Fabrice RIVA Option Pricing and Hedging

Option Pricing and Hedging

Fabrice Riva - Master 224

DYNAMIC HEDGING

Fabrice Riva - Master 224

Master 224 1
Fabrice RIVA Option Pricing and Hedging

The context
• A financial institution selling an option to a client in the OTC market
is faced with the problem of managing its risk
• If the option happens to be the same as one that is traded on an
exchange, the exposure can be neutralized by buying the same
option on the exchange
• When the option has been tailored to the needs of a client, hedging
the exposure is far more difficult
• What are the various alternatives?
– stop-loss strategies
– greeks-based strategies
– static hedging

Fabrice Riva - Master 224

Illustration
• A bank has sold for €300,000 a European call option on 100,000
shares of a nondividend paying stock
• Data:
– risk-free rate: 𝑟 = 5 %
– initial stock value: 𝑆0 = 49
– strike: 𝐾 = 50
– Time to maturity: 𝑇 = 20/52 years
– estimated volatility: 𝜎 = 20 %
– expected return from the stock: 𝜇 = 13 %
• What is this position worth? What is the risk of a loss at expiration?
What is the risk of an overall loss?

Fabrice Riva - Master 224

Master 224 2
Fabrice RIVA Option Pricing and Hedging

Immediate gain and potential loss

• Black-Scholes value of the option:

• Since the institution has sold 100,000 call options, the inflow of
300,000 ensures an immediate "profit" equal to €300,000 –
€240,053 ≃ €60,000
• Yet, the institution still faces the problem of hedging the risk

Fabrice Riva - Master 224

Exercise probability - I
• The option will be exercised if the price of the underlying asset at the
expiration date is greater than (or equal to) 50
• From the stochastic integral:

• Here:

• The probability that 𝑆𝑇 ≥ 50, knowing that 𝑆0 = 49 is equal to the

50
probability that 𝑟𝑇𝑐 ≥ ln = 2.02%
49

Fabrice Riva - Master 224

Master 224 3
Fabrice RIVA Option Pricing and Hedging

Exercise probability - II
• Direct computation (from Excel NORMDIST function) :

• Using standard normal probability table:

Fabrice Riva - Master 224

Probability of losing money

• The institution loses money if the price the underlying asset moves
up by 4€ at least
• In that case, on expiration date, the institution must buy the
underlyng asset at 49+4 = €53 and deliver it at €𝐾=50, thus
incurring a loss of 100,000 × €3 = €300,000 (its initial inflow)
• The probability of losing money is thus equal to 𝑃 𝑆𝑇 ≥ 53
• From previous computations, it is found that 𝑃 𝑆𝑇 ≥ 53 = 38.53%

Conclusion
• There is a 57% probability that part of the initial profit will be lost
• There is a 38% probability of an overall loss

What should be done?

Fabrice Riva - Master 224

Master 224 4
Fabrice RIVA Option Pricing and Hedging

The problem
• Initially, the trader receives the premium (€300,000)
• At maturity (after 20 weeks) he will have to pay €100,000 ×
max 𝑆𝑇 – 50; 0
– If the stock price ends up below €50, the call holder will not exercise the call
option
– If the stock price ends up over €50, the call holder will exercise and the trader will
have to deliver the shares
– In case the shares have to be delivered, they can be either:
– bought on the market at maturity
– previously bought and held until maturity
• What are the hedging alternatives?

Fabrice Riva - Master 224

10

Hedge and forget strategies

• There are 2 static (or hedge and forget) strategies:
1. Do nothing (keep a naked position), i.e. wait until maturity with the hope that the call
option will not be exercised  effective if the terminal stock price is below €50
2. Completely cover the written call (adopt a covered position), i.e. buy 100,000 shares
initially with the hope that the call option will be exercised  works well if the terminal
stock price is over €50
• None of these 2 strategies provides a good hedge since the institution
faces highly risky prospects
Underlying stock terminal price
€40 €60
Strategy 1 €300,000 - €700,000
Strategy 2 - €600,000 €400,000

