Self-fulfilling currency attacks with biased signals

S. Morris and H.S. Shin [(1998). Unique equilibrium in a model of self-fulfilling currency attacks. American Economic Review, 88(3), 587–597] have shown in a model of self-fulfilling currency crises that when removing the assumption of common knowledge among speculators multiplicity of equilibria disappears. In their model, speculators are assumed to receive unbiased private signals about the true state of the fundamentals of the economy. We consider, instead, three different cases: (a) signals are positively biased but speculators are not aware of the bias; (b) the private sector believes that signals are positively biased, although they are not; (c) in presence of a positive bias, speculators are aware of it. While case (c) does not modify substantially Morris and Shin's result, cases (a) and (b) produce respectively a ‘credibility bonus’ and a ‘non-credibility malus’, also obtained in the literature on credibility. A whole range of equilibria then becomes possible, depending not only on the size of the standard error in the distribution of the signal, but also on the size and sign of the bias and, more interestingly, on the expectations of the private sector about the bias, so that a currency crisis might take place even when fundamentals are ‘strong’. Given the potentially destabilizing role of expectations, our results suggest that the introduction of corrective economic policy measures like the imposition of capital controls or of transaction taxes, can be appropriate, at least under this respect, and can work as an insurance device so as to avoid that speculative attacks may take place even when fundamentals are ‘strong’.