The Migration Attractiveness Factors Of The Russian Far East Region

Abstract

Migration processes induce by the uneven development of regions. The Far Eastern region is Russia's largest by land area and the least populated federal district. Despite its economic indicators' steady growth, one of this territory's most significant problems is its population decline. Several econometric models were used to identify causes and factors influencing this region's migration attractiveness. Modeling results allowed for drawing the following conclusions: the most significant factor is per capita income growth and per capita gross regional product adjusted for climatic conditions. Due to its vast territory and severe climate, the Far East's economic growth centers are cities. Such measures as allocation to Russian nationals of one-hectare land plots at no charge and options to buy new housing at low residential mortgage rates are factors that should have promoted the region's migration attractiveness. It was found that the number of issued loans is decreasing simultaneously with the growth of a mean residential loan size, which indicates a reduction of real incomes of the population and a decrease of potential for accumulation of the first installment sum. Therefore, despite adopted development programs, population outflow will continue. Improving the Far Eastern region's migration attractiveness is an objective that requires the engagement of regional (municipal) authorities, business, and population.

Keywords: Econometric modelmigration attractivenessmigration factors

Introduction

The migration processes are often induced by the uneven development of regions: growing differences in quality of life and wage levels, raising the labor market's disproportions. There are other reasons triggering migration processes such as war conflicts, unfavorable climatic or environmental conditions. As for migration processes in Russia, they are primarily motivated by economic considerations.

Analysis of migration flows in Russian regions of the Far Eastern Federal District

are significantly differentiated in terms of economic development and living standards. The Far Eastern Federal District (FEFD) is the largest federal district in Russia, occupying 40.6% of its total area. However, its population accounts for as little as 5.6% of Russia's total population. At first glance, according to the Federal state statistics service of the Russian Federation all key economic indicators are rather stable in the Far Eastern region and their growth rate exceeds relative values for Russia as a whole.

A comparative dynamic of FEFD key economic indicators in 2019 ( % change over the previous year):

• Gross regional product - 119.3;

• Industrial production index - 106.2;

• Production of minerals - 106.6;

• Processing industries - 112.4;

• Electric power generation, gas and steam - 104.1;

• Foreign trade turnover - 106.1.

In 2019, mean nominal wage was above Russia’s average by 24% and 8.2% higher than in 2018. 1,576 billion rubles of capital investments was sunk in 2019 in FEFD’s economic and social development or 103.3% compared with the preceding year.

The Far East is characterized by three migration flows:

– intra-regional flow, i.e., migration from small towns to largest regional centers and to urban areas from rural areas. Such flows are typical of the Republic of Sakha (Yakutia), Khabarovsk Krai and Chukotka Autonomous District (2011–2016) which is explained by commissioning of new facilities requiring additional workforce;

–inter-regional flow is directed mostly out of FEFD to Russia’s southern, central and northwestern regions. Table  1 presents the structure of outgoing inter-regional migration flows from Primorsky Krai which is FEFD’s leader in terms of population mobility (Ivashina & Ryabokon, 2017);

See Full Size >

– external migration flow is typical of Khabarovsk Krai, Primorsky Krai and Sakhalin Oblast. A migration surplus was registered in Primorsky Krai in 2010–2012 in connection with large-scale construction projects on the eve of the Asia Pacific Economic Cooperation (APEC) summit.

Population densities in many Far Eastern administrative regions are extremely low (less than 1 person/sq. km in some of them). Although FEFD is of much importance for Russia’s economic development, according to the Russian Federation's Federal state statistics service, it is a region where a significant population outflow has been observed in the last years (Table  2 ).

See Full Size >

A dynamic of incoming and outgoing migration flows is presented below (Figure  1 ). The migration balance has remained negative since 1997. The graph also shows the ratio of migration balance on the preceding year (right scale, %).

See Full Size >

A maximum population migration outflow was observed in 2018 – 361.6 thousand people. Its intensity considerably decreased in 2019 but migration balance remained negative and amounted to -11.7 thousand people. Some analysts (Kiselyov et al., 2018) note that a trend is distinguishing the Far East from Russia's rest federal districts – long-term population decline. Moreover, migration processes should be considered as an indicator of a region's economic development (Motrich & Izotov, 2018).

Problem Statement

Migration attractiveness is defined as a generalized characteristic of a region's promising outlook for potential migrants based on social, economic and climate-geographic factors. It is important for the redistribution of labor resources and population between regions. We provide in a table below FEFD's key

social indicators in addition to its economic indicators according to the Federal state statistics service of the

Russian Federation and Russian News Agency data to understand the reason for this region’s low migration

attractiveness (Table  3 ). Highlighted by color are the highest/lowest values of FEFD administrative

territories among all Russia’s regions.