A good hedge should ensure that the cost is close to €240,000 (with a €60,000
profit) whatever the terminal price (and the path followed by the price)

Fabrice Riva - Master 224

Master 224 5
Fabrice RIVA Option Pricing and Hedging

STOP-LOSS HEDGING

Fabrice Riva - Master 224

12

The principle
• A simple dynamic alternative to hedge and forget is the
stop-loss strategy
• Objective of the hedge:
– Owning the 100,000 shares if the option closes in the money
– Not owning any share if the option closes out of the money
• Hedging scheme:
– Buy 100,000 shares as soon as the stock price rises above the
strike price
– Sell the shares whenever the stock price drops down below the
strike price
• This strategy seems to produce payoffs that are the
same as those of the option

Fabrice Riva - Master 224

Master 224 6
Fabrice RIVA Option Pricing and Hedging

13

Hedging scheme in practice

• At date 0 (initialization):
– Buy the shares if the share price is above 𝐾, i.e. if the option is initially in the
money. The cost is 𝑆0
– Otherwise do nothing (cost = 0)
• Over the lifetime of the option:
– If 𝑆𝑡−1 < 𝐾 and 𝑆𝑡 ≥ 𝐾, buy the shares on date 𝑡 at price 𝐾 (no position is held on
stocks at 𝑡 − 1)
– If 𝑆𝑡−1 > 𝐾 and 𝑆𝑡 ≤ 𝐾, sell the shares on date 𝑡 at price 𝐾 (long position held on
the stock at 𝑡 − 1)
• At maturity date, the institution:
– Holds no stock if the option closes out of the money
– Holds the stocks if the option closes in the money

Fabrice Riva - Master 224

14

Possible trajectories - I

sale

purchase

purchase

Fabrice Riva - Master 224

Master 224 7
Fabrice RIVA Option Pricing and Hedging

15

Possible trajectories - II

sale
sale sale

purchase

purchase purchase

purchase

Fabrice Riva - Master 224

16

Possible outcomes and hedging cost - I

• At date 0, if the option is in the money:
– The stock is bought to ensure potential delivery at maturity date : initial cost = 𝑆0
– During the lifetime of the option, there are 3 scenarios:

The cost of the hedge, whatever

the scenario, is equal to 𝑆0 − 𝐾

Fabrice Riva - Master 224

Master 224 8
Fabrice RIVA Option Pricing and Hedging

17

Possible outcomes and hedging cost - II

• At date 0, if the option is out of the money:
– No stock is bought at date 0 : initial cost = 0
– During the lifetime of the option, there are 3 possible scenarios:

The cost of the hedge, whatever

the scenario, is equal to 0

Fabrice Riva - Master 224

18

Stop-loss hedging: Theoretical issues

• From previous slides, the cost of the hedge 𝑄 is given by the
expression:

• However (1) is theoretically wrong:

– It does not account for the fact that the cash flows of the hedging strategy occur
at different times and must thus be discounted (initially, you may have to borrow
the amount necessary to set up the hedge)
– It gives rise to an arbitrage opportunity: the speculative value of the option at
date 0 does not show up  (1) only gives the lower bound of the hedging cost
• In addition, purchases and sales cannot be made at the same price:
– When the price is €50, the trader cannot know whether the price will then
increase or decrease  he may thus have to resell (re-buy) the shares at a price
below (above) 𝐾 rather than 𝐾
– Theoretically, this situation can arise an infinite number of times due to the
features of the brownian motion

Fabrice Riva - Master 224

Master 224 9
Fabrice RIVA Option Pricing and Hedging

19

Stop-loss hedging: Practical issues

• The cost of hedge in (1) does not account for transaction costs and
trading frictions (price discreteness, bid-ask spread, price impact,
…). Same in CRR and BS however.
• 𝑆𝑡 = 𝐾  Buy, Sell or do nothing ?
– As a pratical matter, purchases must be made at a price 𝐾 + 𝜀 and sales at a
price 𝐾 – 𝜀
– Every purchase and subsequent sale then involves a cost of 2𝜀 (ignoring explicit
transaction costs)
– Increased monitoring allows lower 𝜀
– Lower 𝜀 lowers the cost borne on each trade but the average number of trades
increases (due to brownian motion characteristics)

