Rewards versus Sanctions in International Relations: A Game-Theoretic Analysis of Bluffing

I use game-theoretic models to compare a sender's expected payoff across two methods of wielding influence under asymmetric information: offering rewards and threatening sanctions. Attempts to influence the actions of another can have the perverse effect of actually encouraging the behavior that one is trying to discourage by creating incentives to bluff, which may differ across foreign policy regimes. Under a reward regime, targets can bluff to extort larger inducements. Under a sanction regime, targets can bluff to deter coercion attempts. I find that while sanctions are a weakly dominant tool of influence under complete information, rewards can be preferable under incomplete information. Specifically, the sender finds promises more beneficial than threats if she expects that (i) coercion would likely fail and (ii) her willingness to offer bribes would not get exploited by the target. I examine the United States' foreign policy reversal from punishments to inducements towards North Korea during the nuclear crisis of 1993--4 to illustrate these theoretical propositions.