FNCE 30007

Derivative Securities

Lecture – Delta hedging

Outline

 Hedging an option
 Delta hedging
 Hedging errors
 Synthetic options
 Rebalancing
Hedging an option
 Risk
 refers to the degree of uncertainty in the future value of
(or change in the value) of a variable – price of a stock (or
some other asset)
 Hedger
 Already has exposure to the future price of a stock and
seeks to reduce this risk – stock price risk
 Hedging reduces or eliminates an exposure
 Perfect hedge = eliminate 100% of the risk
 An option (or other derivative) may be used to hedge an
exposure to a stock
 A stock may be used to hedge an exposure to an option
 Consider a bank that writes an OTC option for a client. Bank wants to
hedge this risk
Hedging an option
 Static hedging
 Simplest way to hedge an option exposure is to take
opposite position and hold to maturity
 eg. Assume you write a call on a stock with strike K
and maturity T
 Diagram

 may not always be available (or may be too

expensive)
 Other ways to hedge an option
 Naked position
 Covered position Don’t work well
 Stop-loss strategy
 Delta hedge
Delta Hedging
 Risk of an option exposure
 Current value of an option on a non-div paying stock is a
function of
f = f (S , K , r, T ,σ )
 A change in any one of these variables change in f.
 K is fixed no need to hedge
 No uncertainty re passing of time no need to hedge
time decay
 BSM assumes only uncertainty affecting option value is
stock price risk, but in practice
 Both r , σ are likely to change over life of option

 Each party exposed to counterparty default

 more than one source of risk

Delta Hedging
 Greeks
 Describe some different dimensions to an options risk
 Measure the sensitivity in option value to changes in value
of each underlying variable
 Delta: change in f due to a change in S

 Rho: change in f due to a change in r

 Vega: change in f due to a change in σ

 Theta: change in f due to a change in T

 Gamma: change in delta due to a change in S

 Each Greek assumes only one variable changed at a time

i.e. all other variable are held constant
 We will focus on delta and gamma
Delta Hedging
 Delta of an option
 Delta is the rate of change of the option price with respect
to the stock price (assuming all other variables are
constant)
∂f
 For small changes in S: ∆ ≈
∂S
 Alternatively: ∂f ≈ ∆∂S

 Technical note: delta is the partial derivative of f with respect to S

 Notation:
 Hull (2002) uses lower case “delta” δ

 Hull (2005,8) uses upper case “delta” ∆

Delta Hedging
 Example
 Consider a call on one Speedy Ltd share.
The option has a delta of 0.6, current price
of $10, and current share price of $100.
 Assume stock price immediately increases by 10
cents.
 Use delta to estimate the price of the option
immediately after the change
Delta Hedging
 Delta of an option as a function of the stock
price
 Long call 0 ≤ ∆ ≤1
 Long put −1 ≤ ∆ ≤ 0
 Diagrams

 Delta of other securities

 Define delta (relative to changes in the stock
price) of any other security (eg futures) in similar
way
∂F
∆ futures =
∂S
Delta Hedging
 Delta of a portfolio
 Portfolio consists of
 n1 units of security 1 (value V1, delta ∆1 )
 n2 units of security 2 (value V2 , delta ∆ 2 )
 Both securities have same source of uncertainty - stock price
 Value of portfolio is
=
V p n1V1 + n2V2
 Delta of portfolio is
∆ p = n1∆1 + n2 ∆ 2
 Long (short) position represented by positive (negative)
values of n1 or n2
Delta Hedging
 Delta Hedging
 Portfolio with a zero delta is delta neutral
 Not affected by stock price risk (or small changes in S)
 Delta hedge an option (or portfolio of securities) by
combining it with the underlying stock so that the delta of
the combined portfolio is zero.
 Required to buy or sell underlying stock

 Required to buy or sell risk less bonds

 Stock purchases funded by selling riskless bonds (borrow
at risk free rate)
 Proceeds from stock sales used to buy riskless bonds
(invest at risk free rate)
 Alternatively, can delta hedge an option/portfolio using
other securities (e.g futures or other options on the
stock)
Delta Hedging
 Delta Hedging (cont’d)
 Example
 Assume you have just written 20 Speedy call options
 i) How can you delta hedge this exposure using the stock?
Security Delta Number
Call 1 0.6 n1 = −20
Stock 2 1 n2 = ?

