The final hypothesis tested in this study concerns the possibility of an extra-value of the
networks linked to the developing countries. In particular, we are interested in the coefficients of
the outward equations and in the possibility that the business networks of immigrants may boost the
investments going from the rich to the poor countries. The results of this partition of the variable
immigration may be of interest in the inward equation also, but for the traditional reasons.
Regarding the outward equations, results are that the skilled non-OECD immigrants have the
strongest positive effects on the FDI to their countries of origin in the case of Italy (Model 5 of
Table 3), while their influence is positive and significant but not statistically different from that of
the OECD immigrants in the regressions regarding Germany, France and the UK (Models 4 of
Tables 2, 4 and 5). In the inward equations, the effects of the skilled non-OECD immigrants in the
U.K. regression are positive and significant (Model 10 of Table 5), while the coefficients of the two
subsets do not differ significantly in the regression regarding Italy (they are both positive and
significant at the 1% level, in Model 11 of Table 3), and are higher for the skilled- OECD
immigrants for Germany and France (Models 10 of Tables 2 and 4).
The dynamic version of the above regressions can now be taken into consideration. The
regressions include the lagged dependent variable and are useful to check the robustness of the
results of the static models. Due to the limited number of time periods covered by the databases
regarding France and the U.K. and the substantial time distance between the observations, the
dynamic regressions will be considered only for Germany and Italy. The above partition of
variables has been maintained: immigrants are split into skilled and unskilled, and the former are
interacted with the OECD and non-OECD dummies.
The results of the dynamic regressions for the outward FDI (Model 5 of Table 2 and 6 of
Table 3) confirm our expectations that the migrants’ networks tied to the developing countries have
a stronger impact on the investments abroad. In particular, a 10% increase in the stock of non-
OECD skilled immigrants in Germany has a long run effect of a 11,1% rise in the country’s
bilateral FDI. A parallel increase of 10% of the Italian emigrants in the non-OECD countries leads
to a long run increase of 6,3% of the Italian FDI in their countries of residence, while the values of
the coefficients of the migration variables to and from the OECD countries are lower and non
significant for both Germany and Italy.
The inward equations depict a different scenario (Models 10 of Table 2 and 12 of Table 3).
The coefficients of the immigrants’ variables are non significant for Germany. The regression on
the Italian data shows that a 10% increase of the emigrants in the OECD countries corresponds to a
long run rise of 5,4% in the inward FDI in Italy, from the emigrants’ countries of residence. Hence,
while the Italian emigrants living in the developing countries attract the Italian investments to their
16