1 / 48

Risk Measures

Risk Measures. IEF 217a: Lecture Section 3 Fall 2002. Risk Measures Assigning a number to risk. Histogram Variance Beta (CAPM) Value at Risk (VaR) Expected utility Time Historical Perspectives. Histogram. Another Histogram. Histogram. Full picture of risk subject to:

agatha
Télécharger la présentation

Risk Measures

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Risk Measures IEF 217a: Lecture Section 3 Fall 2002

  2. Risk MeasuresAssigning a number to risk • Histogram • Variance • Beta (CAPM) • Value at Risk (VaR) • Expected utility • Time • Historical Perspectives

  3. Histogram

  4. Another Histogram

  5. Histogram • Full picture of risk subject to: • Underlying model is correct • Samples big enough • Easy to get with simulations • Problems • Complicated to • Summarize • Evaluate

  6. Risk -> Single Number • Many attempts

  7. Risk MeasuresAssigning a number to risk • Histogram • Variance • Beta (CAPM) • Value at Risk (VaR) • Expected utility • Time • Historical Perspectives

  8. Variance • Expected value • Variance,

  9. VarianceEqual Sample • Expected value • Variance,

  10. Why Variance? • Easy to describe • People learn it in statistics classes • Normal Distribution • Completely described by • Expected Value • Variance (dispersion/risk) • Central limit theorem • Useful properties (adding)

  11. Variances for Two: Covariance • E(x) = expected value of x • Covariance(x,y) = cov(x,y)

  12. Useful Properties If x is independent of y then

  13. Case for Variance • A pictorial example of variance • Matlab example: • varexam1.m

  14. Case Against Variance • Asymmetry is not captured well by variance • For asymmetric distributions variance might not be a good measure

  15. Mean Absolute DeviationClose Relative to Variance • Not as easy to compute • Not used as often • Less often reported with Normal distribution • Less sensitive to outliers than variance

  16. Mean Absolute Deviation • Expected value • MAD,

  17. Risk MeasuresAssigning a number to risk • Histogram • Variance • Beta (CAPM) • Value at Risk (VaR) • Expected utility • Time • Historical Perspectives

  18. CAPM-Beta • r: individual stock return • a: constant for stock I • b: beta • e: idiosyncratic risk

  19. More on Beta

  20. The Importance of Portfolio Thinking (Markowitz) • Previously • Risk of a stock = Variance( r ) • Post CAPM • Risk of a stock = beta (b)

  21. CAPM Example • Matlab: betatest.m • Portfolio problem • 0.8 market, 0.2 individual stock • Experiment • Measure contribution to portfolio std from • Increase beta on individual stock • Increase variance of e, idiosyncratic risk on stock

  22. CAPM Example

  23. CAPM ExampleInterpreting Results • Beta risk contributes more to the overall risk of the portfolio then idiosynchratic • Intuition: • Idiosyncratic risk is uncorrelated with the overall portfolio so it contributes less to overall risk • Market risk is more important since it is correlated with the rest of the portfolio

  24. “Traditional” CAPM • Mean/Variance world • E(return) linear function of beta • We wont dwell on this, why? • Outside of mean/variance world • Nonlinear dependence

  25. Why do we still care about systematic (beta) risk? • Risk depends on • What you are considering to add to your portfolio and, • Its relationship to whatever is in your portfolio • Big picture risk analysis!! • Example: (general factor models)

  26. Risk MeasuresAssigning a number to risk • Histogram • Variance • Beta (CAPM) • Value at Risk (VaR) • Expected utility • Time • Historical Perspectives

  27. What is VaR? • VaR = Value at Risk • Percentile in left tail • Maximum loss, leaving out low probability events • matlab example: varpic.m

  28. 5% VaR

  29. Why VaR? • Nonnormality • Derivatives • More relevant risk measure • Capital requirements • Firm wide risk reports

  30. Close Relation:Expected Tail Loss (ETL) • E( r ) given that r < 5th percentile • Expected loss given that something really bad happens • matlab example: etlex.m

  31. Expected Tail LossExpected Value in the Tail

  32. Risk MeasuresAssigning a number to risk • Histogram • Variance • Beta (CAPM) • Value at Risk (VaR) • Expected utility • Time • Historical Perspectives

  33. Expected Utility • Replace expected value • with

  34. What is u(x)? • u(x) = utility function • Important property • E(u(x)) often decreasing in measures of x riskiness

  35. Example • u(x) = log(x) • x = [1; 2], probs = [0.5; 0.5] • E(x) = 1.5 • E(u(x)) = E(log(x)) = 0.35 • x = [0.5; 2.5], probs = [0.5; 0.5] • E(x) = 1.5 • E(u(x)) = E(log(x)) = 0.11

  36. Why Does Expected Utility Fall? • Function curvature • Moving -0.5 has bigger impact then • Moving +0.5 Log(x) x 1.5

  37. Utility Functions • Many different possibilities • Log, Power • Mean/Variance (Quadratic) • Specific functions (loss aversion, semivariance) • Is this the right way to measure risk? • Yes: under restrictive assumptions • No: In general, but it is often used • Easy to deal with in computer simulations

  38. Risk MeasuresAssigning a number to risk • Histogram • Variance • Beta (CAPM) • Value at Risk (VaR) • Expected utility • Time • Historical Perspectives

  39. Time • Future horizon • 1 Day • 1 Month • 1 Year • 10 Years • Liquidity • Maximum sustainable loss

  40. Common Investor Advice • Long Horizon Investors (young) • More funds in equity (risky) • Short Horizon Investors (old) • More funds in bonds (less risky) • Does this make sense? • Montecarlo test • We’ll check this many times

  41. Portfolio Experiment • Annual return: Normal, mean 0.05, std 0.15 • $100 starting investment • Two Horizons • 5 years • 20 years • Look at histograms and tails • matlab: timehorizon1.m

  42. Liquidity • Good strategy in the long run • High expected return • May sustain large losses in the short run • Needs short term financial backing • Might go bankrupt before big returns are realized

  43. Liquidity Example • FX Trader • Daily Return = [-0.1; 0.1] prob = [0.4;0.6] • Trade for 20 days • Start with $100 • Must stop when portfolio drops below $80 • (Pull the plug) • new matlab commands: next slide

  44. Liquidity Example • Matlab commands • cumprod • Returns: r = [r1 r2 r3 r4] • cumprod(1+r) • (1+r1) • (1+r1)(1+r2) • (1+r1)(1+r2)(1+r3) • (1+r1)(1+r2)(1+r3)(1+r4) • matlab: ruin.m

  45. Risk MeasuresAssigning a number to risk • Histogram • Variance • Beta (CAPM) • Value at Risk (VaR) • Expected utility • Time • Historical Perspectives

  46. Historical Perspective • Risk -> Number: Very old problem • Different approaches • Economists/statisticians • Expected utility • Expected loss • Business/finance/investment • CAPM/Risk neutral pricing • Psychologists • Real people behave in strange ways with risk

  47. Risk MeasuresAssigning a number to risk • Histogram • Variance • Beta (CAPM) • Value at Risk (VaR) • Expected utility • Time • Historical Perspectives

  48. Risk Types (Jorion 1) • Market risk • Credit risk • Operational risk • Model risk • Legal risk

More Related