Fabrice Riva - Master 224

20

Stop-loss hedging: Conclusion

• In some situations, the hedging cost is actually equal to zero, but in
other situations, it is well above 0
• The hedging cost is path-dependent
• How to assess the overall performance of stop-loss hedging?
– Monte-Carlo simulations
– Compute performance measures for the full set of simulations
– A conventional measure is given by the ratio of the standard deviation of the cost
of hedging the option to the theoretical price of the option
• Simulations show that:
– The performance of stop-loss hedging is poor
– The performance of stop-loss hedging does not improve as the size of 𝜀 is
reduced

Fabrice Riva - Master 224

Master 224 10
Fabrice RIVA Option Pricing and Hedging

DELTA-NEUTRAL HEDGING

Fabrice Riva - Master 224

22

Option delta and hedging

• A given value for the ∆ means that when the underlying price
changes by a small amount the option price changes by Δ times this
amount
• Consider the following portfolios:
– long 1 share and short 1/Δ option
– long Δ shares and short 1 option
• What is the result of a €1 increase in the stock price?
• The gain (loss) on the option position tends to be offset by the loss
(gain) on the stock position
• These portfolios are locally riskless

Fabrice Riva - Master 224

Master 224 11
Fabrice RIVA Option Pricing and Hedging

23

Illustration
• Assume that the initial value of delta is 0.522 for an option on one
stock
• As the trader is short a call on 100 000 shares, the delta of his
position is – 52,200
• The institution should thus buy 52,200 shares so that each 1 euro
increase (decrease) in the stock price is offset by a 1 euro decrease
(increase) in the option price
• The portfolio has a delta equal to zero
• Such portfolio is referred to as being delta-neutral

Fabrice Riva - Master 224

24

Hedge and forget? - I

• If the stock price goes up by 1€
– The institution loses €0.522 × 100,000 on the written call
– The institution gains €52,200 × 1 € on the purchased shares
– The value of the overall position is stable…
– … but the delta of the options goes up to 0.586 (see below)
• If the stock price goes down by 1€
– The institution earns €0.522 × 100,000 on the written call
– The institution loses 52,200 × €1 on the purchased shares
– The value of the overall position is stable…
– … but the delta of the options goes down to 0.455

Fabrice Riva - Master 224

Master 224 12
Fabrice RIVA Option Pricing and Hedging

25

Hedge and forget? - II

• Because delta changes, the trader’s position remains delta neutral
for only a relatively short period of time
• To remain delta neutral, the hedge has to be readjusted
(rebalanced) periodically
– As delta increases, the institution has to buy more stock
– As delta decreases the institution has to sell part of its inventory
• At maturity
– If the option closes in the money, the institution can deliver the 100,000 shares it
previously bought
– If the option closes out of the money, the institution has no more shares in its
inventory

Fabrice Riva - Master 224

26

Simulation

Fabrice Riva - Master 224

Master 224 13
Fabrice RIVA Option Pricing and Hedging

27

Delta hedging features

• Delta hedging is a momentum strategy : stocks are purchased after
increases and sold after decreases in stock price
• The average cost of delta hedging (on a risk neutral trajectory) is the
Black-Scholes theoretical option value  comes from the average
difference between the price paid for the stock and the price realized
for it plus the interest borne on the amount borrowed for setting up
the hedge
• The hedging scheme is only locally perfect
– With continuous rebalancing performance is perfect (zero hedging error)
– Discrete time rebalancing gives rise to variations in the cost of hedging
– With transaction costs, delta hedgers are confronted with a tradeoff between
trading costs and hedging error

Fabrice Riva - Master 224

PORTFOLIO INSURANCE

Fabrice Riva - Master 224

Master 224 14
Fabrice RIVA Option Pricing and Hedging

OPTION-BASED PORTFOLIO
INSURANCE (OBPI)

