but those variables turn out to be non-significant. To address a concern that the result are driven by endogeneity of GDP per capita, column (6) reports coefficients of the regression for a cross-section of countries in 2008 with a 5-year-lagged GDP per capita, which does not change the main conclusion. Finally, it might be argued that the transition countries should be compared with a narrower sample of developing countries. I report the results from two restricted samples in 2008: non-OECD countries in column (7) and low and middle income countries in column (8), but the main conclusion stays intact.


While the urbanization rates in transition countries are not statistically different from the urbanization rates in other countries, distribution of city sizes in transition countries deviates considerably from the distribution in other countries at least in two important ways. First, prime cities in transition countries are small relative to prime cities in other parts of the world. Second, the whole distribution of city sizes is skewed towards medium-sized cities. To support the first claim, the following regression is estimated

[prime.sub.i] = [alpha] + [transition.sub.i][gamma]] + [capital.sub.i][delta] + [X.sub.i][beta] + [[epsilon].sub.i] (2)

where [prime.sub.i] is the share of urban population living in the largest city in country i, [capital.sub.i] is a dummy variable that indicates whether the largest city is also the capital of the country, and other controls are the same as in equation 1.

First, the regression model is estimated on a cross-section of data for all countries in 1991, 1998, and 2008; results are presented in columns (1)-(3) of Table 3. Unlike results for the urbanization rates, in 2008 the largest cities in transition countries were 9.5 percentage points smaller than the largest cities in the other countries. (5) The gap declines over time--from 14 percentage points in 1991 to 11 percentage points in 1998 to 9.5 percentage points in 2008--but the rate of convergence in 1998-2008 slows down considerably.

This finding is robust to different model specifications, time periods, and samples. Column (4) presents the random effect estimation of equation 2 on a panel of all countries in 1991-2008 with the transition coefficient of -0.11. In column (5) I control for the shift from agriculture to industry and services. Column (6) reports coefficients of the regression for the cross-section of countries in 2008 with a 5-year-lagged GDP per capita, which does not change the main conclusion. Finally, restricting the sample to non-OECD countries in column (7) and to low and middle income countries in column (8) does not change the conclusion that prime cities in transition countries are smaller than in other parts of the world.

Distribution of cities

Henderson et al. (2004) summarize stylized facts about distribution of cities within a country. First, wide and stable distribution of city sizes is regularly observed in large economies. There is an entry and exit of cities at the bottom of the urban hierarchy, but the top part of the distribution is very stable over time (Eaton and Eckstein, 1997). In particular, the upper part of the distribution is well approximated by a Pareto distribution: Rank(Size) = A x [Size.sup.-[beta]], where Rank is the rank of a city within the country and Size is measured by population. In the log form

ln (Rank) = [alpha] - [beta] ln (size)

where [alpha] = ln A.

Zipf's law states that [beta] = 1. Gabaix (1999) argues that Zipf's law naturally emerges if urban population in a country follows a proportional growth process. The proportional growth hypothesis, a so-called Gibrat's law, was tested for different regions and countries and in general was not rejected (Ioannides and Overman, 2003).

To check how well the distribution of city sizes in transition countries fits Zipf's law, I estimate the following model

ln([Rank.sub.i]) = [alpha] - [beta] ln([Size.sub.i]) + [[epsilon].sub.i],

where [S.sub.i] is size of city i, [R.sub.i] is its rank within a country, and [[epsilon].sub.i] is disturbance term. The model is estimated separately for each country and for 2 years--1979 and 2007. I consider three Eastern European countries--the Czech Republic, Hungary, and Poland--and the former Soviet Union (FSU) countries. The FSU is treated as an integrated system of cities. Data on city sizes in 1979 are from the 1979-1981 census. (6) Data on city sizes in 2007 are from the World Gazetteer database. (7) Finally, I use two cut-off levels of city sizes: 20,000 inhabitants and 50,000 inhabitants. Zipf's law is usually tested for the upper part of the distribution of urban settlements, which is often defined as cities with 50,000 and more (Henderson, 2005), but this threshold is perhaps too large for the Czech Republic and Hungary.

Results are presented in Table 4. Zipf's law is never rejected for Hungary. For the Czech Republic, Zipf's law is rejected at the 20,000 threshold and is not rejected at the 50,000 threshold. Zipf's law hypothesis is always rejected for Poland and is rejected for the FSU at the 50,000 threshold level. As the next step, I analyse the regression residuals. Figure 1 reports a panel of scatterplots of the residuals against In(Ranks) for each country and time period in the 20,000 threshold sample. The finding from the literature is that the largest city is usually a positive outlier--its size is larger than predicted by the Zipf's law (see, for example, Gabaix, 1999; Ades and Glaeser, 1995, for a political economy explanation to this stylized fact). However, it is generally not the case for the transition countries. In particular, the upper part of the distributions for the FSU and for Poland considerably deviate from the power distribution and have negative residuals. The deviation is the most pronounced for the FSU, where the 30 largest cities have population below the levels predicted by Zipf's law. At the same time, the secondary cities in the FSU and Poland tend to be positive outliers, which indicates larger than expected city sizes in the middle of the distribution. Also, the distribution has strong persistence--there are no large and systematic differences between the 1979 and 2007 residuals, which is quite an unexpected result given the dramatic economic and political changes in transition countries that occurred between 1979 and 2007.


These preliminary observations on the peculiarities of the distribution of city sizes in transition economies perhaps indicate a large, long-lasting impact of the socialist urban policies on the city growth rates at the top of the distribution. The concern with under-sized prime cities in transition countries stems from the fact that the urban economies of scale in the region are not fully exploited, resulting in lower productivity per worker and lower wages relative to urban agglomerations in other parts of the world. The loss of competitiveness due to lower urbanization transforms into lower economic growth (Henderson, 2003). This conjecture is supported by results of Deichmann and Henderson (2000), who find that Poland's primacy rate, which is 5% below the optimal level, translates into 0.75% decline in economic growth.


A spatial location model (SLM), which analyses the geographical distribution of economic activities, is built on three indifference conditions: workers are indifferent whether to stay in one location or to move to another, firms are indifferent whether to hire more workers or not, and construction firms are indifferent whether to build more houses in that location or not (Glaeser and Gottlieb, 2009). Therefore, the efficiency of the spatial allocation of resources crucially depends on well-functioning labour, land, and housing markets and on absence of distortive policy interventions. A high mobility of US population, high competition among firms, and elastic supply of housing makes this model suitable for the analysis of the US economy. However, violation of the underlying assumptions of the SLM for transition countries generates distortions to the distribution of population, housing, and firms in transition countries, potentially leading to the inefficient allocation of resources.

These distortions can explain the stylized facts presented in the previous section. Did the shock introduced by the central planners to the spatial equilibrium led to the irreversible shift in the equilibrium distribution of cities within transition countries, which have not disappeared after the policies had been removed? This question is closely linked to the question of the uniqueness of the spatial equilibrium. Several studies look at historically provided sources of exogenous variation that would potentially address the question of the uniqueness of the market equilibrium. Davis and Weinstein (2002) test the multiplicity of equilibria for city location that comes from the new economic geography models by looking at the impact of the bombing of Japanese cities on the spatial distribution of population and find little support for the multiplicity of equilibria. (8) To the contrary, they find that the location fundamentals such as favourable geographic location play a major role in location of cities. Bosker et al. (2007) also present empirical evidence in favour of the unique equilibrium based on German data. However, those studies are far from conclusive due to the short time span of the distortive shocks and some country-specific factors. More evidence from other countries, especially from countries that have large territories, such as Russia, are needed in order to prove the uniqueness of the spatial equilibrium in more general settings. Mikhailova (2010), who studies the effect of the Soviet regional policies, including the system of GULAG prisons and labour camps on development of urban settlements in Russia, finds that unlike Japanese and German urban systems, the USSR urban system has experienced a large and persistent shock, which still lasts in one-third of the urban settlements.

