labor ratio and wages increase. This implies that more labor intensive productions will be adopted
in the developed economy and, symmetrically, more skilled-labor intensive techniques will
implemented in the developing one. From (6*), the stock of capital may either augment or diminish.
Hence, the long run effects on the accumulation of capital are ambiguous in both countries.
The indirect effects on the poor country concern the impact of the networks of the less
skilled migrants on πt* , and hence on w*, kt* and Kt* and on the inflow of investments from abroad.
If these effects are not significant, and if more labor intensive productions in the countries of
destination of the unskilled migration substitute for potential investments abroad, then the effect of
the unskilled migration on the developing country’s investments, both domestic and foreign may be
negative. Unskilled emigration and FDI inflows may substitute for each other.4
The model predictions are that aggregate migration positively affects FDI, but that this
effect is especially due to the presence of skilled immigrants. Besides, as the developing countries’
risk factors, or informal barriers, are higher than those of rich economies, we also expect that the
immigrants from the less developed countries have a higher impact on bilateral FDI than
immigrants from developed economies.
III. The empirical specification
The empirical specification follows the gravity model tradition. The recent literature on
gravity and FDI distinguishes between “vertical” and “horizontal” models of foreign investments.
Horizontal FDI take place between similar countries and have the purpose of selling abroad the
same goods sold at home. Vertical FDI, on the other hand, either go to less developed countries
with relative abundance of labour and natural resources or to other developed countries with relative
abundance of specialized resources. Their aim is to save in production costs. Most horizontal
investments take place among developed countries while vertical investments tend to follow the
North-South direction (Barba Navaretti and Venables, 2004). As discussed above, FDI are expected
to be influenced by the economic characteristics of the home and foreign markets, by the cultural
and institutional features of countries and also by the international networks of immigrants. Hence,
the gravity model we use includes the countries’ demand and supply factors, their institutional and
cultural characteristics, the stocks of immigrants and, for the case of Italy only, the stocks of
emigrants
4 Assuming an equal negative variation in L* for the skilled and unskilled migration, it may be objected that, as
unskilled migrants rise A* the reduction in π* required to boost the accumulation of capital (to increase the equilibrium
level of K*) is inferior to that required for the skilled migration. However, the unskilled migration may reduce the
incentives of firms of the developed countries to invest abroad, because of the reduction of the per capita income (Aka)
at home and because of the narrower wage gap with the poor countries. Besides, the proportion of unskilled migrants in
the total migration from a poor country is generally larger than that of the skilled, so the reduction in L* will in fact be
higher, with negative effects on the equilibrium returns to capital, K*.