دانش عمومی و ارزش های دفاع از یک نرخ مبادله ثابت - توضیح بحران ارز

Based on a framework originally developed by Morris and Shin [Discussion paper, No. 95-24, 1995, University of Southampton], this model shows how a currency crisis may be triggered by a lack of common knowledge regarding government type. Speculators receive noisy differential information concerning the value a government places on maintaining an exchange rate parity. When this value falls in a particular region, the government will abandon the peg if a sufficient number of speculators sell their currency. However, it will maintain the peg if no attack is launched. This paper shows that, in this region, it is always optimal for the speculators to attack the currency, thereby forcing a devaluation.

This paper illustrates how informational events can trigger currency crises using a model based on that of Morris and Shin, 1995 and Morris and Shin, 1998. They show how a self fulfilling speculative attack can be launched on a currency when there is a lack of common knowledge among speculators over the state of the economy. This model extends their analysis to consider the case where the fundamentals are known by each investor. Instead it is the value placed by the government on defending a fixed exchange rate which is viewed with a degree of uncertainty. Arguably, this is a more realistic scenario since figures relating to the state of the economy are widely available. However, the government's degree of commitment to a fixed exchange rate regime is often the cause for debate. This has been seen in terms of the European exchange rate mechanism in 1992 and more recently for the East Asian economies in 1997. In each case, speculators have faced uncertainty regarding the extent to which the national government will intervene in defense of its currency. This adaptation of the Morris and Shin model assumes that the value placed by the government on maintaining an exchange rate parity is viewed by speculators with an error. It is shown that while the value may only be observed with the smallest degree of error, it can nevertheless culminate in a self fulfilling attack on the currency. Furthermore, this framework raises the issue of capital controls. In particular, the Tobin tax is shown to be effective under particular circumstances in acting as a deterrent to a speculative attack. The paper is organized as follows. Section 2 provides an overview of the framework while Section 3 outlines the mathematics of the basic model. Section 4 considers the role of information in the model. It is set up to show what would happen if all information were observed without error and then extended to consider the scenario with noisy differential information. The results of the model are discussed in Section 5 with particular emphasis given to the implications of a Tobin tax and also to the choice of distribution in the framework. The final Section contains concluding remarks.

This paper illustrates how Morris and Shin's model of currency crises may be extended to consider the case where speculators have differential information regarding government type. When there is a lack of common knowledge of the value a government places on maintaining a target rate, it becomes optimal for each speculator to attack the currency when this value lies in the ripe-for-attack region. Conversely, when differential information is observed with an error in the form of a public signal, an attack need not take place in this region. This is because the news received is common knowledge and hence, speculators can agree on how to interpret it. It is only when this differential information becomes noisy that it becomes optimal for each speculator to attack in the ripe-for-attack region. If v lies in this region, the government can avoid a crisis by imposing a Tobin tax which ensures that no attack will be optimal. It follows that, in the context of this model, there is a specific role for the Tobin tax in deterring a speculative attack. However, it has been noted that these results are dependent on the choice of probability distribution as well as the solution concept. In future research, it would be worthwhile investigating the impact of government type given a broader class of distribution. Nevertheless, it is hoped that this paper represents a positive contribution to the growing literature in the area.