In what follows, I first review socialist urban policies and discuss factors that can explain why, unlike in Japan and Germany, urban systems in the post-socialist countries are more prone to multiplicity of equilibria. Second, I look at the effect of housing and land market rigidities and inefficient local governance on the spatial equilibrium.

Socialist urban policies and their impact on spatial equilibrium of cities This section discusses socialist urban policies and their impact on migration and distribution of city sizes. I mostly focus on urban development policies in the Soviet Union, bearing in mind that other countries of the region had similar, although less distortive, policies, taking into account local conditions and shorter period of time under the socialism. Ofer (1976) argues that the central planners intentionally checked rural-to-urban migration and industrial development of large cities and engaged in input substitution policy by keeping the capital-to-labour ratio in urban areas above the level of the market economies and by keeping the capital-to-labour ratio in rural areas below the level of the market economies. They did so to economize on costs that are incurred in urbanization when a migrant moved from a rural to urban area due to higher wages and higher consumption levels of the urban dwellers. All saved resources were further reinvested into the heavy industry production, a development strategy consistent with the idea that the capital accumulation is the major factor leading to the accelerated economic growth.

Migration restrictions in the Soviet Union worked through the system of internal passports and through the residence authorization system (so-called 'propiska'). (9) Both an internal passport and authorization stamp were legally required to get a job in a number of large metropolitan areas, with the strictest enforcement in Moscow, St. Petersburg, and regional capitals. The institution of 'propiska' has created a dualistic structure of population in the restricted cities where the population split into legal and illegal residents with discriminated access to public goods, local amenities, state-provided housing, and jobs. Gang and Stuart (1999) estimate the effectiveness of the migration restrictions by looking at the differences of urban population growth rates between restricted and unrestricted cities in Russia and find that the restricted cities grew approximately at a twice lower rate relative to the unrestricted cities in all decades between 1960s and 1980s. Clayton and Richardson (1989) further find the evidence that the restrictions were more strictly enforced in larger cities, which led to below natural rate of growth in the prime cities and consequent deviation from the Zipf's law. (10) The migration controls created the disequilibrium in the spatial distribution of the population (Gang and Stuart, 1999; Iyer, 2003), which currently slowly corrects itself by the above the average migration flows to the largest cities.

The development and growth in the unrestricted cities--by and large small-to medium-sized cities with population in a range 200,000-500,000 people, often in remote and underdeveloped areas--was promoted through channelling investments in industrial and infrastructure development of those cities. The Soviet urban policy also promoted an eastward migration of population to Ural and Siberia regions. According to Bank (2009), the share of GDP produced by the Eastern regions of the Soviet Union had increased from 4% in 1925 to 28% in the 1980s. The incentives to relocate to cold, distant areas worked through the system of the Northern compensations, preferential system of distribution of housing, creation of urban infrastructure in new locations. Millions of people were subsidized to live in 'cold' (Hill and Gaddy, 2003).

After removal of most of the benefits, the population responded by out-migration from cold areas. More than a million people left the Northern regions since 1990. The population of Magadan and Chukotka, two of the coldest places in Russia, declined by 53% and 66%, respectively, between 1991 and 2001. Still, the self-correction of distortions created by the central planners is slow and a considerable part of population in modern Russia still lives in cities that would have never been built in a market economy.

Mobility in transition

Since the transition started, transition countries have considerably reduced use of distortive urban policies, but the convergence to the market-based equilibrium remains slow, primarily due to low mobility of population. Deichmann and Henderson (2000) find that in Poland the largest cities' growth is slower than would be expected under freely operating post-transition adjustments. They link it primarily to low internal mobility, with rural-to-urban migration declining significantly between 1986 and 1998. Rautio and Tykkylainen (2008) find that '... on average Russians change their place of residence 1.5 times during their lifetime compared to 15 times in the U.S. and 7 times in Britain' (p. 61) Even more disturbing, Andrienko and Guriev (2004), who study intra-Russian migration rates in 1992-1999, find that internal mobility has been declining. The decline in mobility in Russia cannot be attributed to income equalization across regions. To the contrary, Mitra and Yemtsov (2006) report: 'As opposed to relatively stable sectoral and inter-industry wage differentials, regional variation in real wages, relative to the national average, almost tripled in Russia between 1995 and 2003. Segmentation of labour markets is a common feature of many transition economies, but in Russia this dispersion takes particularly extreme forms due to institutional, infrastructure and geographical realities'. Yemtsov (2005), using official per capita income data series, shows that between-regional factors among Russia's 80-plus regions accounted for about a third of the overall inequality in that country by the year 2000, with the increase in the between-regions component being the key driver of the change in inequality between 1995 and 2000. To put between-regional inequality in international perspective, Fedorov (2002) computed that the Gini coefficient of intra-regional inequality in Russia in 1999 was 0.29 compared with the intra-state Gini coefficient in US around 0.10 (Milanovic, 2005).

Other transition countries follow similar migration trends. Bank (2009) reports that migration flows in Eastern Europe and CIS countries have slowed down despite increasing differences in income levels and quality of life. Internal migration in the Czech Republic, Poland, and Slovakia is 0.5% of working population, which is low by the EU standards: it is three times lower than in Germany and five times lower than in France, the Netherlands, and the UK.

The low and declining mobility can be, at least partially, explained by rigidities in still over-regulated and inefficient markets for housing, local government services, utilities, and transportation (Buckley and Mini, 2000). Coricelli and Hagemeyer (1995) estimate that up to 20% of unemployment in Poland is due to housing market rigidities. There was a dramatic drop in housing completion in Poland from 140,000 in 1992 to 60,000 in 1996. More recent analysis of internal migration in Poland by Ghatak et al. (2008) confirms that the shortage of housing remains one of the most important barriers to migration within Poland, which slows down growth of large cities in Poland and lowers productivity and economic growth.

The situation with low mobility in transition countries is worsened by the government policies that preserve the current status quo through unemployment benefits and direct subsidies to depressed regions, job protection regulations, and nationwide minimum wage laws. Transfer of housing ownership from state to residents at low or no cost during the early stages of transition and no tax on land or real estate property in some countries (ie Ukraine) makes the housing and land markets more rigid and less liquid, which further lowers the labour force mobility. Under-provision of local public services and utilities is reflected in the structure of household expenditures. While in market economies the share of those services lays within 43%-63% range, in transition countries it is only 23%-30%.

Rigidities in housing and land markets

Absence of liquid land and housing markets in the centrally planned economies, due to absence of private property on land and housing and lack of institutions supporting transactions in those markets, led to inefficient land use and inadequate housing stock. Evidence on the existence of the housing market in the socialist countries are mixed. Alexeev (1988) finds that housing conditions in the Soviet Union were sensitive to income. This result is considered by the author as an indication that the Soviet households were able to beat the system of non-market distribution of housing: even though the primary allocation of housing was economically inefficient, based on needs and merits system, the secondary housing market allocated housing efficiently, based on such economic characteristics as household income. However, Buckley and Gurenko (1998) using richer data find no effect of income on distribution of housing, which indicates lack of market forces in the allocation of housing under the socialism. Low quality of housing, standardization of demand, and direct restrictions on demand for housing in socialist cities, further distorted the housing market in the socialist countries. (11)

After years of reforms, the housing market liberalized considerably. However, the Soviet legacy of existing stock of housing, peculiarities of housing demand, and monopolization of construction industry create considerable distortions to the market structure that is far from competitive. Becker and Hemley (1998) report a negative impact of housing restrictions on poor demographic situation in Russia. They have estimated that a 15-squared metre increase in living space of the household would lead to an extra birth. The lack of office space in the highly populated areas is the direct consequence of the prevalence of housing units built without space for commercial use. Lack of reforms, poorly defined property rights, and lack of the digitized information on property lead to substantial transaction costs and poor investment climate that hinders investments in real estate.

