مشکلات اساسی رویکردهای مبتنی بر فشار بازار ارز به شناسایی بحران ارز
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|24940||2008||21 صفحه PDF||سفارش دهید||محاسبه نشده|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : International Review of Economics & Finance, Volume 17, Issue 3, 2008, Pages 345–365
This study seeks to demonstrate that the identification of crisis episodes based on commonly applied exchange market pressure (EMP) indices, namely, Eichengreen, Rose and Wyplosz [Eichengreen, B., Rose, A., and Wyplosz, C., 1995, Exchange Market Mayhem: The Antecedents and Aftermaths of Speculative Attacks, Economic Policy 21 (October), 249–312.], Sachs, Tornell and Velasco [Sachs, J.D., Tornel, A., and Velasco, A., 1996, Financial Crises in Emerging Markets: The Lessons From 1995, Brooking Papers on Economic Activity 1, 147–215.], and Kaminsky, Lizondo and Reinhart [Kaminsky, G., Lizondo, S., and Reinhart, C., 1998, Leading Indicators of Currency Crises, IMF Staff Paper 45, 1 (March), 1–48.] are highly sensitive to the choice of: a) the weighting scheme for each component of the EMP index; and b) the statistical parametric assumption used in the constructions of crisis thresholds. To highlight further some of the potential consequences of these two pitfalls in identifying crisis episodes, this paper employs a number of possible alternative approaches to measure the exchange market pressure.
One of the primary tasks of recent research on currency crises is to construct a single composite index that will systematically identify the presence and severity of currency crises or speculative attacks on a currency. Studies such as Eichengreen et al., 1995 and Eichengreen et al., 1996—henceforth ERW, Sachs, Tornell, and Velasco (1996)—henceforth STV, and Kaminsky, Lizondo, and Reinhart (1998), Kaminsky and Reinhart (1999)—henceforth KLR—have proposed different constructions of what is now known as exchange market pressure (EMP) index. This indicator is usually a weighted average of the rate of depreciation of the local currency (mostly against the US dollar in either nominal or real terms), the monthly percentage changes in international reserves, and the monthly changes in the interest rate. 1 The objective of this paper is neither to construct another crises index, nor to conduct another study on the determinants of crises using the different approaches listed in Fig. 1 (discrete, signaling and structural approach). We also do not intend to argue as to which among of the above EMP indices are the most powerful or accurate indicator of exchange market pressures. What we, instead, seek to demonstrate in this paper is that the identification of crisis episodes is highly sensitive to choice of the followings: a) the weighting scheme used for each component of the EMP index; and b) the statistical parametric assumption used in the construction of crisis thresholds.To help illustrate our arguments, the exchange market pressure indices of ERW, STV and KLR are constructed for two groups of countries, namely, the Latin American countries (Argentina, Brazil, Chile and Mexico) and the East Asian countries (Indonesia, Korea, Malaysia, Philippines, Singapore and Thailand) for the period of 1985–2003. As in previous studies, we adopt various conventional thresholds used in the literature, i.e. based on a certain arbitrary number of standard deviations (usually ranging from 1.5 to 3 standard deviations) above the mean of the EMP index.3 An important point to note as far as conventional thresholds is that it implicitly assumes that the EMP indices are normally distributed. However, based on the examination of their basic statistical properties, none of the EMP indices according to the groups of countries examined here are normally distributed. In view of this, to deal with the non-normality, we apply Extreme Value Theory (EVT) based, in particular, to the tail index estimator proposed by Huisman, Koedijk, Kool, and Palm (2001)—henceforth HKKP.4 The HKKP is employed in this study in order to avoid the potential problem with small samples especially encountered when calculating the tail index parameter using the Hill estimator (1975).5 Among a number of contrasting outcomes, we find that the use of three different weighting schemes for each component of the EMP indices leads to variations or differences in identified crisis episodes for all the countries examined, and this outcome holds true irrespective of whether one uses the conventional thresholds or the EVT-HKKP. In other words, in spite of the fact that the components of the EMP indices deals basically with the same macroeconomic variables, the three EMP indices capture different episodes of speculative attacks. This is one clear piece of evidence that weighs against the validity of using arbitrary weights in the construction of any EMP index. An alternative method, which assumes away any question on the validity of arbitrary weights in an EMP index is by Moreno (1995). The basic approach here is to sequentially consider the crisis thresholds for each individual component of the EMP index. We repeat the same procedures of identifying currency crisis episodes as introduced by Moreno (1995), and compare the findings with those of the conventional thresholds as well as that of the EVT-HKKP. Although the work of Moreno (1995) addressed the pitfall of choosing arbitrary weights inherent in an EMP index, the method still, however, relied on the conventional parametric assumption to calculate its crisis thresholds. In an attempt to simultaneously deal with the arbitrary weighting scheme and the conventional parametric assumption, the Multivariate Markov-Switching VAR (MS-VAR) developed by Krolzig (1997) is applied to the same group of countries mentioned earlier.6 The outline of the paper is as follows. Section 2 will briefly review the basic constructions of the three most commonly used exchange market pressure indices (i.e. Eichengreen et al., 1995, Eichengreen et al., 1996, Kaminsky et al., 1998, March, Kaminsky and Reinhart, 1999 and Sachs et al., 1996; ). A brief introduction of the Extreme Value Theory and the HKKP estimator as well as the empirical results using this approach will be discussed in Section 3. The application of the Moreno approach will then follow in Section 4. Section 5 presents and analyses the results using the MSVAR. The paper ends with a brief concluding section (Section 6).
نتیجه گیری انگلیسی
The last two decades have witnessed an increasing frequency of speculative attacks and currency crises in both developing and developed countries. Perhaps due to further acceleration in the openings of the capital accounts in various parts of the world as well as the intensification in integration of global financial markets, which were greatly facilitated by the increased appetites of countries to economic reforms and the vast improvement in technology infrastructures, uncertainties and recurrent volatilities in the financial and currency markets will likely continue to rise and to pose further challenges for policy makers. Efforts in arriving at reliable leading indicators of extreme market pressures, have visibly become more active during the recent years, especially more so in the aftermath of the 1997–1998 East Asian crisis. This study hopes to contribute in this effort by highlighting a number of key findings. First, two fundamental shortcomings of the commonly applied EMP indices, namely, the ERW, STV and KLR, need to be highlighted here again. The first one is with the arbitrary weighting schemes used in the construction of such EMP indices, and the other is with the empirical validity of the statistical parametric assumption used in the construction of conventional thresholds for identifying crises. Second, the identification of the extreme market pressures is also highly sensitive to the techniques employed. In addition to the standard usage of EMP-based indices of ERW, STV and KLR, this study employs a number of alternative approaches, namely, the Moreno approach, the Extreme Value Theory-HKKP technique and the MS-VAR in identifying incidences and episodes of currency crisis. We find that the results vary from one approach to another. One important lesson that comes out of these conclusions is that there is no single approach that can be clearly considered as far superior to other alternatives. This begets a not so encouraging point that the second best option is to consider different constructions of an exchange market pressure (weighted and/or unweighted) and using the various available tools or techniques in identifying currency crisis episodes. More importantly, the identification of different episodes of currency crises should significantly influence the outcomes of the studies on early warning systems, and may provide explanation as to the not so desirable or mixed performance of these models of early warning systems