《供应链外包的dea与风险价值模型概述》由会员分享，可在线阅读，更多相关《供应链外包的dea与风险价值模型概述（33页珍藏版）》请在金锄头文库上搜索。
1、Supply Chain Outsourcing in Enterprise Risk Management: A DEA VaR Model,Desheng Dash Wu University of Toronto Reykjavik University RiskLab dashrisklab.ca,Extracted from Olson D. L. and Wu D. Enterprise Risk Management. World Scientific Publisher. 2007 Wu D. and Olson D. L. A Comparison of Stochastic
2、 Dominance and Stochastic DEA for Vendor Evaluation. Int J of Production Research. 2007 (1).,Nov, 2008,Call for paper,Computers & Operations Research (SCI/EI, Impact factor 1.147) “OR in Risk Management ” Due date: March 31, 2009 Guest editor: Desheng Dash Wu, David L. Olson and John Birge Approxim
3、ate date for final submission of accepted manuscripts: Nov, 2009,Outline,Introduction Enterprise Risk Management (ERM 全面风险管理) Supply Chain Outsourcing, Vendor Evaluation Contribution: ERM steps in Supply Chain Outsourcing Risk Data envelopment analysis (DEA) + Value at Risk (VaR): Intuition Conclusi
4、ons and Future Research (银行链, 金融危机),Review of Risk Management Tools 风险管理工具介绍,Risk Management tools mean-variance framework of portfolio theory i.e., selection and diversification (Markowitz 1952) Capital Asset Pricing Model (Sharpe 1964; Lintner 1965; Mossin 1966) Arbitrage Pricing Theory ( Ross, 19
5、76) Option pricing theory (Black 1972; Black 1973),Value at Risk (VaR), RiskMetrics (Jorion 1997) Prob 1 day Loss VaR=1- Min VaR P (VaR) Enterprise Risk Management Professional organization, Consultant, Rating agency, Academics 31% adopted ERM in Canadian risk & insurance Kleffner 2003 Why ERM? Toyo
6、ta,Review of Risk Management Tools (cont.),Various Risks: $ Measurement,Definition of ERM,Systematic, integrated approach Manage all risks facing organization External Economic (market - price, demand change) Financial (insurance, currency exchange) Political/Legal Technological Internal Human error
7、 Fraud Systems failure Disrupted production,Stochastic OR Models for Risk Management (Beneda 2005, Dash & Kajiji 2005),Multiple criteria analysis Subjective Simulation Probabilistic; Can be subjective (system dynamics) Data envelopment analysis (DEA) Optimization Objective, subjective, probabilistic
8、,ERM Research and Steps,Step 1: Determine the corporations objectives Step 2: Identify the risk factors, exposures Step 3: Quantify the factors, exposures Assess the impact Step 4: Examine alternative risk management tools Step 5: Select appropriate risk management approach Step 6: Implement and mon
9、itor program,More than 80 frameworks: problem-oriented, descriptive, frameworks,Specific ERM: Supply Chain Outsourcing Risk,Supplier,Manufacturer,Retailer,End customer,Warehouse,A Supply Chain Model,Supply Chain Vendor Selection,Supply Chain Vendor Selection goods input bads (risk, uncertainty?) (ri
10、sk, uncertainty?) Efficiency= output / input,Data Envelopment Analysis (DEA) -Deterministic Charnes, Cooper, Rhodes,n Vendors(DMUs) to be evaluated. m different inputs Xij , s different outputs Yrj.,The deterministic DEA model,DEA efficiency for DMUj :,Deterministic DEA (cont.),CCR Multiplier form,D
11、EA VaR -Stochastic model,j :aspiration level ; j : risk criterion; 0 j , j 1 Intuition: 1) At what confidence level, it is efficient to select the ?th Vendor ? 2) At what confidence level, it is enough to reduce the ?th cost in order to make the ?th Vendor efficient?,(1),Stochastic DEA,Assuming mult
12、ivariate normal distribution:,(2),Equivalent linear programming:,(3),Metrics in Vendor Selection Olson & Wu,Data Set Moskowitz, Tang & Lam, 2000, Decision Sciences 31, 327-360,Sample data demonstration,Simulated weights and Parameter Sensitivity,Equal weights Useful to identify dominated solutions V
13、2 0.03, V4 0.08, V6 0.36, V8 0.53 Ordinal weights Reflect decision maker preference More useful to make decision: select nondominated solutions Used centroid weights Olson & Dorai V2 0.71, V4 0.22, V6 0.07, V8 0 Adjusted risk criterion 0 j 1 Adjusted RHSs with j,DEA efficiency scores: equal weight,E
14、xplanation of Results,Self-Rated Score: 95.40=V* Mean(V1) Cross-Rated Score: 94.33=V* Mean(V2) 95% chance, V1 select V6 , V7,Explanation of Results,95% chance, V3 nondominated How? Slacks 4% Quality+ 5.5% Delivery Raw data: Quality 85 (5.1) Delivery 70 (5.5),Rankings: Stochastic DEA,Stochastic effic
15、iency without weight restriction Diagonal V4 V5 V8 V2 V1 V3 V7 V9 V6 Using averages V8 V4 V9 V7 V3 V6 V2 V1 V5 Stochastic efficiency with weight restriction Diagonal V8 V5 V4 V2 V3 V7 V1 V6 V9 Using averages V2 V8 V3 V7 V5 V4 V6 V1 V9,Classical Deterministic DEA Results,Deterministic CCR Without weight restriction All = 1.000 With weight restriction V2 V3 V4 V5 V7 V8 V6 V9 V1 Super CCR Without weight restriction V5 V6 V3 V4 V8 V7 V2 V1 V9 With weight restriction V2 V3 V8 V4 V7 V5 V6 V1 V9 Benchmarking: V2 V4 V6 V8 ,Impl
