Value-at-Risk (VaR) Zvi Wiener 02-588-3049 http://pluto.mscc.huji.ac.il/~mswiener/zvi.html May-00 Risk Management Risk Business Risk Financial Risk – market risk – credit risk – liquidity risk VaR Operational Risk Legal Risk May-2000 slide 2 How much can we lose? Everything correct, but useless answer. How much can we lose realistically? VaR May-2000 slide 3 Derivatives 1993-1995 Shova Shell, Japan Kashima Oil, Japan Metallgesellschaft Barings, U.K. Codelco, Chile Procter & Gamble, US VaR May-2000 ($ million) 1,580 1,450 1,340 1,330 200 157 slide 4 Barings VaR February 26, 1995 233 year old bank 28 year old Nick Leeson $1,300,000,000 loss bought by ING for $1.5 May-2000 slide 5 Public Funds Orange County San Diego West Virginia Florida State Treasury Cuyahoga County Texas State VaR May-2000 ($ million) 1,640 357 279 200 137 55 slide 6 Orange County Bob Citron, the county treasures $7.5B portfolio (schools, cities) borrowed $12.5B, invested in 5yr. notes interest rates increased reported at cost - big mistake! realized loss of $1.64B VaR May-2000 slide 7 Financial Losses Barings Bank Negara, Malaysia 92 Banesto, Spain Credit Lyonnais S&L, U.S.A. Japan VaR May-2000 $1.3B $3B $4.7B $10B $150B $500B slide 8 Metallgesellshaft 14th largest industrial group 58,000 employees offered long term oil contracts hedge by long-term forward contracts short term contracts were used (rolling hedge) 1993 price fell from $20 to $15 $1B margin call in cash VaR May-2000 slide 9 VaR May-2000 slide 10 What is the current Risk? Bonds Stocks Options Credit Forex Total VaR duration, convexity volatility delta, gamma, vega rating target zone ? May-2000 slide 11 Standard Approach VaR May-2000 slide 12 Modern Approach Financial Institution VaR May-2000 slide 13 Risk Management VaR Risk measurement Reporting to board Limits monitoring Diversification, reinsurance Vetting Reporting to regulators Decision making based on risk May-2000 slide 14 Who manages risk? Citibank AIG Nike Bank of England General Re Sony CIBC Swiss Re Dell Computers J. P. Morgan Aetna Philip Morris Bankers Trust Zurich Ford Motor VaR May-2000 slide 15 Regulators VaR BIS FSA SEC ISDA FASB Bank of Israel Galai’s committee May-2000 slide 16 Basic Steps in RM process VaR Identify risks Data base (market + position) Risk measurement Regulators Risk Management Reporting Strategic decisions May-2000 slide 17 Building a RM system VaR Initial study of risks Decision, Risk Manager Risk measurement system Responsibilities and structure Testing Active Risk Management Staff training and maintenance May-2000 slide 18 Risk Management and Risk Measurement VaR May-2000 slide 19 Risk Management System Can NOT Predict future Identify business opportunities Be always right! Risk Management System Can Predict loss, given event Identify most dangerous scenarios Recommend how to change risk profile VaR May-2000 slide 20 Tool, not rule! VaR May-2000 slide 21 Definition VaR is defined as the predicted worst-case loss at a specific confidence level (e.g. 99%) over a certain period of time. VaR May-2000 slide 22 VaR 1 0.8 0.6 0.4 VaR1% 1% 0.2 Profit/Loss -3 VaR -2 -1 1 May-2000 2 3 slide 23 Meaning of VaR A portfolio manager has a daily VaR equal $1M at 99% confidence level. This means that there is only one chance in 100 that a daily loss bigger than $1M occurs, under normal market conditions. VaR 1% VaR May-2000 slide 24 History of VaR VaR 80’s - major US banks - proprietary 93 G-30 recommendations 94 - RiskMetrics by J.P.Morgan 98 - Basel SEC, FSA, ISDA, pension funds, dealers Widely used and misused! May-2000 slide 25 Risk Management Structure Market data Current position Risk Mapping Valuation Value-at-Risk Reporting and Risk Management VaR May-2000 slide 26 Value 8.25 8 7.75 7.5 7.25 10 4.3 4.25 4.2 11 4.15 12 13 Interest Rate 14 4.1 dollar interest rates and dollar are NOT independent VaR May-2000 slide 27 Risk Measuring Software CATS, CARMA Algorithmics, Risk Watch Infinity J.P. Morgan, FourFifteen FEA, Outlook Reuters, Sailfish Kamacura Bankers Trust, RAROC INSSINC, Orchestra VaR May-2000 slide 28 Qualitative Requirements An independent risk management unit Board of directors involvement Internal model as an integral part Internal controller and risk model Backtesting Stress test VaR May-2000 slide 29 Quantitative Requirements 99% confidence interval 10 business days horizon At least one year of historic data Data base revised at least every quarter All types of risk exposure Derivatives VaR May-2000 slide 30 Types of Assets and Risks VaR Real projects - cashflow versus financing Fixed Income Optionality Credit exposure Legal, operational, authorities May-2000 slide 31 Risk Factors There are many bonds, stocks and currencies. The idea is to choose a small set of relevant economic factors and to map everything on these factors. Exchange rates Interest rates (for each maturity and