'Socialist planners made investment and location decisions under a system in which land had no value, capital had no interest opportunity cost, and energy prices were a tiny fraction of loan prices. Since enterprises could not capture any gain from redevelopment or conversion of land to highest and best use, socialist cities often had a pattern of sprawling industrial plants, often using what would be the highest value and highest density office and residential land use under any kind or market system' (Malpezzi, 1999). The price mechanism in the market economy exerts a powerful influence on land-use, replacing inefficient and obsolete economic activities by more efficient and modern. An increase in prices of land in the inner part of city drives out the inefficient businesses and obsolete structures and increases job and population densities in those areas. Under the administrative-command economy, the absence of the land price mechanism eliminates incentives to re-develop. Administrators that do not act as land-use value maximizers respond to the changes in demand for land-use by developing construction-free areas at the outskirts of city because it minimizes construction costs.

The absence of land markets in socialists cities, according to Bertaud and Renaud (1997), lead to: positive population density gradient when the most population-dense areas are located in the outer areas of the city but remoteness of housing is not compensated by better amenities such as larger houses, better environment that are typical for capitalist cities; larger share of city area is allocated to land-intensive industrial use, often occupied by obsolescent industries located in prime areas of the city--build-up land used by industries occupies 31% of all land in Moscow compared with 5% of all land in Paris, 6% of all land in Seoul, and 5% of all land in Hong Kong (Figure 2); residential areas are concentrated in the periphery, which put additional stress on transport infrastructure.

Bertaud and Renaud (1997) compare Paris and Moscow, cities of the similar size, in terms of their land-use and population densities and report that the median distance to the centre is 7 kilometres for Paris and 10 kilometres for Moscow due to higher population density in the inner part of Paris and higher population density in the outer parts of Moscow. Longer commuting distances and concentration of the households in the periphery in Moscow, while jobs are mostly located in the inner areas of Moscow, require more investments in transport infrastructure and creates higher congestion and greater share of labour time waste. High share of land occupied by land-intensive and outdated industries in Moscow means underused land near the city centre, fragmented access due to dense network of rail-roads required to serve the industrial zones, crowding-out of new, technology-intensive industries and services to the outer parts of the city.

Local governance

Development of institutions and efficient governance in transition economies also proved to be a major challenge for reformers. The main challenge of public policy reforms in transition countries is to outline the clear boundaries that separate responsibilities among various branches of government both vertically, between the state and local governments, towards a larger independence, transparency, and accountability, as well as horizontally, between various local jurisdictions to prevent the overlapping responsibilities and resolve a potential conflict of interest between urban and regional administrations that may emerge due to urban sprawl (Buckley and Mini, 2000).

A low accountability and transparency of local governments is another important challenge to local economic development. The local administrations are often appointed by the state or by local legislative branches of the government, which reduces their local accountability.

Buckley and Mini further discuss progress of transition countries in various dimensions of the local government reform--local institution-building, safety net development, and private sector development--which is measured on a scale from 1 to 4 and presented in Table 5. Central and Eastern European and the Baltic countries lead in all dimensions of the local governance reforms, Central and South Balkans have reached considerable progress in several dimensions, including effective real estate market, while the FSU countries considerably lag behind in all dimensions of the local governance reforms.

There are not many studies that analyse the factors determining the effective local governance in transition countries. As an exception, a recent study by Stastna and Gregor (2011) examines the extent of cost inefficiency of local governments in a sample of 202 municipalities of extended scope in the Czech Republic in 2003-2008. The exogenous variables that robustly increase inefficiency are population size and distance to the regional centre. Increase in party concentration and the voters' involvement increases efficiency; local council with a lower share of left-wing representatives also tend to be more efficient. A comparative analysis, conducted for the period 1994-1996, reveals that small municipalities improve efficiency significantly more than large municipalities. As a result, initially low differences in inefficiency between medium-sized and large municipalities have magnified over time. More inefficient local governments in large municipalities can be another factor that explains the smaller size of prime cities in transition countries.

Impact of housing and governance on mobility

In order to assess the impact of market rigidities on mobility within the last decade, I use the LITS II conducted by the European Bank for Reconstruction and Development and the World Bank in 2010. It surveyed almost 39,000 households in 34 countries, including all transition countries. (12) As a reference group, households in France, Germany, Italy, the UK, and Sweden were surveyed, which gives a much needed basis for comparison. I keep the observations on households located in urban and metropolitan areas and aggregate the data to the level of the primary sampling unit (PSU). Panel A of Table 6 reports summary statistics of the key characteristics of urban areas and households located in those areas across countries presented in the survey, grouped in larger geographical areas--Western Europe, Central and Eastern Europe, Southern Europe, and CIS and Mongolia. There is a striking contrast in mobility between the Western Europe and transition countries. While 25.3% of the households in the Western Europe moved into the urban area within the last 10 years, only 8.8%-12.6% of the households in transition countries moved within the last 10 years. Panel B of Table 6 reports housing market conditions within PSUs across the same groups of countries. The Western Europe has the highest shares of households that rent and households with mortgages. In the Western Europe, 41.47% of households currently have a mortgage, while in CIS and Mongolia only 1.78% of households have a mortgage. The most common way of acquiring a dwelling in the Western Europe is to purchase with mortgage (61%), while in the CIS and Mongolia region the most common ways are either to purchase without mortgage (35%) or to privatize for free (33%).

To analyse the determinants of mobility, I use a linear model

[mobility.sub.i,T] = [alpha] + [H.sub.i][beta] + [O.sub.i][gamma] + [R.sub.i][delta] + [G.sub.i][pi] + [D.sub.c] + [[epsilon].sub.i], (3)

where [mobility.sub.i,T] is the share of respondents that moved into the urban area i within the last T years; H is a vector of characteristics of the housing stock in the area i; O is a vector of housing market conditions (share of rented housing and share of households with mortgage); R is a vector of labour demand and supply characteristics of the urban area (average education level, unemployment level, share of self-employed, share of employed by state enterprises), G is a measure of local government quality; [D.sub.c] denotes a set of country fixed effects; and [epsilon].sub.i] is an independently distributed error term. The model is estimated by Tobit, because the dependent variable is censored and takes values between 0 and 1. (13)

Table 7 reports the estimates of model (3). Column (1) reports the estimates with the share of households that moved in the last 10 years and does not include country fixed effects. The specification allows using cross-country variation to identify the effect of variables that do not vary much within a country (structure of the housing stock, education, and quality of local governance) and can be interpreted as the long-run effect estimates. Column (2), the baseline specification, adds country fixed effects, to control for cross-country differences in the level of development, education, institutions, and particularities of the housing stock. Columns (3)-(4) present the baseline specification for mobility measured with T equals to 5 and 15 years consequently. In columns (5) and (6), I split the sample into the households living in urban and metropolitan areas consequently.

Two variables that measure how well the housing market is functioning in the urban area--the share of rented housing and share of households with mortgages--are consistently positive and significant across all model specifications. On the basis of the estimates of the baseline model specification in column (2) of the table, a standard deviation increase in the share of rented housing in a city in transition countries is associated with a 50% closing the gap in mobility between Western Europe and transition countries, while a standard deviation increase in the share of mortgaged dwellings is associated with a 25% closing the gap in mobility. The structure of the housing stock is rarely a significant determinant of mobility. Average approval of the local government, on the other hand, is never significant. However, the variable is a subjective measure and the reference point for what does good local government mean may considerably differ across countries. The economic size of the effect of the housing market is increasing when we look at the longer horizon measures of mobility. However, this result should be taken with caution because the reverse causality problem becomes more severe once we look at longer periods.

