This material is not a solicitation of any offer to buy or sell any security or other financial instrument or to participate in any

trading strategy. This material was not prepared by the
Morgan Stanley Research Department, and you should not regard it as a research report. Please refer to important information and qualifications at the end of this material.
QUANTI TATI VE AND DERI VATI VE STRATEGI ES
QDS Vega Times MA R C H 2 8 , 2 0 1 1
Quantitative and Derivative Strategies (“QDS”) of Morgan Stanley’s Institutional Equity Division Sales Desk have prepared this piece. This is not a
product of Morgan Stanley’s Research Department and you should not regard it as a research report.
IN THE US FOR QUALIFIED INSTITUTIONAL BUYERS (QIBs) ONLY
A Guide to VIX Futures and Options
Interest in trading VIX derivatives has grown substantially, particularly after the
2008 volatility shock. In this note, we provide an in-depth guide to VIX products
and their characteristics, and review use cases for these products.
VIX derivatives behavior is very different from that of regular equity
derivatives
> VIX index is not directly tradable
> Term structure, skew and volatility of volatility characteristics
> VIX futures and options are now one of the most liquid ways to trade
short-dated volatility.
VIX derivatives have broad use cases
> Tail-risk hedging with VIX contracts, particularly for short-term trades
> Alpha strategies with VIX contracts
Equity hedging strategies using VIX derivatives have historically worked
better than S&P500 options
> Historical outperformance over SPX puts, at significantly lower cost
> 1x2 VIX call spreads and structures that trade SPX puts vs VIX calls are
often attractive
> Correct sizing and management of VIX derivatives positions is critical
Listed VIX contracts allow for new systematic alpha sources from the options
market
> Steepness of term structure and richness of SPX implied vol of vol as
alpha drivers
> Limit downside risk through listed option configuration
> Historic profitability of systematic strategies
Diverse offering of VIX futures-linked ETPs
> ETFs and ETNs tracking VIX futures-based indices
> Varying performance and use cases, depending on the product
QUANTI TATI VE AND DERI VATI VE
STRATEGI ES
UNI TED STATES
Daniel Cheeseman
+1 212-761-5832
dar|e|.c|eeserar¿rorçarslar|ey.cor
Simon Emrich
+1 212-761-8254
s|ror.err|c|¿rorçarslar|ey.cor
Boris Lerner
+1 212-761-0759
oor|s.|errer¿rorçarslar|ey.cor
RELEVANT QDS ARTI CLES
“Start Your Hedges: Managing Downside
Risk in an Equity Rally”, Vega Times, Feb
2011.
“Inverse & Leveraged ETF Primer”, Vega
Times, Oct 2010.
“Capital Markets Solutions for Hedging
‘Structural Beta’ Risk”, Global Equity and
Derivative Markets, Aug 2010.
“Volatility Exposure in Hedge Fund Returns-
a Guided Tour”, Vega Times, Jun 2010.
“Tailor-Made Variance Swaps – The Next
Frontier”, Global Equity and Derivative
Markets , Winter 2007

“Strategic and Tactical Use of Variance and
Volatility”, Global Equity and Derivative
Markets, Dec 2005.

2
I. A Roadmap for this Paper
Since its introduction in 1993, the VIX has moved from a
seldom-used, exotic measure of equity risk to the forefront of
investors’ minds as an important indicator of sentiment. The
growth in the popularity of this index has led to it being rebuilt
from the ground up into a new index with derivatives linked to
it.
Investors’ increasing concern for volatility risk and their
desire to control it has led to an explosion of liquidity in VIX
futures and options despite their short trading history. VIX
contracts are now the most liquid way globally to trade short-
term implied volatility.
Since the VIX itself is not investable, the product itself has
many important intricacies that may not be obvious for the
first-time user. Moreover for current users of VIX derivatives,
knowing how to best leverage this product’s dynamics is
crucial.
We therefore provide this paper as a how-to guide on using
this product for all users of VIX contracts, current and
potential. Below is an outline of the content to follow:
Section II. An Introduction to the VIX
Starting on page 4, we give a broad overview of the VIX and
what the index truly represents. The VIX has many properties
that make it unlike any other equity index. The points we
discuss in this section include:
> What the VIX is and an overview of how it is priced
> A history of the index and the growth of global VIX
look-alikes
> An analysis of the key properties of the VIX: (1) mean
reversion, (2) persistent negative correlation, and (3) the
convexity of its returns in a market shock
> How one gets exposure to VIX and the liquidity profile
of VIX futures and options
Section III. VIX Derivatives 101
On page 8, we highlight the unique properties of VIX
derivatives. Because the index is itself not tradable and has
many unique properties, the pricing of VIX derivatives is very
different from regular equity options. Specifically, we cover:
> How VIX derivatives account for mean-reversion
through the VIX futures term structure
> The cost of holding VIX contracts due to the shape of the
term structure
> The time sensitivity of VIX futures versus the roll cost
trade-off
> The slowness of VIX option decay versus regular options
> The shape of the VIX vol surface and how VIX options
decay
> How VIX futures compare to forward variance swaps
Section IV. VIX for Equity Hedgers
We continue on page 12 with a discussion of how VIX
contracts can be used as an equity hedge because of the
negative correlation between volatility and equity returns.
Through backtests and empirical analysis, we discuss:
> How to scale a VIX hedge for an S&P500 portfolio
> How to compare the costs of VIX contracts versus
S&P500 contracts
> Systematic performance of VIX hedges versus S&P500
hedges
> The long-term performance of VIX spread-based hedges
> The potential systematic cheapness of VIX hedges
> How one can use a unit mismatch between the S&P500
and VIX for downside hedges
3
Section V. VIX for Alpha-Seekers
OTC index variance swaps have long been used as a part of
systematic alpha strategies. However, the events of 2008 have
raised questions whether listed VIX contracts can also be used
as a part of an alpha strategy
On page 21, we highlight potential systematic VIX alpha
strategies using the VIX term structure and the VIX option
implied vol of vol. The topics we discuss include:
> How to use VIX derivatives to trade the S&P500 implied
vol term structure
> How to construct term structure trades with reduced
downside risk
> The historical richness of VIX option implied vol versus
subsequent realized vol
> The backtested performance of potential alpha trades
VI. VIX Futures ETPs
For the last major section on page 27, we review the current
landscape for VIX Exchange Traded Products (“ETPs”) and
how the properties of VIX futures affect their performance
The points discussed are:.
> What the current landscape is for ETPs and an overview
of what is currently available
> How VIX ETPs are affected by the VIX futures rolldown
> The long-term performance of VIX futures going back
15+ years
> The performance of inverse & leveraged forms of these
strategies
Section VII. Glossary of Terms
Lastly on page 30, we give a concise glossary of commonly
used terms within this document and their definitions.
4
II. An Introduction to the VIX
The VIX has emerged as an index that is actively tracked by
multiple investor types across virtually all asset classes. The
popularity of this product eventually led to the launch of other
VIX-like indices globally along with VIX derivatives.
However, many misconceptions exist about the index and what
it truly represents. In this section, we aim to cover:
> What the VIX is and how it is priced
> A history of the index and the growth of global VIX look-
alike indices
> An analysis of the key properties of the VIX: (1) mean
reversion, (2) persistent negative correlation, and (3) the
convexity of its returns in a market shock
> Getting exposure to VIX and the liquidity profile of VIX
futures and options
A Real-Time Measure of S&P500 Risk
I MPLYI NG FUTURE EQUI TY RI SK THROUGH OPTI ONS
In the Black-Scholes option pricing model, the largest pricing
input aside from the strike is the forecast of future volatility of
the underlying. Since an option is hedged by actively
managing an amount of an offsetting stock position, higher
volatility requires that the hedge be rebalanced more
frequently. This incurs higher costs and in turn increases the
option premium.
Options are thus a forward-looking instrument, accounting for
expectations of future equity risk. This makes implied
volatility comparable to the yield curve, which accounts for
future expectations of inflation and changes in interest rates.
As option trading has become more prevalent, investors have
increasingly tracked implied volatility as an indication of the
market’s expectation of future risk.
VI X AS A BENCHMARK I NDI CATOR OF EQUI TY RI SK
In 1993, the CBOE introduced the VIX as a benchmark of
implied equity risk. It later modified the methodology in 2003
to make it easier to link derivatives to it. The VIX is an
indication of the implied volatility of the S&P500 over the
subsequent 30 calendar days using the live, intraday prices of
CBOE-listed S&P500 (SPX) options.
The implied volatility of an option is not a directly observable
level, unlike the price of a stock. Computing how much risk
an option is implying is a non-trivial calculation that depends
on other factors such as interest rates and expected dividends
over the life of the option. In addition, implied volatilities
differ across strikes and maturities. This makes it difficult to
quote one single number as ‘the’ level of expected future risk.
This is a significant advantage of the VIX since it is quoted as
a variance swap strike, which is closely linked to implied
volatility. The terms “variance swap strike” and “implied
volatility” are sometimes considered verbal substitutes for
each other depending on the context (Exhibit 1). In this paper,
we often switch from one terminology to the other.
1
Exhibit 1: The historical difference between 1-month S&P500 at-the-
money implied volatility and the VIX, a 1-month S&P500 variance swap
Data from 2 Jan 96 through 23 Mar 11

10º
20º
30º
10º
50º
ê0º
Z0º
80º
J
a
r

9
ê
J
a
r

9
Z
J
a
r

9
8
J
a
r

9
9
J
a
r

0
0
J
a
r

0
1
J
a
r

0
2
J
a
r

0
3
J
a
r

0
1
J
a
r

0
5
J
a
r

0
ê
J
a
r

0
Z
J
a
r

0
8
J
a
r

0
9
J
a
r

1
0
vlX - ATV lrp||ed vo|
1-Vorl| 3&P500 ATV lrp||ed vo|al|||ly
vlX
Source: Morgan Stanley Quantitative and Derivative Strategies
A variance swap is computed using a weighted sum of out-of-
the-money calls and puts, distilling the implied volatilities
across different strikes into a single number representing an
“average” level.
2
Variance swap levels are computed across
listed maturities intraday and are continuously interpolated to
create a constant rolling measure of 30-day risk. This
methodology is largely independent of rate and dividend
assumptions and is easy to quote intraday since it is based on a
basket of listed options.
The success of the VIX has led the CBOE to compute other
equity volatility benchmarks such as VXV (a 3-month VIX),
VXN (NASDAQ 100 VIX), VXD (DJIA VIX), and RVX
(Russell 2000 VIX) along with other VIX-like measures on

1
A variance swap strike can be thought of as the at-the-money implied volatility plus a
skew premium. In fact, if a variance swap is priced using options with the same
implied volatility, the variance swap’s strike and the implied volatility would be equal.
2
The portfolio is weighted so that when delta-hedged has a constant dollar gamma
regardless of the underlying’s path. For more details on the VIX calculation, please
refer to the VIX white paper which can be found at
http://www.cboe.com/micro/vix/vixwhite.pdf.
5
gold (GVXX), oil (OVX), and the Euro (EVZ). The CBOE
has also begun launching VIX-like indices on select US
stocks. As of 23 Mar 11, the universe of single stock VIX
indices was on five names: AAPL, AMZN, GOOG, GS, and
IBM.
Other global options exchanges have also launched similar
benchmark indices on the S&P/TSX 60, Euro STOXX 50,
DAX, SMI, FTSE 100, AEX, CAC 40, BEL 20, Mexican
Bolsa, Nikkei 225, Nifty 50, TOP 40, KOSPI 200, Nifty, ASX
200, and HSI.
Important Properties of the VIX
NEGATI VE CORRELATI ON TO EQUI TY RETURNS
The feature that makes VIX attractive to investors is its
historically negative correlation to equity returns. As equity
markets decline, equities tend to become more volatile.
Hence, the VIX tends to rise, as the market anticipates
increased future volatility.
Exhibit 2: 1-Year rolling correlation between daily changes in the
S&P500 and other asset classes
Based on weekly returns from 3 Jan 77 through 23 Mar 11
-1
-0.8
-0.ê
-0.1
-0.2
0
0.2
0.1
0.ê
0.8
1
1
9
Z
3
1
9
Z
5
1
9
Z
Z
1
9
Z
9
1
9
8
1
1
9
8
3
1
9
8
5
1
9
8
Z
1
9
8
9
1
9
9
1
1
9
9
3
1
9
9
5
1
9
9
Z
1
9
9
9
2
0
0
1
2
0
0
3
2
0
0
5
2
0
0
Z
2
0
0
9
2
0
1
1
1
-
Y
e
a
r

R
o
|
|
|
n
g

6
o
r
r
e
|
a
t
|
o
n
lrlerral|ora| Equ|ly lo 3&P500
Ererç|rç Var|el Equ|ly lo 3&P500
lrveslrerl 0rade F|xed lrcore lo 3&P500
Aççreçale Corrod|l|es lo 3&P500
vlX lo 3&P500
R|s| assel corre|al|ors are c|ose lo 1
vlX's corre|al|or |s c|ose lo -1
We use the following indices: MSCI EAFE (International Equity), MSCI EM (Emerging
Market Equity), Barclays Aggregate (Investment Grade Fixed Income), and DJ UBS
Commodity Index (Aggregate Commodities). Source: Morgan Stanley Quantitative and
Derivative Strategies
Exhibit 2 shows the rolling 1-year correlation between the
VIX and the S&P500, as well as the correlation of other global
risk assets with the S&P500. Throughout its history, the VIX
has exhibited a strong and persistent negative correlation with
S&P500 returns. For example, over the past year, the
VIX/S&P500 correlation was -0.80, one of the strongest
sustained levels of negative correlation in the history of the
index. Historically, there has never been a period where the
correlation has been weaker than -0.40.
3

3
Correlation levels were less negative prior to 2001. S&P500 options were much less
liquid at the time. If S&P500 options back then had been more liquid, it is likely
correlation would have been similar to recent levels.
Other assets classes, on the other hand, have had a mixed
correlation history with the S&P500. International and EM
equities typically show a positive correlation with the
S&P500. Over the past year, correlation was approximately
0.80, one of the highest sustained levels of correlation in its
entire history.
While commodities historically have had relatively little
correlation to US equities, making them popular diversifying
assets, their recent correlation to the S&P500 is similar to that
of International and Emerging Market Equities. Fixed
income’s correlation has varied over time, ranging from
positive in the 1990’s to a negative relationship currently.
The persistent negative correlation between the VIX and the
S&P500 makes the VIX a potential diversifying tool for asset
allocators.
LARGE SPI KES FOR SI GNI FI CANT EQUI TY LOSSES
The VIX also changes asymmetrically to moves in the S&P
500. Volatility expectations tend to spike after large sell-offs
but gradually creep down in a rally (Exhibit 3). This is
consistent with investor behavior – they are more anxious to
purchase protection when equities are falling than they are to
sell volatility when the market is rising.
Exhibit 3: The historical relationship between 1-month S&P500 returns
and 1-month VIX returns
Data from 2 Jan 90 through 23 Mar 11
-100º
-50º

50º
100º
150º
200º
-10º -30º -20º -10º 0º 10º 20º 30º 10º
1-Honth 8&P500 Percent 6hange
1
-
H
o
n
t
h

V
|
X

P
e
r
c
e
n
t

6
h
a
n
g
e
Source: Morgan Stanley Quantitative and Derivative Strategies
This makes the VIX potentially attractive as a tail risk hedge,
due to its negative correlation and its convexity to large
negative equity returns. For example, after September 2008,
the VIX more than tripled from 25 to its peak at 80 as the
S&P500 dropped 40% over the same period.
6
UNLI KE EQUI TI ES, VI X I S MEAN REVERTI NG
Since volatility is a measure of risk, it is bound by zero, which
would indicate no variation in the market. While there is no
theoretical upper bound, in practice volatility resembles
interest rates, which tend to float between zero and an
arbitrarily high number. We typically observe volatility
drifting between ‘low’ and ‘high’ regimes, with the market
often spending extended periods in a regime before switching
to the other.
Volatility thus has mean-reverting properties. While volatility
may be relatively low or high at some point, it eventually will
shift away from that extreme towards a longer-term average.
Getting Exposure to the VIX
THE VI X I S NOT A DI RECTLY I NVESTABLE I NDEX
Unlike equity indices, which can be replicated through a
basket of stocks, the VIX is not a directly tradable index. The
VIX is a theoretical index that is re-interpolated continuously
making a replicating portfolio impossible to construct.
VI X FUTURES AND OPTI ONS EXI ST AS A SURROGATE
The reconstitution of the VIX in 2003 allowed for futures and
options contracts to be developed as a surrogate for exposure
to the unreplicatable VIX.
VIX futures were first listed in March 2004, with options
following in February 2006. Both contracts have monthly
expirations – futures are listed up to eight months out while
options are listed up to six months out. Contracts cash-settle
to a special opening quote of the VIX on a unique expiration
cycle. Consistent with VIX’s 30-day tenor, VIX derivatives
expire on the day that is exactly 30 days prior to the next
months’ SPX option expiration.
4
Typically, this is on a
Wednesday in the third or fourth week of the month.
LI QUI DI TY I N VI X CONTRACTS HAS GROWN TO RI VAL
SPX VARI ANCE SWAP LI QUI DI TY
Trading volume in VIX contracts has grown substantially
since inception. VIX futures and options each trade over 37mn
expiration vega per day (Exhibit 4).
5
Trading volume

