business networks of skilled immigrants have very strong and robust effects on the foreign
investments going to their countries of origin.
The positive and strong relation between skilled networks and FDI is present also in the
inward equations, where an increase of 10% in the stock of skilled immigrants is related to an
increase in the inward FDI of 11% in the U.K., of 10% in Germany, of 8% in France and in Italy (in
all regressions at the 1% confidence level).
The relation between the unskilled immigration and FDI is negative in almost all regressions
of the four databases, both for the outward and the inward equations. The only positive coefficient,
regarding the outward FDI of France, is non significant and has a very low value (Model 3, Table
4). These results are consistent with the very high coefficients of the skilled networks variables and
reveal that the positive results of the aggregate immigration variable of Models 2 and 7 for
Germany, France and the U.K. (and Models 3 and 9 for Italy), depend entirely on the subset of the
skilled immigrants, which compensate for the negative effects of the unskilled. In terms of the
theoretical model of Section II, the negative impact of the low skilled on the host country outward
FDI can be interpreted as due to a weak influence of the less skilled on πt* , the origin country’s risk
factor, together with an impact on the productive structure of the receiving country, where more
low-skilled intensive goods ca be produced at home rather than abroad (Hanson Slaughter, 1999).9
Regarding the inward FDI, the negative relation could be interpreted as a preference of the firms of
the countries of origin of immigrants to invest where the presence of skilled nationals and business
networks is stronger. In sum, as expected, the impact of the networks of skilled immigrants on the
countries FDI is positive, robust and stronger than that of the unskilled. Besides, the networks effect
of the latter on bilateral FDI are not strong enough to compensate for the substitution effects. These
findings are consistent with the results of previous studies. In Kugler and Rapoport (2005)
unskilled immigrants in the U.S. and the country outward FDI are substitutes contemporaneously
but complements in the longer run, in Kugler and Rapoport (2006) unskilled immigration and FDI
are negatively related, especially within the group of the European countries, in Aroca and Maloney
(2005), the relation between the U.S. FDI in Mexico and the Mexican immigration in the U.S. is
negative.
9 More disaggregated data, on the sectors of employment of immigrants, would be useful to analyse these results. At the
present level of aggregation, a strict interpretation of the negative relation is that more unskilled immigrants from a
certain country have a negative impact on the FDI of the host economy that go to their country of origin. It may also be
the case, however, that the low-skilled immigration from a country has an effect on the FDI going to other sending
countries, which are similar in the relative abundance of low skilled labor. More low skilled immigration in Italy from
China may curtail the potential Italian investments in Morocco, and more low-skilled immigration from Morocco can
slower the Italian foreign investments in China. In this case the regression coefficients of the low-skilled immigration
variable would take a negative value, even if the relations of substitution are not strictly bilateral.
15