Elements of financial risk management chapter 11

• We define the gamma of an option, which gives a order approximation of the option price as a function of the underlying asset price second-• We use gamma of an option to construct a qu

Option Risk Management

Elements of Financial Risk Management

Chapter 11Peter Christoffersen

• In this chapter, we try to incorporate derivative securities

into the portfolio risk model

• The chapter is structured as follows:

• We define the delta of an option, which provides a linear

approximation to the nonlinear option price We then present delta formulas from the various models introduced in the

previous chapter

• We establish the delta-based approach to portfolio risk

management The idea behind this approach is to linearize the option return and thereby make it fit into the risk models The downside of this approach is that it ignores the key

asymmetry in option payoffs

• We define the gamma of an option, which gives a order approximation of the option price as a function of the underlying asset price

second-• We use gamma of an option to construct a quadratic model

of the portfolio return distribution We discuss two

implementations of the quadratic model: one relies on the Cornish-Fisher approximation and the other relies on the Monte Carlo simulation technique

• We will measure the risk of options using the full

valuation method, which relies on an accurate but

computationally intensive version of the Monte Carlo

simulation technique

• We illustrate all the suggested methods in a simple

example We then discuss a major pitfall in the use of the linear and quadratic approximations in another numerical example This pitfall, in turn, motivates the use of the full valuation model

Option Delta

• The delta of an option is defined as the partial derivative of

the option price with respect to the underlying asset price, S t.

• For puts and calls, we define

• The option price for a generic underlying asset price, S, is

approximated by

• where S t is the current price of the underlying asset

• To a risk manager, the poor approximation of delta to the true option price for large underlying price changes is

clearly unsettling

• Risk management is all about large price changes, and we will therefore consider more accurate approximations here

Approximation

Black-Scholes-Merton Model

• We refer to this as the delta of the option, and it has the

interpretation that for small changes in S t the call option price

will change by (d)

• Notice that as ( * ) is the normal cumulative density function, which is between zero and one, we have

• so that the call option price in the BSM model will change

in the same direction as the underlying asset price, but the change will be less than one-for-one

Black-Scholes-Merton Model

• For a European put option, we have the put-call parity stating that

• so that we can easily derive

• Notice that we have

• so that BSM put option price moves in the opposite

direction of underlying asset, and again option price will change by less than the underlying asset price

Black-Scholes-Merton Model

• In the case where a dividend or interest is paid on the

underlying asset at a rate of q per day, deltas will be

• where

Figure 11.2: The Delta of a Call Option (top) and a

Put Option (bottom)

Figure 11.3: The Delta of Three Call Options

Out-of-the-money At-the-money

In-the-money

The Binomial Tree Model

• Option deltas can be computed using binomial trees

• This is important for American put options for which early exercise may be optimal, which will impact the current

option price and also the option delta

• The black font shows the American put option price at each node The green font shows the option delta

• The delta at point A (that is at present) can be computed

very easily in binomial trees simply as

Table 11.1: Delta of American Put Option

291.14

F

654.25

445.75

The Binomial Tree Model

• A similar formula can be used for European puts as well as for call options of each style

• Note that delta was already used in Chapter 10 to identify the number of units in the underlying asset we needed to buy to hedge the sale of one option

• Delta changes in each point of the tree, which shows that option positions require dynamic hedging in order to

remain risk free

The Gram-Charlier Model

• As the delta is a partial derivative of an option pricing

model with respect to the underlying asset price, it is

fundamentally model dependent

• The preceding deltas were derived from the BSM model, but different option pricing models imply different

formulas for the deltas

• We saw in the previous chapter that the BSM model

sometimes misprices traded options quite severely

• We therefore want to consider using more accurate option pricing models for calculating the options delta