The LITS survey is representative at the country level, but not at the PSU level. To address the issue of a measurement error, I report the baseline specification results for more aggregated data--at the regional level (roughly equivalent to Nomenclature of territorial units for statistics (NUTS) 2 digit according to the European classification of regions) in column (7) and at the national level in column (8). The share of mortgages retains its positive sign but loses its significance, perhaps due to the small sample size. The share of rented housing, on the other hand, is still significant. Also, the differences in educational level across countries start playing a role in explaining mobility, which is consistent with the results at the PSU level without country fixed effects in column (1).

On the basis of the results, I conclude that there is a strong and robust positive correlation between population mobility and availability of rented housing. Another robust factor that is associated with high mobility is the share of households with mortgages. As a policy implication, stability of macroeconomic environment and development of the financial system might increase population mobility through better provision of affordable mortgages. Of course, these findings should be taken with care due to endogeneity and measurement issues, which are not fully resolved in this analysis, but as the first step it gives a valuable insight into the causes of low population mobility in transition countries.


Agglomeration and localization economies

Underdevelopment of prime cities in transition countries prevents them from capitalizing on agglomeration effects. Marshall (1890) identifies three main micro-foundations of agglomeration economies: labour market pooling (better match and reduced risk), knowledge spillovers (localized learning), and input sharing (internal increasing returns to scale). In addition, the natural advantage, home market effect, consumption opportunities, and rent-seeking all can contribute to agglomeration (Rosenthal and Strange, 2004). Smaller agglomerations have less productive firms, fewer innovative activities, fewer opportunities for human capital development, and less efficient labour markets. On the basis of the literature review summarized in Table 8, Rosenthal and Strange (2004) conclude that doubling of a city size increases productivity of firms located in city by 3%-8%. Separation of localization (within an industry agglomeration) and urbanization (impact of city size across all industries) effects, carried out by Nakamura (1985) for Japan, reveals that doubling of an industry scale leads to a 4.5% increase in productivity, while doubling of a city population leads to a 3.4% increase in productivity. Ciccone and Hall (1996) find a positive effect of population density on productivity in the US--doubling of population density increases productivity by 6%. Ciccone (2002) further finds that the effect is 4.5% for a cross-section of regions in France, Germany, Italy, Spain, and the UK. Moretti (2004) finds that a percentage point increase in the share of college students in a city raises average wages by 0.6%-1.2%, above and beyond the private returns to education.

A meta-analysis of the literature on the relationship between urbanization and productivity by Melo et al. (2009) reveals that the effect is region- and country-specific, with China, Japan, and Sweden having lower returns and the US, France, and Italy having higher returns. Au and Henderson (2006) argue that restrictions on rural to urban migration in China explain insufficient agglomeration of economic activity. Also, the urban agglomeration impact on productivity is stronger in the services sector, which is consistent with the notion that services strongly benefit from proximity to large urban markets.

Research on agglomeration economies in cities of transition countries is scarce. The estimated agglomeration effects are found to be stronger than for the OECD countries. Bekes and Harasztosi (2010), looking at Hungarian manufacturing data from 1992 to 2003, find that firms that are engaged in international trade would gain 16% in total factor productivity as the city size doubles, a number that is twice as large as the upper bound for the consensus estimate presented by Rosenthal and Strange (2004). Bruhart and Mathys (2008) find that the impact of population density on labour productivity in Europe in 1980-2003 has been constantly growing over time, mainly due to higher impact of density on productivity in the Eastern European regions. (14) Vakhitov (2010), looking at the Ukrainian firm level data in 2001-2005, confirms that the agglomeration effect is higher in transition countries. While the higher agglomeration effect in transition countries is expected, due to the dynamic nature of transition from the command economy to the market economy, these results should be taken with care because of noisier data, effect of restructuring that is hard to separate from the effect of urbanization, and higher inflation rates. Still, the results indicate that further growth of the prime cities in the region would generate substantial and positive agglomeration externalities.

Poverty and social instability in urban areas

Declining industrial production in transition countries during the 1990s hits urban population particularly hard, leading to a high incidence of poverty in urban areas. In Moldova, Armenia, Azerbaijan, and Georgia the share of poor in the urban areas is higher than in the rural areas--the fact rarely observed in other developing countries. According to Alam et al. (2005), factors contributing to the high incidence of urban poverty in transition countries are reduction of subsidies to urban infrastructure, several-fold increase in the share of housing and utilities in total expenditures, overall deterioration of urban infrastructure due to poor maintenance, and unequal access of urban population to the quality services and utilities. The secondary cities of the region have even higher incidence of poverty--risk of being poor in the secondary city is two to four times higher relative to the prime cities.

Main factors contributing to probability of being poor at the micro level are a low level of education of the head of the household and large family size. Urban population is also more vulnerable to the macroeconomic shocks--urban poverty in Russia, Moldova, Armenia, and Georgia increased sharply during the economic crisis of 1998.

In addition to the economic problems, the Soviet urban policies of encouraging labour migration of Russian speaking population to the national republics, the Baltic States in particular, have created social tension after the breakdown of the Soviet Union. Balockaite (2010) discusses social and psychological problems the Russian speaking population of Visaginas, Lithuania is facing even after 20 years of the transition period. Lost in transition, previously considered as the elite of the Soviet working class, the workers of the nuclear power station are struggling to find their new identity.


Transition countries have reached high levels of urbanization, comparable with the levels of urbanization in other countries at the similar level of development. However, the evidence indicates that the largest cities of the region are relatively small and probably continue to grow at higher rates to reach a scale consistent with the market-based spatial equilibrium. In 1991 the share of the largest cities' population to total population in transition countries was 14 percentage points smaller on average relative to other countries. In 2008, the difference is still significant--the share of the largest cities' population in transition countries was 9.5 percentage points smaller. In FSU countries, 30 largest cities are smaller in size than what is implied by Zipf's law. This distortion in the distribution of city sizes is very persistence over time.

This paper attributes the source of the spatial distortions to the Soviet policies, which restricted migration to the largest cities and encouraged development of medium-sized cities. Persistence in the distribution of cities, despite massive transformation that occurred in transition countries in the last 20 years, is attributed to low population mobility. Growing regional inequalities coupled with declining labour mobility in transition countries constitutes an empirical puzzle. The paper provides new evidence on the importance of the efficient housing market for high population mobility. A standard deviation increase in the share of rented housing in transition countries is associated with a 50% closing the gap in mobility between Western Europe and transition countries, while a standard deviation increase in the share of mortgaged dwellings is associated with a 25% closing the gap in mobility.

Other important determinants of urban development are properly functioning land markets and efficient local governments. More successful countries of the region in terms of the urban development also have more efficient and transparent local governments. Therefore, decentralization and further democratization of the local administrations is a priority for successful urban development in the region.

Under-populated prime cities in transition countries fail to capitalize on the agglomeration economies to a full extent. This, in turn, compromises the global competitiveness of local firms, reduces attractiveness of the region for investors, and makes the cities in transition countries less attractive to high-skilled workers who prefer more densely populated agglomerations of the US and European Union.


This research was supported by the Global Development Network (GDN) as part of the intra-regional research project on urban development in transition. I thank Tom Coupe, Randall Filer, Ira Gang, Paul Wachtel, and anonymous referees for excellent comments and suggestions.


Acemoglu, D, Johnson, S and Robinson, J. 2002: Reversal of fortune: Geography and institutions in the making of the modern world income distribution. Quarterly Journal of Economics 117(4): 1231-1294.

Ades, A and Glaeser, E. 1995: Trade and circuses: Explaining urban giants. The Quarterly Journal of Economics 110(1): 195-227.

Alam, A, Murthi, M and Yemtsov, R. 2005: Growth, poverty, and inequality: Eastern Europe and the former Soviet Union. World Bank: Washington, DC.

Alexeev, M. 1988: Market versus rationing: The case of Soviet housing. The Review of Economics and Statistics 70(3): 414-420.

