Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

8.1 Introduction

Trends in fertility and mortality, as investigated in previous chapters, determine the total population of a country. The topic discussed in this chapter is internal population migration, which does not affect the increase or decrease of a country’s total population but does determine the territorial allocation of a population within a country. In terms of territory Russia is the world’s largest country, in fact, it is more than 45 times as large as Japan but with a population 1.2 times that of Japan’s. Because of the limited size of population in comparison to its vast territory, population distribution has a greater importance than in the case of a small country.

This chapter presents an analysis of the factors behind population migration between domestic regions in the area covered by the modern Russian Federation during the almost quarter-century period between 1990 and 2013. This period began with the Soviet era, during which interregional migration was restricted under the domestic passport and resident permit systems, followed by the turmoil of the government-system transition period after the collapse of the Soviet Union. The currency crisis of 1998 marked rock bottom for the Russian economy, which recovered and grew steadily throughout the rest of the period. Interregional population migration plays an economic role in evening out the supply and demand for labor between regions as it constitutes the movement of factors of production, and a great deal of research has been conducted on it in both advanced and developing countries (Greenwood 1991, 2010; Greenwood and Hunt 2003). However, interregional population migration under the former planned economy system, which was characterized by the control of population migration, has attracted little interest. It is known that the Soviet Union controlled interregional migration through a system of domestic passports and that residency in large cities required a permit, not just registration (Matthews 1993). 1 If interregional population migration is determined by government policy, the factors behind it are also politically determined. However, verifying whether this was indeed the case has been extremely difficult because data was not made public during the socialist era. Data on interregional population migration since the collapse of the Soviet Union has also been heavily restricted and, in the 1990s in particular, research in a wide range of areas saw limited progress.

The restrictions on data have begun to be eliminated and although access to internal materials at Rosstat (the Russian Federal State Statistics Service) cannot be said to be unrestricted, it is no longer impossible, and a small number of studies employing them have started to appear (Andrienko and Guriev 2004; Kumo 2007; Vakulenko et al. 2011; Guriev and Vakulenko 2015). This analysis has been influenced by this situation, and uses a population migration matrix for origins and destinations at the federal division level (i.e. regional constituents or federal subjects of Russia), recorded for each of the 24 years from 1990 to 2013, to analyse determinants of interregional population migration patterns in a period that includes the tail-end of the Soviet era.

As stated above, interregional population migration constitutes the movement of factors of production, and given Russia’s vast land area and heavily distorted spatial population distribution (Dmitrieva 1996), it is highly significant. Hill and Gaddy (2003) showed that the policy of heavily developing remote regions through distributed resource development and industrial location, the construction of military bases, and so on, caused a distortion in the distribution of population. Because of this, the collapse of the Soviet Union and the transition to capitalism must have wrought major changes to regional population distribution patterns. This phenomenon also hints at the advance of the transition process in Russia. To examine this, it is essential to perform a comparison using detailed population migration statistics, not just for the new Russia but also for the Soviet era. Interregional population migration in the Soviet Union was thought to be affected by government incentives for development. On the other hand, other researchers have stressed the limitation of policy incentives. To discuss this, it is necessary to clarify whether factors regarded as policy incentives had an impact during the Soviet era, and whether that role was lost following the Soviet collapse. Until now, however, previous research performing that kind of analysis has not existed, and the purpose of this chapter is to fill that gap.

8.2 Interregional Population Migration in the Soviet Union and Russia

It has been frequently pointed out that during the Soviet era the obligation to carry a domestic passport and the existence of a permit system rather than a registration system in urban areas affected regional population distribution (Matthews 1993). By designating the work locations of new university graduates and setting high wage rates in specific regions (Ivanova 1973), the Soviet government tried to distribute the labor force in a strategic fashion. This was fairly successful in terms of promoting resource development in the Extreme North 2 and Russia’s far east regions (Perevedentsev 1966). Registration of residence is a condition of applying for various social securities, and because of that the Soviet Ministry of Internal Affairs was aware of what was happening with interregional population migration. 3 Therefore the Ministry’s data is used in this chapter.

The collapse of the Soviet Union changed this situation. The constitution of the Russian Federation afforded freedom of movement, and soon after the collapse of the Soviet Union a federal law abolishing the residence permit system was enacted. 4 This chapter begins by examining what kinds of changes this brought to interregional population migration patterns. If a situation in which the distribution of population was determined by government policy was replaced by one of freedom of movement, a clear contrast in the direction of migration can be expected to have arisen. In fact, as Fig. 8.1 shows, if interregional population migration patterns in 1985, during the final period of stability in the Soviet era, are compared with those following the collapse of the Soviet Union, the differences are clear. During the Soviet era population inflows occurred in the Russia’s far east and the regions in the Extreme North, most of which are located in the Arctic, which demonstrates to a great extent the impact of policy incentives (Fig. 8.1A). Immediately after the collapse of the Soviet Union, however, there was a massive population outflow from the Russia’s far east and northern regions and a population inflow to the southern part of European Russia, which had experienced population outflows during the Soviet era (Fig. 8.1B). In addition, during the 2000s, when the new Russia exhibited sustained economic growth, inflows into regions that are located relatively far north but produce oil, gas, and non-ferrous metals (Tyumen Oblast, Khanty-Mansi Autonomous Okrug, Krasnoyarsk Krai, and so on) were once again observed (Fig. 8.1C).

Fig. 8.1
figure 1figure 1figure 1

Interregional population migration in Russia: net migration rate (/10000 person) (A) 1985: Soviet Era. (B) 1999: Period of Transitional Recession. (C) 2010: Period of Economic Growth (Source: Prepared by the author from Goskomstat/Rosstat, Regiony Rossii (Regions of Russia), various years.)

