to 4). Similarly, also the coefficients of the EU and OECD dummies vary between countries and
specifications, supporting the impression that the world geographic range of our four countries’ FDI
is not homogeneous, with those of the U.K. reaching the farthest countries, including the less
developed (the two dummies are scarcely significant in both the U.K. inward and outward equations
and, as seen, the difference in per-capita incomes is positive in the U.K.’s outward equations), and
those of Italy remaining more tied to the nearest and the more developed economies (the EU
dummy is positive and significant in all specifications and the OECD dummy is mostly positive in
the outward equations of Italy).
The coefficients of the cultural and institutional variables of the augmented base model,
concerning religion and governance for the four countries, and colonial ties and language for France
and the U.K., show that, especially for the outward FDI, governance matters more than culture, i.e.
more than religion, language and colonial ties, and that the similarities with the partner countries are
significant for France and Germany, but less for the U.K.
The coefficients of the variables that interest us more, immigration (IMMIGRANTS) and, in
the case of Italy, also emigration (EMIGRANTS), reveal the more remarkable regularities.
Considering first the outward equations, Tables 2 to 5 show that the coefficients of immigration are
positive and significant in the outward equations of all countries except Italy, where, instead, is
positive and robust the coefficient of the emigration variable. This confirms our prior expectations
on the positive effects of the transnational networks on the bilateral FDI of our four countries. More
precisely, a variation of 10% of the immigrant stock leads to an increase of the outward FDI of
above 5% in the U.K., of 4,6% in France and of about 3% in Germany. A parallel increase of 10%
in the stock of emigrants corresponds to a rise of about 6,5% of the outward FDI in Italy.8
In the inward equations, a 10% increase in the stock of immigrants corresponds to a 3,7%
increase in the France and to a 1,2% increase in Germany, while the coefficient is non significant in
the U.K. In the case of Italy, an increase of 10% of the stock of emigrants corresponds to an
increase of 4,7% of the inward FDI. In both cases, of inward and outward FDI, the emigrants’
coefficients are higher than those corresponding to immigration.
The splitting of the immigrants’ populations into skilled and low-skilled leads to results that
clearly give support to our prior expectation that the networks of skilled immigrants have the
strongest effects on the countries’ foreign investments. In the outward FDI regressions, an increase
of 10% of the stocks of skilled immigrants rise the countries’ investments abroad of 15,5% for the
U.K., of 5,6% for France, of 4,5% for Germany (at the 1% confidence level). The coefficient is
positive but not significant for the Italian regressions. Hence, in three of our four economies the
8 The results on Germany are consistent with the findings of Buch et al. (2006), based on firm-level data.
14