Search

AbstractConventional VAR (Value at Risk) estimation includes historical simulation, variance/covariance, and the Monte Carlo simulation method. This study is the first to present a hybrid method of estimating VAR, combining ARMA and Neural Network. Empirical results demonstrate that the hybrid method obtained superior results to the conventional method in estimating VAR. In terms of accuracy, both the conventional and hybrid methods performed well when applied to the Chinese stock market, with the only poorly performing method being the HS method when applied to the Shanghai A share market. In terms of conservativeness, the hybrid method was superior to the conventional method, while in terms of efficiency, the hybrid method outperformed the conventional method when applied to the Shenzhen stock market. Thus, using hybrid Neural Network with the ARMA method to compare with the conventional method in estimating VAR offers certain advantages Consequently, this study suggests that investors use the hybrid method to estimate VAR.

AbstractThis paper explores a general mathematical expression for approximating the callable bond return in terms of passage of time and yield change. The callable bond can be decomposed into the contributions from the effect of call-adjusted and the effect of passage of time. The study examines each component of the approximation of the callable bond return that is sensitive to changes in the passage of time and changes in market yield. Finally, based on a numerical example, comparisons are made with other approximation approaches based on effective duration and convexity. The results indicate that the proposed approximation measurement is closer to the theoretical measurement than the effective approximation when the market yield is greater than the coupon rate.

ABSTRACTValue-at-Risk (VaR) is a tool widely used by financial institutions to report and measure market risk. There have been a great number of studies in estimating VaR. Most of which are focused on the point estimation of VaR. However, in estimating a quantity, a point estimate can be misleading, because it may or may not be close to the quantity being estimated. So we cannot know the accuracy of estimating the quantity. Confidence interval (CI) is one of the most useful manners of quantifying uncertain due to “sampling error”. Besides, the mathematics of interval estimation and hypotheses testing are closely re-lated. Hence, in practical situations, the interval estimation is more preferred than the point estimation. This paper is aimed at applying the extreme value theory (EVT) to evaluate CIs of the VaR of Taiwan weighted stock index. To assess the efficiency of the proposed method in this paper, comparisons with the methods studied in Chang et al.[9] are also made. The empirical results show that the widths of the CIs obtained by the EVT model are narrower than those obtained by Chang et al.[9]. This indicates that the EVT model is more efficient in estimating the VaR of Taiwan weighted stock index than those in Chang et al.[9].

ABSTRACTMultiple Attribute Decision Making (MADM) problems are commonly encountered in everyday aspect of life. They aim at selecting the optimal alternative among some courses of action in the presence of multiple, usually conflicting, attributes. It is not surprising that, at times, the performance rating cannot be assessed precisely. It may be represent the subjectivity and/or imprecision in human behavior. A remedial means for modeling such uncertainty is the utilization of fuzzy data. Differing from the existing studies in the area of fuzzy MADM, the main purpose of this paper is concerned with the problem of how to efficiently select the optimal alternative for an MADM problem with fuzzy data, where the decision-maker's preference information (i.e., the attributes' weights) is completely known. First, an intuitively appealing selection rule is proposed. Next, under a minimum probability of correct selection of the proposed selection rule, the optimal design (i.e., the number of the values of attributes for each alternative) is obtained by minimizing the total cost of collecting data. A numerical example is provided to illustrate the proposed method.