Full size image

To examine this more closely, the chapter looks at the distribution of birthplaces (origins) and current places of residence (destinations) using federal districts, which are the administrative divisions in modern Russia, at the times of the 1989 (nearly the end of the Soviet era), 2002, and 2010 censuses. This is not ordinary population migration data, which is used for the later analysis, but data that shows the results of life movement at each point in time. According to this data, in 1989, during the final phase of the Soviet era, there were more than 760,000 people living in the Central Federal District (the region centred on Moscow) who had been born in Siberia or the Far East. Conversely, 1.2 million people had been born in the Central Federal District but were now living in Siberia or the Far East (Table 8.1 Panel A). In other words, the number of “people born in Siberia or the Far East but living in European Russia” was far lower than the number of “people born in European Russia but living in Siberia or the Far East.” By the time of the 2002 population census, the number of people born in Siberia or the Far East but living in the Central Federal District had reached one million, while the number of people born in the Central Federal District but living in Siberia or the Far East had shrunk to 600,000 (Table 8.1 Panel B). In the 2010 census, meanwhile, the number of people born in Siberia or the Far East but living in the Central Federal District was 950,000, while the number of people born in the Central Federal District but living in Siberia or the Far East was less than 420,000, meaning that the former figure had reached more than double the latter (Table 8.1 Panel C). 5 In other words, it can be surmised that the opposite to what happened during the Soviet era occurred. People from Siberia and the Far East began moving to European Russia, while a significant proportion of people from European Russia who had been living in Siberia or the Far East returned to European Russia. A comparison of origin-to-destination tables for federal districts reveals that, between 1989 and 2002 and between 2002 and 2010, only the Central Federal District was accepting people from all regions at a higher rate than the average rate of change for all regions or was keeping that decline lower than the average for all regions (Table 8.1 Panel D and Panel E). This indicates that the Central Federal District was attracting relatively large numbers of people, not only from Siberia and the Far East, but from all over Russia.

Table 8.1 Distribution of the place of birth and the place of residence by population census data (in thousand)
Full size table

These tables are not difficult to interpret. Throughout the Soviet era, Russia’s population and economy were concentrated in the European portion of the country (Fig. 8.2; Dmitrieva 1996). During the Soviet era, the socialist government was able, through its development policies, to encourage the flow of labor to remote regions, such as the Russia’s far east and Siberia (Hill and Gaddy 2003). However, after the collapse of the Soviet Union it can be inferred that the direction of the flow reversed, with people moving to the Central Federal District, which contains Moscow, and surrounding parts of European Russia, which was already a very densely populated region. During the Soviet era, regional economic disparities were curtailed through investment policies focused on income redistribution and surrounding regions, but after the beginning of transition to capitalism, a rapid increase in disparities occurred. Figure 8.3 shows that at the same time as the Soviet collapse (in 1991) there was a dramatic increase in regional disparities.

Fig. 8.2
figure 2

Population distribution in Russia, 2002, in thousands (Prepared by the author from Rosstat, Regiony Rossii (Regions of Russia) 2004, 2005, Moscow)

Full size image
Fig. 8.3
figure 3

Income disparity and gross domestic products per capita in Russia, 1980–2013 (Source: Prepared by the author from Braithwaite (1995); Rosstat, Sotsial’noe polozhenie i uroven zhisni naseleniya Rossii (Social Situations and Living Standard of Population in Russia), various years; Rosstat, Regiony Rossii (Regions of Russia), various years)

Full size image

It is possible to describe such inferences as the above. However, the question of what kinds of changes were seen in the determinants of interregional population migration during the Soviet era and in the new Russia following the collapse of the Soviet Union has yet to be studied. Therefore, the analysis in this chapter focuses on that aspect.

It must be added that high-income regions are not concentrated in European Russia. With the exception of the two largest cities in European Russia, namely Moscow and Saint Petersburg, the Extreme North and the Russia’s far east/Siberia actually contain regions with higher incomes. In fact, the distribution of high-income regions has not changed significantly since the Soviet era (Fig. 8.4). Apart from Moscow and Saint Petersburg, all such regions are ones that produce a lot of energy resources, such as oil and natural gas, or non-ferrous metals, such as precious metals (Tyumen Oblast, Yamalo-Nenets Autonomous Okrug, Khanty-Mansi Autonomous Okrug, Krasnoyarsk Krai, Sakhalin Oblast, and Sakha Republic) or ones with extremely small populations (Magadan Oblast, Chukot Autonomous Okrug, Kamchatka Krai, Komi Republic, and Murmansk Oblast).

Fig. 8.4
figure 4figure 4figure 4

Income per capita by region (Source: Prepared by the author from Rosstat, Regiony Rossii (Regions of Russia) in 2010, 2011, Moscow)

Full size image

Because it is not simply the case that incomes are higher in large cities, the explanation may become vague. However, a comparison of Fig. 8.1C with Fig. 8.3C, which illustrates population flows in modern Russia and recent income levels, shows that the population centres of Moscow and Saint Petersburg and resource-producing areas such as Tyumen and Krasnoyarsk are attracting people, whereas the Extreme Northern oblasts, which have traditionally had high nominal per capita incomes but are situated in remote regions, have seen population outflows. The latter saw population inflows during the Soviet era (Fig. 8.1A), but their high incomes were not indicative of the degree of economic development. Instead, it is more appropriate to view the high incomes as meaning that the government targeted them for development and took commensurate measures to attract workers (Perevedentsev 1966; Hill and Gaddy 2003). In modern Russia the government no longer manages population migration, so it is natural that Extreme North regions without resources see population outflows.

