knowledge of its distribution function. At this stage, firms have therefore only
to choose whether to invest in a country that belongs to MAI or in a country
that does not belong to it. MNEs may also randomize their choice, attributing
a probability p to MAI members and a probability (1 — p) to countries that are
not members. Since MNEs are ex-ante identical, this choice will be the same
for all firms. In the third stage, the bargaining power of countries is revealed to
MNEs, and investments are chosen. Finally, goods’ prices are chosen.
Three remarks are in order. First, we restrict the analysis to cases in which
the mass of labor in each country, L; is big enough not to be completely used up
by MNEs located there. Second, note that the sequence of the first two stages
is immaterial. Since countries are atomistic, they will not take into account the
effects of their own actions on the behavior of MNEs. The solution is therefore
as if countries and firms are acting together. Third, it is to be noted that in
the present set-up countries will not be willing to signal their own bargaining
power (for instance, through the use of subsidies) to MNEs.19 The reason is
that a separating equilibrium cannot be realized. The intuition runs as follows.
Subsidy competition when βh is known leads to a full transfer of FDI rents
to MNEs. Countries with low bargaining power are bound to use subsidies of
lower magnitude, if they want to break even. However, when βh is unknown,
countries with high bargaining power will surely mimic “low-beta” countries,
using low subsidies.20
The ob jective of MNEs in the first stage is to choose the location of their
plant (in a MAI or in a non-MAI country) that maximizes expected profits.
Profits at this stage can only be defined in expected value because the bargaining
power of the particular country chosen is still unknown. After having solved for
the equilibrium in the last two stages of the game, the profits ¼h of a MNE
choosing to produce in country h belonging or not belonging to MAI are as
follows
¼h = 0:5(1 — ^h)2 ; (3)
¼h = 0:5(1 — βh + °)2 : (4)
Firms’ profits decrease with βh for two reasons. The higher is countries’
bargaining power, the lower is the share of FDI profits repatriated by MNEs,
and the lower is the amount of investment undertaken. This explains why any
increase in βh produces a more than proportional reduction in ¼h. At given βh ,
the participation in MAI is beneficial for MNEs because of the same reasons:
larger repatriation of FDI profits and more abundant investment. Firms observe
ex-ante MAI membership of all countries and know the distribution of βh across
19Bond and S amuelson (1986) find an opposite result. In that paper, firms are not informed
ab out a productivity parameter characterizing countries. Countries with high productivity can
then signal through subsidies and tax-holidays their superior business environment without
fear of b eing jeopardized by low-productivity countries: a separating equilibrium exists.
20Note also that the only pooling equilibrium is with zero subsidies. The reason is trivial:
with positive subsidies countries will only transfer resources to MNEs without affecting the
probability of receiving FDIs.
11