Uncovered operations against the currency risk are frequent in emerging countries. Using the Obstfeld 1994's model, this paper shows in which conditions an emerging country, during a favorable period, could exit from a fixed exchange rate regime (floatation) with lower costs than those that would be incurred in the case of self-fulfilling currency crisis and in the case of capital inflows controls. The banking system is explicitly introduced in the model and the banking crisis is linked with the currency crisis. Moreover, short term capital inflows controls are considered in favorable period. The non-interest bearing deposit decreases the currency risk of the banking system. This paper is a contribution to the debate about the choice of the appropriate exchange rate regime for accession countries.