The Gram-Charlier Model

• In the case of the Gram-Charlier option pricing model, we have

• and the partial derivative with respect to the asset price in this case is

The Gram-Charlier Model

• which collapses to the BSM delta of (d) when skewness,

11, and excess kurtosis, 21, are both zero

• Again, we can easily calculate the put option delta from

GARCH Option Pricing Models

• Calculating deltas from the general GARCH option pricing model, we face the issue that the option price is not available

in closed form but must be simulated

• We have in general

• which we compute by simulation as

• where is the hypothetical GARCH asset price on

option maturity date for Monte Carlo simulation

path i, where the simulation is done under the risk-neutral

distribution

GARCH Option Pricing Models

• The partial derivative of the GARCH option price with

respect to underlying asset price can be shown to be

• where the function 1(*) takes the value 1 if the argument is

true and zero otherwise

• GARCH delta must also be found by simulation as

• Where is again the simulated future risk-neutral

asset price

GARCH Option Pricing Models

• The delta of the European put option can still be derived

from the put-call parity formula

• In the special case of closed-form GARCH process, we have the European call option pricing formula

• and the delta of the call option is

• The formula for P1 is given in the appendix to the previous chapter

The Portfolio Risk Using Delta

• Consider a portfolio consisting of just one (long) call option

on a stock

• The change in the dollar value of the option portfolio,

DV PF,t+1, is then just the change in the value of option

• Using the delta of the option, we have that for small

changes in the underlying asset price

• Defining geometric returns on underlying stock as

The Portfolio Risk Using Delta

• and combining the previous three equations, we get the

change in the option portfolio value to be

• The upshot of this formula is that we can write the change

in dollar value of the option as a known value t times the future return of the underlying asset, R t+1

• Notice that a portfolio consisting of an option on a stock

corresponds to a stock portfolio with  shares

• Similarly, we can think of holdings in underlying asset as

having a delta of 1 per share of underlying asset

• Trivially, the derivative of a stock price with respect to the

stock price is 1

The Portfolio Risk Using Delta

• Thus, holding one share corresponds to having =1, and

holding 100 shares corresponds to having =100

• And, a short position of 10 identical calls corresponds to

setting =-10c, where c is delta of each call option

• The delta of a short position in call options is negative, and the delta of a short position in put options is positive as the delta

of a put option itself is negative

• The variance of the portfolio in delta-based model is

• where 2

t+1 is the conditional variance of the return on the underlying stock

The Portfolio Risk Using Delta

• Assuming conditional normality, the dollar Value-at-Risk

(VaR) in this case is

• where the absolute value, abs(*), comes from having

taken the square root of the portfolio change variance,

2

DV,t+1

• Notice that since DV PF,t+1 is measured in dollars, we are

calculating dollar VaRs directly and not percentage VaRs

• The percentage VaR can be calculated immediately from

the dollar VaR by dividing it by the current value of the

portfolio

The Portfolio Risk Using Delta

• In case we are holding a portfolio of several options on the

same underlying asset, we can simply add up the deltas

• The delta of a portfolio of options on the same underlying asset is just the weighted sum of the individual deltas as in

• where the weight, m j, equals the number of the particular

option contract j

• A short position in a particular type of options corresponds

to a negative m

The Portfolio Risk Using Delta

• In the general case where the portfolio consists of options

on n underlying assets, we have

• In this delta-based model, the variance of the dollar

change in the portfolio value is again

• Under conditional normality, the dollar VaR of the portfolio

is again just

The Portfolio Risk Using Delta

• Thus, in this case, we can use the Gaussian risk

management framework without modification

• Linearization of option prices through the use of delta,

together with assumption of normality, makes the

calculation of the VaR and other risk measures easy

• Note that if we allow for standard deviations, i,t+1, to be time varying as in GARCH, then the option deltas should ideally be calculated from the GARCH model

• We recall that for horizons beyond one day, the GARCH returns are no longer normal, in which case the return

distribution must be simulated

The Portfolio Risk Using Delta

• When volatility is assumed to be constant and returns are assumed to be normally distributed, we can calculate the

dollar VaR at horizon K by

• where DV is the daily portfolio volatility and where K is

risk management horizon measured in trading days

The Option Gamma

• The linearization of the option price using the delta

approach often does not offer a sufficiently accurate

description of the risk from the option

• When underlying asset price makes a large upward move

in a short time, call option price will increase by more than the delta approximation would suggest

