|
Dynamic hedging is a technique that is widely used by
derivatives dealers to
hedge
gamma or
vega exposures. Because it
involves adjusting a hedge as the
underlier moves—often several times a
day—it is "dynamic." This article discusses the need dynamic hedging
addresses and how it is performed. It identifies an important link between
dynamic hedging and
options pricing theory. It also presents a sophisticated way of
thinking about options (as
volatility bets) that is common
among derivative dealers but alien to most end users of options.
Accordingly, the article is about far more that the simple mechanics of
dynamic hedging.
Traded instruments or positions can generally be broken
down into two types:
linear, and
non-linear.
The former includes spot positions,
forward positions and
futures. Their
payoffs or
market values as a function of some underlier
are either linear or almost linear. Non-linear instruments include
vanilla
options, exotic derivatives and
bonds with embedded options. Their payoffs
or market values as a function of some underlier are highly non-linear.
This distinction is illustrated in Exhibit 1.
|
|
 |
 |
|
Market values as a function of some
underlier value are illustrated for a long future and a long call
option. The future is a linear position. The option is a non-linear
position. |
Derivatives dealers transact in both linear and non-linear
positions with clients. They tend to prefer to transact in non-linear
positions because they are more difficult for counterparties to price,
which means they can make fatter profits on those transactions. There is
another difference in their trading of linear vs. non-linear instruments.
A dealer's clients tend to want to go long or short linear
positions with about equal frequency. For example, an oil company might
want to sell its oil forward to lock in a price. At the same time, a power
plant operator might want to buy oil forward, also to lock in a price. Such
transactions are offsetting, so a derivatives dealer who handles both
transactions can offset them and maintain a largely balanced book of
linear positions.
The same is not true of options or other non-linear
positions. Clients of derivatives dealers routinely want to buy a
call,
buy a put, buy a
cap, or buy some exotic derivative. Rarely does a client
call a derivatives dealer and ask to sell an option. Dealers are left
holding massive short options positions. To
hedge those positions, they
would like to purchase offsetting long options, but there is no one to buy
them from. It makes little sense to buy them from other derivatives
dealers, who are in the same boat with their own massive short options
positions. The solution is to dynamically hedge the short options
positions.
Dynamic hedging is
delta hedging of a non-linear position
with linear instruments like spot positions, futures or forwards. The
deltas of the non-linear position and linear hedge position offset,
yielding a zero delta overall. However, as the underlier's value moves up
or down, the delta of the non-linear position changes while that of the
linear hedge does not. The deltas no longer offset, so the linear hedge
has to be adjusted (increased or decreased) to restore the delta hedge.
This continual adjusting of the linear position to maintain a delta hedge
is called dynamic hedging.
Consider an example. A derivatives dealer sells a client a
put option on STU Corp. stock. At the current stock price of
USD 100, the short option
position has a delta of 22,000 shares. This is evident in Exhibit 2, which
illustrates the market value of the short option position as a function of
the underlying stock's price. A tangent line has been fit to that graph,
and its positive slope indicates the positive delta.
|
|
 |
 |
|
A derivative dealer sells a put option on
STU stock. Its market value is indicated above as a function of the
underlying stock price. The current stock price (indicated by the
gold triangle) is 100. A tangent line has been fit to the graph at
that value. Its positive slope indicates that the position has
positive delta. |
To delta hedge the short put, the dealer sells 22,000
shares of STU stock. The deltas of the short option and the short stock
cancel, yielding an overall delta of zero. The market value of the hedged
position as a function of the stock price is shown in Exhibit 3. A tangent
line fit to that graph has zero slope, indicating zero delta.
|
|
 |
 |
|
The dealer delta hedges the short put
option by selling stock short. The market value of the hedged
position (short put plus short stock) is shown as a function of the
underlying stock's price. A horizontal tangent line indicates that
the hedged position has a zero delta. |
With the underlying stock price at USD 100, the position is
delta hedged, but this doesn't last long. Soon the stock price rises to
USD 103. As indicated in Exhibit 4, at that stock price, the position has
a slightly negative delta. It is no longer delta hedged.
|
|
 |
 |
|
When the underlying stock price rises, the
position is no longer delta hedged. This is indicated by the tangent
line fit to the graph at the new stock price. It has a negative
slope, indicating a negative delta. |
At the new stock price, the derivatives dealer adjusts the
delta hedge, buying back some of the underlying stock he had previously
shorted. The result is a newly delta hedged position at the new stock price
of USD 103. See Exhibit 5.
|
|
 |
 |
|
The dealer adjusts the delta hedge by
buying back some of the underlying stock he previously shorted. The
position is now delta hedged again at the new underlying stock
price. |
The position is once again delta hedged, but not for long.