However, things are not that simple. One point is the distribution of resources. Khanty-Mansi and Yamalo-Nenets autonomous okrugs in Tyumen Oblast, which produce more than 50 % of Russia’s crude oil and over 80 % of its natural gas, are classified as Extreme North regions. At the same time, there are large labor outflows and inflows in such regions, so caution needs to be exercised when conducting an analysis.

8.3 Previous Research

As stated at the beginning, the aim of this chapter is to shed light on the determinants of interregional population migration in the modern Russian Federation, and to compare them with those during the Soviet era. Because not many previous studies have adopted such a perspective, it is possible to discuss them all and to mention general research on population migration in modern Russia. 6

Given that materials that would allow origins and destinations to be specified at the oblast level have not been widely available in the 1990s, research in Russia itself has been conducted based on descriptive statistics in the early stages. Many studies have attempted to explain interregional migration as being due to: the labor market environment; the concentration of economic activity; the accessibility of regions; differences in the degree of infrastructure development; and the impact on the migration rate of the age structure, which results from differences in the propensity to migrate (Moiseenko 2004; Eliseeva 2006; Vishnevskii 2014). However, research has been hindered by a lack of statistics, and very few studies in which quantitative analyses were performed in the period until 2000. 7 Even these studies have had to explain the net migration rate of each region amid an absence of data, and it has been impossible to classify factors in population migration as either push or pull factors.

Brown (1997) showed that factors such as population size and average wage have a positive impact on net inflow, but that factors such as the average temperature in January have a negative impact net outflow. Wages, however, were observed to have a positive impact on net population outflow. This was because, although financial support for the Extreme North in the form of high wages was maintained after the collapse of the Soviet Union, it was insufficient to compensate for the inferior living conditions, resulting in a population outflow from this region. Gerber (2006) also studied net population migration rates, and showed that the population of the region and the average wage have a positive impact, while the rate of unemployment and the average temperature in January have a negative impact. Gerber (2005) used microdata to analyse the determinants of the probability of deciding to migrate, and found that in Russia also a high level of education and a young age increased the migration rate.

Andrienko and Guriev (2004) were the first researchers to analyse both origins and destinations at the oblast level. They obtained origin-to-destination (OD) tables from Goskomstat Russia for 89 regions for the period 1992–1999, and performed a panel analysis with the units being the 78 regions with complete data. Their analysis found that a region’s unemployment rate, population, and level of infrastructure affected population outflows and inflows as intuitively expected. Regarding incomes, Russia was in a recession stemming from the transition to capitalism, and if income levels were extremely low, people got caught in a poverty trap, and a population outflow did not occur. They pointed out that a population outflow from that region occurred as incomes rose; and that if an analysis is performed on all samples, the results become vague, but that if income is divided into bands and an analysis is conducted for each, income gives results that match what would be intuitively assumed. In addition, a distance variable obtained negative and significant coefficients.

Kumo (2007) conducted an analysis using oblast-level OD tables for 89 regions for the year 2003. These tables were obtained directly from an employee at Rosstat, the successor to Goskomstat and Russia’s current statistical organization. Although it is a cross-sectional analysis for a single year, it showed that with the economy growing, the concentration of economic activity in resource-producing areas, the environs of Moscow, and so on, as well as regional factors, such as the location of Extreme Northern areas, all had a conspicuous impact on population fluidity. And, like Andrienko and Guriev (2004), it confirms that the distance variable has a stable and negative impact on the scale of population migration. It seems likely that Vakulenko et al. (2011) made use of oblast-level OD tables from Rosstat for the period 2001–2008. 8 The key finding from their analysis was that the socio-economic variables were significant for migration between regions that were relatively close to each other, but that if the distance between regions was extreme, these variables lost their explanatory power.

Oshchepkov (2007) obtained oblast-level OD tables from Rosstat from the period 1990, at the end of the Soviet era, to 2006, and analysed the causes of migration for 78 regions with complete data. The distance between regions takes a stable and significant negative coefficient for the scale of migration. It was also shown that factors such as the labor market environment (unemployment rate), climate conditions (average January temperature), and the degree of infrastructure development (paved road density) produced results that matched intuitive expectations concerning both outflows and inflows. It was also pointed out that the absolute value of these coefficients becomes larger with the passage of time and that the impact of socio-economic variables becomes stronger. Guriev and Vakulenko (2015) advanced the analysis conducted by Andrienko and Guriev (2004). They used oblast-level OD tables from Rosstat for the period from 1996 to 2010. Regarding the relationship between income and population migration, they showed that while high-income regions indeed saw population inflows, in the poorest regions increases in income resulted in population outflows. They showed that it is likely that in regions with an income level of less than USD 3,000, those classes that wished to move out did not have the capability to do so. In other words, like Andrienko and Guriev (2004), they showed that a geographical poverty trap existed.

Andrienko and Guriev (2004), Kumo (2007), and Oshchepkov (2007) showed that the distance variable had a significant negative impact on the scale of population migration. This is intuitively obvious and a stylized finding from population migration research in advanced countries (Greenwood 2010). In the Soviet Union, however, there have been places that do not fit this description. In other words, as Mitchneck (1991) and Cole and Filatotchev (1992) have pointed out, in the Soviet Union distance did not exhibit a detrimental impact on population migration. Population migration on a larger scale than would be expected was observed, even between areas that were far apart from each other. The fact that the distance variable was stably negative and significant can be said to indicate that compared with the Soviet era, population migration patterns in Russia have changed.

