3
(2) the success of failure of attempted facilitation, and (3) the superpower’s decision
concerning whether or not to coerce.
In this paper, I find evidence that mediator bias serves as one of the significant
elements that distinguishes between a coercive process, in which the superpower chooses
coercion if facilitation fails, and a facilitated one, in which the superpower chooses not to
adversaries, true facilitation is likely because no motivation exists to secure a better settlement
for either adversary. But when the mediator is biased, that is, when its preferences are more
closely aligned with a single disputant’s preferences for a particular arrangement, the parties
are aware that coercion may be the superpower’s response to a failed facilitation effort.
Therefore, the disfavored party is more likely to settle when it can reasonably assume the
superpower is resolved to employ coercion if the negotiations fail.
This paper proceeds as follows. In the first section, I further develop the puzzle and
clarify my argument. In sections two and three, I present an empirical test designed to answer
whether a superpower’s affinity for actors enmeshed in a foreign conflict affects the nature of
its participation in the conflict. Analyzing U.S. involvement in 179 bilateral crises from 1945
to 1990
, I find evidence that successful facilitation is most likely when the mediator favors a
single crisis actor and considers the dispute sufficiently salient. Since these are also the factors
I postulate can influence a superpower’s decision to coerce a settlement, I argue coercion and
mediation are not substitute choices available to a superpower, but rather two consecutive
steps in an essentially coercive process. In this process, facilitation success is more likely
when crisis actors have a rational expectation that the superpower will resort to coercion if
Rosegrant 2001).
5
I solve a formal model of mediation in section three, demonstrating that facilitation always succeeds when the
United States chooses coercion off the equilibrium path.