Research On Assessment Of Banking Operational Risk Based On Evidence Theory

The operational risk in banking system is becoming more series as credit risk and market risk along with the change in banking function and supervising environment in 1990's. Basel Committee on Banking Supervision (BCBS) had proposed the principle of identifying, assessing, inspecting and releasing the operational risk for banks. Whereas the measurement of operational risk is not satisfied the demand for operational risk supervising, which is developed from the models for credit risk and cannot reflect the characteristics of operational risk. Research on assessment approach of operational risk in banking has significant practical meaning in enhancing the risk supervising ability and management efficiency. Aiming at the problems of actual measures, such as model technical matter, data supporting and foundation of system culture in operational risk management, this dissertation focused on incomplete information decision-making based evidence theory, built the opening frame of discernment, and syncretized the public and private inspecting information in operational risk measure. The main research and innovation of this dissertation are:1. Set an assessing indicators system for banking operational riskUsing financial risk micro-mechanism of informational economic, financial system risk theory, static and dynamic analysis revealed the risk driving factors in our banks. This paper built an assessment indicator system for operational risk by anglicizing the model characteristics and assessing rules, designed the workflow-based expert inspecting proceeding and content for operational risk by studying the workflows and management platform in banking, and got the main reasons of operational risk in our banks as personnel factor, system, stratagem structure and workflow by case-based reasoning (CBR) in the operational risk case base. 2. Research on axiom of incomplete informational operational risk assessment and group decision-making rulesAs the process of incomplete informational multi-attribute group decision making, assessment of operational risk in banks has the problems as consistency of individual utility concentrating and amalgamation of incomplete information. This paper proposed the axiom of conversion function of individual utility to group preference by analysis the developmental conditions of Arrow's choice theory:â‘ there is an opening objective frame of discernment;â‘¡there is a integrating rule that respect individual opinion;â‘¢there is the order preserved axiom of alternatives; andâ‘£there is no autarchy rule for any decision-maker. The paper demonstrated the rationality of developmental study of releasing the restrictions of Arrow's theory.Researching on the decision-maker's behavior hypothesis and classification principle in incomplete informational risk assessment, this paper built the rational decision-making route asâ‘ releasing the sequential or partible hypothesis of individual preference and alternatives independence condition in order to satisfy the actual decision-making;â‘¡keeping the optimal result of group decision-making according as the thought of"Binary comparison"of sequence path independence condition (SPI); building the hierarchy structure in opening frame of discernment to increase granularity of discernment.â‘¢reducing complexity of decision-making based on including degree. This paper proposed the grouping reduction strategy of attribute and distilment of decision-making rule in operational risk. 3. Research on banking operational risk assessment based on evidence theoryStudying on the problems of operational risk assessment in Chinese banks as supervising mechanism, risk indicators, model design and data support, this paper gave the essence factors of operational risk model:â‘ risk events definition of model.â‘¡risk character portrayed by model. Operational risk had non-normal distribution, non-linearity and long right-tail.â‘¢inputs of model. This paper made risk drivers by definition of risk events, had the inputs of model as quantitative and qualitative variables.This paper gave the delaminating of assessing indicators structure for experts'judgment information based on DS evidence theory, and studied the strategy of grouping attributes reduction and decision-making rules obtaining method by including degree for operational risk measurement. The test confirmed that the proposed approach had advantages in consonance of conflictive knowledge, reduction of calculating complexity and the legible rules obtaining than traditional rough set.This paper studied the decision-making under uncertainty expressed with fuzzy and language text variables as the inputs of assessment model, proposed the emendatory uncertainty compound rule based on the distance between different evidence. Case study chose four commercial banks and gathered data of risk events during 2003-2008; this paper tested the operational risk model of public quantitative information, private qualitative information and combination of two sources of information, the result revealed that the risk assessment model based evidence theory proposed in this paper made the operational risk management more exactitude and reasonable.