However, it has to be said that a comparative study with the Soviet era has not been performed. In almost all the studies, data on the Soviet era has not been used and cannot be analysed. The only exception is Oshchepkov (2007), but in that study population migration data for 1990 to 2006 is pooled and the year to which the data relates is not specified. As a result, even though migration data for 1990 and 1991, which were during the Soviet era, is used, the analysis cannot interpret it. Although some statistics, such as the unemployment rate and the poverty rate, cannot be obtained for the Soviet era, given that complete time series data that includes the Soviet era exists, an analysis is possible. The factor of whether the region is resource-producing, which was used only by Kumo (2007), will also need to be subject to diachronic verification, not a cross-sectional analysis for a single year. In addition, none of the previous studies, apart from Kumo (2007), have taken into account the scale of migration. In other words, regardless of whether there is only one interregional migrant or tens of thousands of them, an analysis has been performed with this as a single observation. As explained later, this is unusual in the field of population migration research, so the next section expands the analysis period, data observation years, and the explanatory variables to take account of the scale of migration, and so on.

8.4 Empirical Analysis

The insights provided by the accumulation of general population migration research (Greenwood and Hut 2003; Greenwood 2010) and previous research on interregional population migration in Russia can provide hints on what variables should be used. In other words, the size of the population of the origin/destination probably has a positive impact on population flow. Furthermore, unlike in the Soviet era, the distance between regions probably has a stable and significant negative impact. It is also likely that various other socio-economic variables are determinants of the scale of population migration. Therefore, like Andrienko and Guriev (2004), Kumo (2007), and Oshchepokov (2007), this chapter employs the expanded gravity model, which is widely used in the field of population migration research. The formula for this model is:

$$ {M}_{ij}={g}^{*}\frac{{P_i}^{\alpha }{P_j}^{\beta }}{{D_{ij}}^{\delta }}*{\left(\frac{Y_j}{Y_i}\right)}^y $$

where M ij denotes the scale of population migration (number of people) from region i to region j, P i denotes the population of region i, P j denotes the population of region j, and D ij denotes the distance between region i and region j. In addition, Y i denotes an attribute of the origin region i, while Y j denotes an attribute of the destination region j.

8.4.1 Data

This analysis employs: regional data derived from official Soviet and Russian statistics; and origin-to-destination (OD) tables for 1990 to 2013, which are internal materials from Rosstat. The regional data uses statistics that can be accessed by anybody, and are either available online or have been published in paper form by Rosstat or its predecessor organization. The OD tables require a little more explanation, as they have only been used by Russian researchers and the authors of this chapter.

Rosstat publishes “Population and Population Migration in the year of **”, which constitutes widely available population migration data. Until 1999, these statistics contained OD tables for the 11 economic regions in use at the time. From 2000 onwards they contained OD tables for the seven, newly established, federal districts, which were then increased to eight from 2009. However, if one takes account of the diversity seen within the vast area of each region, this regional division is not adequate for analysis, so it was not used for research. Therefore, oblast-level OD tables are used, which are internal materials to Rosstat and were obtained by the author. These materials can be obtained directly from Rosstat employees, probably for a fee. For this analysis, however, the data was received from Rosstat. 9

Kumo (2007) analysed the year 2003 based on a table that related to that year alone, which had been obtained directly from Rosstat. The OD tables used in this analysis are for each year in the 24-year period between 1990 and 2013. Russia’s regional divisions have changed frequently, but the data has been adapted to match each of the 83 federal subject divisions that existed as of 2013—83 × 83 regions -83 (intraregional migration) = 6,806 origins/destination pairs constitute the units of analysis. However, for the Yamalo-Nenets Autonomous Okrug, Khanty-Mansi Autonomous Okrug, Nenets Autonomous Okrug, Chukot Autonomous Okrug, and the Jewish Autonomous Oblast, data is often missing for certain years, particularly for the first half of the period of analysis, which includes the Soviet Union era, so it is often excluded from analysis. In addition, the Chechen Republic and the Republic of Ingushetia were heavily affected by a war that lasted from 1991 to 1997 and which then broke out again in 1999, before finally ending in 2009. There are also numerous gaps in the data for these republics. For these reasons, the authors will exclude them from the analysis. The authors should therefore mention that the number of observations is not as many as 6,806 × 24 years = 163,344. But even if this data is lacking, at the time of writing no other studies exist that have employed such long-term data on interregional migration in Russia. The significance of the fact that these materials can be used to perform a comprehensive analysis of interregional population migration in Russia for a period of approximately a quarter of a century from 1990, before the collapse of the Soviet Union, to 2013, should be emphasized.

The purpose of the analysis is to identify determinants of interregional population migration in Russia. However, that does not mean that it simply backs up the insights confirmed from previous research. It identifies changes in factors behind population migration that occurred between the Soviet era and the emergence of the new Russia, which is only possible with the data obtained. As one can see from Fig. 8.1 population migration patterns in Russia have changed a great deal. It can be expected that during the Soviet era controls and incentives implemented by the central government had an impact, but this ceased to function after the collapse of the Soviet Union. This is identified by using the amount of government investment as an explanatory variable, which indirectly shows the government’s intentions concerning regional development priorities under the socialist regime. The fact that during the Soviet era interregional population migration occurred in line with the development intentions of the government are shown in the population inflow that occurred in Siberia and the Russia’s far east in the 1960s and 1970s. However, it is difficult to imagine that the same thing occurred in the new Russia. Until 1991, therefore, government investment had a positive impact on population migration in Russia, but after the collapse of the Soviet Union, that impact can be expected to have declined. To specify this a cross term for year dummies and the amount of government investment are used. This government investment is described as “basic investment” in the Russian language, and is capital used for production activities. It is not investment in non-production activities, such as healthcare, so it can be expected to serve as one of the development incentives assumed here.