4
This expiration cycle was picked so that if a person were to hold the portfolio of
options comprising the VIX they would have a perfect hedge at expiration for a 30-
day variance swap. This occurs 30 days prior to the next SPX option expiration.
The usual expiration day for SPX options is a Friday, which makes the day 30 days
prior typically a Wednesday.
5
The vega of a VIX future is simply the contract multiplier (1000). For consistency, we
define the vega turnover of VIX options to also be its contract multiplier (100). This
methodology is consistent with convention on how to measure a notional options
turnover.
increased significantly during 2009, after the VIX reached its
all-time highs in 4Q08. This liquidity growth is likely due to
(1) increased demand for tail risk hedging, (2) increased
trading of VIX as a diversifying asset, and (3) greater tactical
usage of the product.
To put VIX derivative liquidity in perspective, the aggregate
trading volume across all listed SPX options during 4Q10 was
around USD 75mm notional vega per day. OTC SPX variance
swaps trade no more than USD 10mn vega notional per day,
according to our traders’ estimates.
VIX futures and options are often traded in spread positions,
which lowers the true net vega traded. Nevertheless, liquidity
in these contracts has increased substantially – we estimate
that VIX derivatives liquidity is now comparable to that of
OTC SPX variance swaps.
It is also important to note that VIX options are the only liquid
vehicle for optionality on volatility. OTC options on realized
variance are available but have scarce liquidity.
Exhibit 4: The 3-month median S&P500 vega traded in VIX derivatives
Data from 3 Jan 06 through 23 Mar 11
0
5
10
15
20
25
30
35
10
J
a
r

0
ê
A
p
r

0
ê
J
u
|

0
ê
0
c
l

0
ê
J
a
r

0
Z
A
p
r

0
Z
J
u
|

0
Z
0
c
l

0
Z
J
a
r

0
8
A
p
r

0
8
J
u
|

0
8
0
c
l

0
8
J
a
r

0
9
A
p
r

0
9
J
u
|

0
9
0
c
l

0
9
J
a
r

1
0
A
p
r

1
0
J
u
|

1
0
0
c
l

1
0
J
a
r

1
1
3
-
H
o
n
t
h

H
e
d
|
a
n

0
a
|
|
y

V
e
g
a

T
r
a
d
e
d

(
U
8
0

H
n
}
vlX Fulures
vlX 0pl|ors
Source: Morgan Stanley Quantitative and Derivative Strategies
LI QUI DI TY I S HEAVI LY FRONT- LOADED WI TH FEW
LONG-MATURI TY TRADES
While the liquidity of VIX contracts is high relative to SPX
variance swaps, the time horizon of this liquidity is different.
The majority of VIX contract liquidity is concentrated within
the first few expirations (Exhibit 5), while SPX variance
trades across the term structure. This makes VIX contracts
potentially the most liquid way to trade short-dated implied
volatility.
The front-loaded nature of VIX liquidity is very persistent
over time, as shown in Exhibit 5. Turnover in the front-month
contract accounts for 42% of the total VIX futures trading
volume currently, with 30% in the second-month contract.
Despite these contracts being listed out to eight months, only
7
16% of the liquidity is in contracts beyond the first three
months. VIX options share the same pattern, with 47% and
29% of daily turnover occurring in the first two maturities.
Exhibit 5: The rolling 3-month average of the percent of daily turnover
each VIX future contract comprises
Data from 23 Mar 07 through 23 Mar 11

10º
20º
30º
10º
50º
ê0º
Z0º
V
a
r

0
Z
V
a
y

0
Z
J
u
|
0
Z
3
e
p

0
Z
N
o
v

0
Z
J
a
r

0
8
V
a
r

0
8
V
a
y

0
8
J
u
|
0
8
3
e
p

0
8
N
o
v

0
8
J
a
r

0
9
V
a
r

0
9
V
a
y

0
9
J
u
|
0
9
3
e
p

0
9
N
o
v

0
9
J
a
r

1
0
V
a
r

1
0
V
a
y

1
0
J
u
|
1
0
3
e
p

1
0
N
o
v

1
0
J
a
r

1
1
V
a
r

1
1
3
-
H
o
n
t
h

A
v
e
r
a
g
e

P
e
r
c
e
n
t

o
f

0
a
|
|
y

T
u
r
n
o
v
e
r
1sl Vorl| vlX Fulures 2rd Vorl| vlX Fulures
3rd Vorl| vlX Fulures 1l| Vorl| vlX Fulures
Source: Morgan Stanley Quantitative and Derivative Strategies
The sensitivity of VIX futures to movements in the SPX is the
likely cause for this front-loaded liquidity. Movements in the
underlying cash index affect equity index futures similarly
across the maturities. VIX futures, on the other hand, tend to
become less sensitive to moves in the VIX the longer-dated
they are. We discuss this phenomenon in more detail on page
9.
Since VIX options are priced off VIX futures, a spike in the
VIX does not necessarily lead to a large P&L swing for long-
dated VIX derivatives. This makes longer-dated contracts less
appealing to hold relative to short-dated contracts, particularly
when used as a tactical way to execute a volatility view.
ETPS BASED ON VI X FUTURES ALSO EXI ST AS A
METHOD FOR VI X EXPOSURE
Starting in January 2009, Exchange Traded Products (“ETPs”)
have been launched as easy-to-access turnkey solutions to give
exposure to strategies based on VIX futures. While these
strategies are not meant to track the VIX, they can deliver a
high correlation to the VIX and exhibit many properties that
are similar to those of VIX futures. However, depending on
the volatility regime, some of these products exhibit
significant negative carry.
In total, assets in these products are now well over $2bn,
spread across 13 different products as of 23 Mar 11. We
highlight the current universe of these products and the factors
affecting their performances on page 27.
The Major Players and Positioning in VIX
VI X DERI VATI VES ARE BEI NG I NCREASI NGLY USED
BY I NSTI TUTI ONS AS A LI STED, LI QUI D VOL VEHI CLE
Prior to 4Q08, VIX futures and options trading interest was
mainly concentrated in the hedge fund community, who used
these contracts as short-term positioning tools. Since then,
institutional and even retail market participants have
increasingly used these contracts, as “tail risk hedging”
became part of investing nomenclature.
The potential benefit of integrating volatility was shown
during the credit crisis, when the VIX rose from 23 the week
before Lehman to a peak of 80. The listed nature and intraday
liquidity of VIX contracts has made them attractive to
investors as a volatility hedge against these types of events.
The rise of VIX futures-linked ETPs can be seen as further
evidence of the broadening of the participant base.
LI QUI DI TY I S LI KELY TO CONTI NUE TO GROW AS
I NVESTORS LOOK TO OWN VOLATI LI TY
Exhibit 6 shows the shift in investor risk perception pre- and
post the credit crisis. After 2008, negative skew (‘tail risk’)
received a greater weighting in investors’ risk assessment. The
increased skew of longer dated SPX options as well as of VIX
call options reflects investors’ response to the new,
unexpected data points observed during 2008.
6
This shift in risk assessment can potentially explain the greater
use of volatility products such as the VIX. As investors further
move to integrate volatility exposure as a tail-risk hedge and
for diversification, we might well see further increases in the
liquidity of VIX-related products.
Exhibit 6: The perception of tail risk is likely higher than it was before
2008 after going through the credit crisis
-50º -10º -30º -20º -10º 0º 10º 20º 30º 10º 50º
3-Honth Return w|th 157 Vo|at|||ty
Percepl|or ol R|s| Pre-Cr|s|s
Percepl|or ol R|s| Posl-Cr|s|s
KroW|rç l|e |rpacl ol la|| everls |as
|rcreased |rveslor aWareress ard
percepl|or ol l|e|r |||e|||ood posl-cr|s|s
Source: Morgan Stanley Quantitative and Derivative Strategies

6
This phenomenon is studied in a field of statistics known as ‘extreme value theory.’ A
probability distribution may have a significant weighting towards a tail but a limited
statistical sample of that distribution is likely to not have a realization from that tail.
Therefore it is difficult to infer from the data alone how likely those tail events really are.
8
III. VIX Derivatives 101
Derivatives on the VIX have many unique properties, because
the VIX itself is not tradable. These contracts resemble
interest rate derivatives, which have a similarly intangible
underlying, more so than they resemble equity derivatives. In
this section, we provide an overview of the mechanics of VIX
futures and options:
> VIX derivatives account for volatility mean-reversion
through the VIX futures term structure
> The shape of the volatility term structure and its impact
on the cost of holding VIX contracts
> The time sensitivity of VIX futures versus roll cost trade-
off
> The slowness of VIX option decay versus that of regular
options
> The shape of the VIX vol surface and how VIX options
decay
> How VIX futures compare to forward variance swaps
The Unique Properties of VIX Contracts
CASH + VI X FUTURES DO NOT MAKE VI X
The arbitrage-based pricing of forward / futures contracts
means that an investment in an equity index can be replicated
through a combination of cash and futures on that index. This
replication arises because the underlying index can be traded
using a managed basket of equities.
VIX futures do not have this arbitrage-based pricing, as the
VIX is not a tradable index. Instead, VIX futures prices
reflect expectations on the VIX level at expiration of the
futures contract. Should VIX spike but expectations are that
the VIX will revert in a month’s time, a 1-month VIX futures
price should remain unchanged.
This property makes VIX futures an imperfect surrogate to
trade the VIX. It is also responsible for many of the other
unique features of VIX contracts.
VI X FUTURES ACCOUNT FOR MEAN REVERSI ON
Just like volatility, the VIX is a mean-reverting index. VIX
futures account for this property and price in expectations of
mean reversion.
Exhibit 7 plots the VIX futures term structure on two extreme
dates – the day of the VIX’s all-time high in 2008, and a day
in January 2007 when the VIX reached its second-lowest level
in history. At these relative extremes, the VIX futures term
structure priced in an expectation of mean reversion. At the
relative low in 2007, the 6-month VIX future was over four
vol points higher than the spot VIX, while at the relative high
in 2008, the 6-month VIX future was over 45 vol points lower
than the spot VIX.
Exhibit 7: The VIX future term structure at relative VIX highs and lows
9
10
11
12
13
11
15
vlX 1sl 2rd 3rd 1l| 5l| êl|
V|X Future
30
10
50
ê0
Z0
80
90 3ecord LoWesl vlX 0ay: 21 Jar 0Z (|ell ax|s)
A||-T|re vlX l|ç|: 20 Nov 08 (r|ç|l ax|s)
Source: Morgan Stanley Quantitative and Derivative Strategies
This changing shape of the term structure is necessary for a
mean-reverting underlying. Otherwise, one could devise a
statistical arbitrage trading strategy that buys the index at its
relative lows and shorts it when it is at its relative highs.
Most of the time, the term structure of the VIX futures is
upwards sloping. This is because VIX futures are priced off
the S&P500 implied volatility term structure, which is usually
upwards-sloping.
The slope reflects a volatility premium – since the VIX and
implied volatility tend to spike in times of market stress, VIX
futures should trade at a premium to spot VIX, accounting for
the probability of such a spike.
VI X FUTURES HAVE A TI ME DEPENDENT BETA TO VI X
The mean reversion priced into the VIX futures term structure
means that different futures maturities show varying
responsiveness to spot VIX. Shorter-dated VIX contracts are
more sensitive to spot VIX, because there is less time for
volatility to mean-revert. The longer the maturity of the
contract, the higher is the probability of mean reversion.
Shocks to spot VIX should thus have less of an impact.
Exhibit 8 plots the sensitivity of VIX futures to changes in the
VIX. We group all historically traded VIX futures by their
time to maturity and compute their betas to daily vol point
changes in VIX. Six-month VIX futures generally move less
9
than 20% of the amount VIX does on any given day. This
sensitivity remains low for most of the life of a VIX future
until just prior to expiration. VIX futures under 20 days to
maturity typically have a beta between 0.6 and 0.8 to spot
VIX.
Exhibit 8: The beta of VIX futures to changes in spot VIX as a function
of time to maturity
Based on daily VIX futures data from 26 Mar 04 through 23 Mar 11
y = -0.15Lr(x) ÷ 0.9ê
R
2
= 0.81
0
0.1
0.2
0.3
0.1
0.5
0.ê
0.Z
0.8
0.9
1
0 20 10 ê0 80 100 120 110 1ê0 180
0ays Unt|| V|X Future Exp|rat|on
ß
e
t
a

t
o

V
|
X
Source: Morgan Stanley Quantitative and Derivative Strategies
This causes front-month futures to be the most volatile while
the back-month futures have only a fraction of the volatility of
the VIX.
THE SHORT-END I S THE COSTLI EST POI NT TO BUY
The shape of the VIX futures term structure gives insight into
the holding costs of going long VIX futures. For example, if
VIX futures are priced two vol points above spot VIX, then
one would expect to pay two vol points at maturity if VIX is
unchanged.
Since VIX futures typically trade at a premium to spot VIX,
particularly in a low-volatility regime, long VIX futures
holders will lose this premium most of the time. They are
compensated for this by substantial gains when volatility
spikes. The tradeoff between holding costs and spikes has
historically been asymmetric, creating potential alpha
opportunities. We discuss strategies around this tradeoff
starting on page 21.
Exhibit 9 plots the median historical term structure of VIX
futures, as well as the marginal contribution of each futures
maturity to the overall term structure. This allows us to assess
the incremental cost of holding VIX futures of longer
maturity, relative to the spot VIX.
The median level of 1-month VIX futures is 1.0 vol points
higher than spot VIX while 2-month futures are 0.68 vol
points above the 1-month VIX futures. This marginal
contribution flattens out quickly with 6-month futures adding
only 0.1 vol points to the term structure.
Exhibit 9: The median VIX futures term structure and the steepness
between each point on average
Based on daily VIX futures data from 26 Mar 04 through 23 Mar 11
1Z.5
18.0
18.5
19.0
19.5
20.0
20.5
21.0
21.5
vlX 1V 2V 3V 1V 5V êV
V|X Future Tenor
0.00
0.15
0.30
0.15
0.ê0
0.Z5
0.90
1.05
1.20
Averaçe Varç|ra| Corlr|oul|or lo
l|e Terr 3lruclure (r|ç|l)
Ved|ar l|slor|ca| va|ue (|ell)
Source: Morgan Stanley Quantitative and Derivative Strategies
This implies that the rolldown cost for VIX futures is greatest
in the front-end while flattening out quickly beyond the 3-
month point. This produces a trade-off between trading the
desired volatility of VIX in the front-end with higher costs
associated with that position.
We highlight this tradeoff in more detail in Exhibit 10 where
we plot these marginal contributions to the term structure
versus the beta of that VIX future to spot VIX (data from
Exhibit 8 and Exhibit 9). This gives a measure of the cost of
holding a VIX future per unit of exposure it gives to VIX.
Exhibit 10: The marginal contribution to the VIX futures term structure
versus the beta of that future to spot VIX
Based on daily VIX futures data from 26 Mar 04 through 23 Mar 11
0
0.1
0.2
0.3
0.1
0.5
0.ê
0.Z
0.8
0.9
1
1.1
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.10 0.15
ßeta of V|X Future to 8pot V|X
A
v
e
r
a
g
e

1
-
H
o
n
t
h

T
e
r
m

8
t
r
u
c
t
u
r
e

R
o
|
|

6
o
s
t

(
V
o
|

P
o
|
n
t
s
}
1-Vorl| Fulures
2-Vorl| Fulures
3-Vorl| Fulures
1 lo ê-Vorl|
Fulures
1-Vorl| lulures
cosl ê0º |ess
l|ar 1-Vorl|
vlX lulures
Source: Morgan Stanley Quantitative and Derivative Strategies
1-month futures have the greatest cost per exposure ratio by
far, while 4 to 6-month future contracts scale much better. For
example, an exposure to a 4-month future that is scaled to give
the same exposure to the spot VIX as a 1-month future would
have only 40% of the holding cost of the 1-month future on
average.
However, the lower cost per unit of spot VIX exposure of
longer-dated futures has to be counterbalanced with the
considerably lower liquidity in these contracts. These liquidity
considerations have to be taken into account when developing
10
strategies that leverage this systematic bias for hedging or
alpha.
THE ATM VI X VOL STRI KE I S BASED ON THE FUTURES
VIX futures represent forward-looking expectations of future
VIX levels at each maturity date. As a consequence, the at-
the-money strike of VIX options is determined with reference
to the VIX future level for a given maturity, rather than to the
cash VIX level. This is in contrast to the cash equities options
market.
Because the VIX futures term structure is generally upwards
sloping, costs of out-of-the-money VIX calls are optically high
relative to spot VIX, while VIX puts appear cheap. However,
this reflects the at-the-money level for each maturity, which
forms a more appropriate base line.
THE VI X OPTI ON I MPLI ED VOL TERM STRUCTURE I S
PERSI STENTLY I NVERTED
The realized volatility of VIX futures declines the longer-
dated they are. Since implied volatility represents the expected
future realized volatility, we find that VIX options implied
volatility has a permanently inverted term structure. The
implied volatility of a short-dated VIX option anticipates
volatility over the near term, which may itself be more
volatile. Given mean-reverting properties of the VIX, a longer
tenor averages in less volatile periods for that VIX future, and
hence a lower implied volatility.
Exhibit 11 plots the rolling time series of ATM 1-month and
5-month VIX implied volatilities, as well as their spread and
the VIX itself. Long-dated implied volatilities are relatively
stable and significantly lower than 1-month implied volatility.
Shorter-dated VIX implied volatility also reacts more to spikes
in the VIX. Similar to the S&P500 option term structure, the
spread between long- and short-dated volatility shifts
significantly when volatility spikes since the VIX and its
implied volatility both tend to rise during times of distress.
7
VI X OPTI ON SKEW I S ALWAYS TI LTED TO CALLS
Unlike S&P500 options, the skew in the VIX implied vol
surface is tilted towards calls instead of puts (Exhibit 12).
This makes sense since VIX implied volatility generally rises
when the VIX does (Exhibit 11).

