dc.description.abstract | Over the last four decades the exchange rate has had an important role on economies. The exchange rate is so volatile that the prediction of it is challenging. Also, there have been several speculative attacks that have caused currency, financial and political crises in several countries. Policy makers had always been late to react against these crises in order to reduce the detrimental effects of them. Therefore, economists have tried to predict currency crises to use as a prior indicator of potential crises. Frankel and Rose (1996) developed the currency crush index. However, this index does not capture unsuccessful speculative attacks due to the construction. So, Girton and Roper (here after GR) developed the exchange market pressure (EMP) index. They criticize some studies that just focus on intervention, while others try to explain foreign exchange rate market (forex). They combine intervention and forex by applying ``the monetary approach to balance of payment''. GR added a change in international reserves and exchange rate appreciation, which is called the EMP index. Now, the index has the ability to capture unsuccessful speculative attacks because it includes a change in international reserves. Hence, if money authorities intervene in the forex market, then the index reveals this involvement. After two decades from GR studies, Eichgreen, Ross and Wyploz (1994, 1995,1996) came to argue that monetary variables do not have a power over exchange rate to explain. They constructed a model independent index, which is called precision index or signaling approach. Moreover, they claimed that money authorities have extensively used policy rate that indirectly effects the exchange rate. This is because a change in interest rate absorbs some excess demand which causes that the depreciation rate to be less than what it would have been. Even if the model independent EMP index is widely used, this precision weighting has some disadvantages because the weight cannot have connection with economy, so it can not be interpreted based on fundamentals. Hence, some economists prefer the model dependent index which is called the ``elasticity approach'', which allows economists make link this weight with fundamentals of an economy. Hence, in my dissertation, I choose model dependent index because economists should make their conclusion and advice according to the economic model. Also, in this approach, it is not required to determine threshold in order to identify economic and politic crises. But, a certain cut off must be set in the model independent index. Generally, this cut off point is defined arbitrarily by economists, and they explain why they prefer this point based on their interpretation. Since the model dependent EMP index is employed in this dissertation, interpretations are based on economic variables and models. A model dependent index is employed in this dissertation, and coefficients of the model is estimated by structural vector auto regression model (SVAR) in the second chapter. The SVAR has many advantages. First of all, we use all data in the estimation. The estimation method is full information maximum likelihood (FIML). Also, the endogeneity problem may be overcame by applying this approach. The second advantage is that contemporaneous matrix allow coefficient to be estimated simultaneously. Since exchange rate, international reserves, interest rate and domestic credit expansion influence each other contemporaneously, the coefficients in the model should be estimated at the same time. To sum up our findings in this dissertation, the weights for international reserve and interest rate differential includes more coefficients. The coefficients of the domestic credit expansion equation directly affects the index. The rational expectations is considered in the model and estimation. By imposing the cross equation restriction, the estimation method now has an ability to expose the effect of the rational expectation. Moreover, even though many studies impose either the PPP or the UIP holds, those assumptions are relaxed in this chapter. In the SVAR estimation, many coefficients are statistically significant. The coefficients’ signs match up with expectations according to economic theories. Furthermore, the effect of the exchange rate over the international reserve is relatively big when we compare it to other coefficients. This result shows how important a variable is in the empirical model. When the EMP index is calculating, it is clearly seen that the elasticity of the international reserve with respect to exchange rate is close to zero. Generally, it is found close to three. But, the method has an ability capture three important points, which may cause lower elasticity. | |