Andrienko, Y and Guriev, S. 2004: Determinants of interregional mobility in Russia. Economics of Transition 12(1): 1-27.

Au, C and Henderson, J. 2006: How migration restrictions limit agglomeration and productivity in China. Journal of Development Economics 80(2): 350-388.

Balockaite, R. 2010: Post-Soviet transitions of the planned socialist towns: Visaginas, Lithuania. Studies of Transition States and Societies 2(2): 63-81.

Bank, W. 2009: World development report 2009: Reshaping economic geography. World Bank Publications: Washington DC.

Becker, C and Hemley, D. 1998: Demographic change in the former Soviet Union during the transition period. World Development 26(11): 1957-1975.

Bekes, G and Harasztosi, P. 2010: Agglomeration premium and trading activity of firms CeFiG Working Papers.

Bertaud, A and Renaud, B. 1997: Socialist cities without land markets. Journal of Urban Economics 41(1): 137-151.

Bosker, M, Brakman, S, Garretsen, H and Schramm, M. 2007: Looking for multiple equilibria when geography matters: German city growth and the WWII shock. Journal of Urban Economics 61(1): 152-169.

Bruhart, M and Mathys, N. 2008: Sectoral agglomeration economies in a panel of European regions. Regional Science and Urban Economics 38(4): 348-362.

Buckley, R and Gurenko, E. 1998: Housing demand in Russia: Rationing and reform. Economics of Transition: 197-210.

Buckley, R and Mini, F. 2000: From commissars to mayors: Cities in the transition economies. World Bank Publications: Washington DC.

Ciccone, A. 2002: Agglomeration effects in Europe. European Economic Review 46(2): 213-227.

Ciccone, A and Hall, R. 1996: Productivity and the density of economic activity. The American Economic Review 86(1): 54-70.

Clayton, E and Richardson, T. 1989: Soviet control of city size. Economic Development and Cultural Change 38(1): 155-165.

Coricelli, FKR and Hagemeyer, K. 1995: Poland. In Commander, S, and Coricelli, FKR (eds). Unemployment Restructuring and the Labor Market in Eastern Europe and Russia. World Bank EDI Development Studies: Washington DC.

Costa, D and Kahn, M. 2000: Power couples: Changes in the locational choice of the college educated, 1940-1990. Quarterly Journal of Economics 115(4): 1287-1315.

Davis, D and Weinstein, D. 2002: Bones, bombs, and break points: The geography of economic activity. The American Economic Review 92(5): 1269-1289.

Davis, J and Henderson, J. 2003: Evidence on the political economy of the urbanization process. Journal of Urban Economics 53(1): 98-125.

Deichmann, U and Henderson, V. 2000: Title: Urban and regional dynamics in Poland World Bank Policy Research Working Paper No. 2457.

Diamond, C and Simon, C. 1990: Industrial specialization and the returns to labor. Journal of Labor Economics 8 (2): 175-201.

Duranton, G and Puga, D. 2001: Nursery cities: Urban diversity, process innovation, and the life cycle of products. The American Economic Review 91 (5): 1454-1477.

Eaton, J and Eckstein, Z. 1997: Cities and growth: Theory and evidence from France and Japan. Regional Science and Urban Economics 27(4): 443-474.

Fedorov, L. 2002: Regional inequality and regional polarization in Russia, 1990-99. World Development 30(3): 443-456.

Gabaix, X. 1999: Zipf's law for cities: An explanation. Quarterly Journal of Economics 114(3): 739-767.

Gang, I and Stuart, R. 1999: Mobility where mobility is illegal: Internal migration and city growth in the Soviet Union. Journal of Population Economics 12(1): 117-134.

Ghatak, S, Mulhern, A and Watson, J. 2008: Inter-regional migration in transition economies: The case of Poland. Review of Development Economics 12 (1): 209.

Glaeser, E and Gottlieb, J. 2009: The wealth of cities: Agglomeration economies and spatial equilibrium in the United States. Journal of Economic Literature 47(4): 983-1028.

Henderson, J. 2005: Urbanization and growth. Handbook of Economic Growth 1(1): 1543-1591.

Henderson, J, Nijkamp, P, Thisse, J, Mills, E and Cheshire, P. 2004: Handbook of regional and urban economics. North Holland, Amsterdam.

Henderson, V. 2003: The urbanization process and economic growth: The so-what question. Journal of Economic Growth 8(1): 47-71.

Hill, F and Caddy, C. 2003: The Siberian curse: How communist planners left Russia out in the cold. Brookings Inst Pr: Washington DC.

Holmes, T. 1999: Localization of industry and vertical disintegration. Review of Economics and Statistics 81(2): 314-325.

Ioannides, Y and Overman, H. 2003: Zipfs law for cities: An empirical examination. Regional Science and Urban Economics 33(2): 127-137.

Iyer, S. 2003: Increasing unevenness in the distribution of city sizes in post-Soviet Russia. Eurasian Geography and Economics 44(5): 348-367.

Jaffe, A, Trajtenberg, M and Henderson, R. 1993: Geographic localization of knowledge spillovers as evidenced by patent citations. The Quarterly Journal of Economics 108(3): 577-598.

Malpezzi, S. 1999: Economic analysis of housing markets in developing and transition economies. Handbook of Regional and Urban Economics lII: 1791-1864.

Marshall, A. 1890: Principles of economics. Macmillan: London.

Melo, P, Craham, D and Noland, R. 2009: A meta-analysis of estimates of urban agglomeration economies. Regional Science and Urban Economics 39(3): 332-342.

Mikhailova, T. 2010: Looking for multiple equilibria in Russian urban system, mimeo, unpublished manuscript.

Milanovic, B. 2005: Half a world: Regional inequality in five great federations. Journal of the Asia Pacific Economy 10(4): 408-445.

Mitra, P and Yemtsov, R. 2006: Increasing inequality in transition economies: Is there more to come? World Bank: Washington DC.

Moretti, E. 2004: Estimating the social return to higher education: evidence from longitudinal and repeated cross-sectional data. Journal of Econometrics 121 (1-2): 175-212.

Nakamura, R. 1985: Agglomeration economies in urban manufacturing industries: A case of Japanese cities. Journal of Urban Economies 17(1): 108-124.

Ofer, G. 1976: Industrial structure, urbanization, and the growth strategy of socialist countries. The Quarterly Journal of Economics 90 (2): 219-244.

Rautio, V and Tykkylainen, M. 2008: Russia's northern regions on the edge: Communities, industries and population from Murmansk to Magadan. University of Helsinki, Aleksanteri Institute, Kikimora Publications: Helsinki, Finland.

Roseuthal, S and Strange, W. 2004: Evidence on the nature and sources of agglomeration economies. Handbook of Regional and Urban Economics 4: 2119-2171.

Stastna, L and Eregor, M. 2011: Local government efficiency: Evidence from the Czech municipalities. Working papers IES 2011/14,

Vakhitov, V. 2010: Are there urbanization economies in a post-socialist city? Evidence from Ukrainian firm-level data. mimeo, unpublished manuscript.

Yemtsov, R. 2005: Quo vadis? Inequality and poverty dynamics across Russian regions. Oxford University Press: USA.


Kyiv School of Economics and Kyiv Economics Institute, Yakira 13, suite 327, Kyiv 0411g, Ukraine.


(1) See for example Alexeev (1988), Clayton and Richardson (1989), Bertaud and Renaud (1997), Buckley and Gurenko (1998), Gang and Stuart (1999), Buckley and Mini (2000).

(2) The survey has been conducted by the European Bank of Reconstruction and Development (EBRD). The report and data are available from

(3) While it is not possible to compare directly the degree of urbanization for some countries that no longer exist, such as Czechoslovakia, the Soviet Union, and Yugoslavia, the table reports the population weighted averages of actual urbanization, estimated urbanization, and residuals for countries that were former members of the above-mentioned countries.