In addition, a factor that is unique to Russia needs to be taken into account. That is the peculiarity of the regions that produce resources such as crude oil and natural gas, but only Kumo (2007) studied its impact on population migration patterns. In Russia mineral resources account for between 50 % and over 60 % of exports, 10 and half of the country’s tax revenue comes from taxes on energy resources. 11 Apart from urban areas such as Moscow, many high-income regions are resource-producing regions, and that probably has an effect on the flow of population migration. This analysis therefore uses a dummy variable to specify regions that produce crude oil or natural gas. This takes account of the fact that regions that produce energy resources tend to attract people. The analysis also explores the impact of Russia’s frigid climate. In Kumo (2007), the dummy variable for “Extreme North region” obtained a significant positive coefficient for both the origin and the destination, and the analysis in this chapter verifies this, and also uses the average temperature in January and investigates its coefficient. It is normal for people to move from places with harsh climates to places with mild climates (Greenwood 1991), and this chapter examines whether this is also a reasonable assumption for Russia. In addition, Russia experienced huge changes in the period from 1990 to 2013, so the ananlysis employed the year fixed effect to control for this.

To confirm the effectiveness of variables that have been used in previous research, they are also used in this analysis. To show economic conditions, average income per capita, average expenditure on charged services per capita, average expenditure on services for living per capita, and the consumer price index are all used. 12 The authors expect migration to occur from regions with lower incomes and expenditures to regions with higher ones. Migration can also be expected to occur from regions with a high price index to regions with a low one. The level of infrastructure is also expected to have an effect on population migration patterns. As measures of the level of infrastructure, the total length of railways, the total length of paved roads per unit of land area, and the number of buses per resident are used. In addition, this analysis uses the number of doctors per resident and the number of hospital beds per resident as indicators of social infrastructure. The analysis also takes account of population density. It can be assumed that regions with better infrastructure or regions that are more densely populated will attract people from regions with poorer infrastructure or regions that are less densely populated. Furthermore, previous research has pointed out the fact that population structure affects interregional population migration patterns, so the proportion of people who live in cities, the proportion of people who have not yet reached working age, and the proportion of people who have reached the age at which they are eligible to receive a pension are used to confirm the effect of these variables.

Just as Andrienko and Guriev (2004), Gerber (2006), and Vakulenko et al. (2011) did, this analysis avoids the problem of endogeneity by giving all the explanatory variables the values of one period (one year) before the interregional population migration. The variables are ratios between origins and destinations of each indicator basically. 13 Regarding the population of regions, the population of origin and the population of destination are employed separately. At the same time, the analysis looks at the dummy for Extreme North regions and the dummy for regions that produce oil or natural gas separately for origin and destination. Variables other than dummy variables are converted into logarithms. Therefore, regional pairs between which no population migration occurred, will not be included in the sample. 14 Definitions of, sources of, and the quantities of descriptive statistics for all the variables are shown in Table 8.2.

Table 8.2 Descriptive statistics
Full size table

8.4.2 Results

The results of the analysis are shown in Tables 8.3A and 8.3B. Table 8.3A uses all observations (total migration: at least one person migrated), while in Table 8.3B regional pairs between which migration on a certain scale occurred have been extracted. 15 In other words, in the latter, the analysis used interregional migration that accounts for 90 %, 80 %, 70 %, and 60 % of the total flow, extracting regions in the order of the scale of migration, and analysed each data set. This is significant because of the following reasons. The data used here is regional level data, and the analysis is attempting to explain the scale of population migration using macro indicators. Therefore, supposing one or two people migrated between two regions, it would probably not be appropriate to explain that using macro data. If interregional migration arises due to differences in the level of economic development, it is difficult to imagine that the volume of migration would be on such a small scale, so it can be said that it is likely that such migration is due to factors that cannot be identified using macro variables. Such migration therefore needs to be excluded, with the analysis only being performed for the main types of migration. However, regardless of the criteria that are applied, there is a risk of criticism that they are arbitrary; therefore a number of criteria were set and an analysis performed for each with the intention of identifying variables that will yield more stable results. The analysis therefore focuses more on Table 8.3B than Table 8.3A, which focuses more on cases in which the sample size is smaller (an analysis that specializes in regional pairs with large-scale migration).

Table 8.3 Results based on all observations
Full size table

Regardless of what criteria for the scale of migration are used to make the partitions, it is shown that fixed-effect models should be chosen. However, to view the impact of factors that do not change over time, such as the distance between regions, reference is made to the results of random-effect models. The distance variable stably obtains a significant negative coefficient, and population size stably obtains a significant positive coefficient for both origin and destination. These match the findings of Andrienko and Guriev (2004), Kumo (2007), and Oshchepkov (2007), and the impact of these variables on population migration patterns could be confirmed. Differences are therefore shown with the results that were observed throughout the Soviet era (Mitchneck 1991; Cole and Filatotchev 1992). Income and expenditure on services for living obtain significant positive coefficients throughout the period, while the price index obtains a significant negative coefficient. These findings are also in line with expectations. The former may indicate that the poverty trap pointed out by Andrienko and Guriev (2004) has been eliminated. The results for the value of consumption of charged services were unstable or obtained a negative coefficient, and this may mean that the price of services is high in regions that are sparsely populated.

Stable results for the number of doctors and hospital beds could not be obtained in the case of Table 8.3B. Attention probably needs to be paid to the fact that the highest numbers for both the number of doctors per capita and the number of beds per capita were observed in regions with extremely small populations. 16

Although these indicators have been used as variables in the economic analysis of the Soviet Union and Russia for many years (Andrienko and Guriev 2004; Oshchepkov 2007; Guriev and Vakulenko 2015), it may be worth re-examining their usefulness as explanatory variables.