7
VIX implied vol actually has a strong relationship to S&P500 skew. The skew of SPX
options is a measure the implied volatility of volatility. Assume that the SPX fixed-
strike implied vol surface is unchanged while the SPX falls. The new ATM implied vol
would have to roll up the SPX put skew to a higher level. The steeper the skew, the
more ATM implied vol changes per 1% change in the S&P500.
Exhibit 11: The implied vol and vol term structure of VIX options
Based on daily data from 23 Mar 08 through 23 Mar 11

20º
10º
ê0º
80º
100º
120º
110º
1ê0º
180º
V
a
r

0
8
V
a
y

0
8
J
u
|

0
8
3
e
p

0
8
N
o
v

0
8
J
a
r

0
9
V
a
r

0
9
V
a
y

0
9
J
u
|

0
9
3
e
p

0
9
N
o
v

0
9
J
a
r

1
0
V
a
r

1
0
V
a
y

1
0
J
u
|

1
0
3
e
p

1
0
N
o
v

1
0
J
a
r

1
1
V
a
r

1
1
10
20
30
10
50
ê0
Z0
80
90
100
1-Vorl| V|rus 5-Vorl| 3pread (||s)
1-Vorl| ATV vlX lrp||ed vo| (||s)
5-Vorl| ATV vlX lrp||ed vo| (||s)
vlX (r|s)
Source: Morgan Stanley Quantitative and Derivative Strategies
Along with a generally upward sloping VIX futures term
structure, a steep VIX call skew further helps to elevate the
costs of VIX calls. Strategies that leverage this, such as OTM
call spreads, are often relatively attractive, particularly in a
high-skew regime. Ever since 2009, VIX option skew has
been steepening and is currently near an all-time high. We
highlight how to leverage this phenomenon on page 18.
Exhibit 12: The implied vol and 130%-100% 1-month VIX option skew
Based on daily data from 23 Mar 08 through 23 Mar 11
20º
10º
ê0º
80º
100º
120º
110º
1ê0º
180º
V
a
r

0
8
V
a
y

0
8
J
u
|

0
8
3
e
p

0
8
N
o
v

0
8
J
a
r

0
9
V
a
r

0
9
V
a
y

0
9
J
u
|

0
9
3
e
p

0
9
N
o
v

0
9
J
a
r

1
0
V
a
r

1
0
V
a
y

1
0
J
u
|

1
0
3
e
p

1
0
N
o
v

1
0
J
a
r

1
1
V
a
r

1
1

10º
20º
30º
10º
50º
ê0º
Z0º
80º
1-Vorl| 130º - 100º 3|eW 3pread (r|s)
1-Vorl| 130º vlX lrp||ed vo| (||s)
1-Vorl| ATV vlX lrp||ed vo| (||s)
Source: Morgan Stanley Quantitative and Derivative Strategies
VI X OPTI ONS DECAY SLOWLY UNTI L EXPI RATI ON
Option prices are largely functions of two inputs: (1) the
moneyness of the option and (2) its non-annualized implied
volatility (ı¥t). In general, the higher the latter component is,
the higher the option price.
8
Since the VIX option implied volatility term structure is
generally inverted, VIX options tend to decay slowly until

8
If we replace the factor
t σ
with a single total implied volatility, σ
~
and zero interest
rates, the Black 1976 model for a VIX call option price reduces to:
»
¼
º
«
¬
ª −
Φ −
»
¼
º
«
¬
ª +
Φ =
σ
σ
σ
σ
~
~
5 . 0 ) / ln(
~
~
5 . 0 ) / ln(
2 2
K F
K
K F
F Call
where F denotes the futures price, K the strike, and ĭ denotes the cumulative density
of the standard normal distribution. This has no independent implied volatility and time
to maturity factors.
11
close to expiration. While the time to maturity t shrinks, the
implied volatility ı rolls up the term structure. These
offsetting factors tend to slow the decay in option prices until
a few weeks before expiration.
We show this in Exhibit 13 where we plot the average shape
of the VIX implied volatility term structure using 1- to 5-
month ATM implied vols. We use these levels to price ATM
VIX calls over time as they roll down the term structure. We
assume zero interest rates. We also plot option prices
assuming a flat implied volatility term structure (at the level of
5-month implied volatility).
Because of the roll-up effect, option prices decay more slowly
than in the case of a flat term structure. For example, the price
of ATM calls drops on average from 14% for 5-month tenors
to 10% for 1-month maturities. Under a flat term structure,
prices would decay to around 6% for 1-month maturity. Since
calls that are not in the money have to converge to zero at
maturity, this implies a rapid decay just before expiration.
Exhibit 13: The average costs of ATM VIX call options versus the
average shape of the VIX implied vol term structure
Based on daily data from 24 Feb 06 through 23 Mar 11



êº

10º
12º
11º
1êº
0 30 ê0 90 120 150
T|me to Hatur|ty (0ays}
15º
50º
55º
ê0º
ê5º
Z0º
Z5º
80º
85º
Averaçe ATV vlX Ca|| Pr|ce (º ol vlX Fulure) (|ell)
Averaçe Ca|| Pr|ce lo|d|rç 5V vo| Corslarl (º ol vlX Fulure) (|ell)
Averaçe vlX lrp||ed vo| Terr 3lruclure (r|ç|l)
Source: Morgan Stanley Quantitative and Derivative Strategies
VIX Contracts versus Variance Contracts
COMPARI NG THE PAYOFF OF VI X VERSUS VARI ANCE
Futures based on VIX deliver a linear payoff with respect to
changes in VIX. For every vol point change in the strike of a
VIX future, the expiration P&L change will be a fixed
multiplier times the strike change (see Exhibit 14).
This is in contrast to OTC index variance swaps. For forward-
starting variance swaps, the payoff at the end of the forward
period is quite similar to that of a VIX future. However, the
payoff of these products is based on variance, the square of
volatility. This creates a convex payoff with respect to
changes in volatility.
9

We compare these payoffs in Exhibit 14 using a hypothetical
1-month VIX future and a 1-month variance swap that forward
starts in one month (1x1 forward variance). The long variance
swap gives a convexity benefit for large changes in volatility.
This benefit comes at the cost of a higher strike than a VIX
future with the same maturity. Since 2004, the strike
difference has averaged around 0.7 vol points for 1-month
maturities.
Exhibit 14: The payoffs of a hypothetical 1-month VIX future and a
hypothetical 1x1 forward starting variance swap in one month’s time
-20
-10
0
10
20
30
10
10º 15º 20º 25º 30º 35º 10º 15º 50º
P
&
L

p
e
r

$
1

o
f

N
o
t
|
o
n
a
|

V
e
g
a
P&L ol a 1-Vorl| vlX Fulure (3lr||e = 21º)
P&L ol a 1x1 ForWard var|arce 3Wap (3lr||e = 2êº)
T|e var|arce sWap |as a corvex|ly
oerel|l lor |arçe c|arçes |r vo|al|||ly
T|e corvex|ly |s pa|d lor
oy a ||ç|er slr||e
Source: Morgan Stanley Quantitative and Derivative Strategies
VI X I S LI QUI D FOR ONLY NEAR- TERM MATURI TI ES
Variance swaps are the preferred vehicle to trade volatility
globally, primarily because of the ease of creation and
hedging. Maturities can exist out to ten years, depending on
the market. Moreover, tenors and forward starting periods can
be fully customized in this OTC product.
VIX products, in contrast, are standardized instruments. Their
liquidity is largely concentrated in the first three months,
while variance swaps are traded across the term structure. For
short-dated volatility views, VIX contracts are by far the most
liquid vehicles, while variance is more liquid for longer-dated
views.
VI X I S THE ONLY LI QUI D VEHI CLE FOR VOL OPTI ONS
VIX options are the most liquid way to obtain optionality on
volatility, opening up more avenues to express volatility views
with precision. While options on variance swaps exist in the
OTC market, they tend to be relatively illiquid with wide
bid/offer spreads.

9
It can be shown that a variance swap can be replicated using a static portfolio of
delta-hedged OTM options, unlike VIX futures which have no semi-static hedge.
This has made variance swaps relatively ease to create - they remain the preferred
vehicle to trade volatility in most markets.
12
IV. VIX for Equity Hedgers
One of the key benefits of integrating volatility exposure in an
equity portfolio is the inherent diversification benefit of the
negative correlation between equity returns and volatility.
Exploiting this ‘tail risk hedge’ benefit requires the
specification of an appropriate strategy – how and when
volatility should be scaled into a portfolio. In this section, we
frame the set of potential strategies, and introduce a number
of backtests of the benefits of VIX-based hedging. In this
section, we cover:
> Scaling a VIX hedge into an S&P500 portfolio
> Costs of VIX contracts versus S&P500 contracts
> Systematic performance of VIX hedges versus S&P500
hedges
> Long-term performance of VIX spread-based hedges
> Potential systematic cheapness of VIX hedges
> Exploiting the unit mismatch between the S&P500 and
VIX for downside hedges
Integrating VIX into Long-Only Portfolios
HEDGI NG AN EQUI TY PORTFOLI O USI NG VOLATI LI TY
Because of the inherent negative correlation between equity
returns and volatility, there are potential diversification
benefits from integrating volatility into an equity portfolio.
Scaling a VIX trade is somewhat more complex than using
regular index options, since the performance of the equity
portfolio and that of the hedge are not directly linked. Based
on our earlier analysis, the sensitivity of a volatility-based
hedge to negative returns in equities will allow us to determine
optimal scaling ratios. We assume an S&P500 portfolio for the
subsequent analysis in this section.
FOCUS ON SHORT-DATED VI X- BASED HEDGES
We focus on hedges using short-dated VIX contracts only.
This is driven by three factors: (1) the liquidity profile of VIX
contracts, (2) the sensitivity of VIX futures to changes in the
spot VIX, and (3) the favorable decay characteristics. Hedges
using equity index options, in contrast, often use three to six
month maturities as these tend to give the most favorable
cost/benefit tradeoff.
SCALI NG A VI X POSI TI ON USI NG REGRESSI ON
To compute the notional size required to scale a VIX hedge
into an equity portfolio, we use a historical regression model.
For the entire history of VIX futures trading, which began in
March 2004, we compute a dataset of daily rolling 1-month, 2-
month, and 3-month VIX futures prices. We then compute
how many vol points each contract changed after one month –
the spread of 1-month futures versus spot VIX one month
later, that of 2-month futures versus 1-month futures one
month later, and so on.
Exhibit 15: The responsiveness of the S&P500 as a function of
changes in the first three VIX futures
Based on daily data from 26 Mar 04 through 23 Mar 11
y = -0.ê3x - 0.0Z
R
2
= 0.ê2
y = -0.89x - 0.11
R
2
= 0.58
y = -1.0Zx ÷ 0.12
R
2
= 0.50
-10
-30
-20
-10
0
10
20
30
-30 -20 -10 0 10 20 30 10 50 ê0
1-Honth 6hange |n V|X Future
1
-
H
o
n
t
h

6
h
a
n
g
e

|
n

8
&
P
5
0
0

(
P
e
r
c
e
n
t
}
1-Vorl| vlX Fulures
2-Vorl| vlX Fulures
3-Vorl| vlX Fulures
Source: Morgan Stanley Quantitative and Derivative Strategies
We then regress the 1-month S&P 500 returns on these
changes in VIX futures, which give us the change in the S&P
500 for a given change in the VIX futures. While this is the
inverse of the typical volatility regression, it allows us a more
direct way to compute the required hedge ratios.
10
Based on
our sample, shown in Exhibit 15, we derive the following
hedge ratios:
> 0.63 1-Month VIX Contracts per $100 of S&P500
> 0.89 2-Month VIX Contracts per $100 of S&P500
> 1.07 3-Month VIX Contracts per $100 of S&P500
As expected, the hedge ratio increases for longer-dated
contracts. Since longer-dated VIX futures are not as volatile as
short-dated futures, we need more of them to overcome their
lower sensitivity to VIX changes.

10
When the regression is performed in the conventional way where the change in
volatility is a function of the market return, the beta has to be inverted to compute the
hedge ratio. Under imperfect correlation, a low beta could be computed and then
inverted, creating a high hedge ratio and overhedging of the portfolio. The opposite
regression form mitigates this.
13
VIX Futures as a Portfolio Hedge
SYSTEMATI C BACKTESTS OF COMMON STRATEGI ES
We backtest a number of systematic hedging strategies using
VIX contracts with rolling 1-, 2- and 3-month maturities. Our
backtests begin in May 2007, when the three months
maturities were first listed regularly. We base our analysis on
listed prices, and add a $0.10 bid/ask spread to account for the
relative lack of liquidity in VIX contracts early in their history.
Endless configurations of VIX hedges are possible. To
illustrate the potential merits of VIX-based hedges, we
consider sample backtests of various VIX futures, ATM VIX
call, and OTM VIX call hedging strategies.
OVERVI EW OF VI X FUTURES HEDGI NG STRATEGI ES
Long VIX futures strategies are the most direct way of
hedging equity risk with VIX contracts. VIX futures will
capture the spike in volatility that typically coincides with an
equity selloff. Conversely, since these are futures contracts,
performance would be negative if volatility were to fall
sharply. We can summarize these effects as follows:
Pros:
> An intuitive and direct way to go long volatility as a
hedge
> Zero up-front cost strategy, unlike long option strategies
(excluding margin requirements)
Cons:
> Full upside and downside exposure to volatility changes
> Long VIX futures is generally a negative carry strategy
due to the upward-sloping shape of the term structure
BACKTESTED STRATEGY PERFORMANCE
We backtest a strategy that combines a long S&P500 position
with long VIX futures overlays, using 1 to 3 month maturities.
We use the hedge ratios derived in the previous section. The
futures are rolled monthly at expiration. Exhibit 16 shows the
historical performance of these overlays.
Exhibit 16: The historical performance of systematic VIX futures overlay strategies
Based on daily data from 16 May 07 through 23 Mar 11
10
50
ê0
Z0
80
90
100
110
120
130
V
a
y