indexation) Spreads VaR Stock indices May-2000 slide 32 How to measure VaR Historical Simulations Variance-Covariance Monte Carlo Analytical VaR Methods May-2000 slide 33 Historical Simulations Fix current portfolio. Pretend that market changes are similar to those observed in the past. VaR Calculate P&L (profit-loss). Find the lowest quantile. May-2000 slide 34 Example Assume we have $1 and our main currency is SHEKEL. Today $1=4.30. Historical data: P&L 4.00 VaR 4.20 4.30*4.20/4.00 = 4.515 0.215 4.20 4.30*4.20/4.20 = 4.30 0 4.10 4.10*4.10/4.20 = 4.198 -0.112 4.15 4.15*4.15/4.10 = 4.352 0.052 May-2000 slide 35 USD NIS 2000 100 -120 2001 200 100 2002 -300 -20 2003 20 30 today VaR 100 200 300 20 2 3 1 0.06 (1 0.061) (1 0.062) (1 0.063) 4 120 100 20 30 2 3 1 0.1 (1 0.11) (1 0.12) (1 0.13) 4 May-2000 slide 36 today Changes in IR 100 200 300 20 2 3 1 0.06 (1 0.061) (1 0.062) (1 0.063) 4 120 100 20 30 2 3 1 0.1 (1 0.11) (1 0.12) (1 0.13) 4 USD: NIS: +1% +1% +1% 0% +1% -1% +1% -1% 100 200 300 20 2 3 1 0.07 (1 0.071) (1 0.072) (1 0.073) 4 120 100 20 30 2 3 1 0.11 (1 0.11) (1 0.11) (1 0.12) 4 VaR May-2000 slide 37 Returns year 1% of worst cases VaR May-2000 slide 38 VaR 1 0.8 0.6 0.4 VaR1% 1% 0.2 Profit/Loss -3 VaR -2 -1 1 May-2000 2 3 slide 39 Weights Since old observations can be less relevant, there is a technique that assigns decreasing weights to older observations. Typically the decrease is exponential. See RiskMetrics Technical Document for details. VaR May-2000 slide 40 Variance Covariance Means and covariances of market factors Mean and standard deviation of the portfolio Delta or Delta-Gamma approximation VaR1%= P – 2.33 P Based on the normality assumption! VaR May-2000 slide 41 Variance-Covariance VaR1% V 2.33 V 1% 2.33 -2.33 VaR May-2000 slide 42 Monte Carlo 1 0.5 -1 0.5 -0.5 1 -0.5 -1 VaR May-2000 slide 43 Monte Carlo VaR Distribution of market factors Simulation of a large number of events P&L for each scenario Order the results VaR = lowest quantile May-2000 slide 44 Monte Carlo Simulation 15 10 5 10 20 30 40 -5 -10 -15 VaR May-2000 slide 45 Real Projects Most daily returns are invisible. Proper financing should be based on risk exposure of each specific project. Note that accounting standards not always reflect financial risk properly. VaR May-2000 slide 46 Example You are going to invest in Japan. Take a loan in Yen. Financial statements will reflect your investment according to the exchange rate at the day of investment and your liability will be linked to yen. Actually VaR there is no currency risk. May-2000 slide 47 Airline company fuel - oil prices and $ purchasing airplanes - $ and Euro salaries - NIS, some $ tickets $ marketing - different currencies payments to airports for services VaR May-2000 slide 48 Airline company VaR loans equity callable bonds May-2000 slide 49 Airline company Base currency - by major stockholder. Time horizon - by time of possible price change. Earnings at risk, not value at risk, since there is too much optionality in setting prices. One can create a one year cashflow forecast and measure its sensitivity to different market events. VaR May-2000 slide 50 Reporting Division of VaR by business units, areas of activity, counterparty, currency. Performance measurement - RAROC (Risk Adjusted Return On Capital). VaR May-2000 slide 51 How VaR is used VaR Internal Risk Management Reporting Regulators May-2000 slide 52 Backtesting Verification of Risk Management models. Comparison if the model’s forecast VaR with the actual outcome - P&L. Exception occurs when actual loss exceeds VaR. After exception - explanation and action. VaR May-2000 slide 53 Backtesting Green zone - up to 4 exceptions OK Yellow zone - 5-9 exceptions increasing k Red zone - 10 exceptions or more intervention VaR May-2000 slide 54 Stress Designed to estimate potential losses in abnormal markets. Extreme events Fat tails Central questions: How much we can lose in a certain scenario? What event could cause a big loss? VaR May-2000 slide 55 Unifying Approach VaR One number Based on Statistics Portfolio Theory Verification Widely Accepted Easy Comparison May-2000 slide 56 Board of Directors (Basle, September 1998) periodic discussions with management concerning the effectiveness of the internal control system a timely review of evaluations of internal controls made by management, internal and external auditors periodic efforts to ensure that management has promptly followed up on recommendations and concerns expressed by auditors and supervisory authorities on internal control weaknesses a periodic review of the appropriateness of the bank’s strategy and risk limits. VaR May-2000 slide 57 Open Questions VaR Risks related to cashflow Non-traded assets Credit information Global Database Liquidity problem May-2000 slide 58 Issues Specific to Israel VaR Indexation Exchange Band Shallow Markets May-2000 slide 59