(10- = (n-l)π,(τ) (10)
dτ
This is unambiguously negative: lower internal tariffs raise the profits on intra-union exports
to each of the n-1 partner countries, making an export platform more attractive. By contrast,
it is clear from (1) and (6) that the profits from exporting and from establishing n union
plants are unaffected by internal trade liberalisation. Putting this differently, χ(t,τ), and hence
Γ(t,τ,f ), is decreasing in τ; whereas γ(τ,f ) is increasing in τ. Hence, as τ falls, FDI with a
single plant becomes more attractive relative to both other options.
Figure 2 illustrates. Relative to Figure 1 (the loci from which are indicated by dotted
lines), the reduction in internal tariffs induces a new region F1 in which foreign direct
investment with only a single plant is profitable. This region corresponds to high but not
prohibitive fixed costs and any level of external tariffs.4 It has emerged at the expense of
all three regions in Figure 1, as indicated by the arrows. Note one interesting sub-region in
particular, denoted by shading, where regime O has been replaced by regime F1. For
combinations of f and t in this region, the union market is not served at all when internal and
external tariffs are equal. However, lowering τ below t encourages the firm to jump to FDI,
without ever passing through a phase of exporting. As the figure shows, this region applies
to firms which face prohibitive external trade costs and which have relatively high fixed costs.
2.3 No Internal Tariffs
Finally, consider what happens when internal tariffs are abolished altogether. Foreign
direct investment is now even more attractive relative to exporting or not serving the union
4 Since t is greater than τ by assumption, the region to the left of τ in Figure 2 can be
ignored.