![Sample our Urban Studies journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/110826/TOC-Subject-Sample-2021-Urban-Studies.jpg"})
Abstract
Kubis A. and Schneider L. Regional migration, growth and convergence – a spatial dynamic panel model of Germany, Regional Studies. This paper empirically analyses the question of how regional migration affects regional convergence and growth in post-reunification Germany. Addressing the endogeneity of migration and human capital, a dynamic panel data model within the framework of β-convergence is applied, accounting for spatial effects. The regressions indicate that out-migration has a negative but modest effect on regional growth; the expected effect of skill selection is only partly confirmed. In the East German subsample, in-migration increases growth independently of its human capital effect; in West Germany, in-migration lowers growth per se, but this negative impact is offset by the growth-stimulating forces of migrants’ skills.
Kubis A. and Schneider L. 区域移民、成长与聚合—德国的空间动态面板模型，区域研究。本文经验性地分析东西德统一后，区域移民如何影响区域的聚合与成长之问题。为了处理移民与人力资本的内生性，本研究在β-聚合的架构中应用动态面板数据模型，并考量空间效应。迴归指出，向外移民对于区域成长具有负面却轻微的影响；被期待的技术选择效应，仅有部分被証实。在东欧的子样本中，向内移民增加了成长，却和人力资本影响无关；在西欧，向内移民降低了成长本身，但此一负面效应则由刺激成长的移民技术之驱力所抵消。
Kubis A. et Schneider L. La migration, la croissance et la convergence régionales – un modèle dynamique spatial en données de panel pour l'Allemagne, Regional Studies. Cet article analyse de manière empirique la question de savoir comment la migration régionale influe sur la convergence et la croissance régionales après la réunification de l'Allemagne. Abordant l'endogénéité de la migration et du capital humain, on applique un modèle dynamique spatial en données de panel dans le cadre de la convergence-β, tenant compte des effets spatiaux. Les régressions indiquent que le solde migratoire négatif a un effet défavorable, bien que modeste, sur la croissance régionale; elles ne confirment l'effet attendu, à savoir la sélection des compétences, que partiellement. Dans le sous-échantillon est-allemand, le solde migratoire positif augmente la croissance indépendamment de son effet sur le capital humain; dans la zone ouest allemande, le solde migratoire positif réduit la croissance en soi, mais les compétences des migrants en tant que moteur de croissance compensent cet impact négatif.
Kubis A. und Schneider L. Regionale Migration, Wachstum und Konvergenz – ein räumliches dynamisches Paneldatenmodell von Deutschland, Regional Studies. In diesem Beitrag wird auf empirische Weise die Frage untersucht, wie sich regionale Migration auf die regionale Konvergenz und das regionale Wachstum im wiedervereinigten Deutschland auswirkt. Zur Analyse der Endogenität von Migration und Humankapital kommt innerhalb des Rahmens der β-Konvergenz ein dynamisches Paneldatenmodell unter Berücksichtigung räumlicher Effekte zur Anwendung. Aus den Regressionen geht hervor, dass sich die Auswanderung negativ, aber mäßig auf das regionale Wachstum auswirkt; die erwartete Auswirkung der Qualifikationsauswahl bestätigt sich nur teilweise. In der ostdeutschen Teilstichprobe erhöht die Einwanderung das Wachstum unabhängig von ihrer Auswirkung auf das Humankapital; in Westdeutschland verringert die Einwanderung das Wachstum an sich, wobei diese negative Auswirkung aber durch die wachstumsfördernden Kräfte der Qualifikationen der Migranten ausgeglichen werden.
Kubis A. y Schneider L. Migración, crecimiento y convergencia regionales: un modelo de panel dinámico espacial de Alemania, Regional Studies. En este artículo analizamos empíricamente cómo influye la migración regional en la convergencia y el crecimiento regionales en Alemania desde la reunificación. Estudiando la endogeneidad de la migración y el capital humano, aplicamos un modelo dinámico de datos de panel en el marco de convergencia β, teniendo en cuenta los efectos espaciales. Las regresiones indican que la emigración tiene un efecto negativo, pero modesto en el crecimiento regional; el efecto esperado de la selección de habilidades se confirma solo parcialmente. En la submuestra de Alemania del este, la inmigración aumenta el crecimiento independientemente de su efecto en el capital humano; en Alemania del oeste, la inmigración reduce el crecimiento como tal, pero su impacto negativo se compensa mediante las fuerzas de las habilidades de los inmigrantes que estimulan el crecimiento.
