Do Institutions Play a Role in Skilled Migration? The Case of Italy

Abstract

Nifo A. and Vecchione G. Do institutions play a role in skilled migration? The case of Italy, Regional Studies. This paper is based on the idea that the migration choices of the highly skilled are driven by the search for favourable socio-economic contexts able to ensure higher income, better employment opportunities but also better institutions. By using data on a sample of 47 300 Italian graduates, the impact of provincial institution quality on the probability of resident graduates migrating is studied. A Heckman probit estimation indicates that institutions do matter for migration decisions and their importance is comparable with that of per capita income provincial differences.

Nifo A. and Vecchione G. 制度是否在技术移民中扮演一定的角色？意大利的案例，区域研究。本文以下列概念为基础：高技术移民的选择除了由追求能够确保更高收入的更佳社会经济脉络，以及更好的就业机会所驱动之外，亦为了追求更佳的制度。本研究运用47, 300 位意大利毕业生的样本资料，研究省层级的制度品质，对于居住该地的毕业生的迁徙可能性所具有的影响。海克曼机率（Heckman Probit）评估指出，制度对于迁徙决策而言的确有所影响，且其重要性可以与各省的个人平均所得差异相比拟。

Nifo A. et Vecchione G. Les institutions, est-ce qu'elles ont un rôle à jouer dans la migration des travailleurs qualifiés?: étude de cas de l'Italie, Regional Studies. Cet article est fondé sur la notion que les choix de migration des travailleurs hautement qualifiés sont motivés par la recherche de milieux socio-économiques favorables, susceptibles de garantir non seulement des revenus plus élevés et de meilleures possibilités d'emploi mais aussi de meilleures institutions. À partir des données auprès d'un échantillon de 47.300 diplômés italiens, on étudie l'impact de la qualité des institutions provinciales sur la probabilité que les diplômés résidants vont migrer. Une estimation probit d'Heckman indique que les institutions jouent un rôle important en ce qui concerne les décisions liées à la migration et que leur importance est comparable à celle des écarts de revenu provinciaux par tête.

Nifo A. und Vecchione G. Spielen Institutionen für die Migration von Fachkräften eine Rolle? Der Fall Italien, Regional Studies. Dieser Beitrag beruht auf der Idee, dass die Migrationsentscheidungen von hochqualifizierten Fachkräften durch die Suche nach günstigen sozioökonomischen Kontexten, die ein höheres Einkommen gewährleisten, nach besseren Berufschancen, aber auch nach besseren Institutionen motiviert werden. Anhand der Daten einer Stichprobe von 47300 italienischen Hochschulabsolventen untersuchen wir die Auswirkung der Qualität von provinziellen Institutionen auf die Wahrscheinlichkeit einer Migration der ansässigen Hochschulabsolventen. Aus einer Heckman-Probit-Schätzung geht hervor, dass Institutionen für Migrationsentscheidungen in der Tat eine Rolle spielen und dass sich ihre Bedeutung mit der von den Unterschieden zwischen dem Pro-Kopf-Einkommen in den einzelnen Provinzen vergleichen lässt.

Nifo A. y Vecchione G. ¿Revisten importancia las instituciones en la migración de personal cualificado? El caso de Italia, Regional Studies. En este artículo nos basamos en la idea de que los trabajadores altamente cualificados deciden emigrar para buscar contextos socioeconómicos favorables que les garanticen ingresos más altos y mejores oportunidades laborales, pero también mejores instituciones. A partir de los datos de una muestra de 47.300 licenciados italianos, analizamos el impacto de la calidad de las instituciones provinciales en cuanto a la probabilidad de que los licenciados residentes opten por emigrar. Una estimación según el modelo Heckman Probit indica que las instituciones son importantes para las decisiones migratorias y que su importancia es comparable a la de las diferencias entre los niveles de renta per cápita en las provincias.

• Skilled migration
• Institutional quality
• Southern Italy

• 高技术移民
• 制度品质
• 南意大利

• Migration des travailleurs qualifiés
• Qualité institutionnelle
• Italie du Sud

• Migration von Fachkräften
• Qualität von Institutionen
• Süditalien

• Migración de personal cualificado
• Calidad institucional
• Sur de Italia

JEL classifications:

• J24
• O15

Acknowledgements

The authors gratefully thank three anonymous referees; Andrea Filippetti, Leonid Polishchuk and Hillel Rapoport; and the participants at the Centre for Research and Analysis of Migration (CReAM), University College London, London, UK, 10–13 April 2013; and the International Society for New Institutional Economics (ISNIE), Florence, Italy, 20–22 June 2013, conferences for useful suggestions. Special thanks are due to Domenico Scalera who reviewed an earlier version of this paper.

Notes

1. By Mezzogiorno is meant all the regions in Southern Italy, including the islands, namely Abruzzo, Molise, Campania, Puglia, Basilicata, Calabria, Sicily and Sardinia.

