resistance, thereby receiving W
BW
which is greater than V
BW
= 0. In weak sanctions,
the non-allied targets prefer non-compliance despite the observable messages that the
consequence of non-compliance would be deadlock. In strong sanctions, however, all
parameters related to the target’s deadlock payo¤ (W
BS
) are insigni…cant. Then,
the target becomes almost indi¤erent between compliance (C
B
= 0) and deadlock
(W
BS
0). This con…rms the previous …nding that the predicted probabilities of
compliance and deadlock in strong sanctions between non-allied dyads are similar.
In sum, the Dyadic Relationship Hypothesis is con…rmed by the results. Whether
or not the target is an ally signi…cantly a¤ects the sender’s decisions. The sender is
less likely to extend the sanctions against her allies. Moreover, the target’s decision to
either comply or resist is di¤erent from what the conventional model of crisis bargaining
would expect. This …nding gives valuable information for understanding the way in
which sanctions fail. While the allied targets tend to comply even when they can
win the sanctions, the non-allied targets tend to resist even when the sender credibly
signals that she would continue the sanctions.
Two important implications can be derived from the dyadic relationship analysis:
From the sender’s perspective, economic coercion against a closely allied target would
produce more favorable outcomes than one would expect in a general crisis bargain-
ing situation. However, the sender would need to …nd coercion measures other than
economic sanctions when she is confronting an adversarial (or even neutral) target.
Despite the expectation that the sender would not lift the sanctions if the target re-
sists, the target is less likely to concede. Sanctions are mostly likely to fail in these
cases. Even when the sender imposes strong sanctions and demonstrate her willingness
to continue them, success is far from guaranteed. More importantly, sanctions take
places disproportionately between non-allied dyads (66% of all sanctions cases). Since
those are predominantly likely to fail, we can begin to understand why sanctions have
such a low success rate.
from zero. For the sake of argument here, I put zero for that part of payo¤. Second, since institution1
lies between -0.9 and 1, the minimum value of W
BW
is 0.2201, whereas the maximum value is 0.7065.
27
Convention