Soon the underlying stock price moves again, and the delta hedge is thrown
off. The dealer readjusts the delta hedge. The price moves again, and the
dealer readjusts again, and so on. This ongoing process of a market move
throwing off the delta hedge and the dealer readjusting the delta hedge is
illustrated through several cycles of the process in Exhibit 6.
|
|
 |
 |
|
With each move in the underlying stock
price, the dealer readjusts the delta hedge, either buying or
selling shares to achieve a net delta of zero at the new underlying
stock price. This exhibit illustrates three iterations of that
process. |
The continual readjustment of the delta hedge ensures that
the portfolio always has a zero delta—and that it loses only a little
value each time the underlying stock price moves. Note however, that the
portfolio
always loses value. It never gains it. A delta-hedged negative gamma
portfolio loses money irrespective of whether the underlier rises or
falls. Look again at Exhibits 2 through 6. As your eyes run from one chart
to the next, you see the portfolio slowly but inexorably losing money.
Each time the underlier moves, the portfolio suffers a small loss. The dealer
readjusts the delta hedge, locking in that loss. The underlier moves
again, causing another loss. The dealer readjusts the delta hedge again,
locking in that loss as well. The process continues until the option
expires and the dealer can stop dynamic hedging. This is an important observation with
implications that go well beyond the immediate problem of dynamically hedging a
short put option. Let's briefly explore some of those implications.
First of all, the portfolio loses money with dynamic
hedging because it has negative gamma—something the dynamic hedging cannot
change. How that negative gamma came about is immaterial. It could have
been achieved by shorting a put, or shorting a call, or shorting some
exotic derivative. The fact that the portfolio has negative gamma means
that the dealer is going to lose money dynamically hedging it. If the
portfolio had positive gamma, the opposite would be true. The dealer would
make money dynamically hedging it. Each time the underlier moved, the
portfolio would make a small profit. By readjusting the delta hedge, the
dealer would lock in this small profit ... and so on.
To recap, you lose money dynamically hedging a negative
gamma position. You make money dynamically hedging a positive gamma
position. To make sense of this observation, note that negative gamma
positions arise when you sell options. You receive a
premium for selling
the options but lose money dynamically hedging the negative gamma
position. Positive gamma positions arise when you buy options. You pay a
premium for the options but make money dynamically hedging the long
options position.
Wouldn't it be interesting if the amount of money you
could expect to lose dynamically hedging a short option position to
expiration is precisely the (accumulated value of the) option premium you receive for selling the
option in the first place? Actually, this isn't a new idea. When Black and
Scholes published their famous
option pricing formula,
they asserted that the price of an option should be the (discounted
value of the) cost of
dynamically hedging it to expiration. With this novel idea,
they launched the field of
option pricing theory.
Black and Scholes' analysis assumed that the underlier's
volatility is constant over time, but what happens if it is not? Suppose
you are dynamically hedging a short options position. How would you feel
if the underlier's volatility suddenly increased? Take a moment and think
about this ...
In fact, you would be concerned if the volatility
increased. At a higher volatility, the underlier will fluctuate more, and
you will need to adjust the delta hedge more frequently. You will lose
money more rapidly dynamically hedging. The opposite would be true if the
underlier's volatility suddenly fell. You could readjust the delta hedge
less frequently, and you would lose money more slowly dynamically hedging.
These concepts are illustrated in Exhibit 7. It shows the
cash position of a derivatives dealer who sells an option and then
dynamically hedges it until expiration. The dealer originally prices the
option at 25% volatility, but the exhibit considers three volatility
scenarios:
The underlier
experiences 20% volatility.
