International Journal of Information Technology & Decision Making

This paper first reviews empirical evidence and estimation methods of structural credit risk models. Next, an empirical investigation of the performance of default prediction under the down-and-out barrier option framework is provided. In the literature review, a brief overview of the structural credit risk models is provided. Empirical investigations in extant literature papers are described in some detail, and their results are summarized in terms of subject and estimation method adopted in each paper. Current estimation methods and their drawbacks are discussed in detail. In our empirical investigation, we adopt the Maximum Likelihood Estimation method proposed by Duan [Mathematical Finance10 (1994) 461–462]. This method has been shown by Ericsson and Reneby [Journal of Business78 (2005) 707–735] through simulation experiments to be superior to the volatility restriction approach commonly adopted in the literature. Our empirical results surprisingly show that the simple Merton model outperforms the Brockman and Turtle [Journal of Financial Economics67 (2003) 511–529] model in default prediction. The inferior performance of the Brockman and Turtle model may be the result of its unreasonable assumption of the flat barrier.

Credit scoring models are very important tools for financial institutions to make credit granting decisions. In the last few decades, many quantitative methods have been used for the development of credit scoring models with focus on maximizing classification accuracy. This paper proposes the credit scoring models with the area under receiver operating characteristics curve (AUC) maximization based on the new emerged support vector machines (SVM) techniques. Three main SVM models with different features weighted strategies are discussed. The weighted SVM credit scoring models are tested using 10-fold cross validation with two real world data sets and the experimental results are compared with other six traditional methods including linear regression, logistic regression, k nearest neighbor, decision tree, and neural network. Results demonstrate that weighted 2-norm SVM with radial basis function (RBF) kernel function and t-test feature weighting strategy has the overall better performance with very narrow margin than other SVM models. However, it also consumes more computational time. In considering the balance of performance and time, least squares support vector machines (LSSVM) with RBF kernel maybe a better choice for large scale credit scoring applications.

In this paper, a modified least squares support vector machine classifier, called the C-variable least squares support vector machine (C-VLSSVM) classifier, is proposed for credit risk analysis. The main idea of the proposed classifier is based on the prior knowledge that different classes may have different importance for modeling and more weight should be given to classes having more importance. The C-VLSSVM classifier can be obtained by a simple modification of the regularization parameter, based on the least squares support vector machine (LSSVM) classifier, whereby more weight is given to errors in classification of important classes, than to errors in classification of unimportant classes, while keeping the regularized terms in their original form. For illustration purpose, two real-world credit data sets are used to verify the effectiveness of the C-VLSSVM classifier. Experimental results obtained reveal that the proposed C-VLSSVM classifier can produce promising classification results in credit risk analysis, relative to other classifiers listed in this study.

To evaluate whether the investments of e-logistics systems may increase financial risks, models of Bayesian networks are constructed in this study with the mechanism of structural learning and parameter learning. Empirical findings from the transport and logistics sectors suggest that the e-logistics investments generally do not increase the financial risks of companies except the implementation of computer aided picking systems and radio frequency identification. Meanwhile, only the investment of enterprise resource planning can reduce the financial risks and enhance the profitability at the same time. Generally speaking, most advanced e-logistics investments do not yield financial advantages for the transport and logistics companies from the perspective of Bayesian inference. Empirical study based on the proposed models also demonstrates the practicability of Bayesian models.

Following the Basel II Accord, with the increased focus on operational risk as an aspect distinct from credit and market risk, quantification of operational risk has been a major challenge for banks. This paper analyzes implications of the advanced measurement approach to estimate the operational risk. When modeling the severity of losses in a realistic manner, our preliminary tests indicate that classic distributions are unable to fit the entire range of operational risk data samples (collected from public information sources) well. Then, we propose a piecewise-defined severity distribution (PSD) that combines a parameter form for ordinary losses and a generalized Pareto distribution (GPD) for large losses, and estimate operational risk by the loss distribution approach (LDA) with Monte Carlo simulation. We compare the operational risk measured with piecewise-defined severity distribution based LDA (PSD-LDA) with those obtained from the basic indicator approach (BIA), and the ratios of operational risk regulatory capital of some major international banks with those of Chinese commercial banks. The empirical results reveal the rationality and promise of application of the PSD-LDA for Chinese national commercial banks.

Software development involves plenty of risks, and errors exist in software modules represent a major kind of risk. Software defect prediction techniques and tools that identify software errors play a crucial role in software risk management. Among software defect prediction techniques, classification is a commonly used approach. Various types of classifiers have been applied to software defect prediction in recent years. How to select an adequate classifier (or set of classifiers) to identify error prone software modules is an important task for software development organizations. There are many different measures for classifiers and each measure is intended for assessing different aspect of a classifier. This paper developed a performance metric that combines various measures to evaluate the quality of classifiers for software defect prediction. The performance metric is analyzed experimentally using 13 classifiers on 11 public domain software defect datasets. The results of the experiment indicate that support vector machines (SVM), C4.5 algorithm, and K-nearest-neighbor algorithm ranked the top three classifiers.

Project risk management, as a subjective and dynamic decision method of avoiding project risk, plays an important role in selecting appropriate control measures (i.e. measure set) to handle multiple emerging risks (i.e. risk set). This paper focuses on the multi-risk control problem under the condition of finite risk management resources. Based on the description of a Multi-HUB network and a collaboration mechanism within project risk management system, a multi-risk control system is formulated as a non-cooperative complete information game model between the project risk set and the measure set. In the proposed model, each risk within the risk set is described as a player in the game, who is competing with other players for the allocation of risk control resources that are available in limited quantities within a given measure set. Then an efficient algorithm is proposed to solve the allocation solution based on Nash equilibrium, and an experiment is presented to illustrate the effectiveness of the proposed game model. The proposed modeling and solution methods can be used to support decision-making in project risk management.

In this paper, a copula-based correlation measure is proposed to test the interdependence among stochastic variables in terms of copula function. Based on a geometric analysis of copula function, a new derivation method is introduced to derive the Gini correlation coefficient. Meantime theoretical analysis finds that the Gini correlation coefficient tends to overestimate the tail interdependence in the case of stochastic variables clustering at the tails. For this overestimation issue, a fully new correlation coefficient called Co is developed and extended to measure the tail interdependence. Empirical study shows that the new correlation coefficient Co can effectively solve the overestimation issue, which implies that the proposed new correlation coefficient is more suitable to describe the interdependence among stochastic variables than the Gini correlation coefficient.

Oil economies in the Former Soviet Union (FSU) region, with geographical proximity to each other, are usually impacted by some common risk factors, which make their country risks closely correlated. This paper focuses on correlation between country risks and investigates the spillovers of country risk returns (CRR). Taking Russia and Kazakhstan for example, firstly, this paper identifies the structural breaks in CRR series, using iterated cumulative sums of squares (ICSS) algorithm. Secondly, on the assumption that there may be similarity in structural breaks of CRR series of the two countries, Vector Autoregression (VAR) process and Granger causality test are used to identify whether there are mean spillovers of CRR series. Finally, the volatility spillovers are captured by using multivariate conditional volatility models in the framework of the BEKK models. Empirical results show that (1) there are significant unidirectional mean spillovers from Russia to Kazakhstan; (2) there are asymmetric bidirectional volatility spillovers between Russia and Kazakhstan; and volatility spillover effects from Russia to Kazakhstan are stronger.