Immigrants’ Networks, Distance, and Trade Creation Effects: An Study Employing Province-Level Data for Italy, Spain and Portugal

Abstract

Neoclassical trade theory assumed international flows of goods (commodities) to be substituting for people (factor) flows under certain circumstances. However, recent empirical evidence shows a complementary relationship between these two types of flows, with migration creating new trade exchanges. Immigrants tend to form networks across borders, reducing fixed trade costs. They also retain some preference for their home-produced goods. These two channels provide the rationale of the immigration trade-enhancing linkage. In this study we investigate that issue for the cases of Italy, Spain and Portugal, employing province-level data for the period 2002–2010. Results show that the first channel (network channel) is the most important in this case. In addition, we observe that the larger the distance between trade partners (in terms of geography, culture, income per capita, or institutions), the bigger the trade creation effect found. All these findings are relevant for prescriptions in terms of EU Common Policies of Migration and Trade.

Appendix

1.1 Data Description

We construct a new trade-immigration database using regional data for Italy, Portugal and Spain over the period 2002–2010 using three sets of variables: (1) Bilateral exports and imports between the provinces of Italy, Portugal and Spain and a particular country; (2) Bilateral stocks of foreigners residing in a province in Italy, Portugal or Spain; (3) A number of observed characteristics at both country level and province level, including the standard gravity variables (GDP and distance) and other variables required specifically to examine the trade-migration relationship.

The database contains information on bilateral trade flows and immigration for 103 Italian provinces that existed until 2006 (the 4 provinces created after 2006 have been excluded), 18 Portuguese inland districts (the islands of Azores and Madeira have been excluded) and 50 Spanish provinces (the African territories of Ceuta and Melilla have been excluded).

• (1) Trade data: Trade data are taken from the publicly available database of the Italian Institute of Statistics (http://www.coeweb.istat.it), the Portuguese Institute of Statistics (http://www.ine.pt), and the Spanish Customs (http://www.aeat.es). Trade flows refer to the value of exports and imports of 107 Italian provinces (NUTS-III), 30 Portuguese provinces (NUTS-III) and 52 Spanish provinces (NUTS-III) with around 200 trading partners around the world. Data are measured in such a way that exports and imports are associated with the province of shipment, i.e. the province where the custom transaction was registered. Data on country bilateral trade flows are taken from UN COMTRADE in US current dollars and then import and export shares from each province are applied to scale trade flows for each province. For Portugal we have matched the 30 NUTS-III provinces with the 20 districts in the following way: 1. Lisboa (Gran Lisboa), 2. Leiria (Oeste, Pinhal Litoral), 3. Santarém (Medio Tejo, Leziria Do Tejo), 4. Setúbal (Setúbal), 5. Beja (Alentejo Litoral, Baixo Alentejo), 6. Faro (Algarve), 7. (Evora, Alentejo Central), 8. Portalegre (Alta Alentejo), 9. Castelo Branco (Cova de Beira, Beira Interior Sul, Pinhal Interior Sul), 10. Guarda (Serra de Estrella, Beira Interior Norte), 11. Coimbra (Baixo Mondego, Pinhal Interior Norte), 12. Aveiro (Entre Douro e Vouga, Baixo Vouga), 13. Viseu (Dao Lafoes), 14. Braganza (Douro), 15. Vila Real (Alto Tras os Montes), 16. Oporto (Gran Oporto, Tamega), 17. Braga (Ave, Cávado), 18. Viana do Castelo (Minho-Lima), 19. Azores (Azores), 20. (Madeira).

• (2) Immigration data: Foreign-born residents data are taken from the public available database of the Italian Institute of Statistics (http://demo.istat.it/), the Portuguese Servico de Estrangeiros e Fronteiras (Anuario de Extranjeria, Annual Report, http://sefstat.sef.pt/) and Spanish Institute of Statistics (http://www.ine.es). Data on foreign-born residents at the end of the year by province are taken from 2002 to 2010.

• (3) GDP and population: Data on country Gross Domestic Product and population are taken from the World Development Indicators, and are expressed in current US dollars and thousands, respectively. The GDP and population of Italian, Portuguese and Spanish provinces are taken from EUROSTAT and then rescaled to match the value of national GDP and population of each country, as reported in WDI.

• (4) Bilateral distance: We follow Head and Mayer (2000) to construct the distance variable between each province and each foreign country. We calculate a weighted average of the great circle distance (in kilometres) from the capital of each province to the five most important cities of each partner country, in which the weights are the respective populations of the latter. The great circle distance between i’s and j’s cities is calculated as follows. First we transform the latitude φ and the longitude λ into radians (×π/360). Second, the formula used to calculate the distance between the pair of cities is Δ _ij  ≡ λ _j  − λ _i , d _ij  = arccos[sin φ _i  sin φ _j  + cos φ _i  cos φ _j  cos Δ _ij ]z, with z = 6,367 for km. Third, we calculate the population-weighted average distance between the capital of the province and the cities of the foreign countries using the formula D _i,cou  = ∑  _{j ∈ cou} w _j d _ij , w _j  = pop _j /pop _cou .

• (5) Quality of institutions (governance): The governance indicators of the World Bank reflect the statistical compilation of responses on the quality of governance given by a large number of enterprise, citizen and expert survey respondents in industrial and developing countries, as reported by a number of survey institutes, think tanks, non-governmental organizations, and international organizations. The indicators are constructed using the unobserved components methodology described in detail in the paper of Kaufmann et al. (2010), “The Worldwide Governance Indicators: A Summary of Methodology, Data and Analytical Issues”. World Bank Policy Research. Here we use the rule of law index as a measure of the quality of institutions. Rule of law captures perceptions of the extent to which agents have confidence in and abide by the rules of society, and in particular the quality of contract enforcement, property rights, the police, and the courts, as well as the likelihood of crime and violence. The index is decreasing in the quality of institutions and stands between −2.5 and 2.5.

• (6) Cultural distance. We have created a formative index based on five of the major dimensions included in Dow and Karunaratna (2006), which are differences in language, religion, industrial development, education and degree of democracy. The specific scores for the five variables are publicly available (Dow 2010) and have been converted in to a single composite index using the same methodology as for the Hofstede index:

  $$ {\mathrm{CD}}_{\mathrm{DK}}={\Sigma}_{\mathrm{k}}{\left({\mathrm{I}}_{\mathrm{ijk}}\right)}^2/{\mathrm{V}}_{\mathrm{k}}/5 $$

  where I_ijk is the distance between countries i and j for the k^th dimension of cultural distance, and V_k is the variance of the k^th dimension of cultural distance across 120 countries.