• If the underlying price today is $100 and it moves to $115, then the nonlinear option price increase is substantially

larger than the linear increase in the delta approximation

The Option Gamma

• Risk managers care deeply about large moves in asset

prices and this shortcoming of the delta approximation is therefore a serious issue

• A possible solution to this problem is to apply a quadratic rather than just a linear approximation to the option price

• The quadratic approximation attempts to accommodate part of the error made by the linear delta approximation

The Option Gamma

• The Greek letter gamma, , is used to denote the rate of change of  with respect to the price of the underlying

asset, that is,

• The model option price is approximated by the

second-order Taylor expansion

Figure 11.4: Call Option Price (blue) and the Gamma

Approximation (red)

The Option Gamma

• For a European call or put on an underlying asset paying a cash

flow at the rate q, and relying on the BSM model, the gamma can

be derived as

• and where (*) as before is the probability density

function for a standard normal variable,

Figure 11.5: The Gamma of an Option

The Option Gamma

• When option is at-the-money, the gamma is relatively large and when option is deep out-of-the-money or deep in-the-

money gamma is relatively small

• This is because the nonlinearity of the option price is highest when the option is close to at-the-money

• Deep in-the-money call option prices move virtually

one-for-one with the price of the underlying asset because the

options will almost surely be exercised

• Deep out-of-the-money options will almost surely not be

exercised, and they are therefore worthless regardless of

changes in the underlying asset price

The Option Gamma

• For these options, the linear delta-based model can be

highly misleading

• Finally, we note that gamma can be computed using

binomial trees as well

• The formula used for gamma in the tree is simply

• and it is based on the change in the delta from point B to C

in the tree divided by the average change in the stock price

when going from points B and C

Portfolio Risk Using Gamma

• In the previous delta-based model, when considering a

portfolio consisting of options on one underlying asset, we have

• where  denotes the weighted sum of the deltas on all the

individual options in the portfolio

Table 11.2: Gamma of American Put Option

F

654.25

445.75

Portfolio Risk Using Gamma

• When incorporating the second derivative, gamma, we

instead rely on the quadratic approximation

• where the portfolio  and  are calculated as

• where again m j denotes the number of option contract j in

the portfolio

Cornish-Fisher Approximation

• If we assume that the underlying asset return, R t+1, is

normally distributed with mean zero and constant variance

2, and rely on the preceding quadratic approximation, then the first three moments of the distribution of changes in the value of a portfolio of options can be written as

Cornish-Fisher Approximation

• For example, we can derive the expected value as

Cornish-Fisher Approximation

• In particular, we note that even if underlying return has mean zero, the portfolio mean is no longer zero

• More important, the variance formula changes and the

portfolio skewness is no longer zero, even if the underlying asset has no skewness

• The asymmetry of the options payoff itself creates

asymmetry in the portfolio distribution

• The linear-normal model presented earlier fails to capture the skewness, but quadratic model considered here captures the skewness at least approximately

• In this way, the quadratic model can offer a distinct

improvement over the linear model

Cornish-Fisher Approximation

• The approximate Value-at-Risk of the portfolio can be

calculated using the Cornish-Fisher approach

• The Cornish-Fisher VaR allowing for skewness is

• Unfortunately, the analytical formulas for the moments of options portfolios with many underlying assets are quite

cumbersome, and they rely on the unrealistic assumption of normality and constant variance

• We will therefore now consider a much more general but simulation-based technique that builds on the Monte Carlo method

Simulation Based Gamma Approximation

• Consider again the simple case where the portfolio consists

of options on only one underlying asset and we are interested

in the K-day $VaR

• We have

• Using the assumed model for the physical distribution of

the underlying asset return, we can simulate MC pseudo

K-day returns on the underlying asset

Simulation Based Gamma Approximation

• and calculate the hypothetical changes in the portfolio value as

• from which we can calculate the Value-at-Risk as

• In the general case of options on n underlying assets, we

have

Simulation Based Gamma Approximation

portfolio with respect to the ith return

• If we in addition allow for derivatives that depend on several underlying assets, then we write

• which includes the so-called cross-gammas, ij

• For a call option, for example, we have

Simulation Based Gamma Approximation

• Cross-gammas are relevant for options with multiple sources