Fabrice Riva - Master 224

30

OBPI: Objective
• OBPI is a class of portfolio insurance methods
• The objective is to garantee a fund's subscriber that he will be
returned a given % of his initial investment at the end of his
investment horizon
• This method has been successfully implemend by Leland, O'Brien
and Rubinstein (LOR) until the collapse of their company during the
October 19, 1987 market crash

Fabrice Riva - Master 224

Master 224 15
Fabrice RIVA Option Pricing and Hedging

31

OBPI and put option

• Guaranteeing a given amount over horizon can be achieved using a
put option
– Identify portfolio insurance date 𝑇
– Buy puts with maturity 𝑇 and strike price 𝐾 where 𝐾 corresponds to the
guaranteed floor
– With a one-to-one relationship between the number of risky assets in the portfolio
to insure and the number of put options to purchase
• There are practical problems however:
– For some indices traded options are only of American style
– Desired strike prices and maturities may not be available or liquid enough
– The underlying asset of the options may differ from the portfolio to be insured
– Legal constraints might forbid investing in derivatives

Fabrice Riva - Master 224

32

The synthetic put - I

• Rather than buying an option, dynamic portfolio insurance consists
in creating the put option synthetically
• The starting point of the method is the Black-Scholes formula for the
value of a put option:

• The first term in the right-hand side of (1) corresponds to the amount
invested in the underlying asset: this is a short position
• The second term in the right-hand side of (1) corresponds to the
amount invested in the risk-free asset: this is a long position
• At any time 𝑡, a portfolio combining the risky asset and a put with
strike 𝐾 on this asset ensures a guranteed floor of 𝐾 (ignoring the
premium paid to buy the option), i.e. 𝑆 𝑡 + 𝑃(𝑡) ≥ 𝐾

Fabrice Riva - Master 224

Master 224 16
Fabrice RIVA Option Pricing and Hedging

33

The synthetic put - II

• Replacing 𝑃(𝑡) by its expression in (1) , the asset + put value
expresses as:

• From (2), it is possible to infer the proportions 𝛼𝑡 and (1 − 𝛼𝑡 ) that

have to be held in the risky asset and the risk-free bond
(respectively) to achieve the guaranteed floor:

Fabrice Riva - Master 224

34

The synthetic put - III

• The quantities 𝛼𝑡 and (1 − 𝛼𝑡 ) must be regularly (ideally
continuously) updated over the lifetime of the fund for two reasons:
– Changes if the underlying asset value induce change in 𝛼𝑡 (and thus 1 − 𝛼𝑡 )
– Even with no change in 𝑆(𝑡) the effect of time on 𝑑1 and 𝑑2 induces changes in
𝛼𝑡

Fabrice Riva - Master 224

Master 224 17
Fabrice RIVA Option Pricing and Hedging

35

Implications of OBPI
• As the value of the original portfolio increases
– The delta of the put becomes less negative
– Risk-free bonds are sold and the position in the stock is increased (the
proportion of the portfolio sold must be decreased, i.e. some of the
original portfolio must be repurchased)
– the maximum limit will be reached when the entire original portfolio has
been repurchased (∆𝑝 = 0), that is, when the portfolio is made up of only
risky assets.
• As the value of the original portfolio declines
– The delta of the put becomes more negative
– The position in the stock portfolio is decreased (i.e. some of the original
portfolio must be sold) and risk-free bondsare purchased
– Rhe maximum limit will be reached when the entire original portfolio has
been sold (∆𝑝 = −1), that is, when the portfolio is made up only of risk
free bonds
• OBPI is a momentum strategy

Fabrice Riva - Master 224

36

Illustration: Guaranteed floor = 90

Fabrice Riva - Master 224

Master 224 18
Fabrice RIVA Option Pricing and Hedging

37

Illustration: Guaranteed floor = 90

Fabrice Riva - Master 224

38

Overall performance

Fabrice Riva - Master 224

Master 224 19
Fabrice RIVA Option Pricing and Hedging

CONSTANT PROPORTION
PORTFOLIO INSURANCE
(CPPI)