(4) See for example Acemoglu et al. (2002) for a discussion on links between GDP per capita and urbanization.

(5) Davis and Henderson (2003) also find that the prime cities in the centrally planned economies are under-urbanized by 5 % -11% depending on the regression specification.

(6) For Hungary and the Czech Republic the earlier samples are for 1980, for Poland the earlier sample is for 1981.

(7) 2007 data are available at For Georgia, the sample is for 2008. For Moldova, the sample is for 2004. For the Kyrgyz Republic, the sample is for 2009. For Uzbekistan, the sample is for 2006.

(8) A system of cities exhibits multiplicity of equilibria if a shock leads to irreversible changes in location of cities or to changes in distribution of city sizes.

(9) Rural residents did not have passports until 1974 and could not officially migrate to urban areas.

(10) In the estimation of the rank-size regression for S00 Soviet cities, 10 of the 11 largest cities are major outliers--their actual sizes are well below the predicted sizes,

(11) A couple with minor children could have owned only one dwelling, with not more than 60 square metres of living space (Alexeev, 1988).

(12) The LITS II is available online at

(13) I also estimated the model by the ordinary least squares, which does not require the assumption of normality of the error term. The results are similar and available upon request.

(14) Inclusion of the Eastern European regions rises the impact from 4% to 13%.
Table 1: Urbanization in transition countries relative to OECD
countries in 1960 and 2008

Country            Urbanization rate in 1960, %

                  Actual   Estimated   Difference

Bulgaria            38        50          -12
Czechoslovakia      57        62           -5
East Germany        72        61           11
Soviet Union        48        57           -9
Hungary             40        55          -15
Poland              47        53           -6
Romania             35        50          -15
Yugoslavia          28        48          -20

Overall             47        56           -9

Country            Urbanization rate in 2008, %

                  Actual   Estimated   Difference

Albania             47        56           -9
Bulgaria            71        63            8
Czech Republic      73        73            0
Slovak Republic     57        71          -14

Average             68        72           -4

Armenia             64        55            9
Azerbaijan          52        60           -8
Belarus             73        64            9
Estonia             69        69            0
Georgia             53        53            0
Kazakhstan          58        63           -5
Kyrgyzstan          36        43           -7
Lithuania           67        68           -1
Latvia              68        67            1
Moldova             42        45           -3
Russia              73        67            6
Turkmenistan        49        57           -8
Tajikistan          27        41          -14
Ukraine             68        58           10
Uzbekistan          37        43           -6

Average             64        61            3

Hungary             67        69           -2
Poland              61        67           -6
Romania             54        65          -11
Bosnia              47        57          -10
Croatia             57        67          -10
Macedonia           67        58            9
Serbia              52        60           -8
Slovenia            48        74          -26
Average             53        62           -9

Overall             63        62            1

Note: Data on urbanization rates for 2008 are from the urban
development database by the World Bank.

Real GDP per capita is from the Penn World Tables 7.0. The OLS
regression is estimated on a sample of OECD countries and then
predicted for the former socialist countries based on the following
regression models.

Year 1960: UR = -35.71 + 13.93 x ln(Y) Adjusted [R.sup.2] = 04.25
(Ofer, 1976).

Year 2008: UR = -53.11 + 12.44 x In(Y) Adjusted [R.sup.2] = 0.25
(Author's calculations).

Table 2: Urbanization rates

                                      (1)         (2)         (3)

                                     1991        1998         2008

Transition                          0.024       0.042 *    -0.030
                                   (0.026)     (0.023)     (0.022)

Log GDP per capita                  0.15 ***    0.14 ***    0.12 ***
                                   (0.010)     (0.010)     (0.012)

Log population                      0.0017      0.0037      0.0037
                                   (0.010)     (0.010)     (0.011)

Log country area                    0.010       0.011       0.013
                                   (0.0088)    (0.0091)    (0.0094)

Land locked                        -0.055 **   -0.060 **   -0.076 ***
                                   (0.024)     (0.024)     (0.024)

Openness to trade                   0.028       0.036       0.042 *
                                   (0.033)     (0.033)     (0.025)

Voice and accountability           -0.018      -0.017      -0.0055
                                   (0.014)     (0.013)     (0.014)

Agriculture to GDP ratio

Services to industry ratio

Five-year Lag Log GDP per capita

Constant                           -0.93 ***   -0.84 ***   -0.70 ***
                                   (0.11)      (0.11)      (0.12)

Adjusted [R.sup.2]                  0.644       0.593       0.529

Overall [R.sup.2]

Observations                          170         183         184

                                      (4)          (5)         (6)

                                                   Ec.       Lag GDP
                                   1991-2008    structure   per capita

Transition                          0.043       -0.010      -0.0080
                                   (0.034)      (0.024)     (0.022)

Log GDP per capita                  0.027 ***    0.10 ***
                                   (0.0088)     (0.023)

Log population                      0.12 ***    -0.0031      0.0029
                                   (0.013)      (0.013)     (0.011)

Log country area                   -0.067 ***    0.024 *     0.014
                                   (0.010)      (0.013)     (0.0096)

Land locked                        -0.092 **    -0.046      -0.075 ***
                                   (0.043)      (0.029)     (0.024)

Openness to trade                   0.0039       0.011       0.041
                                   (0.0044)     (0.036)     (0.025)

Voice and accountability            0.074 ***    0.0031     -0.013
                                   (0.018)      (0.019)     (0.014)

Agriculture to GDP ratio                        -0.20

Services to industry ratio                       0.012

Five-year Lag Log GDP per capita                             0.13 ***

Constant                            0.067       -0.61 **    -0.72 ***
                                   (0.11)       (0.25)      (0.12)

Adjusted [R.sup.2]                               0.479       0.540

Overall [R.sup.2]                   0.18

Observations                          3457         150         183

                                      (7)           (8)

                                                 Poor and
                                   Non-OECD    middle income

Transition                         -0.035       0.0025
                                   (0.026)     (0.026)

Log GDP per capita                  0.12 ***    0.11 ***
                                   (0.014)     (0.017

Log population                      0.0055     -0.023
                                   (0.014)     (0.015)

Log country area                    0.013       0.034 ***
                                   (0.011)     (0.011)

Land locked                        -0.076 **   -0.11 ***
                                   (0.031)     (0.030)

Openness to trade                   0.040      -0.0058
                                   (0.028)     (0.040)

Voice and accountability           -0.0049     -0.012
                                   (0.018)     (0.018)

Agriculture to GDP ratio

Services to industry ratio

Five-year Lag Log GDP per capita

Constant                           -0.72 ***   -0.60 ***
                                   (0.14)      (0.15)

Adjusted [R.sup.2]                  0.464       0.435

Overall [R.sup.2]

Observations                          157         134

* p<0.1, ** p<0.05, *** p<0.01.

Robust standard errors in parentheses.

Notes: Regressions of the urbanization rates on the level of
development, population, openness to trade, institutions, and
geographical characteristics. Data sources: urbanization rates are
from the urban development data provided by the World Bank; GDP per
capita, population, and openness to trade are from the Penn World
Tables 7.0; area, landlocked are from the CEPII Geo Data; voice and
accountability is from the worldwide governance indicators (WGI)
provided by the World Bank; share of agriculture in GDP and ratio of
services to industry are from the world development indicators (WDI)
provided by the World Bank.

Transition dummy is equal to one for the transition countries and
zero otherwise.