Regarding railway density and the number of buses per resident, though not the case in Table 8.3A, in most cases in Table 8.3B a significant positive coefficient was obtained, which is what was expected. The density of paved roads was strongly correlated with the density of railways (r = 0.73), and this may be the reason that results could not be obtained. The Extreme North dummy obtained a positive and significant coefficient for both origins and destinations, which is the same finding as in Kumo (2007). The fact that it is not significant for the origin alone may mean that resource-producing regions in the Extreme North play a certain role not only in sending people but also in receiving them. This may be a coincidence with the fact that the coefficient for the average temperature in January was significant and positive. In other words, it may match the fact that people migrate to colder places. 17 The same explanation may be used for the fact that similar results were found for population density.

Regarding population structure, stable results could not be obtained for the proportion of people living in cities, the proportion of the population who were children, or the proportion of the population of an age eligible to receive a pension. Moiseenko (2004) pointed out the effect of age structure on population migration, namely that in Russia also the propensity to migrate is higher the younger the people are. This could be a factor that ought to be taken account of at the individual level. Alternatively, because resource-producing regions, many of which are situated in the Extreme North, attract people, the proportion of the population that is of working age and the proportion of the population that are children is high. On the other hand, remote regions in the Extreme North, such as Magadan Oblast and the Chukot Autonomous Okrug, have experienced large population outflows. Such diversity among regions lead to ambiguous findings such as these.

The analysis employed the dummy for oil/gas-producing regions to find out about conditions unique to Russia, and per capita government investment, which takes account of changes that have occurred since the collapse of the Soviet Union. The dummy for oil/gas-producing regions, obtains a significant and positive coefficient for the destination with every sample. For the origin, meanwhile, although it is insignificant in some cases, in cases where it is significant, it always obtained a negative coefficient. This matches the predictions made before the analysis, and demonstrate a result that is even clearer than Kumo (2007), the only previous study to have employed similar indicators. As the authors mentioned earlier, from the 1990s to 2010, minerals accounted for between 40 % and over 60 % of the value of exports. In addition, 50 % of federal government revenue came from oil and natural gas. As a result, there is no question that mineral and resource production affects the Russian economy as a whole (Kuboniwa 2014). Furthermore, these results show that it also affects the direction of interregional population migration.

Per capita government investment exhibited clear results. With the explanatory variable for 1989 (which is supposed to explain interregional migration patterns in 1990) as the base, it can be seen that the coefficient was significantly smaller, or that it was negative, throughout the 1990s. This means that interregional population migration patterns at the end of the Soviet era were significantly different from those following the collapse of the Soviet Union. During the Soviet era, the main targets of government investment and the direction of migration matched each other, and this probably indicates that government investment was effective as an incentive for regional development. At the same time, although Sonin (1980) and Milovanov (1994) have pointed out that during the Soviet era people were seen to migrate in a manner unrelated to government policy, this can also be said to suggest that the regional allocation of population through policy incentives was effective to a certain extent. It also shows that during the 1990s, after the collapse of the Soviet Union, government policy was no longer significant as policy incentive in the context of regional development. 18

The changes that occurred during the 2000s need to be mentioned here. Whichever results are used, in the middle of the 2000s at the at the earliest, the interaction term of the amount of government investment with 1989 as the base and the year dummy ceased to be significant. In other words, as was the case in the Soviet era, regions that were intensively targeted for government investment and the direction of population migration tended to match each other. However, it should be borne in mind that this does not mean that the same phenomena that occurred during the Soviet era had re-emerged. This is because there was a big difference between the regional distribution of per capita government investment in the Soviet era and in the new Russia (see Table 8.4). In other words, even if government investment in the Soviet era was implemented as a development incentive for remote areas in regions such as the Extreme North, it is likely that the regional allocation of government investment in the new Russia was conducted in such a way that a conclusion like that cannot be drawn. If, from the 2000s onwards, money was allocated with more of a focus on resource development, such a change would obviously have occurred. Note that government investment as used here refers to basic investment, which generally denotes capital for production purposes. It should therefore be borne in mind that the above interpretation is consistent with the nature of that investment.

8.5 Conclusions

As had been confirmed in previous research (Andrienko and Guriev 2004; Oshchepkov 2007), the analysis in this chapter showed that to analyze interregional population migration patterns in Russia, standardized techniques can be adequately applied. Regions with higher populations and income levels attract people. This is obvious, but it needs to be stressed that during the Soviet era it was not the case (Mitchneck 1991). Outflows from remote regions and inflows into resource-producing regions situated in the Extreme North occurred simultaneously. Therefore the results are not straightforward, but the overall trends are generally understandable. It could be assumed that because Russia possesses a wealth of mineral and energy resources, oil/gas-producing regions attract people from other regions. Kumo (2007) also pointed out that interregional migration patterns in Russia are partially shaped by such regions, as confirmed in this chapter by using a much broader set of data. On the other hand, it can be said that the fact that climatic conditions yield ambiguous results is indicative of a phenomenon unique to Russia, namely that resources are located in regions with harsh climate conditions. Government investment affected population migration patterns in the Soviet era, but its impact waned conspicuously after the collapse of the Soviet Union. Either that or it ceased to function as an explanatory variable. That phenomenon was in itself predictable, but the analysis conducted in this chapter was the first to employ data from the Soviet era to show that change clearly.