The territory of the Far Eastern Federal District is characterized by sharp environmental contrasts casing a broad variance of data in Table  4 . The lack of labor resources hampers its social and economic development and aggravates disproportions between FEFD regions.

Research Questions

Industrial production growth in FEFD has been 23% since 2013 when the target-specific federal program “Economic and Social Development of the Far East and Baikal Region through 2025” was adopted and launched, while growth in Russia as a whole was 8.3% (Kozlov, 2019). In particular, this program aims to solve one of the region's greatest problems – population decline. It is expected that population inflows will also be driven by development of Rapid Development Areas and Special Economic Zones. To attract people to FEFD and ensure the Far East's social and economic development, the "Far Eastern Hectare" program was launched. It provides for allocation of one-hectare land plots to Russian nationals at no charge in any FEFD's administrative territory. Its key purpose is to attract people to FEFD as permanent residents. Also in process is a governmental residential mortgage program providing benefits for people buying or building homes in the Far Eastern Federal District, which offers residential mortgage loans at rates 4.5 times lower than Russia's average. Why do people still keep leaving this territory?

Purpose of the Study

This study aims to investigate population outflow trends using econometric analysis and identify which specific factors or their combinations influence migration processes in FEFD.

Research Methods

This paper makes use of econometric models described in (Degtyareva, 2018) and (Tikhomirov, 2017).

We suggest using the Holt — Winters adaptive model (Hyndman et al., 2008), a modification of the exponential smoothing method, for short-term forecasting and assessment of population outflows from the Far Eastern region. Its representation in the form of an additive combination of the linear trend with a seasonal component will be taken as a series model.

A Holt ― Winters model-based forecast for p periods in the future is determined by expression (1):

${\hat{\mathit{X}}}_{\mathit{t}\mathbf{+}\mathit{p}}\mathbf{=}\mathbf{(}{\mathit{F}}_{\mathit{t}}\mathbf{+}\mathit{p}{\mathit{C}}_{\mathit{t}}\mathbf{)}{\mathit{M}}_{\mathit{t}\mathbf{+}\mathit{p}\mathbf{-}\mathit{k}}$ (1),

where k – number of phases in a full seasonal cycle,

X _t – original time series.

The updating of the coefficients is performed as follows (Formulas 2):

$\left\{\begin{array}{c}{F}_t={\alpha }_F\frac{X_t}{M_{t-k}}+(\mathrm{1}-{\alpha }_F)(F_{t-\mathrm{1}}+C_{t-\mathrm{1}})\\ C_t={\alpha }_C(F_t-F_{t-\mathrm{1})}+(\mathrm{1}-{\alpha }_C)C_{t-\mathrm{1}}\\ M_t={\alpha }_M\frac{X_t}{F_t}+(\mathrm{1}-{\alpha }_M)M_{t-k}\end{array}\right.\mathrm{}\mathrm{}\mathrm{}\mathrm{}\mathrm{}\mathrm{}\mathrm{}\mathrm{}\mathrm{}\mathrm{}$ (2),

where α_F, α_M and α_С are parameters of adaptation. Each parameter belongs to [0;1] interval. The closer is a parameter to 1, the larger weighting is assigned to the latest observations. A criterion for parameter selection is minimization of the model’s mean relative error. Initial values F ₀ and С ₀ are estimated using the least square method (3):

X_t=F₀+C₀t+ε_t(3).

In order to forecast population outflows by FEFD administrative regions and address factors influencing migration, we suggest using a fixed effects model that rests on a panel data structure thus allowing for taking account of non-measurable individual differences of objects. The fixed effects model treats each economic unit as “unique” which cannot be treated as a result of random selection from some general assemblage. Such approach is quite justifiable when considering countries or large regions.