 ii) Assume you have just delta hedged your exposure and
the stock price immediately increases by 10 cents. Use
the delta to estimate the impact on the value of the
hedged portfolio.
 Delta Hedging
 Calculating delta of an option
 Delta is model specific
 If binomial model is being used to price the option, delta
at start of each one period sub-tree is
f − fd
∆= u
uS − dS
 If BSM is being used delta at each instant is Non-div paying stock

=
Call : ∆ N ( d1 ) =
Put : ∆ N ( d1 ) − 1

 If Merton’s model is being used delta at each instant is

Call : ∆ e − qT N ( d1 ) =
Put : ∆ e − qT  N ( d1 ) − 1
..
Calculate d1 using St =

 BSM with discrete div is not examinable.

Delta Hedging
 Gamma
 Is the rate of change of the options delta with respect to
the stock price, assuming all other variables constant.
 For small changes in S:
∂∆
Γ≈ ∂∆ ≈ Γ∂S
∂S

 Technical note: Exact definition is the second partial

derivative of f with respect to S
∂∆ ∂ 2 f
=
Γ =
∂S ∂S 2
Delta Hedging
 Gamma (cont’d)
 Using Merton’s model, the gamma for a call or put at each
instant is: − qT − d 2 /2
e 1

Γ= Γ call =
Γ put > 0
.. Sσ 2π T
St not S t
 The BSM gamma is derived by setting q=0.
 Gamma is greatest when at the money and decreases the
further the option goes either into or out of the money.
Hedging errors
 In theory, delta hedging is perfect
 Binomial & BSM models assume we can form a perfect
riskless hedge
 This assumption is based on assumptions that are unlikely
to hold in practice
 Binomial model assumes stock price can only take on one of
two possible values at the end of each period – many prices
are available.
 BSM assumes
i) can trade & hence hedge continuously – can only trade at
discrete points in time
ii) stock prices change continuously (i.e at any time they
can change by a very small amount) – in reality stock prices
may jump even in the absence of divs
 Delta hedging likely to be imperfect
 Hedging errors
 Assume you wish to hedge a call option against an
immediate change in the stock price. Let
∂S = S '− S = actual change in S
∂f = f '− f = actual change in f

 Delta hedging implies that for a small change in S

Actual change in f ∂f ≈ ∆∂S Estimated change in f

 Or equivalently
∂f = ∆∂S + ε1
 i.e the hedging error is
ε1 = ∂f − ∆∂S
Hedging errors
 In a plot of option price versus stock price, delta is the
slope of the tangent to the curve at the current stock price
 Delta hedging assumes that f is a linear function of S,
when it is in fact curved (convex) function of S
 hedging error
 Delta hedging always underestimates the price of a call
after a stock price change
 Underestimates ∂f if stock price increases

 Overestimates ∂f if stock price decreases

 Delta hedging always underestimates the price of a put

after a stock price change and so
 Overestimates ∂f if the stock price increases

 Underestimates ∂f if the stock price decreases

Hedging errors

 Gamma measures some of the curvature between the

option and stock prices
 Taking account of gamma gives better estimate of ∂f

 For a small ∂S

Γ ( ∂S ) Γ ( ∂S )
2 2

∂f ≈ ∆∂S + → ∂f = ∆∂S + + ε2
2 2
Taylor series expansion
 where
Γ ( ∂S )
2

ε 2 = ∂f − ∆∂S − < ε1
2
Positive
Delta hedge error
Synthetic options
 Recall, you hedge an option by taking the opposite
position ( and holding to maturity)
 e.g you hedge a short call by buying a long call
 Key idea behind delta hedging is that rather than buying
(or selling) an option to hedge your exposure, you create
(or replicate) the required option position
 e.g. you hedge a short call by creating a synthetic long call
 Delta hedging creates a synthetic option by trading the
underlying stock and riskless bonds over the life of the
option (also referred to as a “replicating portfolio”)
 Fundamental rule: You create a synthetic option position
in the same way as you would hedge the opposite option
position
 Technical note: It may be better to delta hedge by trading in futures (lower
transaction costs, higher liquidity)
Synthetic options
 Assume you seek to protect the value of a portfolio
(portfolio insurance)
 Buy a put or create a long put synthetically
 Create a long put same way as you would hedge a
short put
f =∆S − π
Sell put Long, but remember ∆<0
π
short ∆S and invest bonds

 Why not just buy a put?