0
Z
A
u
ç

0
Z
N
o
v

0
Z
F
e
o

0
8
V
a
y

0
8
A
u
ç

0
8
N
o
v

0
8
F
e
o

0
9
V
a
y

0
9
A
u
ç

0
9
N
o
v

0
9
F
e
o

1
0
V
a
y

1
0
A
u
ç

1
0
N
o
v

1
0
8
t
r
a
t
e
g
y

V
a
|
u
e
-25º
-20º
-15º
-10º
-5º


10º
15º
20º
0
u
a
r
t
e
r
|
y

R
e
t
u
r
n
3&P500
3&P500 ÷ 1V vlX Fulures
3&P500 ÷ 2V vlX Fulures
3&P500 ÷ 3V vlX Fulures
3&P500
3&P500 ÷ 1V vlX
Fulures
3&P500 ÷ 2V vlX
Fulures
3&P500 ÷ 3V vlX
Fulures 3&P500
3&P500 ÷ 1V vlX
Fulures
3&P500 ÷ 2V vlX
Fulures
3&P500 ÷ 3V vlX
Fulures
300Z 2.0º 1.8º ê.5º ê.3º 3009 15.êº 10.1º 8.Zº 11.2º
100Z -3.3º -3.8º 0.9º 1.8º 1009 ê.0º -2.0º -1.5º 1.1º
1008 -9.1º -9.5º -Z.2º -Z.2º 1010 5.1º -1.1º -3.1º -0.Zº
2008 -2.Zº -8.êº -ê.2º -5.5º 2010 -11.1º -3.1º -3.2º 0.5º
3008 -8.1º 0.1º -1.5º -1.Zº 3010 11.3º -2.3º -1.2º 1.8º
1008 -21.9º -2.5º ê.1º 3.5º 1010 10.8º 3.9º -2.2º -0.8º
1009 -11.0º -8.1º -3.9º -3.1º 1011 3.êº Z.9º 0.êº 0.0º
2009 15.9º 0.êº 1.5º 1.1º
Source: Morgan Stanley Quantitative and Derivative Strategies
14
All three variations of the strategies worked as effective
downside hedges. During 4Q08, when the S&P500 dropped
22%, the performance of the S&P500 portfolios with futures
overlays ranged from a 3% loss to a 6% gain. In addition, as
expected, the strongest S&P500 quarter (2Q09, +16%) caused
the futures overlay strategies to underperform, with the
portfolios with overlays delivering a gain between 1% to 2%.
The most obvious difference between the three overlays is in
their long-term performance. The strategy utilizing 3-month
futures significantly outperforms the strategy using 1-month
futures. This is consistent with our earlier finding that per unit
of spot VIX exposure, shorter-dated VIX futures were more
expensive. Through holding higher notional exposure of 3-
month futures, this difference leads to more efficient hedge.
A steep VIX futures term structure, such as in the period after
4Q09, is particularly beneficial for strategies using longer-
dated futures maturities. During this time, the 3-month VIX
futures strategy minimized the carry cost of the position,
compared to strategies using shorter-dated futures.
Overall, the 3-month futures strategy outperformed the 1-
month strategy in 12 of 14 full quarters in our sample, with an
average quarterly outperformance of 2.1%. During the four
worst S&P500 quarters (4Q08, 2Q10, 1Q09 and 1Q08), the
S&P500 lost -13% on average. Adding a VIX futures overlay
improved the average returns during these quarters to -6.1%
for 1-month, -2.0% for 2-month, and -1.6% for 3-month
futures – this again demonstrates the potential benefit of using
longer-dated futures.
During bull markets, on the other hand, VIX futures can
introduce a drag on portfolio performance. During the four
best full S&P500 quarters (2Q09, 3Q09, 3Q10 and 4Q10), the
S&P500 returned 13% on average. Meanwhile, 1-, 2-, and 3-
month futures overlays reduced the average returns to 3%, 2%,
and 4% per quarter, respectively.
RETURN AND RI SK PROFI LE OF FUTURES HEDGES
Exhibit 17 shows the annualized return and risk characteristics
of the different hedging strategies across the entire backtesting
period. The 3-month futures overlay strategy was the only one
with a positive return over this period. Risk reduction was
similar across all overlay strategies, whether measured by the
standard deviation or by the 5% CVaR.
Exhibit 17: The return and risk profile of VIX futures hedges
Based on daily data from 16 May 07 through 23 Mar 11
S&P500
S&P500 +
1M VIX
Future
S&P500 +
2M VIX
Future
S&P500 +
3M VIX
Future
Return (Ann.) -1.8% -4.3% -2.4% 1.9%
Std. Dev. (Ann.) 27.9% 17.5% 17.1% 17.4%
5% CVaR -4.4% -2.6% -2.6% -2.7%
Sharpe Ratio -0.12 -0.33 -0.23 0.03
Source: Morgan Stanley Quantitative and Derivative Strategies
VIX Calls as a Portfolio Hedge
OVERVI EW OF VI X CALL HEDGI NG STRATEGI ES
VIX call options are potentially another approach to hedge
downside risk in equity positions. Compared to VIX futures,
call options give similar upside volatility exposure but reduced
downside exposure. Depending on the VIX implied volatility
regime (the ‘vol of vol’), such a strategy can be attractive:
Pros:
> Reduced downside risk in falling vol environments
> Calls can reduce potential rolldown costs, depending on
the VIX implied vol regime
Cons:
> Up-front premium payment for options exposure, unlike
futures which have no up-front premium outside of
margin
> The mark-to-market of VIX options may not be equal to
changes in the spot VIX prior to expiration
BACKTESTED STRATEGY PERFORMANCE USI NG ATM
VI X OPTI ONS
We start by backtesting strategies using ATM VIX call
overlays, with 1, 2 and 3 months to expiry. Exhibit 18 shows
the historical performance of these overlays. Compared to
VIX futures overlays, which hedge against both increases and
decreases in volatility, the one-sidedness of VIX options
means that they only reduce risk in a single direction. As a
result, the payoff profile of an S&P500 position with VIX
options overlays is much closer to that of the S&P500 hedged
with protective puts.
15
Exhibit 18: The historical performance of systematic ATM VIX call overlay strategies
Based on daily data from 16 May 07 through 23 Mar 11
10
50
ê0
Z0
80
90
100
110
120
130
V
a
y

0
Z
A
u
ç

0
Z
N
o
v

0
Z
F
e
o

0
8
V
a
y

0
8
A
u
ç

0
8
N
o
v

0
8
F
e
o

0
9
V
a
y

0
9
A
u
ç

0
9
N
o
v

0
9
F
e
o

1
0
V
a
y

1
0
A
u
ç

1
0
N
o
v

1
0
8
t
r
a
t
e
g
y

V
a
|
u
e
-25º
-20º
-15º
-10º
-5º


10º
15º
20º
0
u
a
r
t
e
r
|
y

R
e
t
u
r
n
3&P500
3&P500 ÷ 1V ATV vlX Ca||s
3&P500 ÷ 2V ATV vlX Ca||s
3&P500 ÷ 3V ATV vlX Ca||s
S&P500
S&P500 + 1M
ATM VIX Calls
S&P500 + 2M
ATM VIX Calls
S&P500 + 3M
ATM VIX Calls S&P500
S&P500 + 1M
ATM VIX Calls
S&P500 + 2M
ATM VIX Calls
S&P500 + 3M
ATM VIX Calls
3Q07 2.0% 3.7% 4.2% 4.5% 3Q09 15.6% 12.7% 11.5% 12.0%
4Q07 -3.3% -4.4% -2.2% -1.6% 4Q09 6.0% 1.4% 1.9% 3.0%
1Q08 -9.4% -12.7% -10.3% -10.2% 1Q10 5.4% 2.4% 1.8% 2.4%
2Q08 -2.7% -5.7% -4.6% -4.2% 2Q10 -11.4% -3.9% -4.4% -3.7%
3Q08 -8.4% -1.7% -6.6% -7.2% 3Q10 11.3% 3.6% 3.6% 4.9%
4Q08 -21.9% -1.3% 0.8% -4.4% 4Q10 10.8% 7.0% 4.4% 5.5%
1Q09 -11.0% -15.7% -10.9% -10.7% 1Q11 3.6% 8.0% 2.3% 2.2%
2Q09 15.9% 8.8% 9.0% 9.7%
Source: Morgan Stanley Quantitative and Derivative Strategies
For example, during the four worst full quarters for the
S&P500 in our sample, the S&P500 lost on average -13%,
while the three strategies with the VIX options overlay lost
between -6% and -8%, in part due to the call premiums paid.
During the four best quarters for the S&P500, in contrast,
while the S&P500 gained 13%, the three overlay strategies
gained between 7% and 8%. Compared to the overlays using
VIX futures, this represents greater upside participation during
S&P500 rallies.
In the most extreme quarter (4Q08), when the S&P500 lost
22%, these hedging strategies were either flat or slightly down
(-4% to +1% return). During the best S&P 500 quarter in
2Q09, the S&P 500 gained 16%, while the hedged strategies
returned between 9% and 10%.
Strategies using longer-dated VIX calls generally slightly
outperform those using shorter-dated calls. On average, the 2-
month strategy outperformed the 1-month strategy by 27bps a
quarter while the 3-month strategy outperformed by 41bps a
quarter. This shows that scaling between the different
maturities can give very similar volatility exposure, due to the
low decay properties of VIX calls.
However, this slight outperformance comes at the cost of
significantly higher premium to purchase these calls due to the
(1) longer-dated nature of the options and (2) a higher hedging
ratio.
RETURN AND RI SK PROFI LE OF ATM CALL HEDGES
Exhibit 19 shows the annualized return and risk characteristics
of the ATM VIX call overlays. Risk reduction was generally
smaller than in the case of VIX futures overlays. On the other
hand, the average returns of the 1- or 2-month overlays are
generally more attractive than for VIX futures overlays.
Exhibit 19: The return and risk profile of ATM VIX call hedges
Based on daily data from 16 May 07 through 23 Mar 11
S&P500
S&P500 +
1M ATM
VIX Calls
S&P500 +
2M ATM
VIX Calls
S&P500 +
3M ATM
VIX Calls
Return (Ann.) -1.8% -0.9% -0.8% -0.4%
Std. Dev. (Ann.) 27.9% 21.3% 20.9% 21.5%
5% CVaR -4.4% -3.1% -3.1% -3.3%
Sharpe Ratio -0.12 -0.11 -0.11 -0.08
Source: Morgan Stanley Quantitative and Derivative Strategies
16
WHAT ABOUT OUT OF THE MONEY CALL OPTI ONS?
Like hedges using regular equity puts, VIX call hedges can be
configured in any combination of maturities and strikes. To
indicate the potential benefits of using OTM VIX calls, we
show backtests of various 1-month OTM VIX call strategies,
using 10-, 20-, and 30-delta calls.
Since the calls are OTM and thus have a lower mark-to-
market to spot VIX changes, we double the notional exposure
for these calls. A larger notional exposure in OTM VIX calls
can still be cheaper than a single ATM VIX call of equal
maturity. While this exposure in OTM VIX calls will be
beneficial when volatility rises sharply, it would give much
less protection in gradually increasing volatility scenarios.
Exhibit 20 shows the results of our historical backtest. In
4Q08, the quarter with the worst S&P500 performance, the
hedges performed quite well, with the strategies returning
anywhere from 68bps (10-delta calls) to 5.6% (30-delta calls).
Meanwhile, the ATM call strategies had returns between
+80bp and -4.4%, depending on maturity (see Exhibit 18).
Similarly, during the four best S&P500 quarters, OTM VIX
call strategies outperformed the ATM VIX call strategies, with
returns ranging from 7% (30-delta calls) to 11% (10-delta
calls). This compares to an average 13% return for the
S&P500, and 7%-8% returns for ATM call strategies. This
reflects the lower premiums for OTM VIX call options than
for ATM options, even after the scaling of the notional
exposure.
On the other hand, OTM call strategies underperform ATM
strategies during small declines in the S&P 500. During the
two quarters in which S&P500 returns were between 0% and -
5%, ATM calls generally performed better than OTM calls,
particularly longer-dated ATM calls
.
Exhibit 20: The historical performance of systematic OTM VIX call overlay strategies
Based on daily data from 16 May 07 through 23 Mar 11
10
50
ê0
Z0
80
90
100
110
120
130
V
a
y

0
Z
A
u
ç

0
Z
N
o
v

0
Z
F
e
o

0
8
V
a
y

0
8
A
u
ç

0
8
N
o
v

0
8
F
e
o

0
9
V
a
y

0
9
A
u
ç

0
9
N
o
v

0
9
F
e
o

1
0
V
a
y

1
0
A
u
ç

1
0
N
o
v

1
0
8
t
r
a
t
e
g
y

V
a
|
u
e
-25º
-20º
-15º
-10º
-5º


10º
15º
20º
0
u
a
r
t
e
r
|
y

R
e
t
u
r
n
3&P500
3&P500 ÷ 10-0e|la vlX Ca||s
3&P500 ÷ 20-0e|la vlX Ca||s
3&P500 ÷ 30-0e|la vlX Ca||s
S&P500
S&P500 10-
Delta VIX Calls
S&P500 20-
Delta VIX Calls
S&P500 30-
Delta VIX Calls S&P500
S&P500 10-
Delta VIX Calls
S&P500 20-
Delta VIX Calls
S&P500 30-
Delta VIX Calls
3Q07 2.0% 0.2% 2.5% 4.4% 3Q09 15.6% 14.2% 12.9% 11.9%
4Q07 -3.3% -4.6% -5.9% -7.1% 4Q09 6.0% 4.0% 2.7% 1.1%
1Q08 -9.4% -10.8% -11.8% -13.0% 1Q10 5.4% 3.8% 2.8% 1.7%
2Q08 -2.7% -3.8% -4.3% -5.1% 2Q10 -11.4% -6.7% -1.6% 1.1%
3Q08 -8.4% -7.6% -4.5% -3.7% 3Q10 11.3% 8.3% 5.7% 2.3%
4Q08 -21.9% 0.1% 4.6% 6.0% 4Q10 10.8% 9.3% 7.9% 6.6%
1Q09 -11.0% -12.6% -14.1% -15.6% 1Q11 10.8% 9.3% 7.9% 6.6%
2Q09 15.9% 12.9% 10.1% 8.0%
Source: Morgan Stanley Quantitative and Derivative Strategies
17
VIX Calls or SPX Puts for Hedging?
HI STORI CALLY, VI X CALLS HAVE BEEN CHEAPER
To compare the relative costs of using VIX calls versus SPX
puts, Exhibit 21 shows the historical premiums for both. We
construct a set of systematic strategies that use either SPX puts
or a scaled notional of VIX calls as per our previous scaling
ratios. On each roll date, we value the cost of the strategies
using 10- through 50-delta options.
Exhibit 21: The cost of 1-month SPX put hedges versus scaled 1-
month VIX call hedges
Based on data from 16 May 07 through 23 Mar 11
0.0º
0.5º
1.0º
1.5º
2.0º
2.5º
3.0º
3.5º
1.0º
1.5º
0º 1º 2º 3º 1º 5º êº Zº 8º 9º
6ost of 1-Honth 8PX Put hedge
6
o
s
t

o
f

8
c
a
|
e
d

1
-
H
o
n
t
h

V
|
X

6
a
|
|

h
e
d
g
e
50-0e|la 0pl|ors
30-0e|la 0pl|ors
10-0e|la 0pl|ors
15 deçree ||re
Source: Morgan Stanley Quantitative and Derivative Strategies
Based on these scaling ratios, VIX calls with a given delta
have been persistently cheaper than SPX puts with the same
delta. Moreover, this cheapness has been relatively consistent
and ranges from a 55% cheaper (50-delta options) to 60%
cheaper (30-delta options).
However, this cheapness is based on the assumption of a
constant scaling or hedging ratio. In practice, the optimal
hedge ratio is variable, as indicated by the R-squared of the
regressions in Exhibit 15. This may lead to over- or under-
hedging. Using SPX puts does not carry that risk.
In addition, while VIX calls may seem optically cheap based
on their up-front premium, an investor in these instruments
also implicitly pays for the shape of the term structure. Since
the term structure is generally upward-sloping, ATM VIX call
strikes are generally higher than spot VIX levels. Unless this
gap is overcome by expiry through rising spot VIX levels, the
VIX call cannot be monetized.
I N-LI NE PERFORMANCE VERSUS S&P500 PUTS
We perform an empirical analysis of the true potential
‘cheapness’ of VIX calls versus SPX puts through a set of
backtests. We construct sets of 1-month option overlays on
the S&P500 using 10-delta to 50-delta VIX calls and SPX
puts. All options are held to expiration and then rolled into the
next month’s contract.
Exhibit 22: The returns of the S&P500 hedged with 30-delta S&P500
puts and VIX calls versus the S&P500 return
Based on daily data from 16 May 07 through 23 Mar 11
-15º
-10º
-5º


10º
-20º -15º -10º -5º 0º 5º 10º 15º 20º
1-Honth 8&P500 Return
1
-
H
o
n
t
h

8
t
r
a
t
e
g
y

R
e
t
u
r
n
30-0e|la 3&P500 Puls
30-0e|la vlX Ca||s
lr-||re |edç|rç poWer
|r 0clooer 2008
lr-||re perlorrarce lor
sra|| dec||res
vlX |edçes s||ç|l|y
oulperlorr W|er l|e
3&P500 |s up
1009 W|er 3&P500 ard
vlX Were ool| la|||rç
10
50
ê0
Z0
80
90
100
110
V
a
y

