Пи = π[O,(n-l)d (1 + (n-l)π[r,(n-l)d (21)
From (19), this can be rewritten as:
∏f7 = ∏x + γ(r∕), where: γ(r,/) ≡ π[O,(n-l)r]-/-π[r,(n-l)r] (22)
+ -
As in Section 2, the function γ(t,f ) measures the net gain from tariff-jumping.9 But now it
applies to the case where n-1 of the rival firms with which the multinational competes in the
market in question also face an access cost of t.
What about the gains from establishing extra plants? It turns out that the argument
from Section 2 continues to apply: it never pays to establish less than n plants. Just as in
Section 2, the total profits from m plants equal:
j1Frn = 1jF,m-l + (23)f} (23)
Hence, if γ(t,f ) is positive, so tariff-jumping is profitable, then profits are maximised by
establishing a plant in each union country.
Now, return to the case where the tariff rate exceeds one third. If it also exceeds t
(which equals one half), then there can be no intra-union trade whatsoever, just as in Section
2. In that case, the multinational earns duopoly profits only in the markets in which it locates
a plant. Hence, if it earns profits from one such plant, it does even better by establishing n
plants. More subtly, the same outcome arises for tariff rates between one third and one half.
In this range, the multinational enjoys the role of a duopolist in markets where it locates a
plant, and would also find it profitable to export to other markets. Its total profits are now:
9 The derivative of γ(t,f ) with respect to t is proportional to 1+(n-3)t. Even when n equals
2, this cannot be negative, since the expression for γ is only relevant for external tariffs less
than 0.5.
13