 Option markets may not have enough liquidity to
absorb trades of large funds
 May require strikes & maturity dates that are different
from exchange traded options
Hedging an option
 Example (continued from Binomial Part I)
 Current stock price of a non div paying stock is $20 and at
end of period will be either $22 or $18. Risk free rate is
3% per period and markets are frictionless. Show how to
create a synthetic long European call option on the stock
with a strike of $21 and maturity of one period.

Cash now Cash at end of year

ST = 18 ST = 22
Buy a call
Long 1 call

Create a call
Long 0.25 stock
Short $4.367 bonds
Net
Rebalancing
 Delta hedging is an example of dynamic hedging
 Delta of an option is not constant but changes with
changes in the stock price.
 Option price will also change with time (even if stock price does not
change)
 Delta hedge (delta neutral portfolio) will need to be
rebalanced from time to time to remain delta neutral
 Binomial at the start of each period, BSM each instant
 A measure of importance of rebalancing is the gamma
 Greater the gamma greater the change in delta
for a change in stock price greater potential for
hedging error from not rebalancing
 Gamma is greatest when at the money and therefore
re-balancing is most important here.
Rebalancing
 Example
 Consider a European call with a strike of $80, and
maturity 4 weeks from now on a non div paying stock.
Current stock price $80, risk free rate 5% p.a, stock price
volatility is 25% p.a.
 Assume a bank has just sold this option to a client for $3
and wants to immediately hedge the exposure.
 Describe the dynamic hedging strategy assuming the bank
employs the BSM model, the hedge is rebalanced weekly
and the stock price path is
End of week 0 1 2 3 4
Stock price $80.00 $77.86 $83.10 $83.67 $82.00
Rebalancing
Short position in riskless bonds
(3)
End of Delta # of shares Open Interest Shares Option Close
week bought/sold
St (1) (2) (4) (5) (6)

0 80.00 0.5359 0.5359 0.00 0.00 42.87 42.87

1 77.86 0.3557 -0.1802 42.87 0.04 -14.03 28.88
2 83.10 0.7995 0.4439 28.88 0.03 36.88 65.79
3 83.67 0.9099 0.1103 65.79 0.06 9.23 75.09
4 82.00 -0.9099 75.09 0.07 -74.61 2.00 2.55

Cumulative cost of hedging i.e how much is owed at the end.

If a perfect hedge, PV(2.55) should equal the options fair value.
Rebalancing
 Calculations (previous slide)
 1) Based on BSM (don’t forget to allow for decline in
maturity of option)
 2) Positive (negative) amounts represent stock purchases
(sales)
 3) Positive (negative) amounts represent borrowing
(investing) at risk free rate
 4) Based on opening balance
 5) Positive (negative) amounts represent stock purchases
(sales)
 6) Payoff on short option position at maturity
Rebalancing
 Notes
 1) Closing short position in riskless bonds (end of week 4)
represents the cumulative cost of hedging the option
 2) Difference between the initial fair value of option & the
present value of the cumulative cost of hedging is the
cumulative hedging error over the life of the option
 2.39-2.55exp(-0.05*4/52)
 If a perfect hedge is not possible then the cost of creating an option
will not equal the cost of buying the option
 3) Hedging error depends on the assumed path followed
by the stock over the life of the option
 Different stock paths will have different hedging errors
 Error may be positive or negative
 Could estimate expected error by simulation
Rebalancing
 Notes (cont’d)
 4) A delta hedge is ex-ante perfect only if there is a zero
hedging error for all possible stock price paths
 5) The more frequent you rebalance the lower the
expected hedging error
 In the limit, the BSM assumption of continuous rebalancing leads to a
perfect hedge
 6) In practice, rebalancing involves transaction costs –
increasing the frequency of rebalancing involves a trade-
off between lower expected hedging errors and higher
transaction costs
 7) If stock price falls, then so does the delta
 Delta hedging requires sell (buy) stock when price falls (rises)
 Larger hedging errors if everyone want to rebalance at same time
 e.g. portfolio insurance strategies during 87 crash
 ↓ S →↓ ∆ , synthetic puts needed to re-balance by selling more stock but limited/no
buyers