2. A common approach to the determinants of migration is the one denominated ‘push and pull approach’: on the one hand, push factors, i.e. the characteristics of the country of origin that drive the individual to emigrate; and, on the other hand, pull factors, i.e. that set of opportunities offered by the destination region attracting workers (Kwok and Leland, 1982; Pissarides and McMaster, 1990; Miyagiwa, 1991; Haque and Kim, 1995; Dustmann, 1999; Güngör and Tansel, 2009).

3. This approach is consistent with the central idea of human capital theory (Schultz, 1961; Becker, 1964; Mincer, 1984): building capacities through education bear costs and benefits. These latter are modelled to accrue mostly in the future in the form of higher earnings (Marinelli, 2011).

4. This also explains the observed tendency of skilled workers to concentrate geographically in urban richer and more innovative areas (Ritsila and Ovaskainen, 2001; Giannetti, 2001, 2003; Florida, 2002a, 2002b; Pekkala, 2003; Rutten and Gelissen, 2008).

5. To keep the same hierarchy framework as the WGI, a weight assignment technique called the analytic hierarchy process (AHP) is used, which gives the possibility to adopt a multilayer preconceived framework (in the present case the hierarchy framework used by the WGI).

6. The degree of openness of the economy is derived as the sum of imports and exports for each province. Data are drawn from the dataset of Istituto Tagliacarne (2001).

7. The Italian term used for this practice is commissariamento. When there is substantive evidence of administrative mishandling or Mafia infiltration, the local administration is dissolved and a commissario or special commissioner is appointed by the Ministry of the Interior to run the administrative unit in question.

8. The ISTAT survey on graduate employment refers to those who graduated in 2004 and were interviewed in 2007.

9. The AHP technique is widely used in a multi-criteria decision environment (Malczewski, 1999). Applications of AHP are also reported in numerous fields such as supply chain management (Akarte et al., 2001), healthcare (Li et al., 2008), manufacturing (Wang et al., 2004), experimental economics (Ishizaka et al., 2010), and many others. An advantage of AHP is in the possibility it gives to adopt from the beginning a preconceived framework (in this case the same hierarchy framework used by the WGI). In addition, it allows one to exploit verbal judgments on the relative importance of elementary indexes made by researchers or panellists, by supplying a procedure to transform these judgments in a vector of weights.

10. For example, to assess the relative importance of the elementary index ‘Crime against the public administration’ with respect to ‘Golden–Picci Index’ and ‘Special commissioners’ in determining the ‘Corruption’ dimension, AHP starts from pairwise comparisons between each of these indexes and each of the other two (that is, ‘Crime against the public administration’ versus ‘Golden–Picci Index’; ‘Crime against the public administration’ versus ‘Special commissioners’; ‘Golden–Picci Index’ versus ‘Special commissioners’).

11. Classifications of each dimension are available from the authors upon request: (1) Voice and accountability, (2) Government effectiveness, (3) Regulatory quality, (4) Rule of law, and (5) Control and corruption.

12. The survey was conducted periodically with the computer-assisted telephone interviewing (CATI) technique, administering a detailed questionnaire to a large number of recently graduating individuals (in this case the survey was in 2007 and concerned 47 300 individuals graduating in 2004). The interviewed sample represents about 18% of the reference population (260 070 graduates in 2004; Student Registration Service, Ministero dell'Università e della Ricerca (MIUR)). Data processing was carried out at the ADELE ISTAT laboratory in Naples in compliance with regulations on personal data protection. The results and opinions expressed are the exclusive responsibility of the authors and do not constitute official statistics.

13. Measuring the migration phenomenon with changes in residence leads to a very cautious estimate of the actual number of migrations. In practice, there are many cases in which the migrant maintains residency, at least for a certain number of years, in his/her area of origin.

14. As illustrated above, IQI is expressed at a provincial scale. The regional values were determined by appropriately weighting for the resident population.

15. The procedure à la Heckman assumes that the errors of the two equations are normally distributed with zero mean and variance, and are correlated among themselves:independent of the set of covariates X and Z. It is possible to test the null hypothesis that the two errors are not correlated: with a specific Wald test. Rejecting the hypothesis H ₀ of zero correlation, it can be stated that in the model there is no problem of self-selection and the estimators are not biased. Finally, for the goodness of the estimates, as suggested by Heckman (1979), it is necessary that in the selection equation there is at least a variable included in X_kj and not present in Z_kj of the first-stage equation.

16. The results of the first-stage equation are available from the authors upon request.

17. In all, 3456 (7.3%) students in the ISTAT sample participated in the Erasmus project, while 43 844 (92.7%) went through a regular degree course.

18. In this specification the effect of the unemployment rate variable in the province of origin loses significance. Instead, there are no major changes as regards the other unemployment rate, GDP, and net monthly salary in the provinces of origin and destination.

19. The same work was carried out by drawing up different scenarios in the context of the weights assigned to the elementary indexes. Despite the different proposals, no notable variations were detected when considering the indirect impact of these elementary indexes on the final index. In practice, the investigations confirmed that the real source of variation in the final index is in the weights of dimensions which, for the way the model is constructed, have a direct impact on the same index.