0
Z
J
u
|

0
Z
3
e
p

0
Z
N
o
v

0
Z
J
a
r

0
8
V
a
r

0
8
V
a
y

0
8
J
u
|

0
8
3
e
p

0
8
N
o
v

0
8
J
a
r

0
9
V
a
r

0
9
V
a
y

0
9
J
u
|

0
9
3
e
p

0
9
N
o
v

0
9
J
a
r

1
0
V
a
r

1
0
V
a
y

1
0
J
u
|

1
0
3
e
p

1
0
N
o
v

1
0
J
a
r

1
1
V
a
r

1
1
8
t
r
a
t
e
g
y

V
a
|
u
e
3&P500
3&P500 ÷ 1-Vorl| 30-0e|la 3PX Puls
3&P500 ÷ 1-Vorl| 30-0e|la vlX Ca||s
Source: Morgan Stanley Quantitative and Derivative Strategies
The first chart in Exhibit 22 compares the monthly
performance of S&P 500 overlaid with 30-delta VIX calls and
30-delta SPX puts, versus the performance of the S&P 500.
The results for other delta levels were generally comparable.
In rising market environments, VIX hedges tended to
outperform SPX put hedges slightly, due to their relative
cheapness. During the 26 months of positive S&P500 returns
in our backtest sample, the VIX hedge outperformed the SPX
put hedge by 1.1% per month on average.
In months with small declines in the S&P500 (0% to -5%), the
performance of VIX call hedges outperformed the S&P500
put hedge by 32bps per month on average. For larger declines
– up to -10% - SPX put hedges generally outperformed VIX
call hedges by 0.9% per month.
For very large declines, returns were in-line again. During
Oct08 when the S&P500 was down 17%, the VIX call-hedged
strategy performed nearly in-line with the SPX put-hedged
strategy.
Over the long-run, a systematic VIX call hedge would have
outperformed an SPX put hedge, while providing nearly the
same hedging power in S&P500 drawdown periods (see lower
chart in Exhibit 22).
18
Since the upfront premium costs of scaled 30-delta VIX calls
were only about 40% of the price of 30-delta SPX puts,
Exhibit 23 shows the results of a hedging backtest where we
scale the VIX call exposure to the same premium outlay as the
SPX puts.
Exhibit 23: The cumulative returns of a systematic 30-delta S&P500
overlays including one with amplified VIX call exposure
Based on daily data from 16 May 07 through 23 Mar 11
10
50
ê0
Z0
80
90
100
110
V
a
y

0
Z
A
u
ç

0
Z
N
o
v

0
Z
F
e
o

0
8
V
a
y

0
8
A
u
ç

0
8
N
o
v

0
8
F
e
o

0
9
V
a
y

0
9
A
u
ç

0
9
N
o
v

0
9
F
e
o

1
0
V
a
y

1
0
A
u
ç

1
0
N
o
v

1
0
F
e
o

1
1
8
t
r
a
t
e
g
y

V
a
|
u
e
3&P500
3&P500 ÷ 1-Vorl| 30-0e|la 3&P500 Puls
3&P500 ÷ 1-Vorl| 30-0e|la vlX Ca||s
3&P500 ÷ 1-Vorl| 30-0e|la vlX Ca||s (250º Exposure)
Source: Morgan Stanley Quantitative and Derivative Strategies
Over the longer term, the performance of this scaled VIX call
hedge is similar to that of the ‘regular’ VIX call hedge. While
the performance in periods of rising S&P500 was generally
lower (6bps outperformance over SPX put hedge per month,
versus 1.1% on average for the ‘regular’ VIX call hedge),
performance in declining S&P 500 periods was substantially
higher.
In 4Q08, for example, our scaled VIX call hedge plus an
S&P500 position returned 31%. In contrast, ‘regular’ VIX call
hedges or SPX equity put hedges were up by 1-2%, while the
S&P500 alone was down by around -17% over this period. In
May 2010, with the S&P500 down -8%, the scaled VIX hedge
was up 6% while the other two hedging strategies were down
-1% to -2%.
On the other hand, the higher notional VIX exposure also
highlights periods of hedging mismatch. In Nov 08, the VIX
started to decline very quickly after reaching its all-time high,
even though the S&P 500 did not bottom until Mar 09. VIX
hedges did not provide downside protection during this period,
since VIX and S&P500 were both falling at the same time.
During Nov 08, the S&P500 was down -7% while the 30-delta
S&P500 put strategy fell approximately -2%. The regular-
sized and higher notional VIX call strategy actually fell -9%
and -13%, respectively, because of the sharp decline in
implied volatility during this period.
These results highlight the tactical nature of many volatility-
based hedges. The appropriateness of VIX-based hedges
depends on the level of the VIX as well as on the recent
direction of volatility changes.
Performance of Spread-Based VIX Hedges
BACKTESTI NG OTHER SPREAD- BASED VI X HEDGES
Similar to SPX put-spread based hedging strategies, we can
also consider VIX call spread hedging strategies to target
changes in volatility more precisely. The upper chart in
Exhibit 24 shows the results of four backtests involving
commonly used VIX option hedging strategies:
> Buy 1x 1-month 40-delta VIX call
> Buy 1x 1-month 40-delta VIX calls, Sell 1x 1-month 20-
Delta VIX call
> Sell 1x 1-month 40-delta VIX call, Buy 2x 1-month 20-
Delta VIX call
> Buy 1x 1-month 40-delta VIX call, sell 1x 2-month VIX
call with the same strike
The lower chart shows the same strategies, but pushes out all
maturities by one month. All strategies are rolled on the
monthly VIX expiration dates. We assume a $0.10 bid-ask
spread for every transaction.
Our benchmark is the 40-delta call strategy. The call spread
strategy shows the impact of selling some upside to cheapen
the hedge, with the deltas chosen to result in approximately a
50% premium reduction. The 1x2 call spread strategy is thus
approximately zero premium, and is designed to gain when the
VIX is rising sharply while accepting a small loss when the
VIX rises slightly. The mark-to-market profile of this trade is
attractive – the position only tends to lose significantly right
near expiration if VIX is near the long call strike. This is
because of the slow decay of VIX options. However, full
upside is given up in case of a volatility shock.
The calendar call spread purchases a near-dated call while
attempting to fund the trade by selling a longer-dated VIX
call. The rolldown of the term structure for the longer-dated
call will typically offset some of the cost of holding the near-
dated call. If volatility spikes sharply, the shorter-dated call
should increase in value faster than the longer-dated call.
However, this trade takes on other term structure risks such as
(1) a persistently flat term structure and (2) parallel shifts in
the term structure.
19
Exhibit 24: The cumulative P&L of various systematic alternate
hedging strategies using VIX calls
Based on daily data from 16 May 07 through 23 Mar 11
-20
-15
-10
-5
0
5
10
15
20
25
30
35
10
V
a
y

0
Z
A
u
ç

0
Z
N
o
v

0
Z
F
e
o

0
8
V
a
y

0
8
A
u
ç

0
8
N
o
v

0
8
F
e
o

0
9
V
a
y

0
9
A
u
ç

0
9
N
o
v

0
9
F
e
o

1
0
V
a
y

1
0
A
u
ç

1
0
N
o
v

1
0
F
e
o

1
1
6
u
m
u
|
a
t
|
v
e

8
t
r
a
t
e
g
y

P
&
L
1-Vorl| 10-0e|la Ca||
1-Vorl| 20/10 0e|la Ca|| 3pread
1-Vorl| 20/10 0e|la 1x2 Ca|| 3pread
1-Vorl| 10 0e|la Ca|erdar Ca|| 3pread
-20
-15
-10
-5
0
5
10
15
20
25
30
35
10
V
a
y

0
Z
A
u
ç

0
Z
N
o
v

0
Z
F
e
o

0
8
V
a
y

0
8
A
u
ç

0
8
N
o
v

0
8
F
e
o

0
9
V
a
y

0
9
A
u
ç

0
9
N
o
v

0
9
F
e
o

1
0
V
a
y

1
0
A
u
ç

1
0
N
o
v

1
0
F
e
o

1
1
6
u
m
u
|
a
t
|
v
e

8
t
r
a
t
e
g
y

P
&
L
2-Vorl| 10-0e|la Ca||
2-Vorl| 20/10 0e|la Ca|| 3pread
2-Vorl| 20/10 0e|la 1x2 Ca|| 3pread
2-Vorl| 10 0e|la Ca|erdar Ca|| 3pread
Source: Morgan Stanley Quantitative and Derivative Strategies
1X2 CALL SPREADS DELI VER HI GH PAYOUTS I N
CRI SI S SCENARI OS WI TH LOW HOLDI NG COSTS
While simple VIX call spreads generally reduced the cost of
the hedge, upside participation in volatility spikes was
significantly reduced as well, making the tradeoff between
cost and upside participation unattractive. The strategy
generally decays over time, while underperforming other
hedging strategies during volatility spikes.
Calendar spread structures are particularly attractive during
periods of steep, upward-sloping term structures, such as the
one in force since late 2009. In such an environment, calendar
call spreads can minimize holding costs due to decay.
However, in relatively flat term structure environments – like
in 2007 – these strategies are generally unattractive as the
decay in front-month options is much faster than that for
longer-dated options. Moreover, P&L of calendar spreads
tends to be more volatile than that of other strategies, since
calendar spreads involve two separate maturities.
The 40-delta call worked the best in a 2008-style scenario with
sharply rising volatility due to its lack of an upside cap. When
factoring long-run holding costs, a 1x2 call spread worked
better as a tail risk hedge. The up-front premium for this trade
is generally close to zero. This benefits performance in a
falling volatility environment, while still providing upside
exposure when volatility spikes.
On the other hand, this trade suffers during small volatility
spikes, such as March 2011. On the back of Middle East
protests and the Japanese earthquake, VIX rose but not to
crash levels. When our systematic 1-month rolled in
February, it traded a March 2011 20-24 1x2 call spread. The
options settled to a March opening print of 25.14 and created a
loss.
A VI X 1X2 CALL SPREADS NEED TO BE ACTI VELY
MANAGED TO AVOI D POTENTI AL LOSSES
The mark-to-market profile of the 1x2 call spread is greatly
affected by the slow decay of VIX options. We display in
Exhibit 25 the mark-to-market profile of a 20-24 1x2 call
spread which was nearly zero cost on 23 Mar 11.
Exhibit 25: The hypothetical future mark-to-market profile of an April
VIX 20-24 1x2 calls spread as of 23 Mar 11
-5
-1
-3
-2
-1
0
1
2
3
1
5
ê
Z
12 1ê 20 21 28 32 3ê
V|X Futures Pr|ce
1
-
H
o
n
t
h

V
|
X

6
a
|
|

8
p
r
e
a
d

P
&
L
1-Vorl| url|| Valur|ly
2 wee|s url|| Valur|ly
1 wee| url|| Valur|ly
Al Valur|ly
Source: Morgan Stanley Quantitative and Derivative Strategies
The structure on a mark-to-market basis gives full upside in
the case of a volatility spike. However there is a 8 volatility
point gap in the middle at expiration that could produce a loss.
The calls decay slowly though which mitigates most of this
risk. This position should be rolled at the latest a week before
expiration to avoid a loss due to rapid decay. This requires the
hedge to be actively managed to avoid this expiration loss.
Using a Scaling Mismatch for a VIX Hedge
UTI LI ZI NG THE SCALI NG MI SMATCH OF I NDI CES
VIX and the S&P500 are denominated in different units. $1 is
about 5% (1/20) of the VIX, but only about 8bps of the
S&P500 (1/1300). Therefore, a VIX call option costing $1
should purchase more VIX upside than a $1 SPX put option
would purchase in S&P500 downside.
A trade that sells a $1 S&P500 put to finance a $1 VIX call
would be selling very deep OTM S&P500 exposure to buy a
more near-the-money VIX hedge for zero upfront premium. If
the S&P500 declines, there is a positive probability that the
20
VIX call will be ITM at expiration. As long as the S&P500
does not cross the SPX put strike, this strategy will yield a
positive payoff in this scenario.
However, because the S&P500 is denominated in larger units,
an X% sell-off in the S&P500 causes a larger S&P500 point
shift than the VIX point shift under an X% rise in the VIX.
For example, a 10% drop in the S&P500 corresponds to
approximately 130 index points (1300 * 10%) while a 10%
rise in the VIX corresponds to 2 index points (20 * 10%).
If an equity sell-off is substantial and the S&P500 breaches
the put strike, the SPX put would thus rise faster in value than
the VIX call. In a 1987-style scenario, this scaling mismatch
would likely cause such a strategy to be substantially
unprofitable.
SI NCE VI X OPTI ONS WERE LI STED I N 2006, THE
S&P500 HAS NOT CROSSED I TS PUT STRI KE…YET
Exhibit 26 shows the backtest of a strategy that buys the listed
1-month VIX call with a bid closest to $1, using prices at the
close of every listed VIX expiration. We then construct a set
of 1-month theoretical SPX put prices using the volatility
surface of listed options. We assume a $1 bid/offer spread for
the SPX puts and compute a put strike that would have had a
theoretical bid with the same price as the VIX call.
We plot the cumulative P&L of this strategy, the level of the
S&P500, and the strike of the $1 SPX put over time. In our
backtest, the S&P500 never crossed the put strikes. As a
result, a loss was never realized at roll time, while the strategy
provided upside participation through the VIX call, resulting
in positive P&L in times of rapid VIX spikes.
Exhibit 26: The cumulative P&L of a strategy that sells 1-month
S&P500 puts costing $1 and buys 1-month VIX calls costing $1
Based on daily data from 24 Feb 06 through 23 Mar 11
-20
-10
0
10
20
30
10
50
ê0
Z0
F
e
o

0
ê
V
a
y

0
ê
A
u
ç

0
ê
N
o
v

0
ê
F
e
o

0
Z
V
a
y

0
Z
A
u
ç

0
Z
N
o
v

0
Z
F
e
o

0
8
V
a
y

0
8
A
u
ç

0
8
N
o
v

0
8
F
e
o

0
9
V
a
y

0
9
A
u
ç

0
9
N
o
v

0
9
F
e
o

1
0
V
a
y

1
0
A
u
ç

1
0
N
o
v

1
0
F
e
o

1
1
0
200
100
ê00
800
1000
1200
1100
1ê00
1800
3lraleçy P&L (|ell) 3&P500 (r|ç|l) 3&P500 Pul 3lr||e (r|ç|l)
Source: Morgan Stanley Quantitative and Derivative Strategies
On a mark-to-market basis, the trade did become unprofitable
whenever the S&P500 neared the put strike. For example on
10 Oct 08, the S&P500 reached 899 while the put for this
period was struck at 894. The mark-to-market value of the
SPX put rose much faster than the value of the VIX call. If the
S&P500 had continued to fall, the value of the SPX put would
have continued to rise more rapidly. However, the S&P500
rose 105 points on the next day, mitigating the mark-to-market
loss.
SUDDEN MARKET CRASHES HOWEVER WI LL
PRODUCE A LOSS WI TH THI S STRATEGY, SUCH AS ‘ 87
In our backtest, 1-month SPX puts with around a $1 premium
were on average 15% OTM, ranging from 5% to 46% OTM. If
market were in a low implied volatility regime and equities
were to suddenly drop, such as in the October 87 crash, it is
likely this trade would have been very unprofitable.
We can highlight this possibility using the dynamics of the
1987 crash. During the 1987 crash, the S&P500 dropped by
32%, while the old VIX rose from 23 to 150, an increase of
127 points.
At current S&P 500 and volatility levels, our strategy would
lead us to short an S&P 500 put that is 15% OTM, and a VIX
call with about a 27.5 strike. A crash like in 1987 would lead
to an increase in the intrinsic value of the S&P 500 put of
1300* (32%-15%) = $221, while the VIX call would have
increased in intrinsic value by 150 – 27.5 = $122.5. This
would product a mark-to-market loss of approximately $221 -
$122.5 = $98.5.
Note that this might be a conservative estimate of the mark-to-
market loss, since we did not consider the time value of the
options. The VIX call would be so far in the money that it
would trade with very little time value – it would be
equivalent to a cash + VIX futures position. The S&P 500 put,
on the other hand, would have a substantial short volatility
exposure due to its higher time value. This would tend to
exacerbate the mark-to-market loss of this position. We can
see this in the volatility of the P&L of our strategy in Exhibit
26, for example during October 2008.
21
V. VIX for Alpha-Seekers
Volatility has long been an alpha source for many absolute
return strategies using index OTC variance swaps. After the
events of 2008, many question whether liquid, listed VIX
derivatives can complement or replace index variance swaps
in these systematic alpha strategies
This section provides a framework for using VIX futures &
options as volatility alpha vehicles, highlighting advantages
as well as potential risks and drawbacks:
> Using VIX derivatives to trade the S&P500 implied vol
term structure
> Constructing term structure trades with reduced
downside risk
> VIX options implied volatility has historically traded rich
versus subsequent realized vol
> We backtest the performance of potential alpha trades
A Listed Vehicle to Arb the Term Structure
VI X FUTURES VS. THE SPX VOL TERM STRUCTURE
The SPX implied volatility term structure is generally upward
sloping – the implied volatility of longer-dated options is
higher than that of shorter-dated options. Exhibit 27 shows
three sample term structures – upwards sloping, flat and
inverted. Over the past five years, the term structure was
upwards sloping on 78% of trading days. Inversions of the
term structure typically only occur in distressed market
scenarios. The shape and slope of the term structure during
inversions can be more extreme than when the term structure
is upwards sloping – see Exhibit 28.
Exhibit 27: Three single day examples of potential term structure
shapes: upward sloping, downward sloping, and flat