Nevertheless, the analysis in this chapter remains insufficient. Materials relating to economic variables in the Soviet era are still impossible to obtain fully, so some of the analysis is based on estimates. Furthermore, it was in 1987 that the Gorbachev administration, the final government of the Soviet era, implemented the perestroika (restructuring) reforms. Turmoil followed, and the Soviet Union was dissolved on December 25, 1991. In light of that, in order to compare the Soviet era with modern Russia it is necessary to use interregional migration statistics dating back to before 1990. Efforts need to be made to secure additional data. In addition, when analysing Soviet and Russian economic dynamics diachronically, it is usual to come up against inconsistent definitions of indicators, so it will be necessary to try to identify convincing variables.

The introduction to the chapter pointed out that one of the issues with interregional population migration would be whether it would result in a narrowing of disparities between regions in terms of the level of economic development. Vakulenko (2014) studied the relationship between population migration and the narrowing of disparities but did not obtain clear results. In light of the findings of this chapter, namely that population migration patterns in Russia have become similar to those seen in other countries, long-term inflows into regions with higher levels of economic development could serve to narrow regional economic disparities. However, if the concentration of population in Moscow continues it may result in a short-term increase in disparities, and this confusing situation may have led to the unclear results. The usability of the data has been confirmed to some extent, and from now on it would be desirable if efforts are made to deepen the analysis.

8.6 Appendix

Fig. 8.A
figure 5

Interregional migration and migration within regions in Russia, 1990–2013 (Source: Prepared by the author from the Internal Material offered by Rosstat.)

Full size image
Table 8.4 Changes in regional distribution of governmental investment per capita (based on 84 regions)
Full size table

8. Notes

  1. 1.

    On December 27, 1932, the Central Executive Committee and the People’s Commissar of the Soviet Union formalized “the establishment of a unified system of passports and the obligation to obtain residence permits” (Postanovlenie VtsIK i SNK ot 27.12.1932, «Ob ustanovlenie edinoi pasportnoi systemy po Soyuzu SSR i obyazatelnoi propiske pasportov»). Initially, the residence permit system was applied on a priority basis to the major cities of Moscow, Leningrad, Rostov, Kiev, Kharkov, and Minsk, but later it was introduced in almost every medium-sized and large city.

  2. 2.

    Refers to regions situated in the Arctic and other regions with similarly harsh living conditions. They were designated for the preferential allocation of resources and higher wages. Since the collapse of the Soviet Union the government has continued to provide assistance to Extreme North regions, but it is not of the type that would encourage the inflow of labor into these regions. In fact, the government has adopted policies that encourage the outflow of population from these regions (Thompson 2005). There are many laws and regulations, but see the Russian Federal Law “National Social Security and Subsidy Programs for Workers/Residents in Extreme North Regions and Similar Regions” (December 31, 2014) («O gosudarstvennykh garantiyakh i kompensatsiyakh dlya lits, rabotayushchikh i prozhivayushchikh v rayonakh Kraynego Severa i priravnennykh k nim mestnostyakh (s izmeneniyami na 31 dekabrya 2014 goda) »).

  3. 3.

    However, it was only in 1974 that passports began to be issued to residents of farming villages. Until then such residents were basically not allowed to move to cities (“Approval of Rules Concerning the Passport System in the Soviet Union”, Soviet Cabinet Decision No. 677, 28 August 1974) (Postanovlenie Sovmina SSSR ot 28 avgusta 1974 goda No.677 «Ob utverzhdenii polozheniya o pasportnoi sisteme v SSSR»). A look at the interregional population migration matrix (paper version) for the 1950s and 1960s from the Russian State Archive of the Economy (RGAE) shows that information about city-to-city migration was obtained, but adequate information about city-to-village, village-to-city, or village-village migration may not have been. In 2007–2008 the authors studied archived materials at the RGAE, but only documents on city-to-city migration had been filed, and there were not even any statistics recording origins/destinations for other types of migration.

  4. 4.

    With the passage of “Freedom of Movement and Rights Concerning the Selection of Resident Location within the Russian Federation by Citizens of the Russian Federation,” Russian Federal Law, October 1, 1993 (Zakon RF ot 1 oktyabrya 1993 «O prave grazhdan Rossiiskoi Federatsii na svobodu peredvizheniya, vybor mesta prebyvaniya i zhitelstva v predelakh Rossiiskoi Federatsii»), the residence permit system was formally abolished. This has been cited as a problem because authorities such as the city and oblast of Moscow have continued to require the residence permission (Moskovskie novosti, March 25, 2005; The Moscow Times, January 17, 2013). At the same time, however, there are apparently numerous ways to avoid registration, and this chapter does not consider the impact of the residence permit system in Russia after the breakup of the Soviet Union.

  5. 5.

    Given that Russia’s total population declined continuously from 1992 onwards, the fact that the number of people from Siberia and the Far East residing in the Central Federal District dropped between 2002 and 2010 is not in itself surprising. Given that the total number of people who left a federal district and moved from their birthplace to their current place of residence declined by an average of 14 % during this eight-year period (Table 8.1 Panel E), the key point must be that this number fell by a much lower rate than the trend for the population as a whole.

  6. 6.

    Refer to Lewis (1969), and chapter 3 of Kumo (2003), a survey relating to population migration research in the Soviet era.

  7. 7.

