Early Warning Model Of Currency Crises

Since 1990s, the foreign exchange reserves of a lot of countries reduced sharply, and the currency devalued rapidly, which are the characteristics of the monetary crises. As the emergence of the crises is more and more frequent and destroyed to the international economy, this kind of phenomenon impelled economists and governments of various countries to launch further investigation in the monetary crises theory and early warning system of monetary crises. Though the economic growth in China is steady at present, with the exchange rate system reform liberalization, and the accelerated open process of currency under the capital account, there may be some risks in crises erupting. So it is necessary to set up an early warning system of currency crises for healthy development of China's economy.Chapter one of this thesis briefly introduces the background of research. For the currency crises frequently happened in developing countries, and had influenced the economy of other developed countries seriously, there should be more attention to the setting-up of the early warning system of the monetary crises. Then this chapter stated the purpose of this paper is to establish a warning model form new angle----using the fuzzy mathematics theory.Chapter two of this paper states the research documents of monetary crises and the early warning model of monetary crises.Firstly the thesis introduces the research process of the three generations of monetary crises theory in detail. The theoretical literature on crises has flourished following Krugman's seminal paper of 1979. Initially, the literature stressed that crises were caused by weak economic fundamentals. This generation of models had well explained the crises of Mexico (1992). More recently, some papers argued that a crises might develop without a significant change in the fundamentals. In these models, economic policies were not predetermined but respond to changes in the economy. But after Asian financial crises in 1997, the economists constructed the third generation of crises models.Secondly the paper introduces the research literatures of early warning system of the monetary crises, with analysis of the prevailing five models. And then it summarizes the pluses and defects of these models in order to set up an improved model. Frankel's probit model uses discrete choice model to estimate the probability of crises. The characteristic of this model is that the dependent variable is an attribute variable, and its value is 1 when crises occurs, and its value is 0 when crises does not occurs. According to the distribution of error term, discrete choice model is divided into probit model and logit model. Kaminsky's signal approach is to monitor a set of leading indicators of exchange crises, and to predict whether crises will occur according to whether these indicators exceed their thresholds. Sachs's cross-country regression model pays attention to countries in difference areas at the same time. This model introduces two dummy variables, and they represent that the exchange reserve is high or low and the fundamental is strong or weak separately. These models well explained the first two generations of monetary crises, but still were not efficient in recent crises. With certain defects in these models, this text tries to explore a new angle to set up the early warning model of monetary crises which is different from the researches described above. Chapter three is a key part of the article, explaining the construction of the early warning comprehensive evaluation model of monetary crises based on the fuzzy theory. First of all, the thesis defines the monetary crises index formula by the index of exchange rate and foreign exchange reserve according to the definition of monetary crises. Then this paper chooses the data of China, Mexico, and South Korea to verify the validity of the formula.Then according to the principle of validity, prospective, and availability, and based on the above analysis, the paper chooses 5 indexes as the early warning system index of the crises, at the same time, utilizes the monetary crises index to verify the indexes. The early warning model index system finally covers the following five indexes: the real effective exchange rate, domestic credit expansion, trade balance, M2-to- reserves expansion, and real interest rate. The third part of this chapter constructs a new early warning system, using fuzzy theory and comprehensive evaluation method. The process covers the following steps:1, Establish the early warning index system, and then collecting the data.2, Confirm the weight of each index in the system adopting the Principal Component Analysis approach.3, Establish warning periods of each index for index appraisal. The index exceeding the standard deviation of mean value falls in the warning period. With referred to deviation range, there are three early warning blocks --safe area, early warning area, and seriously early warning area. The single index appraisal result then is collected to construct the fuzzy relation matrix of the index system, which is the comprehensive appraisal matrix of the early warning index system.4, Set up the fuzzy comprehensive appraisal model of early warning system to confirm the grade or warning period of the warning index system. The paper multiply the comprehensive appraisal matrix with index weight vector quantity, and then get the vector quantity of appraisal results for the early warning index system.5, Judge the early warning signal block of index system under the Self-Adapting principle. The thesis uses the historical data of the early warning indexes during 1990-2005 in China to examine the effect of the model. As the result, it is found to be effective for earning warning of monetary crises.Chapter four mainly concluded the advantages and limitations for the fuzzy comprehensive appraisal model of early warning system of monetary crises. According to the verification results of the model, it can be concluded that: when the early warning signal is sent out by the model, the country can adopt the appropriate retrenchment policy, with increased interest rate to prevent the flowing out of capital, and make the current account improved.The main contribution of this paper lies in innovative research approach. Because it is an uncertain incident that monetary crises takes place, so the thesis introduces the fuzzy mathematics thought, and adopts the method of comprehensive appraisal to set up early warning model.