Let’s introduce the following denotations: $\mathit{i}=\mathit{1},\dots ,\mathit{}\mathit{n}$ – numbers of objects; $\mathit{t}=\mathit{1},\dots ,\mathit{}$T – moments of time;– number of features; x _it – set of independent variables (dimension vector ); y _it – dependent variable for economic unit i at the moment of time t; ε _it – respective error. Lets’ also denote the following:

$${\mathit{y}}_{\mathit{i}}=\left[\begin{array}{c}{\mathit{y}}_{\mathit{i}\mathbf{1}}\\ \dots \\ {\mathit{y}}_{\mathit{i}\mathit{T}}\end{array}\right],\mathrm{}{\mathit{X}}_{\mathit{i}}=\left[\begin{array}{c}{\mathit{x}}_{\mathit{i}\mathbf{1}}^{\mathrm{\text{'}}}\\ \dots \\ {\mathit{x}}_{\mathit{i}\mathit{T}}^{\mathrm{\text{'}}}\end{array}\right],\mathrm{}{\mathit{\epsilon }}_{\mathit{i}}=\left[\begin{array}{c}{\mathit{\epsilon }}_{\mathit{i}\mathbf{1}}\\ \dots \\ {\mathit{\epsilon }}_{\mathit{i}\mathit{T}}\end{array}\right],\mathit{y}=\left[\begin{array}{c}{\mathit{y}}_{\mathbf{1}}\\ \dots \\ {\mathit{y}}_{\mathit{T}}\end{array}\right],\mathit{X}=\left[\begin{array}{c}{\mathit{X}}_{\mathbf{1}}\\ \dots \\ {\mathit{X}}_{\mathit{n}}\end{array}\right],\mathit{\epsilon }=\left[\begin{array}{c}{\mathit{\epsilon }}_{\mathbf{1}}\\ \dots \\ {\mathit{\epsilon }}_{\mathit{n}}\end{array}\right]$$

where y and X are panel (“integrated”) observations and errors; y and ε – vectors with a dimension of nT '1; X – regressor matrix with a dimension of nT'k . The fixed effects model for panel data denoted as above is described by the following equation (4):

$${\mathit{y}}_{\mathit{i}\mathit{t}}={\mathit{\alpha }}_{\mathit{i}}+{\mathit{x}}_{\mathit{i}\mathit{t}}^{\mathrm{\text{'}}}\bullet \mathit{\beta }+{\mathit{\epsilon }}_{\mathit{i}\mathit{t}}\mathrm{}\mathrm{}\mathrm{}\mathrm{}\mathit{\beta }\in {\mathbb{R}}^{\mathit{k}},\mathrm{}{\mathit{\alpha }}_{\mathit{i}}\in \mathbb{R}\mathbb{}\left(\mathit{i}=\mathbf{1},\dots \mathit{n}\right)\mathrm{}\mathrm{}\mathrm{}\mathrm{}\mathrm{}\mathrm{}\mathrm{}\mathrm{}\mathrm{}\mathrm{}\mathrm{}\mathrm{}\left(\mathbf{4}\right)$$

The value α _i expresses an individual effect of object i , not dependent on time t , with regressors x _it containing no constant.

We used a combined model looking as (5) to model volumes of residential mortgage loans – this is an additive model of multiple linear regression and time series model for residuals.

Y_vif =F1_t+F2_t+ε_t= a₀+a₁x₁+…+a_nx_n+ARIMA(p,i,q)+ε_t (5)

F1 _t = a ₀ +a ₁ x ₁ (t)+…+a _n x _n (t) – multiple regression where x ₁ (t)…x _n (t) are factors significantly influencing Y _vif .

We used the variance inflation factor (VIF) to identify linearly independent factors. This indicator allows for evaluation of variance growth of a given regression rate, occurring due to high correlation of factors. The higher is this indicator for a factor, the stronger is linear relation between it and other factors. This indicator is calculated by formula (6), where ${\mathbf{R}}_{\mathbf{j}}^{\mathbf{2}}$ is the factor’s determination coefficient j relative to all the rest ones.

$\mathit{V}\mathit{I}\mathit{F}=\frac{\mathbf{1}}{\mathbf{1}-{\mathit{R}}_{\mathit{j}}^{\mathbf{2}}}$(6)

To eliminate multi-collinearity, we excluded factors with too high VIF values from the model. Generally, a critical VIF value is 5; however, this indicator is equal to 10 in some sources. Data are subject to pre-standardization.

