A Study on Causality Relationships Among the Leading Indicators of Currency Crisis─An Evidence from Thailand

• Main Authors: Po-Wen Hsueh, 薛博文
• Other Authors: Chin-Shien Lin
• Format: Others
• Language: zh-TW
• Published: 2004
• Online Access: http://ndltd.ncl.edu.tw/handle/39230137662397766230

碩士 === 靜宜大學 === 會計學系研究所 === 92 === Although currency crises have impacted economics seriously in many countries since 1970, the authorities concerned still do not know what the proper and timely hedging policies are according to the existing economic theories and early warning models. Based on the empirical results of Kaminsky and Reinhart (1996, 1999), we select the leading indicators as our explanatory variables of currency crisis. In addition to applying vector autoregressive model, we are also trying to apply path model to study the causal relationships among the variables in order to know how the variables affected each other in depth during the crisis. As Thailand was the first Asian country encountered currency crisis in 1997, this research is trying to explain how the currency crisis happened in this country by using its monthly data. The empirical results are summarized as follows: 1. The deviation of the ratio of domestic credit to GDP has a significant positive effect on both exchange market pressure (EMP) and the deviation of bank deposit, but a significant negative impact on the deviation of the international reserves. 2. The deviation of bank deposit has a significant positive influence on EMP, but a significant negative effect on the deviation of the international reserves. 3. The deviation of the international reserves has a significant effect on EMP in positive direction. 4. Among the leading indicators, the deviation of the ratio of domestic credit to GDP has the most intense influence on EMP and the strongest exogeneity relative to the other indicators. 5. The goodness-of-fit test, GFI AGFI, RMR, RMSEA and CFI, assured us that the path model obtained can represent the relationship among the variables to a significant level of extent.