بحران ارز و مشاهده آن با ابزار فازی C

In this paper, we attempt to analyze currency crises within the decision theory framework. In this regard, we employ fuzzy system modeling with fuzzy C-means (FCM) clustering to develop perception based decision matrix. We try to build a prescriptive model in order to determine the best approximate reasoning schemas. We use the underlying behavior of the market participants during the crisis. With this analysis, we form the dictionary catalogs to construct a perception based payoff matrix. As an illustrative example, we used data from Turkish economy that covers two currency crises. The results show that market participants’ dictionary catalogs based on perception knowledge extracted from the first crisis help participants to perceive the rise in market uncertainty. When the expectations are revised accordingly a speculative attack becomes inevitable.

Since the early 1970s, the world economy witnessed an increase in frequency and severity of currency crises originated from industrial as well as from developing countries. Each passing crisis increased the variety and the frequency of currency crises and hence generated a plethora of theoretical models. In spite of their many insightful contributions, these theoretical models of currency crises cannot be considered fully successful in determining the causes and the timing of currency crises.1 In historical perspective, the dynamics of currency crises appears to be elusive due to the evolutionary nature of the market participants. While it is already difficult to understand and predict daily fluctuations in foreign exchange markets, currency crises understandably present additional difficulties. The difficulties lie in the very nature of the crises. By definition, crises’ periods are “times of turbulence and excessive volatility”. Hence the models developed to explain market behavior under ordinary periods may not serve well during those chaotic times. One may choose to refine the existing models and gather more data to increase the explanatory and predictive power of the current theoretical and empirical models of currency crisis. On the other hand, one may choose to develop newer models with the help of emerging data analysis techniques. This paper in conjunction with our previous work [30] can be considered examples of such attempts. Also, an application of support vector machines to determine the most significant factors in explaining the consequences of currency crises on the economy [32] can be considered as similar example of the usage of new data analysis technique. A particularly important aspect of FSM is its power to capture underlying behavior of historical data with proper analysis and without excessive ad hoc axioms. There are at least two advantages of FSM that attracts researchers: (i) its power of linguistic explanation with resulting ease of understanding, and (ii) its tolerance to imprecise data which provides flexibility and stability for prediction. Because of these features, FSM has been increasingly applied to problems in various areas such as computer science, system analysis, electronic engineering, pharmacology, finance and more recently social sciences (some related examples are [41], [39], [34] and [33]). To our knowledge this paper represents the first attempt to analyze currency crisis within a decision theory with an application of FSM framework. To be more precise, in our analysis of currency crises, we adopt Zadeh’s perception based decision approach [45] with an application of the rule based fuzzy system modeling. Accordingly, we attempt to capture the underlying behavior of market participants during the crisis as part of perceptions. Then we analyze how a payoff matrix can be constructed by integrating these perceptions. The rest of the paper is organized in four sections. In Section 2 we explain why we choose to use FSM to investigate currency crises. In Section 3, we obtained dictionary catalogs (fuzzy clusters) by using FSM with FCM using Turkish data from 1990 to 2002 which covers two currency crises. This section also includes the results of the model and payoff matrix construction. In Section 4 we present our conclusions.

Currency crises are the results of human actions which are the results of the human decisions. Perception is an essential part of the human decision process. Uncertainty is the inherent part of any decision problem. At any moment of a decision problem, economic agents have to form expectations about uncertain future events based on their perception which is built on the past experiences of the perceivers. One can relate some recent discussions on similarity based decision and case based reasoning in the theory of decision making under uncertainty [15] and [16] to perception based decision. In this paper, we build a decision model to analyze currency crises under perception based decision theory framework. Within this framework, we were able to link fuzzy modeling methodology to the problem of decision making under uncertainty. Our results suggest that speculative currency attack is inevitable when increased uncertainty is perceived, and the expectations are revised accordingly.