Riskmeasurement and Decomposition - PowerPoint PPT Presentation

About This Presentation
Title:

Riskmeasurement and Decomposition

Description:

The potential loss in market value on financial assets that results from an ... Quantile of the Confidence level. Mag. Eugen PUSCHKARSKI ... – PowerPoint PPT presentation

Number of Views:44
Avg rating:3.0/5.0
Slides: 23
Provided by: mageugenpu
Category:

less

Transcript and Presenter's Notes

Title: Riskmeasurement and Decomposition


1
Riskmeasurement and Decomposition
2
Taxonomy of Risk
  • Market Risk
  • The potential loss in market value on financial
    assets that results from an adverse movement in
    market prices or rates
  • Credit Risk
  • The risk that a counterparty will fail to
    perform on an obligation owed to the firm
  • Liquidity Risk
  • Operational Risk
  • Legal Risk
  • ...

3
Market Risk
  • FX - Risk
  • Exposure
  • Value at Risk
  • Interest Rate Risk
  • mod. Duration
  • PVBP
  • Value at Risk

4

Value at Risk (VaR)
  • the portfolio loss which is not exceeded with
  • a certain probability (e.g. 95 )
  • over a specific time horizon (e.g. 1 month)

5
Risk Terminologies

Time
0
Value
t
o
6
Calculating VaR
  • Positions - Exposures
  • Volatilities
  • Correlations

Quantile of the Confidence level ?
Market value of Portfolio
7
Positions - Exposures - RiskMetrics Cashflow
Mapping
  • 1. The positions are stripped to the individual
    cashflows.

8
Volatilities
Written recursively
9
Volatilities
  • By varying the decay factor recent observations
    can be given more weight then older ones
  • The RiskMetrics Research Group has concluded that
    for short periods (e.g. up to 10 days) a decay
    factor of 0,94 is optimal and for longer ones (
    one month and more) 0,97 is optimal in predicting
    future volatility

Volatility clustering
10
Volatilities
  • A higher decay factor corresponds to considering
    a longer period of historical observations.
  • A decay factor of one is equal to a simple moving
    average.

11
Risk Attribution
Business Date (t)
Business Date (t-1)
Pricing Date (t)
Pricing Date (t-1)
Position (t-1)
Position (t)
12

Stress Tests
The Problem VaR does not show how large a
possible loss is beyond the confidence level!!!
13

Stress Tests
Solution Find plausibel Szenarios of Market
Stress which result in large losses.
  • Historical Stress Szenarios
  • Hypothetical Stress Szenarios
  • Factor Push Method
  • Extreme Value Theory
  • Monte Carlo Methods

14
Stress Tests
Simple Stress Test Stressed Risk Factor has no
influence on other Risk Factors Predictive
Stress Test Stressed Risk Factor influence other
Risk Factors consistent with observed
correlations
15
Backtesting
General Issues
  • Backtesting refers to the testing of VaR models
    to ensure that VaR estimates are sufficiently
    accurate
  • Concerned about under- and over-prediction of VaR
  • Under-prediction implies firm riskier than it
    seems
  • Over-prediction implies firm has excessive risk
    capital

16
Backtesting
Procedure
  • Clean Backtesting static portfolio holdings
    corresponding to the VaR assumption
  • holding period one day in order not to bend the
    above assumption to much
  • Step 1 calculate VaR (potential PL) over the
    next day
  • Step 2 the next day revalue the positions and
    compare with VaR from the day before
  • Continue with step 1 and 2

17
Statistical Tests
  • These are the most popular backtests
  • Usually applied to frequency of excessive losses
  • Based on whether number of losses in excess of
    VaR is consistent with what we would expect
  • 4 main tests in this class
  • Kupiecs frequency of failures test
  • Textbook proportions test
  • Crnkovic-Drachman VaR percentile test
  • Christoffersens interval forecast test

18
Backtesting
Kupiecs frequency of failures test
  • If for example 281 observations
  • and 14 exceptions
  • gt4,98 of exceptions versus 5 predicted
  • Teststatistic is the Loglikelyhood Ratio

19
Backtesting
Kupiecs frequency of failures test
  • Loglikelyhood Ratio is Chi2(1) distributed
  • Result We can be 98,907 sure, that 4,98 does
    not differ from 5 significantly!

20
Comparison of methods
21
Reference List
  • Value at Risk A New Benchmark for Measuring
    Derivatives Risk by Philippe Jorion Hardcover -
    332 pages (August 1996) Irwin Professional Pub
    ISBN 0786308486 Dimensions (in inches) 1.20 x
    9.33 x 6.34
  • Managing Financial Risk A Guide to Derivative
    Products, Financial Engineering and Value
    Maximization (Irwin Library of Investment
    Finance) by Charles W. Smithson, Clifford W.
    Smith Hardcover - 620 pages 3rd edition (July
    1998) McGraw-Hill ISBN 007059354X Dimensions
    (in inches) 2.05 x 9.76 x 7.86
  • Mastering Value at Risk A Step-By-Step Guide to
    Understanding and Applying Var by Cormac Butler
    Paperback - 288 pages (April 1999)
    Trans-Atlantic Publications, Inc. ISBN
    0273637525 Dimensions (in inches) 0.97 x 9.83
    x 6.82

22
Internet Resources
  • All About Value-at-Risk
  • http//www.gloriamundi.org/
  • RiskMetrics Technical Document
  • http//www.riskmetrics.com/research/techdoc/
  • Risk Waters Group
  • http//www.riskwaters.com/home.htm
Write a Comment
User Comments (0)
About PowerShow.com