An integrated moving average model is applied to residuals of the multiple regression model: F2 _t = Y _t -F1 _t = ARIMA(p,d,q). The autoregressive integrated moving average model ARIMA(p,d,q) can be represented by equation (7), where p and q are whole numbers governing its order and d is a finite difference operator order.

$${∆}^{\mathit{d}}{\mathit{X}}_{\mathit{t}}=\mathit{c}+\underset{\mathit{i}=\mathbf{1}}{\overset{\mathit{p}}{\sum }}{\mathit{a}}_{\mathit{i}}{∆}^{\mathit{d}}{\mathit{X}}_{\mathit{t}-\mathit{i}}+\underset{\mathit{j}=\mathbf{1}}{\overset{\mathit{q}}{\sum }}{\mathit{b}}_{\mathit{j}}{\mathit{\epsilon }}_{\mathit{t}-\mathit{j}}+{\mathit{\epsilon }}_{\mathit{t}},\mathrm{}\mathrm{}\mathrm{}\mathrm{}\mathrm{}\mathrm{}\mathrm{}\mathrm{}\mathrm{}\mathrm{}\mathrm{}\mathrm{}\mathrm{}\mathrm{}\mathrm{}\mathrm{}\mathrm{}\left(\mathbf{7}\right)$$

whereε_t is a stationary time series, c is a constant, a _i – AR model parameters (autoregression rates), b _j – MA model parameters (moving average model coefficients), and ∆ ^d is a finite difference operator of the time series of d order.

Findings

A source of data for model construction was materials posted on the websites run by the Federal State Statistics Service in 1998–2018.

Identification of migration attractiveness factors

We used a fixed effects panel regression (4) to identify factors influencing the region's migration attractiveness. The region's administrative territories made up the panel for the period of 1997–2017. In this period, FEFD did not include Trans-Baikal Krai and Republic of Buryatia as yet. Values of outgoing migration flows by each administrative territory were considered as a dependent variable. To obtain a significant model, we constructed an artificial variable Vrp_Climat = Vrp_Persan*Climat that characterizes per capita GRP adjusted for a “correction” reflecting the territory’s climate severity. Climat is a categorical variable characterizing climatic conditions of FEFD administrative territories and has the following values: 1 – for Chukotka Autonomous District; 2 – for Magadan Oblast and Republic of Sakha (Yakutia); 3 – for Khabarovsk Krai; 4 – for Sakhalin Oblast and Amur Oblast, Kamchatka Krai and Jewish Autonomous Oblast; 5 – for Primorsky Krai. Vrp_Persan is per capita GRP. The model includes a factor Dohod_pers – “region’s mean per capita income”. The model was implemented using the Gretl econometric software package (Figure  2 ).

See Full Size >

The dependent variable l_out is the logarithm of outgoing migration flows by each administrative region of FEFD, and factor variables d_Dohod_pers and d _ Vrp_Climat are absolute increments of variables “region’s mean per capita income” and “per capita GRP adjusted for climatic conditions”. All other factors were excluded from the panel regression as non-significant, having a non-interpretable plus-minus sign or high value of vif-factor. The minus sign before coefficients indicates at an inverse relation between variables. The most significant value is per capita income increment in each of the administrative territories under consideration. The coefficients are less than 1 and, therefore, are well interpretable: as the dependent variable is in logarithms, antilogs need to be calculated for regression rates for correct interpretation of results. The outcome will be as follows: an increase of income increment by 1 ruble will reduce population outflow by 5.8e-05 per cent, and an increase of per capita GRP adjusted for climatic conditions by 1 ruble will reduce population outflow by 0.15 per cent. The constant value shows that the impact of non-observed factors will keep migration outflows at a level of 9.79% annually. The panel regression has identified economic factors that curb migration outflows. An annual increment of per capita GRP adjusted for climatic conditions is the most significant factor for improving the Far Eastern region's migration attractiveness.

Conclusion

Migration outflow is a systemic problem caused both by economic and social reasons. A high wage level is proportional to high prices for goods and public utilities which make impossible any savings, qualified medical services or education of children, while severe climate hardships are not compensated by personal income of the region’s residents. Taken together with growing mobility of people, all this drives the migration outflow. According to RosStat data, the highest cost-of-living index in 2017 was registered in Petropavlovsk-Kamchatski and Anadyr – 1.58. Comparison with other regions would form an idea of ​ ​ prices in the Far East. For reference, this index is 1.27 in Moscow and 0.99 in Krasnodar. Real disposable personal incomes have dropped to 94.2% since 2015 and the number of people with incomes below subsistence level has accordingly grown. Given high housing prices, an effective behavior pattern for the region’s residents is to sell their rather expensive real estate in the Far East and search for more suitable options in other regions.

