A model of interregional migration under the presence of natural resources: theory and evidence from Russia

Abstract

Internal net-migration rates in Russia are negatively correlated with regional labour shares in mining. In order to explain this phenomenon theoretically and empirically, Crozet’s (J Econ Geogr 4:439–458, 2004) theoretical model is augmented by the mining of natural resources to allow for exogenous market developments and spatially bounded production. The model is directly transformed into an econometric panel specification and tested for 78 Russian regions for the observation period 2004–2010. The empirical results show that the mining of natural resources attracts internal migrants, while regional price-indexes have unexpected positive effects.

Appendices

Appendix A: List of regions

Adygeya republic, Altai krai, Altai republic, Amur oblast, Arkhangelsk oblast (including Nenets autonomous district), Astrakhan oblast, Bashkortostan republic, Belgorod oblast, Bryansk oblast, Buryat republic, Chelyabinsk oblast, Chita oblast (Zabaykalsk krai) (including Agin-Buryat autonomous district), Chukotka autonomous okrug, Chuvash republic, Dagestan republic, Evrei autonomous oblast, Irkutsk oblast (including Ust-Orda Buryat autonomous district), Ivanovo oblast, Kabardino-Balkar republic, Kaliningrad oblast, Kalmyk republic, Kaluga oblast, Kamchatka krai (including Koryak autonomous district), Karachaevo-Cherkess republic, Karelia republic, Kemerovo oblast, Khabarovsk krai, Khakasia republic, Kirov oblast, Komi republic, Kostroma oblast, Krasnodar krai, Krasnoyarsk krai (including Taimyr and Evenk autonomous districts), Kurgan oblast, Kursk oblast, Leningrad oblast, Lipetsk oblast, Magadan oblast, Mari-El republic, Mordovia republic, Moscow city, Moscow oblast, Murmansk oblast, Nizhny Novgorod oblast, North Osetiya republic, Novgorod oblast, Novosibirsk oblast, Omsk oblast, Orenburg oblast, Oryol oblast, Penza oblast, Perm krai (including Komi-Permyak autonomous district), Primorskii krai, Pskov oblast, Rostov oblast, Ryazan oblast, Sakha (Yakutia) republic, Sakhalin oblast, Samara oblast, Saratov oblast, Smolensk oblast, St. Petersburg city, Stavropol krai, Sverdlovsk oblast, Tambov oblast, Tatarstan republic, Tomsk oblast, Tula oblast, Tuva republic, Tver oblast, Tyumen oblast (including Khanty-Mansi and Yamalo-Nenets autonomous districts), Udmurtia Republic, Ulyanovsk oblast, Vladimir oblast, Volgograd oblast, Vologda oblast, Voronezh oblast, Yaroslavl oblast.

Appendix B: Estimations of Tables 2 and 3 with complete samples

See Tables 4 and 5.

Table 4 Gravity-type specification with complete sample

Table 5 Model transformation with complete sample

Appendix C: Additional results

Table 6 Alternative model transformation

Equation (20) can be rewritten as

$$\begin{aligned} \ln M_{ij,t} -\ln \sum _{j\ne i}^N {M_{ij,t} }= & {} \left( {1-\mu } \right) \ln w_{j,t-1} -\phi \ln P_{C,j,t-1}\\&+\frac{\mu }{\sigma _{_D} -1}\ln L_{D,j,t-1} +\psi _1 \ln \left( {\frac{\dot{p}_{B,t} }{p_{B,t} }} \right) \\&+\psi _1 \ln \left( {1-{\frac{\dot{p}_{B,t} }{p_{B,t} }}/{\frac{{\dot{w}}'_t }{{w}'_t }}} \right) +\psi _2 \ln L_{B,j,t-1} \\&-\psi _3 \ln u_{j,t-1} +\psi _4 \ln S_{j,t-1} -\lambda _1 \ln \delta _{ij} \\&-\lambda _2 \ln \left( {1+\vartheta \Delta _{ij} } \right) +\psi _1 \ln \frac{x_i }{x_i -1}-\tilde{\alpha }_{i,t} +\tilde{\beta }_{_D} \end{aligned}$$

and transformed into the econometric specification

$$\begin{aligned} \ln M_{ij,t} -\ln \sum _{j\ne i}^N {M_{ij,t} }= & {} \alpha _i +\beta _1 \ln w_{j,t-1} +\beta _2 \ln L_{B,j,t-1} +\beta _3 \ln L_{D,j,t-1} \\&+\beta _4 \ln P_{C,j,t-1} +\beta _5 \ln u_{j,t-1} +\beta _6 \ln S_{j,t-1} \\&+\beta _7 \ln \delta _{ij} +\beta _8 \ln \xi _{ij} +\beta _9 \ln \varpi _t +\beta _{10} \ln \varsigma _t +\varepsilon _i \end{aligned}$$

with $\ln \varpi _t =\ln ({26+100{( {p_{B,t} -p_{B,t-1} } )}/{p_{B,t-1} }} )$, where 26 has been added to avoid undefined values, and $\ln \varsigma _t =100\ln ( 1-{{( {{w}'_t -{w}'_{t-1} } )}/{{w}'_{t-1} }}/{( {p_{B,t} -p_{B,t-1} } )}/{p_{B,t-1} })$, where the multiplication by 100 is undertaken for scaling reasons. The coefficients $\beta _1 ,\ldots ,\beta _8$ and their expected signs correspond to the ones as given below Eq. (22).

Table 7 Alternative model transformation with complete sample

Table 8 Uppermost quintiles results

The results can be found in Tables 6 and 7. They are very similar to Tables 3 and 5, with the positive coefficient for $\varpi$ underling the positive effect of commodities price-changes. The negative coefficient for $\varsigma$ indicates that commodities price-changes have a stronger effect than internal wage-increases.

Furthermore, the estimation results as they correspond to Eq. (22) and Table 3 have been rerun for migration patterns between the uppermost quintile of resource-rich regions, as well as the remaining regions. The results can be found in Tables 8 and 9 and are commented in Sect. 4.

Table 9 Lower four quintiles results

Table 10 Summary statistics

Table 11 Correlation coefficients

Table 12 Comparison statistics

Appendix D: Summary statistics and correlation coefficients

See Tables 10, 11 and 12.