FDIit = α + β1TGDPit + β2sqjGDPDIFFit + β3PCGDPDIFFt + β4DISTi + β5OPENNESSit +
β6DUMMYEU15 + β7DUMMYOECD + β8CHRISSHAREi + β9GOVERNANCEi +
β10IMMIGRANTSit + β11EMIGRANTSit + β12Dt + ut
All variables are at time t and, except dummies and indexes, in logarithms.5 FDIit is the stock of
foreign direct investments from the country of origin to the country of destination, i is the foreign
country, ut is the error term. We use the pooled OLS method of estimation.
Following Markusen and Maskus (2002), we use the sum of the two countries’ GDP
(TGDP) as a measure of the aggregate volume of demand, and the squared difference between the
countries’ GDPs (sq_GDPDIFF) as an indicator of similarity between them. The horizontal model
is consistent with a positive coefficient of TGDP and a negative coefficient of sq_GDPDIFF. We
add the difference in countries’ per-capita GDPs (PCGDPDIFF) proposed by Gao (2003), as a
proxy for relative factor endowments differences. The vertical model of FDI predicts a positive
coefficient of the difference in the per-capita GDP of countries, while the horizontal model predicts
a negative coefficient. Given that we consider four developed economies, most of their FDI are
likely to take place with other rich economies. Hence we expect the horizontal model to prevail.
The great circle distance (DIST) between capital cities of the countries of origin and
destination of the FDI is meant to capture all the measurable and invisible transaction costs related
to travel and communication and has a negative expected sign. The base model is completed with a
measure of the commercial openness of countries (OPENNESS), which is the share of exports plus
imports in each country’s GDP. In principle, bilateral FDI and trade can be complements or
substitutes, hence no assumptions are made on the sign of this variable.
To control for cultural and institutional similarities between partner countries we augment
the base model with a set of indicators. We use a composite index(GOVERNANCE), to represent
the quality of countries’ institutions, we include the share of Christian religion in each country
(CHRISSHARE) as a proxy for religion and culture and, in the France and U.K. regressions, we add
two dummies, indicating the presence of past colonial ties (DUMMYCOLOTIE) and of a common
language (DUMMYLANG) with the foreign economies. Two other dummies indicate countries’
membership to economic and political blocks. The first refers to the European Union of 15
members (DUMMYEU15) and the second to the OECD (DUMMYOECD). They denote trade and
political agreements, but may also capture similarities among member countries that escape from
5 Since taking the logarithm would lead to having negative values for observations for which the total stock of FDI take
a value lower than unity, we add one before taking the log.