Jianming Chen

Operational Risk Aggregation across Business Lines Based on Frequency Dependence and Loss Dependence

In loss distribution approach (LDA), the most popular approach in operational risk modeling, frequency dependence and loss distribution dependence across business lines are two dependences which banks should consider. In practice, mainly for simplicity, many banks only model frequency dependence although they think that the impact of frequency dependence is insignificant. In this study, two approaches, respectively, models frequency dependence and loss distribution dependence, are introduced. Both approaches are modeled by copula function, which is capable of capturing nonlinear correlation. Based on the most comprehensive operational risk dataset of Chinese banking as far as we know, the operational risk capital charge of the overall Chinese banking is calculated by the two approaches. The results show that there is an obvious distinction between the capital calculated by modeling frequency dependence and the capital calculated by modeling loss dependence. The approach with very limited attention exactly yields a much larger capital result. So it is advised in this paper that banks should not just rely on the approach to modeling frequency dependence for it is natural and easy to deal with. A safer and more effective way for banks is to comprehensively take the results of the two kinds of approach into consideration.

The Mutual-Information-Based Variance–Covariance Approach: An Application to Operational Risk Aggregation in Chinese Banking

Most advanced measurement approaches cannot simultaneously capture the overall dependence between operational risk components and be easy to use and understand. This paper proposes a mutual-information-based variance–covariance approach that is able to capture the overall correlation and is also highly tractable. Specifically, we replace the linear correlation coefficient with the global correlation coefficient in the framework of the variance–covariance approach. Originating from the theory of mutual information, the global correlation coefficient is able to capture both linear and nonlinear correlation relationships. The value-at-risk (VaR) of each individual risk component is calculated; these VaRs are then aggregated by using the global correlation coefficient. In empirical analysis, the proposed approach is employed to aggregate the operational risk of Chinese banking across business lines, based on the most comprehensive (to the best of our knowledge) operational risk data set. After an overall comparison with results from other correlation assumptions and the actual capital allocation of Chinese banking in 2013, we conclude that the actual capital allocation in China at present is not effective, and the aggregate VaR calculated from our approach is more reasonable.

Integrating Credit and Market Risk: A Factor Copula based Method☆

Abstract This paper presents a factor copula model for the integration of Chinese commercial banks’ credit risk and market risk. By defining the dependence structure through a set of common factors reflecting the macro-economic situation, this model reveals the intrinsic correlation between credit risk and market risk. We derive the integration process with factor copula and generate common factors by performing a principal component analysis on 4 different macro-economic indicators that have impact on banks profit, namely the GDP growth, M2 growth, benchmark for loan rate, and the ratio of new loans to GDP. In the empirical study, 15 Chinese listed banks are chosen to construct the model. The results are compared with that of elliptical copulas and Archimedean copulas, we find that factor copula gives a more prudential result in risk integration.

Copula based Change Point Detection for Financial Contagion in Chinese Banking

Abstract In this paper, a change point detection approach based on copula with two notable advantages is put forward. One is that the approach can deal with the common but special unbalanced panel data. The other is that it can detect multiple change points. Firstly, a proper copula that most accurately describes the dependence structure of the data is chosen. Then, the chosen copula is fitted to the data dynamically by adding new data. Finally, the change points are located by analyzing the trends o f fitted parameters of the copula. Based on the quarterly financial data of 16 listed Chinese commercial banks, we empirically use the proposed approach to detect the subprime crisis contagion period in Chinese banking. The results show that the contagion starts in 2007Q2 and ends in 2009Q1, which is reasonable according to relevant researches.

A Nonparametric Operational Risk Modeling Approach Based on Cornish-Fisher Expansion

It is generally accepted that the choice of severity distribution in loss distribution approach has a significant effect on the operational risk capital estimation. However, the usually used parametric approaches with predefined distribution assumption might be not able to fit the severity distribution accurately. The objective of this paper is to propose a nonparametric operational risk modeling approach based on Cornish-Fisher expansion. In this approach, the samples of severity are generated by Cornish-Fisher expansion and then used in the Monte Carlo simulation to sketch the annual operational loss distribution. In the experiment, the proposed approach is employed to calculate the operational risk capital charge for the overall Chinese banking. The experiment dataset is the most comprehensive operational risk dataset in China as far as we know. The results show that the proposed approach is able to use the information of high order moments and might be more effective and stable than the usually used parametric approach.

Comparison of Different Methods to Design Risk Matrices from The Perspective of Applicability

Abstract The design of risk matrices is a topic that has not reached a consensus, although risk matrices are widely used in practice. Several methods have been proposed to help design risk matrices. However, all the methods seem to have their own advantages, and it is difficult for the decision makers to choose one. In this paper, we compare two different risk matrix design methods from the perspective of applicability. Specifically, we give three detailed scenarios where different settings of the risk matrices are given, and then compare the performance of the methods. Results show that both the two methods have their own advantages, but they will fail to give an effect design sometimes.

Has China’s oil-import portfolio been optimized from 2005 to 2014? A perspective of cost-risk tradeoff

Abstract Optimizing the oil-import portfolio has become extremely effective for enhancing energy security. The focus of this paper is on whether China’s oil-import portfolio has been continuously optimized since 2005. Firstly, a multi-objective programming problem was constructed based on cost–risk tradeoff, and the optimal results were obtained. The oil-import portfolio of China since 2005 was then analyzed by comparing the optimal import strategy with the actual strategy. Next, the cost fluctuation was decomposed to demand, structure, and price effects with LMDI index decomposition. Finally, the driving force behind China’s actual oil-import strategy was further explored by setting up different scenarios that consider the trade relations as well as risk exposure of exporting countries and import diversification. The optimization results and discussion show that a non-linear negative correlation exists between oil-import risk and cost, and that a Pareto curve offers an alternative decision set of oil-import portfolio optimization. Although the oil-import portfolio seems to have been optimized gradually since 2005, the deviation between the actual and optimal import decisions has grown since 2009. For China, the actual optimization space is limited due to the constraints of trade relationships and diversification, which makes the risk-optimization space larger than the cost-optimization space.

Operational Risk Measurement: A Nonparametric Approach Using Cornish-Fisher Expansion

The severity loss distribution is the main topic in operational risk estimation. In this paper, we propose a novel model for quantifying operational risk in the framework of the loss distribution approach (LDA) as suggested by the Basel II. We use Cornish¡VFisher Expansion, which is non-parameter method, to fit operational risk loss severity, and then we use simulation technique to measure the operational risk in the framework of LDA. We use this approach to measure the operational risk of Chinese commercial banking. Empirical analysis shows that this approach allows the allocation of capital in an efficient way.