    Quite a few studies have also pointed out problems with the statistical record. This shows that the change in systems has had a major impact on migration statistics (Eliseeva 2006; Vishnevskii 2014; Shcherbakova 2015). Refer to Fig. 8.A. It shows total interregional population migration from the end of the Soviet era in 1990 to 2013, with figures based on data from Rosstat. It appears that total population migration declined continuously following the Soviet collapse. In addition, from 2011 onwards this trend seems to have increased rapidly. However, the change in systems has played a role. The residence permit system in the Soviet Union made it easy to grasp what was happening with interregional migration. However, after the collapse of the Soviet Union, its formal abolition inevitably reduced the proportion of identifiable cases of migration (Vishnevskii 2014). Another point is that definitions used in migration statistics changed in 2011. Until then, a migrant was defined as someone who changed their permanent domicile (i.e. a place in which they had resided for one year or more), but from 2011 the period was changed to nine months or more (Shcherbakova 2015), because of this it is impossible to discuss the scale of total population migration in the later period.

  8. 8.

    Scant explanation concerning the data was provided, making it difficult to know what sort of materials had been used. Because their analysis could not be conducted without the distance between regions, there can be no doubt that they used OD tables.

  9. 9.

    It was confirmed that the 2003 figures obtained from it matched the figures used in Kumo (2007).

  10. 10.

    Rosstat, Rossiiskii statisticheskii ezhegodnik (Russian Statistical Yearbook), Moscow, various years. (in Russian)

  11. 11.

    Ministerstvo finansov rossiyskoy Federatsii (2014), «Byudzhet dlya grazhdan», k Federal’nomu zakonu o federal’nom byudzhete na 2015 god i na planovyy period 2016 i 2017 godov (Ministry of Finance of the Russian Federation, Budget for the Citizens by the Federal Law on the Federal Budget for 2015 and the planned period for 2016 and 2017), Moscow. (in Russian)

  12. 12.

    Average expenditure on charged services per capita and average expenditure on services for living per capita are Soviet/post-Soviet categories of expenditure. The former involves expenditure on transport, communication, education, travel, healthcare, cultural activities (museums, theatres, and so on); the latter is expenditure on shoes, clothing, machine repairs, cleaning, home renovations, saunas, and so on. Variables that denote monetary amounts such as incomes and expenditure result in serious problems. Refer to Note 13 for more information on this.

  13. 13.

    This is to avoid problems that could be generated by monetary indicators. In 1992–1995 hyperinflation occurred, and no reliable deflator exists. In addition, a redenomination was carried out in 1998. To avoid such problems, Andrienko and Guriev (2004), for example, used the ratio of nominal income to minimum living expenses as the income variable. This chapter employs the ratio of incomes in the origin and the destination and the ratio of the amount of government investment in the two regions directly as explanatory variables. This should eliminate problems stemming from the units of measurement.

  14. 14.

    As methods for dealing with these missing figures, previous research has set the population migration figure as 1 or 0.5 (Guriev and Vakulenko 2015). This cannot escape criticism as being arbitrary. Regardless of whether 1 or 0.5 is set for the number of migrations for calculation purposes for the regional pairs with zero migrations (a total of 8,824), the results of analysis for the entire sample were qualitatively the same as when zero migrations was treated as a missing value (when excluded from the sample; as shown in Table 8.3A).

  15. 15.

    Total interregional migration (excludes migration within regions) was more than 30.53 million persons in 159,290 regional pairs over the 24 years; 58,308 regional pairs saw migration of 91 people or more, and these regional pairs accounted for migration of 27.47 million people (90 % of the total). Similarly, a total of 34,477 regional pairs saw migration of 178 people or more, and these regional pairs accounted for migration of 24.43 million people (80 %); 21,207 pairs saw migration of 305 people or more, and these regional pairs accounted for 21.37 million people (70 %). Finally, 13,202 regional pairs saw migration of 484 people or more, and these accounted for 18.32 million people (60 %). These were the sub-sets of each analysis. However, even if migration of at least one person occurred, there were cases in which the other data was missing, so the actual number of observations used in the analysis was smaller than this. Refer to Table 8.3B.

  16. 16.

    For example, in 2008 the regions with the most hospital beds per capita were the Chukot Autonomous Okrug, Magadan Oblast, Tyva Republic, Sakhalin Oblast, Jewish Autonomous Oblast, and Murmansk Oblast. Regarding the number of doctors, the city of Saint Petersburg was at the top throughout the period, followed by the Chukot Autonomous Okrug and the city of Moscow. These were followed by regions that are far away from European Russia, namely the Republic of North Ossetia, Tomsk Oblast, Astrakhan Oblast, and Amur Oblast.

  17. 17.

    No region had an average January temperature of more than zero degrees Celsius.

  18. 18.

    There are a number of problems with the data used here. First, some of the explanatory variables for 1989 are estimates (see Table 8.2 for details). The figure for the amount of government investment in 1989, in particular, was extrapolated from the figures for 1990 and 1991. The authors also performed an analysis based on data for 1990, the oldest year for which actual figures could be used. According to that, either the interaction term of government investment and the year dummy ceased to be significant at an earlier stage (from the beginning of the 2000s or the end of the 1990s), or a positive and significant coefficient is obtained depending on sub-sets that limit the number of observations. However, a similar explanation can be made when the estimated 1989 data is used as the base. Furthermore, it is difficult to imagine that during the Soviet era development policy changed all that much from year to year, and the regional distribution of government investment in 1989, for which the figure is an estimate, government investment in 1990, and government investment in 1991 are all highly correlated with each other (see Table 8.4). As a result, rather than excluding the data for 1989 from the analysis, the authors emphasize the use of interregional population migration in 1990, during the Soviet era, which is rare data. Second, as the authors mentioned in Note 7, there is the problem that in 2011 the definitions used in population migration statistics changed. With regard to this point, the authors performed an analysis using only migration data for the period to 2010 and confirmed that the results were qualitatively indifferent from the ones obtained in this chapter.