Table 3: Primacy

                                      (1)          (2)          (3)

                                      1991         1998         2008

Transition                         -0.14 ***    -0.11 ***    -0.095 ***
                                   (0.031)      (0.025)      (0.029)

Capital                             0.12 ***     0.11 ***     0.10 ***
                                   (0.029)      (0.027)      (0.026)

Log GDP per capita                 -0.023       -0.025 *     -0.022
                                   (0.016)      (0.014)      (0.014)

Log population                     -0.043 ***   -0.044 ***   -0.044 ***
                                   (0.014)      (0.013)      (0.012)

Log country area                   -0.018       -0.018 *     -0.017
                                   (0.011)      (0.011)      (0.011)

Land locked                        -0.040       -0.029       -0.022
                                   (0.034)      (0.028)      (0.029)

Openness to trade                   0.062        0.053        0.050
                                   (0.048)      (0.057)      (0.042)

Voice and accountability           -0.013       -0.0093      -0.0078
                                   (0.019)      (0.018)      (0.017)

Agriculture to GDP ratio

Services to industry ratio

Five-year lag log GDP per capita

Constant                            1.03 ***     1.07 ***     1.03 ***
                                   (0.20)       (0.17)       (0.16)

Adjusted [R.sup.2]                  0.476        0.479        0.444

Overall [R.sup.2]

Observations                          165          178          179

                                      (4)          (5)          (6)

                                                   Ec.        Lag GDP
                                   1991-2008    structure    per capita

Transition                         -0.11 ***    -0.064 **    -0.098 ***
                                   (0.022)      (0.026)      (0.029)

Capital                             0.095 ***    0.070 ***    0.10 ***
                                   (0.027)      (0.023)      (0.026)

Log GDP per capita                  0.0049      -0.013
                                   (0.0089)     (0.018)

Log population                     -0.074 ***   -0.064 ***   -0.044 ***
                                   (0.020)      (0.012)      (0.012)

Log country area                   -0.0034      -0.011       -0.017
                                   (0.014)      (0.013)      (0.011)

Land locked                        -0.013       -0.037       -0.024
                                   (0.027)      (0.026)      (0.029)

Openness to trade                   0.0059      -0.044        0.052
                                   (0.010)      (0.034)      (0.042)

Voice and accountability           -0.034 **    -0.025       -0.0049
                                   (0.015)      (0.022)      (0.018)

Agriculture to GDP ratio                         0.095

Services to industry ratio                       0.0019

Five-year lag log GDP per capita                             -0.025 *

Constant                            0.95 ***     1.14 ***     1.05 ***
                                   (0.13)       (0.23)       (0.16)

Adjusted [R.sup.2]                               0.506        0.446

Overall [R.sup.2]                   0.46

Observations                          3359         145          178

                                      (7)            (8)

                                                  Poor and
                                    Non-OECD    middle income

Transition                         -0.11 ***    -0.072 **
                                   (0.030)      (0.032)

Capital                             0.12 ***     0.096 ***
                                   (0.030)      (0.032)

Log GDP per capita                 -0.019       -0.020
                                   (0.016)      (0.015)

Log population                     -0.036 ***   -0.063 ***
                                   (0.013)      (0.012)

Log country area                   -0.020 *     -0.0059
                                   (0.012)      (0.012)

Land locked                         0.019       -0.0081
                                   (0.032)      (0.030)

Openness to trade                   0.083 **     0.0020
                                   (0.035)      (0.035)

Voice and accountability            0.0031      -0.022
                                   (0.020)      (0.019)

Agriculture to GDP ratio

Services to industry ratio

Five-year lag log GDP per capita

Constant                            0.93 ***     1.09 ***
                                   (0.17)       (0.17)

Adjusted [R.sup.2]                  0.450        0.509

Overall [R.sup.2]

Observations                          153          130

* p<0.1, ** p<0.05, *** p<0.01.

Robust standard errors in parentheses.

Notes: Regressions of the share of population living in the largest
city on the level of development, population, openness to trade,
institutions, and geographical characteristics. Data sources: primacy
rates are from the urban development database provided by the World
Bank; GDP per capita, population, and openness to trade are from the
Penn World Tables 7.0; area, landlocked are from the CEPII Geo Data;
voice and accountability is from the worldwide governance indicators
(WGI) provided by the World Bank; share of agriculture in GDP and
ratio of services to industry are from the world development
indicators (WDI) provided by the World Bank. Transition dummy is
equal to one for the transition countries and zero otherwise.

Table 4: Zipf's law for selected transition countries in 1979 and 2007

                         Czech      Hungary      Poland        FSU


Panel A: Zipf's law in 1979. Dependent variable is ln

[[beta].sub.1979]       1.119 ***   1.119 ***    1.113 ***   1.008 ***
                       (0.057)     (0.121)      (0.017)     (0.012)

Constant               15.09 ***   15.15 ***    16.34 ***   17.06 ***
                       (0.593)     (1.267)      (0.179)     (0.126)

[R.sup.2]               0.970       0.950        0.994       0.984

N                         58          63           184         1116

H0: [beta] = 1          4.375       0.977       45.11        0.424

p-value                 0.041       0.327        0.000       0.515

Panel B: Zipf's law in 2007. Dependent variable is ln

[[beta].sub.2007]       1.165 ***   1.139 ***    1.156 ***   0.986 ***
                       (0.067)     (0.112)      (0.015)     (0.012)

Constant               15.63 ***   15.32 ***    16.94 ***   16.94 ***
                       (0.701)     (1.173)      (0.157)     (0.127)

[R.sup.2]               0.972       0.953        0.994       0.981

N                         63          61           219         1154

H0: [beta] = 1          6.084       1.539      112.1         1.358

p-value                 0.016       0.220        0.000       0.244

                         Czech      Hungary      Poland        FSU


Panel A: Zipf's law in 1979. Dependent variable is ln

[[beta].sub.1979]       1.021 ***   0.932 ***    1.198 ***   1.108 ***
                       (0.079)     (0.162)      (0.023)     (0.019)

Constant               13.93 ***   12.91 ***    17.36 ***   18.29 ***
                       (0.893)     (1.825)      (0.269)     (0.220)

[R.sup.2]               0.923       0.867        0.993       0.976

N                         22          21           75          489

H0: [beta] = 1          0.073       0.177       70.81       31.95

p-value                 0.790       0.679        0.000       0.000

Panel B: Zipf's law in 2007. Dependent variable is ln

[[beta].sub.2007]       1.008 ***   0.982 ***    1.238 ***   1.083 ***
                       (0.080)     (0.163)      (0.021)     (0.018)

Constant               13.755 ***  13.464 ***   17.915 ***  18.132 ***
                       (0.901)     (1.836)      (0.242)     (0.211)

[R.sup.2]               0.930       0.874        0.993       0.972

N                         21          21           87          536

H0: [beta] = 1          0.010       0.012      125.1        20.75

p-value                 0.920       0.913        0.000       0.000

* p<0.05, ** p<0.01, *** p<0.001.

Robust standard errors are in parentheses.

Table 5: Taxonomy of progress in transition countries


                               Political      Transparency
                             accountability   and economic
                                of local        autonomy
                              governments       of local

Central and Eastern Europe        3.0             2.0
The Baltics                       3.0             2.0
Central and South Balkans         1.5             1.0
Former Soviet Union               1.0             1.0

                                   Safety net

                             Pricing of   Effective
                              services    allocation
                                          of social

Central and Eastern Europe      2.5          2.5
The Baltics                     2.5          2.5
Central and South Balkans       2.0          2.0
Former Soviet Union             1.0          1.0

                                 Private sector        Overall
                                   development         rating

                             Effective    Financial
                               real        sector
                              estate     development

Central and Eastern Europe      3.0          3.0         2.7
The Baltics                     2.5          2.5         2.5
Central and South Balkans       2.5          2.0         1.8
Former Soviet Union             1.5          1.0         1.1

Notes: The scale is from 1 to 4 where higher numbers represent better

Source: Buckley and Mini (2000), From commissars to mayors: Cities in
transition countries

Table 6: Summary statistics

                          Western  Central  Southern  CIS and   Total
                          Europe     and     Europe   Mongolia

A: Summary statistics

Share of households        0.149    0.068     0.044     0.065    0.075
moved in within last 5    (0.15)   (0.10)    (0.06)    (0.08)   (0.10)

