The puzzle is obvious. Most countries, including LDCs, are likely to benefit
from investment liberalization. Even when LDCs estimate losses from entering
an agreement on investment liberalization, they are free to opt out. Neverthe-
less, we observe countries (some LDCs especially) that are against the prospect
of MAI, even when they do not consider to enter. What they do object is the
eventuality that other countries form a club with liberal rules on investment.
In the following sections we present a model in which the choice of MAI mem-
bership for the single country involves the basic trade-off identified in Markusen
(1998) and the puzzle mentioned above finds a possible explanation.
3 The Model
3.1 The World Economy
We assume a world with many countries. For simplicity, we index countries by
h on a continuum of unit measure. In the world, there is also a unit continuum
of goods, indexed by i. Labor is the only production factor, and each country
has the same endowment L. Labor is perfectly mobile across sectors or firms,
and perfectly immobile across countries. Consumers are identical everywhere.
They have Cobb-Douglas preferences described by the following utility function
where C (i) is consumption of good i. All goods are freely tradable, so, their
price is equalized worldwide.
U=
1 ln C (i)
0
di;
(1)
Production technologies are identical across sectors and countries. Each
good can be produced either out of a CRS technology by competitive firms, or
with an IRS technology by a monopolist.12 Moreover, monopolistic firms can
exploit their knowledge capital internationally by choosing the location of their
operations. Consequently, we call these firms MNEs, henceforth.
3.2 MNEs
There is a one-to-one correspondence between sectors and MNEs. So, in the
world there is a unit mass of MNEs, and the distribution of their ”headquarters”
across the world is exogenous and fixed. The share of MNEs having their home
country in country h is denoted by μh. MNEs repatriate all their profits to the
home country.
Each MNE operates from a single plant, and is required to make a plant-
level, cost reducing, investment I before starting production. Plants have a
fixed size.13 We assume that in each plant one unit of labor is turned into
12This set up is analogous to that used in Murphy, S hleifer and Vishny (1989).
13The assumption that firms’ size is given can be justified, for instance, on the ground of
monitoring and coordination costs (see, e.g., Holmstrom and Tirole (1989)).