Title: Book Review: Energy Derivatives: Pricing and Risk Management by Clewlow and Strickland, 2000 Chapter
1Book Review Energy Derivatives Pricing and
Risk Management by Clewlow and Strickland,
2000Chapter 3 Volatility Estimation in Energy
Markets
- Anatoliy Swishchuk
- Math Comp Lab
- Dept of Math Stat, U of C
- Lunch at the Lab Talk
- November 28th, 2006
2Chapter 3
3Chapter 3 (cntd)
4Outline
- Intro
- Estimating Volatility
- Stochastic Volatility Models
5Intro
- Volatility can be defined and estimated in the
context of a specific stochastic process for the
price returns - Volatility definition and measure should capture
the key features of energy markets, such as the
seasonal dependence
6Intro II (most important issues)
- Investment Assets vs. Consumption Goods
(Commodities cannot be treated as purely
financial assets) - Prices of Energy Commodities Display Seasonality
- Commodity Prices Often Display Jump Behaviour
- Prices Gravitate to the Cost of Production
7Estimating Volatility (EV)
- EV From Historical Data
- EV For a Mean-Reverting Process
- EV Special Issues
- Intraday Price Variability
- EV for a Basket
- Implied Volatility
8EV from Historical Data
- Step 1 Calculate Logarithmic Price Returns
- Step 2 Calculate Standard Deviations of the
Logarithmic Price Returns - Step 3 Annualize the St. Dev. By Multiplying it
by the Correct Factor
9EV from Historical Data II
- Step 1 log price returns (lpr)-log(1r)
- Step 2 st. dev. of lpr
- Step 3 annualization
- \sigmasqrt(n)\sigma(lpr)
- Standard usage
- Seasonality effect
10EV for a Mean-Reverting Process
- Ornstein-Uhlenbeck process (OU)
- Solution
- Discrete analogue (autoregressive process)
- OU is the limiting case for
- (dt-gt0)
- \nu_t-zero mean and variance
11EV for a Mean-Reverting Process II
- Recovering of the initial parameters from
discrete version
12EV Special Issues
- The choice of the annualisation factor and use of
intra-period data (intraday prices) - Posibilities sqrt(266)52x(41.107)
- Sqrt(273)52x(41.25)
13EV Intraday Price Variability
14EV Basket Options (Sum of 2 (weighted) or more
prices)
- The Call Option Payoff
- The Put Option Payoff
15EV Basket Options (Sum of 2 weighted or more
prices) II
- Two Commodities (GBM)
- PDE
- Volatility
16Implied Volatility (IV)
- IV Vol. that is used as an input to an option
pricing formula that equates the model price with
the market price - Existence of fat tails (leptokurtic) its
described by the kurtosis (4th moment around the
mean) (for normal 3)
17Stochastic Volatility Models (SVM)
- Ornstein-Uhlenbeck
- Vasicek
- Ho Lee
- Hull-White
- Cox-Ingersoll-Ross
- Heath-Jarrow-Morton
- Continuous-time above
18Stochastic Volatility Models (SVM) II
- Engle (1982) ARCH(q)
- Price returns
- Variance
- Bollerslev (1986) GARCH(p,q)
- GARCH(1,1)
19Stochastic Volatility Models (SVM) IV
20Stochastic Volatility Models (SVM) III
21EV Estimation and Testing
- Parameters Estimation
- Usefulness of a parameter estimator
- Unbiased and Efficient
- Unbiased is good
- Biased but Efficient may be preferable to an
unbiased
22Estimation and Testing Least Squares
- Stochastic equation
- Minimization
23Estimation and Testing Least Squares II
- Example I
- Estimation of Mean
24Estimation and Testing Least Squares II
- Example II
- Estimation of Standard Deviation
- Unbiased, consistent, efficient
25Maximum Likelihood Estimation (MLE)
- Equation
- Probability density function
- Joint distribution
- Likelihood function
26MLE I
27MLE II
- MLE for St. Dev.
- Consistent
- But biased
- Unbiased (LSE)
28Testing
29Testing II
- Skewness
- Kurtosis
- Jarque-Bera Statistic
- Goldfeld-Quandt test
30Testing (Example from Energy Commodity Markets)
31Testing (Example from Energy Commodity Markets I)
32Testing (Goodness of Fit)
- Likelihood Ratio Test
- Schwartz Criterion (SC)
- (the most probable model-with the smallest SC)
33Testing (Goodness of Fit)
34Testing (Goodness of Fit)
35Figures (Simulated vs. Actual Data) PD
36Figures (Simulated vs. Actual Data) JD
37Figures (Simulated vs. Actual Data) JDGARCH
38The End