The underlier
experiences 25% volatility.
The underlier
experiences 30% volatility.
At expiration, the dealer's profit on the transaction is
the (accumulated value of the) option premium received when selling the option less the total cost of
dynamically hedging the short option to expiration. The Exhibit shows how,
if the option is priced at 25% volatility but the underlier experiences
20% volatility during the life of the option, the dealer ends up with a profit. If
the option is priced at 25% volatility and actual volatility also turns
out to be 25%, the dealer breaks even. Finally, if the option is priced at
25% volatility but actual volatility turns out to be 30%, the dealer
suffers a net loss on the transaction.
|
|
 |
|
 |
|
A dealer sells an option priced at 25%
volatility and then dynamically hedges the position until
expiration. This exhibit considers how the dealer's cash balance
evolves over time under three scenarios. Under all scenarios, we
assume an initial cash balance of zero. When the option is sold, the
dealer receives a premium, so the cash balance jumps. Next, the
dealer dynamically hedges the short option, gradually losing cash as
he does so. Under the first scenario, the underlier experiences 20%
volatility. Dynamic hedging costs less than it would have had the
underlier experienced the 25% volatility used to price the option.
The dealer ends up with a profit. Under the second scenario, the
underlier experiences 25% volatility. This is the volatility at
which the option was priced, so the dealer breaks even on the
transaction. Finally, under the third scenario, the underlier
experiences 30% volatility. This is higher than anticipated, and the
dealer ends up with a loss. |
When a dealer is dynamically hedging a short options
position, he doesn't care whether the underlier goes up or down. Because
he is always delta hedged, he is neither long nor short the underlier. He
does care whether the underlier's volatility goes up or down. In a very
real sense, he is short volatility. This is the same thing as having
negative vega (or short vega), so the
phrases negative vega, short vega and short
volatility all mean the same thing. A dealer dynamically hedging a long
options position is in the opposite situation. He benefits if volatility
increases, so he is long volatility.
Synonyms would be long vega or positive vega.
This is how derivatives dealers perceive options. Because
they routinely dynamically hedge their options positions, they don't think
of options as bets on the direction of the underlier. They think of them
as bets on the direction of volatility.
This has implications for
gamma and
theta as well. While contrived counterexamples
are possible (see if you can think of one!) it is generally true that a
negative vega position is also negative gamma. A derivatives dealer is
typically in the position of having sold options, and he is dynamically
hedging a position that is delta neutral, short gamma, short volatility
and long theta.
|
|
 |
|
Give
an example of how vanilla options might be combined in a portfolio
to achieve the atypical combination of positive gamma but negative vega.
[solution] |
|
|
|
|
 |
|
Black-Scholes (1973) option pricing
formula Used for pricing options on non-dividend paying stocks.
delta and gamma Factor sensitivities measuring
a portfolio's
first and second order (linear and quadratic) sensitivity to the value of an
underlier.
delta
hedge A type of hedge that is widely used by derivative
dealers to reduce or eliminate a portfolio's exposure to some
underlier.
derivative
instrument An instrument
which derives its value from the value of other financial
instruments. Article includes a list of vanilla and exotic derivatives.
Greeks A set of
factor sensitivities, which includes delta and gamma.
hedging
and diversification Standard techniques for reducing risk.
option
A type of derivative instrument.
option pricing theory The
body of financial theory used by financial engineers to value options and other
derivative instruments.
option spreads
Positions combining one or more options in a single underlier.
put-call
parity
A formula that relates the price of a put to the price of a
corresponding call.
rho Factor sensitivity measuring a portfolio's first order
(linear) sensitivity to the risk-free rate.
theta Factor sensitivity measuring a portfolio's first
order (linear) sensitivity to the passage of time
time value and
intrinsic value
The two components that comprise an option's market value.
vega Factor sensitivity measuring a portfolio's first
order (linear) sensitivity to the implied volatility of an underlier
volatility A metric of
variability in a stochastic process.
|
|
|
|
|
 |
|
|
|
|
 |
|
|
|
|
|
|