Fabrice Riva - Master 224

40

CPPI: Principle
• Black and Jones introduced this method in 1986 in a research paper
from the Goldman Sachs Bank with the express goal of simplifying
portfolio insurance
• CPPI is also called the « cushion method »
• The strategy is very both simple and flexible
– It does not require the estimation of the volatility of tje underlying asset
– It can continue indefinitely as it has no time-horizon
– The floor can easily be modified as time goes by or after significant increases

Fabrice Riva - Master 224

Master 224 20
Fabrice RIVA Option Pricing and Hedging

41

CPPI: the "two" portfolios

• We can envision the CPPI portfolio according to two viewpoints
– the portfolio is split between a risky position (exposure), 𝑆𝑡 , and a risk-free
position, 𝑀𝑡 . The expression of the value of the portfolio 𝑉𝑡 is thus:

– The portfolio can be seen as well as the sum of the guaranteed floor, 𝐹 and the
cushion, 𝐶𝑡 , so that 𝑉𝑡 = 𝐹 + 𝐶𝑡
• Graphically:

Fabrice Riva - Master 224

42

The multiple
• The multiple 𝑚𝑡 is defined as the ratio of the exposure to the
𝑆
cushion, i.e. 𝑚𝑡 = 𝑡
𝐶𝑡
1
• The inverse of the multiple 𝜆𝑡 = is called the management ratio
𝑚𝑡
• At the inception of the fund, the manager must choose the value of
the target multiple 𝑚∗ or the target management ratio 𝜆∗
• The CCPI method consists in:
– Keeping a constant proportion of exposure to the risky asset
– By keeping the amount in the risky asset proportional to the value of the cushion
by the factor 𝑚∗
• At any time, we must have thus:

Fabrice Riva - Master 224

Master 224 21
Fabrice RIVA Option Pricing and Hedging

43

CPPI management step by step

Step 1 : Choose the floor and the multiple

Step 2 : Calculate cushion size as 𝐶 = 𝑉 – 𝐹
Step 3 : Calculate the exposure as 𝑆 = 𝑚 × 𝐶
Step 4 : Calculate the amount invested in the risk-free asset as 𝑀 =
𝑉– 𝑆
Step 5 : Dynamically rebalance to keep the exposure constant

Fabrice Riva - Master 224

44

CPPI: Example - I
• Suppose that 𝑉0 is equal to 1,000
• 𝐹 is set to 900
• Initially, the cushion is 𝐶0 = 1,000 − 900 = 100
• Consider a target multiple set to 𝑚∗ = 4, or 𝜆∗ =.25
• The amount that has to be invested in the risky asset is 𝑆0 = 𝐶0 ×
𝑚∗ = 400
• Subtracting this value to the portfolio value, we get that 𝑀0 = 𝑉0 −
𝑆0 = 600
• The above process describes the initialization of the CPPI method.
Next, based on fluctuations in 𝑆, rebalancements will take place to
maintain the exposure of the portfolio equal to 𝑚∗

Fabrice Riva - Master 224

Master 224 22
Fabrice RIVA Option Pricing and Hedging

45

CPPI: Example - II
• Assume that the risky asset increases by 10%
– 𝑆1 = 440
– The cushion is now worth 𝐶1 = 1040 − 900 = 140
440
– The multiple decreases to 𝑚1 = = 3.14
140

• In order to bring 𝑚 back to its target value 4, we must adjust the

risky position
– The objective is 𝑆1∗ = 140 × 4 = 560
– We must thus buy extra units units of the risky asset for an amount equal to
560 − 440 = 120
– Conversely, the risk-free position must decrease by 120 (CPPI is a self-financing
strategy)

Fabrice Riva - Master 224

46

Graphical illustration
After an increase After a decrease in
in the value of the the value of the
risky asset: risky asset:
• The portfolio • The portfolio
value 𝑉′ value 𝑉′
increases decreases
• Since the floor 𝐹 • Since the floor 𝐹
is constant, the is constant, the
cushion 𝐶 ′ cushion 𝐶 ′
increases decreases
• The exposure 𝑆′ • The exposure 𝑆′
must be must be
increased and decreased and
the amount the amount
invested in the invested in the
risk-free asset risk-free asset
𝑀′ must be 𝑀′ must be
decreased increased