10º
15º
20º
25º
30º
35º
1-Vorl| 3-Vorl| ê-Vorl| 1-Year
|mp||ed Vo| Tenor
A
T
H

|
m
p
|
|
e
d

V
o
|
21 Jar 0Z - Vu|l|-Year 1V ATV vo| LoW
Z Vay 10 - l|ç|esl vo| 0ay 0ur|rç 3over|eçr R|s| Rel|al|or
12 3ep 08 - T|e 0ay Pr|or lo Le|rar 8ar|ruplcy
Source: Morgan Stanley Quantitative and Derivative Strategies
The upward-sloping term structure of implied volatility is
often attributed to supply-demand dynamics. Demand for
volatility comes from tactical option positions as well as
downside hedging strategies. Volatility suppliers, such as
covered call writers or outright variance swaps sellers, fill this
demand. Further out the term structure, there are fewer natural
sellers of volatility. The premium for longer-dated tail risk
increases, helping to push the term structure upwards.
VIX futures are priced off the SPX implied volatility term
structure, and are related to the 1-month forward starting
implied volatility.
11
As a result, the basis of VIX futures tends
to track the shape of the S&P500 term structure (Exhibit 28).
Exhibit 28: The VIX futures term structure has a similar shape as the
S&P500 implied vol term structure
Based on daily data from 26 Mar 04 through 23 Mar 11
-15
-10
-35
-30
-25
-20
-15
-10
-5
0
5
10
15
V
a
r

0
1
J
u
|

0
1
N
o
v

0
1
V
a
r

0
5
J
u
|

0
5
N
o
v

0
5
V
a
r

0
ê
J
u
|

0
ê
N
o
v

0
ê
V
a
r

0
Z
J
u
|

0
Z
N
o
v

0
Z
V
a
r

0
8
J
u
|

0
8
N
o
v

0
8
V
a
r

0
9
J
u
|

0
9
N
o
v

0
9
V
a
r

1
0
J
u
|

1
0
N
o
v

1
0
|
m
p
|
|
e
d

V
o
|

8
p
r
e
a
d

(
V
o
|

p
o
|
n
t
s
}
ê-Vorl| vlX Fulures - vlX
ê-Vorl| ATV 3PX lrp||ed vo| - 1-Vorl| ATV 3PX lrp||ed vo|
Source: Morgan Stanley Quantitative and Derivative Strategies
YI ELD FROM UPWARD-SLOPI NG TERM STRUCTURES
One of the most common volatility alpha strategies is to
exploit the generally upward-sloping term structure through
risk premium extraction, for example by systematically selling
forward starting variance swaps. The term structure of forward
starting implied volatilities tends to be upwards sloping as
well, driven by the shape of SPX implied volatilities. Forward
starting implied volatility can be traded via VIX futures.

11
One can express the price of the VIX at time t=T as an expectation of future realized
volatility from time T to time T+30:
] [ 30
2
+ −
=
= T T
T t
T
RVol E VIX
.
The price of a VIX future, however, is an expectation of where the expectation of
realized vol will be in the future. So at time t=0, the price of a VIX future can be
thought of as:
] ] [ [ ] [ _ 30
2 0 0
+ −
= = =
= = T T
T t t
T
t
T
RVol E E VIX E Fut VIX
.
Using Jensen’s Inequality, it can be shown that:
]] [ [ _ ]] [ [ 30
2 0
30
2 0
+ −
= =
+ −
= =
≤ ≤ T T
T t t
T
T T
T t t
RVol E E Fut VIX RVol E E
which simplifies to:
30 30
_
+ − + −
≤ ≤
T T T T T
wap FowardVarS Fut VIX Swap ForwardVol
For more information, please refer to “A Tale of Two Indices,” Carr & Wu, The
Journal of Derivatives, Spring 2006.
22
For example, if spot VIX remained unchanged over the
lifespan of a VIX future, the price of the VIX future has to
converge towards the spot VIX as it moves towards expiry. In
an upwards-sloping term structure environment, this ‘rolldown
effect’ can be exploited by selling the VIX future.
Spikes in volatility would typically lead to a loss in this
scenario, which can often be multiples of the potential risk
premium gain. Successful premium extraction strategies of
this type therefore require a hedging strategy that mitigates the
impact of volatility spikes.
FOUR METHODS TO GENERATE HEDGED ALPHA
VIX- based alpha strategies are often based on exploiting the
richness of the SPX implied volatility term structure. In
addition, VIX options tend to trade at a premium relative to
the realized volatility of VIX futures in low-risk environments.
Through selling (delta-hedged) VIX options, we can also
exploit this ‘vol-of-vol’ effect as another potential alpha
source.
Utilizing all the payoff profiles that combinations of VIX
futures and options allow, we can structure different trades
that can extract this rolldown P&L while minimizing potential
losses. We examine four basic types of trades as examples of
pure alpha trades around the VIX:
> VIX future calendar spreads
> VIX put spreads
> VIX calendar put spreads
> Delta-hedged VIX straddles
VIX future calendar spreads are the most direct way to
leverage the term structure steepness. Put spreads can reduce
the exposure to volatility spikes. Delta-hedged VIX straddles
are pure plays on the ‘vol-of-vol’ effect.
In addition to analyzing each of these strategies individually,
we will also highlight the potential benefits of aggregating
them into a single strategy.
VIX Futures Calendar Spreads
MOST DI RECT WAY TO TRADE THE TERM STRUCTURE
Selling short-dated VIX futures is the simplest way of
exploiting the steepness of the upward-sloping term structure.
However, the strategy is exposed to sudden spikes in the VIX.
Calendar spreads can mitigate the impact of spikes in the VIX.
Exhibit 10 earlier in the paper showed that the rolldown cost
per unit of VIX exposure was not uniform across all
maturities. Longer-dated contracts tend to have a smaller
rolldown cost than a 1-month future. By using some of the
rolldown yield from selling a 1-month VIX future to purchase
an amount of longer-dated VIX futures, the overall strategy
can be effectively hedged against term structure inversions
and spikes in the VIX.
Based on historical rolling VIX futures data, we find that 2- to
6-month futures have a sensitivity to 1-month futures of 1.4,
1.7, 1.9, 2.1 and 2.4 times, respectively. Using these ratios,
we can scale a hedging strategy that would profit in steep term
structure environments. If the term structure remains upward-
sloping, and if the historical hedge ratio is accurate, such a
strategy could be profitable. Two major risks remain: (1)
parallel shifts inconsistent with these hedge ratios and (2) a
sustained inversion in the term structure. An inverted term
structure would cause the entire strategy to work in reverse as
the short 1-month future rolls up the futures curve rather than
down.
1-MONTH / 3- OR 4-MONTH SPREADS WORKED BEST
Using these historical hedge ratios of longer-dated VIX
futures to 1-month futures, we backtested a series of
systematic strategies that execute this trade monthly. For
every $100 of strategy value, we sell one 1-month VIX future
while buying 2- through 6-month futures according to our
hedge ratios. The cash component earns the Effective
Overnight Federal Funds rate and the strategies are rolled on
the listed VIX futures expirations. All futures trades assume
execution at their daily closing prices.
We have extended the history of VIX futures from their Mar
2004 first listing back to January 1996, using a regression
model.
12
This longer period covers three bull markets and two
recessions, allowing us to evaluate our strategies more
robustly.
Exhibit 29 shows the performance of these VIX futures
calendar spreads, comparing them to that of a strategy that is
short the 1-month VIX futures without a hedge. The table

12
Because VIX futures are strongly linked to forward starting volatility, we construct a
two-factor regression model. We regress in-sample the VIX futures basis on (1) the
spread between the equivalent ATM forward volatility level and spot 1-month implied
vol and (2) the square root of time until maturity. In sample, this model has a
regression R
2
of 98% and we use the in-sample results to extrapolate the data using
S&P500 volatilities.
23
separates the performance of the strategies during the
backfilled period from that over the whole period.
Strategies using longer-dated VIX futures as hedges yielded a
higher return in general, since the longer-dated vol term
structure tends to be quite flat. However, volatility and CVaR
are higher for these strategies as well, compared to shorter-
dated hedges, since longer-dated futures are a less effective
hedging tool.
Most strategies outperformed the unhedged trade on both an
absolute and a relative basis. Using historical VIX futures data
since 2004, the highest Sharpe Ratio (1.15) strategy was based
on hedging with 4-month VIX futures. Over the whole period
since 1996, which includes our backfilled data, using 3-month
VIX futures yielded the highest Sharpe Ratio (0.85).
Exhibit 29: The historical performance of hedged VIX futures spread
strategies and a naked short VIX futures strategy
Based on daily data from 4 Jan 96 through 23 Mar 11
0
200
100
ê00
800
1000
J
a
r

9
ê
J
a
r

9
Z
J
a
r

9
8
J
a
r

9
9
J
a
r

0
0
J
a
r

0
1
J
a
r

0
2
J
a
r

0
3
J
a
r

0
1
J
a
r

0
5
J
a
r

0
ê
J
a
r

0
Z
J
a
r

0
8
J
a
r

0
9
J
a
r

1
0
J
a
r

1
1
3|orl 1x 1V Fulure, Lorç 1.1x 2V Fulures
3|orl 1x 1V Fulure, Lorç 1.Zx 3V Fulures
3|orl 1x 1V Fulure, Lorç 1.9x 1V Fulures
3|orl 1x 1V Fulure
Unhedged 2-Honth 3-Honth 4-Honth 5-Honth ô-Honth
Dur-ol-$amµ|e Pesu|rs. 4 Jan 96 - 2J Var 11
Relurr (Arr.) 13.2º 11.2º 15.2º 11.1º 11.2º 11.êº
R|s| (Arr.) 20.2º 13.Zº 13.9º 13.Zº 15.0º 1Z.êº
5º CvaR -3.1º -2.0º -2.1º -2.0º -2.2º -2.êº
3|arpe Ral|o 0.18 0.Z9 0.85 0.Z8 0.Z2 0.ê1
ln-$amµ|e Pesu|rs. 24 Var 04 - 2J Var 11
Relurr (Arr.) 13.9º Z.8º 12.5º 11.1º 11.3º 11.1º
R|s| (Arr.) 21.0º 8.5º 9.5º 10.1º 10.êº 12.1º
5º CvaR -3.5º -1.1º -1.3º -1.1º -1.5º -1.Zº
3|arpe Ral|o 0.55 0.ê3 1.0ê 1.15 1.11 0.9ê
Long V| X Future Hatur|ty
Source: Morgan Stanley Quantitative and Derivative Strategies
During the 4Q08 vol spike, the hedged strategies were able to
minimize losses and recovered relatively quickly, while the
unhedged strategy required nearly two years to recover to its
pre-4Q08 highs. During the higher-volatility period between
2000 and 2002, all strategies suffered. The implied volatility
term structure oscillated from relatively flat to inverted for a
sustained period, which tends to make these volatility
premium extraction strategies unprofitable.
Short-Dated VIX Ratio Put Spreads
SELLI NG THE TERM STRUCTURE THROUGH PUTS
To mitigate the downside risk from volatility spikes that is
inherent in a VIX futures selling strategy, we can hedge using
long positions in longer-dated VIX futures. Another strategy
involves trading short-dated puts on the VIX instead, which
have limited downside.
For VIX options, the ATM strike is defined as the level of the
corresponding VIX future. When the term structure is
upwards-sloping, this ATM strike will be higher than the spot
VIX level. If the term structure remains unchanged, an ATM
put becomes increasingly ITM as the VIX futures price
approaches the VIX spot.
Unlike strategies involving VIX futures, buying VIX puts
requires an up-front premium. The level of the VIX at
expiration has to drop by more than the option premium for
the strategy to break even. We would expect strategies
involving VIX puts to underperform those using VIX futures
in moderately upwards-sloping term structure regimes, as the
P&L gained on the put may not be enough to cover the initial
premium. We investigate selling other options to fund the VIX
put premium.
SHORT- DATED STRATEGI ES TEND TO WORK BEST
Similar to the VIX futures backtest, we backtest four VIX put
strategies to highlight various characteristics of these trades:
> Buy a 1-month ATM put
> Buy a 1-month ATM put, sell an 80% 1-month put
> Buy a 1-month ATM put, sell 2x an 80% 1-month put
> Buy a 2-month ATM put, sell 2x an 80% 2-month put
and roll monthly
For every $100 of capital, the strategy buys a single VIX put
or put spread on each roll date. The remaining cash accrues
the overnight Federal Funds rate. All option transactions
assume a bid-ask spread of $0.10.
Exhibit 30 shows the historical backtest of the strategies since
September 2006. The period until March 2009 was generally
unprofitable for long ATM put strategies, as expected – rising
volatility and inverted term structures were drags on the
performance. Since the S&P500 low in March 2009, when the
index began one of its strongest rallies in history, long ATM
24
put strategies have been profitable. This was driven both by
declining volatility and by a strongly upwards-sloping term
structure.
Selling OTM VIX puts to raise partial funding for the long
ATM put position can help to overcome the drag in
performance of long ATM put strategies. Both the 1x1 and the
1x2 put spreads showed higher returns and higher Sharpe
Ratios than the long ATM put strategy in our backtest.
Exhibit 30: The historical performance of VIX put spread strategies
Based on daily data from 20 Sep 06 through 23 Mar 11
80
90
100
110
120
130
110
3
e
p

0
ê
0
e
c

0
ê
V
a
r

0
Z
J
u
r

0
Z
3
e
p

0
Z
0
e
c

0
Z
V
a
r

0
8
J
u
r

0
8
3
e
p

0
8
0
e
c

0
8
V
a
r

0
9
J
u
r

0
9
3
e
p

0
9
0
e
c

0
9
V
a
r

1
0
J
u
r

1
0
3
e
p

1
0
0
e
c

1
0
V
a
r

1
1
Lorç 1-Vorl| ATV Puls
Lorç 1-Vorl| ATV Puls, 3e|| 1x 1-Vorl| 80º Puls
Lorç 1-Vorl| ATV Puls, 3e|| 2x 1-Vorl| 80º Puls
Lorç 2-Vorl| ATV Puls, 3e|| 2x 2-Vorl| 80º Puls
1-Vorl| Puls
1-Vorl| Pul
3pread
1-Vorl| 1x2
Pul 3pread
2-Vorl| 1x2
Pul 3pread
Relurr (Arr.) 1.5º 5.Zº ê.8º 3.1º
R|s| (Arr.) 8.5º ê.8º 5.Zº 2.êº
5º CvaR -1.3º -1.0º -0.8º -0.1º
3|arpe Ral|o 0.30 0.51 0.85 0.11
Source: Morgan Stanley Quantitative and Derivative Strategies
Extending the maturity of the put spreads has a detrimental
effect on returns and Sharpe Ratio. Since the term structure is
not as steep for longer maturities, the roll-down benefit to
performance is less strong, even in falling volatility
environments.
Weighted VIX Calendar Put Spreads
VI X FUTURES CALENDAR SPREADS THROUGH PUTS
We can introduce an additional time dimension into the put
spread strategy, similar to the VIX futures calendar spread we
considered earlier. In this trade, we purchase a short-dated
ATM or ITM VIX put while selling a longer-dated put with
the same strike. In effect, this is equivalent to a VIX futures
calendar spread, where we buy a short-dated VIX call and sell
a longer-dated VIX call to reduce exposure to spikes in the
VIX.
If the term structure remains unchanged, the long ATM put
would become increasingly ITM over time as its
corresponding VIX future rolls down the term structure. This
benefit is counterbalanced by the cost of the initial put
premium. To help fund this, an investor can sell a weighted
amount of a longer-dated put with the same strike as the long
put. The rolldown of the longer-dated put is less than the
shorter dated put, given the shape of the term structure. The
long put should increase in value faster than the short put –
this difference in decay helps to pay for the long put position.
The key risk in this strategy comes from the shape of the term
structure. Should the term structure become relatively flat or
inverted, the long put would not be ITM at maturity,
contributing to potential losses. Moreover, a breakdown in the
regression relationship between changes in different volatility
tenors would also produce a mishedge and potentially losses.
LONGER- DATED, I TM CALENDAR PUTS WORK BEST
We consider four representative sample strategies for VIX
calendar put spreads to evaluate historical benefits and risks.
In all cases, we set the short put strike to be the same as the
long put strike.
> Buy a 1-month ATM put, sell 1.4x a 2-month put
> Buy a 1-month 10% ITM put, sell 1.4x a 2-month put
> Buy a 1-month 20% ITM put, sell 1.4x a 2-month put
> Buy a 1-month 20% ITM put, sell 1.7x a 3-month put
The 1.4x and 1.7x hedging ratios in the short puts are
consistent with the ratios used for VIX futures calendar
spreads in the previous sections.
Like the other backtests, this strategy buys one VIX option for
every $100 of strategy value, with the remaining cash accruing
Overnight Fed Funds. Strategies are rolled at the monthly VIX
expirations, and a $0.10 bid-ask spread was assumed on all
trades.
Our backtests start in May 2007, when expirations up to 3
months were listed on a consistent basis. Because VIX option
liquidity is concentrated in the first three months, we did not
backtest strategies using longer-dated options.
Exhibit 31 shows the historical performance of the weighted
calendar VIX put spread strategies. The performance of ITM
put spread strategies was consistently above that of the ATM
put spread strategy. We attribute this to the lower time value
embedded in ITM puts, which reduces the net premium of the
put positions that has to be overcome by the term structure
25
slide.
13
While using ITM strikes increases potential exposure
to a volatility spike, the risk of a volatility spike is offset by
the premium gain from the short put position.
Moreover, performance strongly picked up when the longer-
dated set of 3-month options was used. This is consistent with
our findings from the VIX futures calendar spreads that
showed stronger returns when extending to longer maturities.
Exhibit 31: The historical performance of weighted calendar VIX put
spread strategies
Based on daily data from 16 May 07 through 23 Mar 11
80
90
100
110
120
130
110
150
1ê0
1Z0
V
a
y

0
Z
A
u
ç

0
Z
N
o
v

0
Z
F
e
o

0
8
V
a
y

0
8
A
u
ç

0
8
N
o
v

0
8
F
e
o

0
9
V
a
y

0
9
A
u
ç

0
9
N
o
v

0
9
F
e
o

1
0
V
a
y

1
0
A
u
ç

1
0
N
o
v

1
0
F
e
o

1
1
Lorç 1-Vorl| ATV Puls, 3|orl 1.1x 2-Vorl| Puls
Lorç 1-Vorl| 10º lTV Puls, 3|orl 1.1x 2-Vorl| Puls
Lorç 1-Vorl| 20º lTV Puls, 3|orl 1.1x 2-Vorl| Puls
Lorç 1-Vorl| 20º lTV Puls, 3|orl 1.Zx 3-Vorl| Puls
1-Vorl| / 2-Vorl|
ATV 3lraleçy
1-Vorl| / 2-Vorl|
10º lTV 3lraleçy
1-Vorl| / 2-Vorl|
20º lTV 3lraleçy
1-Vorl| / 3-Vorl|
20º lTV 3lraleçy
Relurr (Arr.) 3.Zº 5.9º ê.êº 11.Zº
R|s| (Arr.) 5.1º ê.1º Z.3º 8.5º
5º CvaR -0.8º -1.0º -1.1º -1.2º
3|arpe Ral|o 0.12 0.Z1 0.Z1 1.21
Source: Morgan Stanley Quantitative and Derivative Strategies
The strategy benefited greatly during the 4Q08 volatility
spike. During this rapid rise in volatility, both of the puts
became deep OTM puts that converged towards a zero price.
Since this strategy is short more of the longer-dated puts than
it owns in short-dated puts, this produced a net premium
benefit for the strategy.
Delta-Hedged VIX Straddles
TRADI NG THE RI CHNESS OF VI X I MPLI ED VOLATI LI TY
Much like S&P500 options, VIX options generally trade rich –
their implied volatility is typically greater than subsequent
realized volatility. Exhibit 32 shows the 1-month ATM
implied volatility of VIX options as well as the subsequent
realized volatility of a theoretical, corresponding VIX future.
14

13
An easier way to think of this is to decompose these puts into positions in VIX futures
and VIX calls. Being long a 2-month put and short a 1-month put is almost like being
long (1) a 2-month minus 1-month futures spread and long (2) a 1-month minus 2-
month calendar call spread. When the call strikes become further OTM, the price of
this call spread reduces to zero.
14
Unlike regular equity index futures, VIX futures’ realized volatility is heavily
dependent on time to maturity. Using a set of futures data, we create a strip of
theoretical VIX futures that have daily expirations. This way for each day’s ATM
Unless VIX volatility is spiking (e.g. in 4Q08 or 2Q10), VIX
implied volatility is priced at a premium to subsequent VIX
futures realized volatility. This is driven by the same structural
forces that cause SPX options to trade at a premium to realized
volatility. Investors are generally net demanders of optionality
– for example through OTM SPX puts for hedging purposes.
Market makers supply this optionality, which makes them
short volatility. The implied/realized volatility premium
compensates for the risk of this short volatility position.
Similarly, investors are likely net demanders of VIX options,
for example for VIX upside calls as tail risk hedges. Again,
the implied/realized VIX volatility premium is a compensation
for the risk to the suppliers of this volatility.
Exhibit 32: The implied – realized volatility spread in VIX options
Based on daily data from 31 Dec 07 through 23 Mar 11
-150º
-100º
-50º

50º
100º
150º
200º
0
e
c

0
Z
F
e
o

0
8
A
p
r

0
8
J
u
r

0
8
A
u
ç

0
8
0
c
l

0
8
0
e
c

0
8
F
e
o

0
9
A
p
r

0
9
J
u
r

0
9
A
u
ç

0
9
0
c
l

0
9
0
e
c

0
9
F
e
o

1
0
A
p
r

1
0
J
u
r

1
0
A
u
ç

1
0
0
c
l

1
0
0
e
c

1
0
lrp||ed - Rea||zed
1-Vorl| 3uosequerl vlX Fulures Rea||zed vo|
1-Vorl| vlX ATV lrp||ed vo|
0erera||y a prer|ur ur|ess vlX
lulures rea||zed vo| sp||es
Source: Morgan Stanley Quantitative and Derivative Strategies
STRONGER RETURNS THAN SHORT SPX STRADDLES
Exhibit 33 shows the backtested performance of a strategy that
systematically shorts 1-month ATM VIX straddles, delta-
hedged. We compare this strategy to one without a delta
hedge, as well as to a delta-hedged straddle using SPX options
instead of VIX options
The VIX strategy sells one VIX straddle for every $100 of
index value every month; the SPX strategy sells $100 of
notional straddles for every $100 of strategy value. We
assume a $0.10 bid-ask spread on each option for the VIX
straddle, and a 20bps bid-ask spread for SPX options. Trades
are executed at closing prices, and cash collateral accrues
overnight Fed Funds.
The risk profile of the delta-hedged strategies is relatively
similar – 5% CVaRs are nearly identical (-1.0% vs. -1.1%)
while the annualized standard deviations were within 63bps of

implied volatility level, there is a VIX future with exact 30 days until maturity. We
then compute the realized volatility of this future as it becomes increasingly volatile
over time.
26
each other. However, VIX strategy returns were much
stronger over the backtest period, particularly given our
aggressive bid-ask spread assumptions. The delta-hedged
VIX strategy has outperformed the SPX strategy by over
228bps per year despite its lower risk. The unhedged strategy
has performed poorly in comparison, combining near-zero
returns with higher risk.
Exhibit 33: The historical performance of delta-hedged & unhedged
short VIX straddles and delta-hedged short SPX straddles
Based on daily data from 24 Feb 06 through 23 Mar 11
50
ê0
Z0
80
90
100
110
120
130
110
150
F
e
o

0
ê
V
a
y

0
ê
A
u
ç

0
ê
N
o
v

0
ê
F
e
o

0
Z
V
a
y

0
Z
A
u
ç

0
Z
N
o
v

0
Z
F
e
o

0
8
V
a
y

0
8
A
u
ç

0
8
N
o
v

0
8
F
e
o

0
9
V
a
y

0
9
A
u
ç

0
9
N
o
v

0
9
F
e
o

1
0
V
a
y

1
0
A
u
ç

1
0
N
o
v

1
0
F
e
o

1
1
3|orl 1-Vorl| vlX ATV 3lradd|es, No 0e|la ledçe
3|orl 1-Vorl| vlX ATV 3lradd|es, 0e|la-ledçed
3|orl 1-Vorl| 3PX ATV 3lradd|es, 0e|la-ledçed
3|orl 1-Vorl| ur|edçed
vlX 3lradd|es
3|orl 1-Vorl| ledçed
vlX 3lradd|es
3|orl 1-Vorl| ledçed
3PX 3lradd|es
Relurr (Arr.) 0.1º ê.êº 1.3º
R|s| (Arr.) 1Z.5º 5.9º ê.êº
5º CvaR -2.Zº -1.0º -1.2º
3|arpe Ral|o -0.13 0.Z2 0.30
Source: Morgan Stanley Quantitative and Derivative Strategies
One potential driver for the performance differential of VIX
straddles versus SPX straddles may be a greater mismatch
between the supply and demand of VIX optionality. VIX
options were introduced to the market much more recently.
Moreover, trading the richness of VIX options requires an
active delta hedge. SPX variance swaps, in contrast, are a
common way of trading SPX volatility without the need to
manage delta actively. Since there is no equivalent product
for the VIX, the set of suppliers of VIX volatility is likely to
be much smaller, which may account for VIX implied
volatility staying rich.
Aggregating VIX Options Strategies
PERFORMANCE OF ALLOCATI NG TO ALL STRATEGI ES
In this section, we have reviewed a number of different alpha
strategies using VIX instruments. The strategies exploited the
shape of the term structure of the VIX, the skew, as well as the
implied-to-realized volatility premium. To gauge the potential
attractiveness of these alpha strategies, we take the best
performing strategy from each of the strategy types we
considered. Exhibit 34 shows the historical performance of a
strategy that allocates to each of those strategies on every VIX
expiration date. For every dollar allocated to the VIX futures
calendar spread, we allocate two dollars to the other three
strategies to make the risk contributions comparable.
Exhibit 34: The historical performance of all VIX alpha strategies and a
monthly-rebalanced strategy that allocates between them
Based on daily data from 16 May 07 through 23 Mar 11
Z5
100
125
150
1Z5
200
225
250
2Z5
V
a
y

0
Z
J
u
|

0
Z
3
e
p

0
Z
N
o
v

0
Z
J
a
r

0
8
V
a
r

0
8
V
a
y

0
8
J
u
|

0
8
3
e
p

0
8
N
o
v

0
8
J
a
r

0
9
V
a
r

0
9
V
a
y

0
9
J
u
|

0
9
3
e
p

0
9
N
o
v

0
9
J
a
r

1
0
V
a
r

1
0
V
a
y

1
0
J
u
|

1
0
3
e
p

1
0
N
o
v

1
0
3|orl 1-Vorl| vlX Fulures, Lorç 1.9x 1-Vorl| vlX Fulures
Lorç 1-Vorl| ATV Puls, 3|orl 2x 1-Vorl| 80º Puls
Lorç 1-Vorl| ATV Pul, 3|orl 1.Zx 3-Vorl| Puls
3|orl 1-Vorl| 0e|la ledçed vlX 3lradd|es
Equa||y-we|ç|led 3lraleçy
vlX Fulures
Ca|erdar 3pread
1x2 vlX Pul
3preads
vlX Ca|erdar Pul
3preads
0e|la-ledçed vlX
3lradd|es
Equa||y-we|ç|led
A||ocal|or
Relurr (Arr.) 1Z.9º 12.0º 22.3º 8.1º 15.5º
R|s| (Arr.) 12.8º 12.0º 1ê.9º 11.3º 9.0º
5º CvaR -1.9º -1.Zº -2.3º -2.5º -1.1º
3|arpe Ral|o 1.29 0.88 1.23 0.1Z 1.5ê
Source: Morgan Stanley Quantitative and Derivative Strategies
The strategies clearly show correlation with each other,
particularly during 4Q08 and during the recovery phase that
started in March 2009. Combining the strategies nevertheless
provided diversification benefits resulting in a superior risk
and return profile. The annualized standard deviation of the
combined strategy is significantly below that of individual
strategies, as is the CVaR.
27
VI. VIX Futures ETPs
In the previous two sections, we analyzed how the properties
of VIX derivatives affected the performance of systematic VIX-
based portfolio hedging and alpha extraction strategies. In
this section, we overview the emerging VIX ETP market and
how the properties of VIX derivatives also affect their
behavior. The points we cover include:
> Current ETP landscape - what is currently available?
> Impact of VIX futures rolldown on VIX ETPs
> The long-term performance of these indices going back
15+ years
> The hypothetical performance of inverse & leveraged
strategies
Overview of US-Listed VIX Futures ETPs
CURRENT LANDSCAPE & AVAI LABI LI TY OF VI X ETPS
ETFs and ETNs linked to VIX futures have grown
substantially since the first were listed in January 2009. At the
time of this publication, 14 ETPs were listed in the US, with 7
products linked to the Short-Term VIX Futures index and 6
linked to the Medium Term VIX Futures index. The
remaining one is an arbitrage strategy that trades these two
indices against each other to collect a risk premium.
Three products are also listed in Europe while two products
are linked to equivalent indices using VSTOXX futures.
Exhibit 35 provides an overview of basic characteristics of the
strategies backing these products. We omit other listed
products that contain VIX futures amongst other instruments,
and focus on pure volatility products.
The two ETPs that were listed first – the VXX and VXZ –
have the majority of AUM. As of 23 Mar 2011, all other
products have less than USD 100mm AUM. These two ETPs
track the short- and medium-term VIX indices, respectively.
While the other ETPs have yet to gain significant assets, they
have increased the variety of products, with some products
offering -1x inverse and 2x long exposure to these indices.
THE STRATEGI ES BEHI ND VI X FUTURES I NDI CES
The indices that back this new class of ETPs are based on the
returns of VIX futures rather than spot VIX. The portfolios of
VIX futures underlying these indices are designed to have a
high correlation with the spot VIX. However, they do not
track the spot VIX – as discussed earlier, spot VIX does not
equal cash plus VIX futures.
Exhibit 35: US-listed VIX futures-linked volatility indices as of 23 Mar 2011