Share of households        0.253    0.121     0.088     0.126    0.138
moved in within last 10   (0.19)   (0.14)    (0.08)    (0.13)   (0.14

Share of households        0.325    0.155     0.130     0.163    0.181
moved in within last 15   (0.20)   (0.15)    (0.11)    (0.15)   (0.17)

Share of detached          0.374    0.334     0.378     0.288    0.336
houses                    (0.32)   (0.39)    (0.38)    (0.39)   (0.38)

Share of townhouses        0.255    0.057     0.074     0.028    0.085
                          (0.30)   (0.15)    (0.15)    (0.09)   (0.19)

Share of rented housing    0.375    0.187     0.113     0.101    0.174
                          (0.25)   (0.20)    (0.11)    (0.11)   (0.19)

Share of households        0.263    0.080     0.042     0.021    0.083
with mortgage             (0.18)   (0.11)    (0.06)    (0.04)   (0.13)

In (Monthly rent) US       6.825    5.540     5.245     4.948    5.497
dollars                   (0.36)   (0.68)    (0.55)    (0.80)   (0.92)

Share of unemployed        0.090    0.121     0.168     O.148    0.135
                          (0.11)   (0.15)    (0.14)    (0.17)   (0.15)

Education level           (4.210   (4.003    (4.138    (4.856   (4.348
                          (0.82)   (0.75)    (0.66)    (0.62)   (0.79

Share of self-employed     0.120    0.111     0.127     0.172    0.136
                          (0.14)   (0.15)    (0.15)    (0.17)   (0.16)

Share of employed by       0.330    0.381     0.437     0.563    0.445
state enterprises         (0.27)   (0.29)    (0.26)    (0.28)   (0.29)

Level of satisfaction      3.327    3.134     2.844     3.052    3.074
with local government     (0.49)   (0.53)    (0.59)    (0.50)   (0.55)

Observations                181      294       244       349     1068

B: Housing conditions

                                      Type of dwelling (%)

Detached house            36.6     42.98     55.13     50.97

Townhouse                 22.45     8.09      7.63      2.78

Apartment                 40.46    48.54     36.93     43.23

                                     Type of ownership (%)

Rented                    34.61    14.67      8.02      6.39

Owned                     63.99    82.5      90.25     90.99

                                 How dwelling was acquired (%)

Privatized                15.9     16.7        10      33.02

Purchased with mortgage   60.59    14.52     15.48      3.55

Purchased without         11.27    35.48     38.3      34.86

Cooperative                0.31     5.82      1.04      1.99

Inherited                 10.02    24.87     34.03     25.23

                                     Mortgage currently (%)

Yes                       41.57     8.65      3.87      1.78

No                        58.38    91.35     96.13     98.2

Table 7: Determinants of population mobility

                                 (1)          (2)          (3)

Share of detached houses       0.057 ***     0.017        0.0026
                              (0.02)        (0.02)       (0.01)

Share of townhouses            0.0055       -0.061 *     -0.067 *
                              (0.03)        (0.03)       (0.04)

Share of rented housing        0.36 ***      0.35 ***     0.29 ***
                              (0.04)        (0.04)       (0.04)

Share of mortgages             0.32 ***      0.25 ***     0.15 ***
                              (0.05)        (0.06)       (0.05)

Unemployment                   0.013         0.0094      -0.0058
                              (0.03)        (0.03)       (0.03)

Average education level in     0.026 ***     0.0064       0.0010
locality                      (0.007)       (0.009)      (0.008)

In (Monthly rent)             -0.0034        0.0038       0.0041
                              (0.007)       (0.009)      (0.008)

Share of self-employed         0.099 ***     0.057 *      0.024
                              (0.03)        (0.03)       (0.03)

Share of employed by          -0.0020       -0.0055      -0.012
state-owned enterprises       (0.02)        (0.02)       (0.02)

Average approval of local      0.0047        0.0044       0.0030
government in locality        (0.009)       (0.009)      (0.009)

Country effect                   No           Yes          Yes

Log like.                    181.9         309.8        173.5

Observations                    1068          1068         1068

                                (4)          (5)          (6)

Share of detached houses       0.026        0.014      -0.0085
                              (0.02)       (0.02)      (0.05)

Share of townhouses           -0.052       -0.097 **    0.13
                              (0.04)       (0.04)      (0.08)

Share of rented housing        0.36 ***     0.39 ***    0.21 ***
                              (0.04)       (0.05)      (0.06)

Share of mortgages             0.27 ***     0.25 ***    0.23 **
                              (0.06)       (0.06)      (0.1)

Unemployment                  -0.013        0.013      -0.028
                              (0.04)       (0.04)      (0.08)

Average education level in     0.0041       0.011      -0.00071
locality                      (0.010)      (0.01)      (0.02)

In (Monthly rent)              0.0081       0.017      -0.056 ***
                              (0.009)      (0.01)      (0.02)

Share of self-employed         0.086 **     0.070 *     0.022
                              (0.04)       (0.04)      (0.07)

Share of employed by          -0.0011      -0.0030     -0.021
state-owned enterprises       (0.02)       (0.02)      (0.04)

Average approval of local      0.0039       0.0061     -0.011
government in locality        (0.010)      (0.01)      (0.02)

Country effect                  Yes          Yes          Yes

Log like.                    322.1        250.7        84.8

Observations                    1068         828          226

                                (7)          (8)

Share of detached houses       0.0086      0.15 **
                              (0.02)      (0.06)

Share of townhouses           -0.076 *     0.14
                              (0.04)      (0.10)

Share of rented housing        0.35 ***    0.29 **
                              (0.06)      (0.1)

Share of mortgages             0.30 ***    0.13
                              (0.09)      (0.2)

Unemployment                   0.10 **    -0.070
                              (0.05)      (0.1)

Average education level in     0.0094      0.063 ***
locality                      (0.01)      (0.02)

In (Monthly rent)              0.016       0.0074
                              (0.01)      (0.02)

Share of self-employed         0.090 *     0.27 *
                              (0.05)      (0.1)

Share of employed by          -0.0023     -0.0014
state-owned enterprises       (0.02)      (0.09)

Average approval of local     -0.0088     -0.035
government in locality        (0.01)      (0.04)

Country effect                  Yes           No

Log like.                    348.0        53.6

Observations                    503           35

* p<0.05, ** p<0.01, *** p<0.001.

Robust standard errors in parentheses

Table 8: Literature on sources of agglomeration

Micro-foundation   Paper                      Main finding

Input sharing      Holmes (1999)              More purchased input in

Labour market      Diamond and Simon (1990)   Workers compensated
pooling                                       with higher wages

                   Costa and Kahn (2000)      Well-educated married
                                              prefer large cities

Knowledge          Jafe et al. (1993)         More citations in the
spillovers                                    same location

                   Duranton and Puga (2001)   Cities are 'nurseries'
                                              for new ideas

                   Moretti (2004)             More college graduates
                                              raises wages

Source: Rosenthal and Strange (2004)

Figure 2: Industrial zones as percent of total build-up.

Industrial land as percent of build-up area

Industrial land (% of built-up land)

St. Petersburg   43.8
Moscow           31.6
Cracow           28
Ljubjana         27.4
Sofia            27.1
Warsaw           15.1
Prague           13.4
Seoul             6.9
Hong Kong         5.4
Paris             5.2
London            4.7
Atlanta           4.1

Source: Bertaud, 2004

Note: Table made from bar graph.
COPYRIGHT 2012 Association for Comparative Economic Studies
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2012 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Regular Article
Author:Shepotylo, Oleksandr
Publication:Comparative Economic Studies
Article Type:Report
Geographic Code:4E
Date:Sep 1, 2012
Previous Article:Inter-regional insurance and redistribution--a non-parametric application to Russia.
Next Article:Employment: global and country perspectives.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters