Speculative Attacks

Models with multiple equilibria are a popular way to explain
currency attacks. Morris and Shin (1998) have shown that, in the
context of those models, unique equilibria may prevail once noisy
private information is introduced. In this paper, we generalize the
results of Morris and Shin to a broader class of probability
distributions and show - using the technique of iterated
elimination of dominated strategies - that uniqueness will hold,
even if we allow for sunspots and individual uncertainty about
strategic behavior of other agents. We provide a clear exposition
of the logic of this model and we analyse the impact of
transparency on the probability of a speculative attack. For the
case of uniform distribution of noisy signals, we show that
increased transparency of government policy reduces the likelihood
of attacks.