Acknowledgements
The authors are grateful to Wolfgang Dauth for his comments made on an earlier draft of this paper. They also thank several anonymous reviewers for their constructive comments, which helped improve this paper.
Disclosure statement
No potential conflict of interest was reported by the authors.
Supplemental data
Supplemental data for this article can be accessed at http://dx.doi.org/10.1080/00343404.2015.1059932
Notes
1. See the review by Ozgen et al. (Citation2010). The exceptions are the most advanced papers of Ostbye and Westerlund (Citation2007) and, recently, Fratesi and Percoco (Citation2014) (see the second section).
2. For a review of theoretical papers, see Drinkwater et al. (Citation2003), Fratesi and Riggi (Citation2007), and Fratesi and Percoco (Citation2014).
3. A Solow-style model explicitly considering the role of migration can be found in Barro and Sala-i-Martin (Citation1995). Because the model considers the human capital that migrants typically bring in, it somewhat transcends the homogenous framework.
4. Various agglomeration mechanisms – in addition to the ‘home market effect’ – might drive diverging patterns of spatial economic distribution, particularly because of differences in regional innovativeness due to Marshallian or Jacobian externalities.
5. For a similar argument, see Arntz et al. (Citation2014).
6. In essence, this paper extends their approaches by accounting for spatial dependence. Surprisingly, both papers found only modest support for the existence of spatial autocorrelation. However, the reliability of the applied Moran's I statistic is disputable in cases of substantial spatial dependence generated by a spatial autoregressive data-generating process (Li et al., Citation2007).
7. Fratesi and Percoco (Citation2014) do not explicitly test the convergence hypothesis.
8. A more recent study by the same authors (Granato et al., Citation2014) reveals that the equalizing neoclassical effect is restricted to low- and medium-skilled worker migration. High-skilled worker migration tends to increase regional unemployment disparities.
9. Due to the strong impact of wage-setting institutions across the German regions, estimating the effect of mobility on the regional wage level is a very different task than analysing the productivity effect of migration. In Niebuhr et al. (Citation2012), the insignificant results regarding the regional wage regressions are explained by the impact of these institutions. From this perspective, it helps to compare the wage regression results with the present approach using regional productivity as a dependent variable.
10. For a comprehensive description of the estimation technique as well as the implementation of spatial effects in the dynamic panel model, see Technical Appendix B in the Supplemental data online.
11. If the stock of human capital of the regional workforce is regressed on migration rates (see Appendix Table A2 in the Supplemental data online), the expected positive skill bias of migration is found. In-migration – as well as positive net migration – improves the regional skill level, and out-migration lowers the human capital basis. However, the effects are small and sometimes not significant. Particularly noteworthy is the insignificant effect of in-migration on human capital in the sample containing all German regions. Thus, human capital accumulation might be more affected by educational or employment decisions rather than by migration.
12. In addition, the positive effect of net-migration on regional human capital (see Appendix Table A2 in the Supplemental data online) supports an opposing perspective. Nonetheless, even if migrants are positively selected in terms of human capital and increase the average stock of human capital in the receiving region, the average stock of capital defined in a broad sense might decrease by migration.
13. The extremely high β-parameter in the East might be at least partly attributable to the low number of regions.
14. Due to geographic reasons, the core–periphery regime overlaps with the East–West regime to some extent. Only three of 24 core regions are located in the East; 15 of the 22 East regions are peripheral areas, whereas in West Germany there are 22 peripheral regions. Thus, the higher speed of convergence within the periphery sample reflects the high share of Eastern regions to some extent.
15. However, the difference in the speed of convergence substantially reduces when GVA per capita instead of GVA per worker is used (see Appendix Table A4a in the Supplemental data online).
16. The test for the absence of spatially correlated residuals when allowing for a spatially lagged dependent variable cannot be rejected at the 1% level, whereas the test for the absence of spatial correlations for the dependent variable when allowing for spatially correlated residuals must be rejected.
17. The correlation between the log productivity level and its spatial lag is approximately 0.8.
18. For an analogous result on the level of the European regions, see Bouayad-Agha and Védrine (Citation2010).
19. To test the robustness of results, various estimations applying different specifications were performed. First, a different lag structure is used, exploiting one deeper lag as an instrument. Second, the number of observations per region is extended by shortening the panel period from four to two. Third, the reliability of the human capital measure is tested by weighting the in-migration flows with the similarity in the human capital structure between the origin and the destination region. In sum, those tests confirm the results of the main specifications.