A low unemployment rate (4-5%) shows that there is no deficit of jobs because everybody willing to work even for small money is working. Any further population outflow will give rise to a serious workforce deficit. The Interestedness of developers and borrowing of residential mortgage loans by population might help to reduce population outflows from FEFD. In general, residential mortgage crediting is a promising factor of improvement of the region's migration attractiveness. Still, population decline in FEFD requires the solution for economic, ecological, and social problems and cannot be solved only by reducing residential mortgage loan rates. The key factor reducing the region's migration attractiveness is the lack of an infrastructure capable of ensuring migrants' social adaptation willing to join the Far Eastern Hectare program. FEFD administrative regions and municipalities have to cope with a budget deficit and cannot provide financing through local infrastructure projects. A target-specific priority-level governmental program must meet the Far East's demand for all transport services at the federal budget's expense.

A comprehensive approach should be applied to the problem of improving FEFD's migration attractiveness: not only efficient migration policies with a focus on practical rather than theoretical aspects of their implementation but also improvement of the quality of life, major investments in the development of the region's social sphere and infrastructure.

Acknowledgments

This paper has been prepared with support of the Russian Foundation for Fundamental Research. Grant number 19-010-00085 A.

References

1. Danilenko, T., Demidova, O., & Signorelli, M. (2018). Unemployment Clubs in Russian Regions. Emerging Markets Finance and Trade, 54(6), 1337-1357.
2. Degtyareva, N. (2018). Time series analysis and forecasting models. Chelyabinsk: Cicero.
3. Demidova, O., Daddi, P., Medvedeva, E., & Signorelli, M. (2018). Modeling the employment rate in Russia: A spatial-econometric approach. Economy of Region, 14(4), 1383-1398. https://doi.org/10.17059/2018–4–25
4. Frisman, E., Kulakov, M., Revutskaya, O., Zhdanova, O., & Neverova, G. (2019). The key approaches and review of current researches on dynamics of structured and interacting populations. Computer Research and Modeling, 11(1), 119–151.
5. Gerber, T. (2006). Regional economic performance and net migration rates in Russia, 1993–2002. International Migration Review, 40(3), 661–697.
6. Hyndman, R., Koehler, A., Ord, J., & Snyder, R. (2008). Forecasting with Exponential Smoothing. The State Space Approach. Heidelberg: Springer-Verlag.
7. Ivashina, N., & Ryabokon, K. (2017). Analiz iskhodyashchikh migratsionnykh potokov iz Primorskogo kraya [Analysis of outgoing migration flows from the Primorsky region]. Cross-border markets of goods and services: problems of research. Proceedings of II International Scientific and Practical Conference, 43–47.
8. Khavinson, M. & Kulakov, M. (2017). Gravitational Model of Population Dynamics. Bulletin of the South Ural State University. Series: Mathematical Modelling. Programming and Computer Software, 10, 80-93.
9. Kiselyov, N., Gurshalov, P., & Oleinik, E. (2018). Primorsky Region: Economic Growth and Migration Attractiveness Issues, Economics and Law Issues, 7(121), 114-118.
10. Kozlov, A. (2019, November 14). O zakonodatel'nom obespechenii operezhayushchego sotsial'no-ekonomicheskogo razvitiya Dal'nego Vostoka i Arktiki [On legislative support to rapid social and economic development of the Far East and Arctic]. https://minvr.gov.ru/press-center/news/23774/?sphrase_id=1449200
11. Lifshits, M. (2016). Forecasting of the global migration situation based on the analysis of net migration in the countries. Applied Econometrics, 41, 96-122.
12. Motrich, E., & Izotov, D. (2018). Modern Trends and Problems of Migration in a Russian Border Region: The Far East. Studies on Russian Economic Development, 29, 245-251.
13. Prokhorov, P., & Epstein, N. (2016). Statistical analysis of the migration situation and factors influencing Russia's demographic expansion. Statistics and Economy, 6, 9-18.
14. Ryazantsev, S. (2019). Modern Migration Policy of Russia: Challenges and Approaches to Improvement. Sotsiologicheskie Issledovaniya, 9, 117-126.
15. Ryazantsev, S., & Khramova, M. (2018). Factors of emigration from Russia: Regional features. Economy of Region, 14(4), 1298-1311. https://doi.org/10.17059/2018–4–19
16. Sardadvar, S., & Vakulenko, E. (2017). A model of interregional migration under the presence of natural resources: theory and evidence from Russia.  Annals of Regional Science, 59(2), 535-569.
17. Tikhomirov, N. (2017). Methods of econometrics and multidimensional statistical analysis. Moscow: Economy.
18. Vakulenko, E. (2016). Econometric analysis of factors of internal migration in Russia. Regional Research of Russia, 6, 344–356.
19. Vorobyeva, O., Rybakovsky, L., & Rybakovsky, O. (2016). Migration Policy of Russia: history and modernity. Moscow: Econ-Inform.