|
terms of trade effect, tracing it to the demands made of rich governments by export-
platform FDI. Importantly, however, she offers two reasons to suspect that both
investment treaties and trade agreements play the roles we posit, explaining that this
effect “cannot persist under threat of expropriation,” and second, it can be undermined by
measures that discriminate between the country’s own exporters operating abroad, and
the developing country’s exporters (2005, 34-35). Our point is that a BIT gets at the first
concern, and a PTA at the second.
Two further points bear mentioning. First, our hypothesis that a pair of countries
with a BIT is more likely to sign a PTA is specific to North-South investment and trade.
BITs are signed by pairs of rich and poor states; pairs of rich states seldom worry about
expropriation, and sometimes embed more technical investment obligations in special
trade provisions, whereas pairs of poor states are usually ill-suited, given their factor
endowments, to benefit from entering into investment (Egger, Larch and Pfaffermayr
2007) or trade (Venables 1999; Berthelon 2004; Mayda and Steinberg 2006) agreements
with each other.
Second, we argue that having many BITs with other wealthy states does not boost
the odds that a developing country will sign a PTA with a developed one. A growing
6
number of studies stress the role of international influences on the propensity for states to
join investment treaties. For example, Elkins, Guzman and Simmons (2006) show that
developing countries are more likely to sign a BIT if competing jurisdictions have
already done so. Likewise, Egger and Larch (2006) observe that countries are more
likely to join a PTA if neighboring countries are members as well. Blonigen et al. (2005)
note that these spatial interdependencies are also evident in FDI decisions, although their
magnitude varies with the sample of countries studied. These international influences are
undoubtedly important, but they usually underpin the conjecture that more FDI (Büthe
and Milner 2006), or more trade (Ingram, Robinson and Busch 2005; Goldstein, Rivers
and Tomz 2007), follows from more institutional memberships. We dissent from this
view, at least with respect to explaining the likelihood of signing North-South PTAs.
We contend that having many BITs congests the special market access conferred
by a trade agreement, making a PTA less preferential for the developed country. This is
likely to undermine political support in the wealthy country for pursuing a trade agree-
ment. This logic is well rehearsed in economics, but has received scant attention from
political scientists, and almost no empirical scrutiny in either field. As Grossman and
Helpman (1995) explain, industries have incentives to support PTAs that are trade divert-
ing, since these promise private gain, making them more politically viable for govern-
ments to pursue. Krishna (1998) arrives at the same conclusion, observing that trade-
diverting PTAs create rents for producers that are tied to the agreement’s preferences, and
warns that, unless liberalization provides sufficient rents to offset the loss of these prefer-
ences, producers are unlikely to support multilateral trade talks. Even in the case of
PTAs offered for non-economic reasons (i.e., rewarding an ally), maintaining higher
7
preferences is crucial to the bargaining power of the country granting a trade pact, the
upshot being that these, too, are subject to congestion, which is why they are just as likely
to impede multilateral liberalization (Limão 2002). Our argument about having multiple
BITs follows this same logic: some investment will substitute for trade, while inflows of
FDI will increase competition in the economy more generally, congesting exporter rents
and undermining the (preferential) benefits to be had from signing a PTA.
Our argument can thus be summarized as follows:
Hypothesis 1: If a developing country has a BIT with a developed country, the two states
are more likely to sign a PTA
Hypothesis 2: If a developing country has relatively few BITs with other wealthy states,
it is more likely to sign a PTA with a given developed country, but if it has many BITs
with other wealthy states, it is less likely to sign a PTA with a given developed country
III. Research Design
We estimate a model of the determinants of PTAs, focusing on the explanatory
leverage of BITs, in particular. While there is no generalizable model of PTA formation,
our specification follows from our argument and includes controls identified in the litera-
ture as being especially influential in this regard.
Our base model takes the following form: a PTA (PTA) depends on a BIT in force
between the country pairs (BIT), dyad-specific characteristics (d), host country-specific
characteristics (z), regional characteristics (r),
fixed time effects (τ), and error term (ε).
8
)
1
(
,
4
,
3
,
2
2
,
1
,
t
ij
t
i
t
i
t
ij
t
ij
o
t
ij
BIT
PTA
ε
τ
+
+
Β
+
Β
+
Β
+
Β
+
Β
=
−
Do'stlaringiz bilan baham: | 