Fabrice Riva - Master 224

Master 224 23
Fabrice RIVA Option Pricing and Hedging

47

CPPI properties
• For a systematically increasing risky asset
– The cushion is always increasing
– For a €1 increase in the cushion, the exposure increases by €𝑚∗
– The amount invested in the risky-free asset tends to become negative
• For a systematically decreasing risky asset
– The cushion converges to 0
– The exposure converges to 0
– The amount invested in the risk-free asset converges to the present value of the
floor
• Even for a continuous decrease in 𝑆 the portfolio value always
remains above the floor because the exposure is continuously
adjusted (so that it remains proportional to the cushion)
• For a discrete decrease in 𝑆 the portfolio value breaks the floor if the
decrease is higher than the cushion. The negative return the
portfolio can absorb between two rebalancements is equal to the
management ratio
Fabrice Riva - Master 224

48

CPPI limitations - I
• Rebalancements may be costly in volatile markets
• Conventionally, a third parameter, the tolerance, is set to limit the number
of portfolio adjustments
– The portfolio is not rebalanced unless the observed multiple moves by a given
percentage (20%, for example)
– When the portfolio is rebalanced, adjustments should be such that the multiple is
brought back to its target value
• Higher tolerance leads to higher gap risk
– The strategy can guarantee the floor only for a negative return inferior to the
management ratio minus its tolerance
• A higher multiple results in
– A higher exposure
– A portfolio that better benefits from market rises
– A portfolio that comes closer to the floor more rapidly in bear markets
• It translates into a lower management ratio  in the case of discrete
rebalancements the probability to break the floor (gap risk) is higher
Fabrice Riva - Master 224

Master 224 24
Fabrice RIVA Option Pricing and Hedging

PORTFOLIO INSURANCE:
COMPARING THE METHODS

Fabrice Riva - Master 224

50

Common features - I
• PI strategies are momemtum strategies
– It is necessary to have this kind of buy high/sell low scheme to obtain convex
payoffs but may be seen as contradictory with respect to basic investment theory
– Some argue that such schemes tend to accentuate markets tendencies or even
cause market crashes
• Transaction costs
– Lead to returns that are negatively related to volatility
– Make it necessary to have some tolerance in the rebalancing scheme, but:
• The effective performance is modified (replication error)
• The gap-risk increases
– May be reduced by the use of futures, but:
• Margin calls must be handled
• Basis risk must be accounted for

Fabrice Riva - Master 224

Master 224 25
Fabrice RIVA Option Pricing and Hedging

51

Common features - II
• Stochastic interest rates may cause a premature cash-out for CPPI
– A sharp decrease in 𝑆 leads to wholly invest the portfolio in safe assets
– Guarantee may not be achived in the case of a subsequent decrease in interest
rates
• Gap risk
– Theoretically, portfolios must be rebalanced continuously which is not the case in
reality
– OBPI may break the floor at intermediairy dates
• Results are path-dependant
– due to discrete rebalancing (OBPI)
– due to the adjustement rule per se (CPPI)

Fabrice Riva - Master 224

52

Which method?
• Path dependancy
– CPPI results are more path-dependent than dynamic OBPI results
– Static OBPI is path independant
• Volatility
– P&L of static OBPI only depends on implied volatility
– Volatility estimation is critical in the case of dynamic OBPI
– CPPI results depends on realized volatility
• Performance
– Under no arbitrage, no strategy should dominate the others in all states of nature
– It is not straightforward to compare the performance of the different strategies:
• Various comparison criteria can be used (expected returns, volatility of
returns, gap-risk, conditionnal returns, semi-variance, …)
• Oonly comparable strategies should be compared (with at least the same
floor and expected return)
– CPPI tends to outperform OBPI in markets with marked tendencies (either bull or
bear)

Fabrice Riva - Master 224

Master 224 26