EXCHANGE TRADED
FUNDS
EXCHANGE TRADED NOTES
VIXY VIXM VXX XXV IVO VXZ VZZ XVIX VIIX TVIX XIV VIIZ TVIZ ZIV
Issuer ProShares Barclays UBS VelocityShares
Description
VIX Short-
Term
Futures ETF
VIX Mid-
Term
Futures ETF
VIX Short-
Term
Futures ETN
Inverse VIX
Short-Term
Futures ETN
Inverse
January
2021 Short-
Term
Futures ETN
VIX Mid-
Term
Futures ETN
Long
Enhanced
VIX Mid-
Term
Futures ETN
Daily Long-
Short VIX
ETN
VIX Short-
Term
Futures ETN
Daily 2x VIX
Short-Term
Futures ETN
Daily Inverse
VIX Short-
Term ETN
VIX Mid-
Term
Futures ETN
Daily 2x VIX
Mid-Term
Futures ETN
Daily Inverse
VIX Mid-
Term ETN
Suggested
Usage
Tactical
Tactical /
Hedging
Tactical
Tactical /
Yield
Generation
Tactical /
Yield
Generation
Tactical /
Hedging
Tactical
Tactical /
Yield
Generation
Tactical Tactical
Tactical /
Yield
Generation
Tactical /
Hedging
Tactical
Tactical /
Yield
Generation
Underlying
Index
1x Long
SPVXSP
1x Long
SPVXMP
1x Long
SPVXSTR
1x Short
SPVXSP
1x Short
SPVXSP
1x Long
SPVXMTR
2x Long
SPVXMTR
1x Long
SPVXTSER
1x Long
SPVXSP
2x Long
SPVXSP
1x Short
SPVXSP
1x Long
SPVXMP
2x Long
SPVXMP
1x Short
SPVXMP
Long or Short
Vol Exposure
Long Long Long
Short
(No Reset)
Short
(No Reset)
Long
Leveraged
Long
(Daily Reset)
Long / Short Long
Leveraged
Long
(Daily Reset)
Short
(Daily Reset)
Long
Leveraged
Long
(Daily Reset)
Short
(Daily Reset)
Leverage 1x 1x 1x 1x 1x 1x 2x
2x Long / 1x
Short
1x 2x 1x 1x 2x 1x
Launch Date 3-Jan-11 3-Jan-11 29-Jan-09 19-Jul-10 14-Jan-11 29-Jan-09 30-Nov-10 1-Dec-10 29-Nov-10 29-Nov-10 29-Nov-10 29-Nov-10 29-Nov-10 29-Nov-10
AUM
(on 23 Mar 11)
25mn 7mn 1,340mn 29mn 40mn 592mn 18mn 70mn 7mn 30mn 73mn 4mn 4mn 6mn
Term Structure
Held
1-Month 5-Month 1-Month 1-Month 1-Month 5-Month 5-Month
Long 5m /
Short 1m
1-Month 1-Month 1-Month 5-Month 5-Month 5-Month
Reactivity to
Volatility
High Medium High High High Medium High Low High Very High High Medium High Medium
Rolldown
Exposure
High Medium High High High Medium High Medium High Very High High Medium High Medium
Management
Fee
0.85% 0.85% 0.89% 0.89% 0.89% 0.89% 0.89% 0.85% 0.89% 1.65% 1.35% 0.89% 1.65% 1.35%
Source: Morgan Stanley Quantitative and Derivative Strategies
28
The short-term futures index holds a portfolio of two VIX
futures that are time-weighted to deliver a constant maturity 1-
month VIX future. For example if the first two contracts were
15 and 45 days from maturity, the strategy would hold both in
approximately equal amounts. Every day, as these contracts
become shorter-dated, an amount of the front-month contract
is sold to buy the second month contract. To limit downside
risk to the amount invested, notionals are set such that if the
entire VIX futures term structure were at zero, so would be the
value of the index.
Almost the same exact strategy is used for the medium-term
VIX futures indices. The strategy holds 4- to 7-month futures
for a time-weighted exposure of approximately 5½ months.
Each day, an amount of the 4-month future is sold to buy the
7-month future.
EXPOSURE TO THE VI X FUTURES ROLLDOWN
Currently, all the strategies underlying these ETPs are backed
by VIX futures. They therefore exhibit the properties one
would expect with VIX futures, namely (1) less volatility than
the VIX and (2) rolldown losses most of the time.
While we would expect the short-dated VIX futures index to
be more volatile than the medium-term index, we also would
anticipate higher rolldown costs in the short-dated index. The
term structure is particularly steep for short-dated futures in
upward sloping term structure environments. This is the crux
of many of the VIX alpha trades in Section V.
Exhibit 36 shows the performance of the short- and medium
term volatility indices since their inception (bottom chart). We
have extended the performance in the top chart using the
regression-fitted VIX futures we constructed in Section V.
The effect of the rolldown cost is significant – the short-term
VIX futures index has lost virtually its entire value over both
sample periods. The medium-term VIX futures index, which
targets a part of the term structure that is typically less steep,
has been able to maintain its value. In very volatile periods, on
the other hand, the short-term index performed much stronger
than the medium-term index.
We therefore see different use cases for these two products.
While the medium-term index is more attractive for systematic
hedging strategies, the rapid decay of the short-term index
suggests usage for tactical volatility positioning with short
expected holding periods.
Exhibit 36: Extended short- and medium-term VIX futures index
performance (top) and actual performance (bottom)
All data through 23 Mar 11. Extended performance begins on 3 Jan 96 while actual
performance begins on 20 Dec 05
0
50
100
150
200
250
300
J
a
r

9
ê
J
a
r

9
Z
J
a
r

9
8
J
a
r

9
9
J
a
r

0
0
J
a
r

0
1
J
a
r

0
2
J
a
r

0
3
J
a
r

0
1
J
a
r

0
5
J
a
r

0
ê
J
a
r

0
Z
J
a
r

0
8
J
a
r

0
9
J
a
r

1
0
J
a
r

1
1
Exlerded 3|orl-Terr vlX Fulures lrdex
Exlerded Ved|ur-Terr vlX Fulures lrdex
0
50
100
150
200
250
300
0
e
c

0
5
V
a
r

0
ê
J
u
r

0
ê
3
e
p

0
ê
0
e
c

0
ê
V
a
r

0
Z
J
u
r

0
Z
3
e
p

0
Z
0
e
c

0
Z
V
a
r

0
8
J
u
r

0
8
3
e
p

0
8
0
e
c

0
8
V
a
r

0
9
J
u
r

0
9
3
e
p

0
9
0
e
c

0
9
V
a
r

1
0
J
u
r

1
0
3
e
p

1
0
0
e
c

1
0
V
a
r

1
1
3|orl-Terr vlX Fulures lrdex
Ved|ur-Terr vlX Fulures lrdex
Source: Morgan Stanley Quantitative and Derivative Strategies
I NVERSE & LEVERAGED VI X FUTURES I NDI CES?
The expansion of VIX futures ETPs into the inverse &
leveraged space has opened up new opportunities for short-
term volatility trading. Our recent report on inverse and
leveraged ETFs
15
highlighted that these products are best used
for very short-term, tactical views. The directional view of the
investor has to be substantial relative to the volatility of the
underlying. Otherwise, the compounding effects inherent in
the daily resetting of leverage can cause the value of these
ETFs to decay quickly, particularly for volatile underliers.
Exhibit 37: Replicated performance of daily-resetting 1x inverse and
2x long strategies on the short- and medium-term VIX futures indices
Based on daily data from 31 Dec 07 through 23 Mar 11
0
50
100
150
200
250
300
350
100
150
500
0
e
c

0
5
V
a
r

0
ê
J
u
r

0
ê
3
e
p

0
ê
0
e
c

0
ê
V
a
r

0
Z
J
u
r

0
Z
3
e
p

0
Z
0
e
c

0
Z
V
a
r

0
8
J
u
r

0
8
3
e
p

0
8
0
e
c

0
8
V
a
r

0
9
J
u
r

0
9
3
e
p

0
9
0
e
c

0
9
V
a
r

1
0
J
u
r

1
0
3
e
p

1
0
0
e
c

1
0
V
a
r

1
1
1x lrverse 3|orl-Terr vlX Fulures
1x lrverse Ved|ur-Terr vlX Fulures
2x Lorç 3|orl-Terr vlX Fulures
2x Lorç Ved|ur-Terr vlX Fulures
Source: Morgan Stanley Quantitative and Derivative Strategies

15
See ‘Vega Times: Inverse & Leveraged ETF Primer’, Morgan Stanley Quantitative
and Derivative Strategies, October 2010.
29
Exhibit 37 shows the replicated performance of -1x inverse
and 2x leveraged strategies on short- and medium-term VIX
futures indices. The indices generally perform poorly,
especially the 2x long strategy based on the short-term VIX
futures index. This is expected, as the strategy combines the
compounding effect of leveraged ETFs, the rolldown decay of
the index, and the index’s high realized volatility. On the
other hand, during 4Q08, the 2x long medium-term VIX
futures strategy would have produced significantly higher
returns due to (1) its naturally low decay and (2) increased
leverage.
30
VII. Glossary of Terms
A Reference on Commonly Used Terms
At-the-money (ATM): An ATM option is any option whose
strike price is equal to the current underlying spot price.
Basis: The basis of a future is the difference between its price
and the underlying spot price. It is typically quoted as the
futures price minus the spot price. The basis of VIX futures is
a measure of the steepness of the futures term structure.
Beta: The beta of an asset represents the sensitivity of its
expected return to the return in another asset. In this paper, we
often express the beta of volatility to the S&P500 as the as the
change in vol points for a given percentage change in the S&P
500. For example, a beta of -2 in this case would mean for a
1% fall in the S&P500, a volatility measure would be expected
to rise by 2 vol points.
Call Option: An option that gives an investor the right but
not the obligation to buy an underlying security at a set price
in the future.
Convexity: Within the volatility space, this typically refers to
the non-linear variance swap payoff for a linear change in
volatility. Compared to a (linear) VIX future, a variance swap
will outperform for a given change in volatility. Variance
swaps trade at a premium to linear vol instruments to
compensate for this more attractive payoff.
CVaR: VaR is a threshold such that returns at or above that
level occur with a certain probability. For example, a 5% VaR
equal to -2% means that we expect daily returns of less than -
2% to occur 5% of the time.
The CVaR measures the expected return given that the
threshold is crossed. It is therefore a measure of the fat-
tailedness of a return distribution. For example, a 5% CVaR
of -3% means that if the 5% VaR level is crossed, we expect a
daily return of -3%.
Delta: The option delta represents the expected change in the
value of an option for a $1 change in the underlying. For
example if an option has a delta of 0.50, it is expected to
increase or decrease in value by $0.50 for every $1 change in
the underlying. By extension, delta also represents the number
of shares of the underlying required to delta-hedge the option.
The deltas of a call range from 0 to 1 while the deltas of a put
range from -1 to 0.
Delta-Hedge: To hedge out the risk to an option’s value due
to movements in its underlying, we can use a dynamic trading
strategy known as delta-hedging. Hedge ratios are set such
that for a $1 change in the value of an option, a -1$ opposing
change in the value of the underlying is gained. For example,
if an investor was long a call option with a delta of 0.50, they
could short sell 0.50 shares of the underlying against it for a
delta hedge. For small changes in the underlying, the net
position has no exposure to price movements. However, the
hedge will have to be reset after larger changes in the
underlying to keep the delta exposure at zero – hence the need
for a dynamic trading strategy. .
Exchange Traded Product (“ETP”): An ETP is a general
term that encompasses exchange traded funds, notes, and
commodities.
In-the-money (ITM): An ITM option is one with a strike
such that if the option expired immediately, the option would
have a positive payout. In the case of puts, this means the spot
prices is lower than the strike while the opposite is true for
calls.
Mean Reversion: Mean-reversion is a phenomenon when a
certain index or underlying tends to be range bound, has
pressure to rise when it gets to a relative low, and tends to fall
when it reaches a relative high. Volatility is well known to be
mean-reverting.
Moneyness: Moneyness is a term colloquially used to express
the relative value of an option strike versus the spot. A
moneyness of 95% indicates that the strike price is 5% lower
than the current spot price.
Out-of-the-money (OTM): An OTM option is one with a
strike such that if the option expired immediately, the option
would have no payout. In the case of puts, this means the spot
prices is higher than the strike while the opposite is true for
calls.
Put Option: An option that gives an investor the right but not
the obligation to sell an underlying security at a set price in the
future.
Rolldown: With an upward-sloping term structure, the
rolldown is the expected loss (in volatility points) of a long vol
trade if the term structure remained unchanged until expiry of
31
the trade. For example if 1-month VIX futures were 3 vol
points higher than spot VIX, we would be expected to lose 3
vol points in rolldown if spot VIX were unchanged at
expiration.
Sharpe Ratio: The Sharpe Ratio is a measure of the relative
attractiveness of an investing strategy, by comparing the
annualized excess return of an asset / strategy over a risk-free
rate with the strategy’s annualized volatility. The Sharpe
Ratio is a well-defined when returns are normally distributed.
However, for heavy-tailed returns, this measure fails to fully
measure the strategy’s true downside risk. Other indicators
such as CVaR or the Sortino Ratio aim to measure this risk.
Skew: Skew measures how much higher OTM implied
volatilities are relative to an ATM implied volatility. For
regular index options, generally put implied vols are higher
than call implied vols, while the opposite is true for VIX
options. One of the most common and simple measures of
skew is the different of the implied vols of a 90% moneyness
option and a 110% moneyness option.
Straddle: An option strategy that goes long both a put and a
call. The strategy benefits if the underlying either goes up or
down, but loses if the underlying stays close to its current
value. A delta-hedged straddle can be used for pure implied
volatility exposure.
Term Structure (Futures): The futures term structure refers
to the strip of futures prices for different maturities. This
gives a market-implied view on where the price of the
underlying will be at maturity. The shape of the term structure
is often quoted as the basis.
Term Structure (Volatility): The implied vol term structure
is the strip of ATM implied volatilities for different maturities.
Typically, the S&P500 implied term structure is upward
sloping while the VIX option implied vol term structure is
inverted. The shape of the term structure is a visual indicator
or the cost of holding onto a volatility position due to
rolldown.
Variance Swap: A variance swap is a contract whose long
position receives the square of the difference between the
realized volatility over the life of the variance and a certain
strike. A forward starting variance swap is similar, except that
the realized volatility does not begin calculation until some
point later on in the future.
Variance swaps have attractive replication properties – they
can be semi-statically hedged using a strip of options. This is
unlike a volatility swap which cannot be directly replicated.
This replication property has made variance swaps the most
popular way to trade volatility globally.
Vega: The vega of an option represents the expected dollar
change in the price of an option per point change in its implied
volatility. A vega of $1 means if an option’s implied vol goes
from 15% to 16%, the option would increase in value by $1.
The vega of a variance swap represents the expected P&L
swing per 1 vol point change from the swap’s strike. For
example, a vega of $100 on a variance swap means we would
expect to gain $200 if realized volatility is 2 vol points higher
than the variance swap strike.
Volatility Swap: A long volatility swap is an OTC contract
that receives the future realized volatility of an underlying in
exchange for a pre-set strike. This product is much less
common in the US because it cannot be directly replicated,
unlike a variance swap.
© 2011 Morgan Stanley. All rights reserved.
Important Information and Qualifications
The information provided herein was prepared by sales, trading, or other non-research personnel of one of the following: Morgan Stanley & Co. Incorporated, Morgan Stanley & Co.
International PLC, Morgan Stanley MUFG Securities Co., Ltd, Morgan Stanley Capital Group Inc. and/or Morgan Stanley Asia Limited (together with their affiliates, hereinafter
“Morgan Stanley”), but is not a product of Morgan Stanley's Equity Research or Fixed Income Research Departments. This communication is a marketing communication and is not a
research report, though it may refer to a Morgan Stanley research report or the views of a Morgan Stanley research analyst. We are not commenting on the fundamentals of any
companies mentioned. Unless indicated, all views expressed herein are the views of the author’s and may differ from or conflict with those of the Morgan Stanley Equity Research or
Fixed Income Research Departments or others in the Firm. For additional information, research reports and important disclosures, see https://secure.ms.com.
Morgan Stanley is not acting as a municipal advisor and the opinions or views contained herein are not intended to be, and do not constitute, advice, including within the meaning of
Section 975 of the Dodd-Frank Wall Street Reform and Consumer Protection Act.
This material is not (and should not be construed to be) investment advice (as defined under ERISA or similar concepts under applicable law) from Morgan Stanley with respect to an
employee benefit plan or to any person acting as a Fiduciary for an employee benefit plan, or as a primary basis for any particular plan investment decision.
The information provided herein has been prepared solely for informational purposes and is not an offer to buy or sell or a solicitation of an offer to buy or sell the securities or
instruments mentioned or to participate in any particular trading strategy. These materials have been based upon information generally available to the public from sources believed to be
reliable. No representation is given with respect to their accuracy or completeness, and they may change without notice. Morgan Stanley on its own behalf and on behalf of its affiliates
disclaims any and all liability relating to these materials, including, without limitation, any express or implied representations or warranties for statements or errors contained in, or
omissions from, these materials. Morgan Stanley and others associated with it may make markets or specialize in, have or may in the future enter into principal or proprietary positions
(long or short) in and effect transactions in securities of companies or trading strategies mentioned or described herein and may also perform or seek to perform investment banking,
brokerage or other services for those companies and may enter into transactions with them. We may at any time modify or liquidate all or a portion of such positions and we are under no
obligation to contact you to disclose any such intention to modify or liquidate or any such modification or liquidation. Morgan Stanley acts as “prime broker” and lender for a number of
hedge funds. As a result, Morgan Stanley may indirectly benefit from increases in investments in hedge funds.
Unless stated otherwise, the material contained herein has not been based on a consideration of any individual client circumstances and as such should not be considered to be a personal
recommendation. We remind investors that these investments are subject to market risk and will fluctuate in value. The investments discussed or recommended in this communication
may be unsuitable for investors depending upon their specific investment objectives and financial position. Where an investment is denominated in a currency other than the investor’s
currency, changes in rates of exchange may have an adverse effect on the value, price of, or income derived from the investment. The performance data quoted represents past
performance. Past performance is not indicative of future returns. No representation or warranty is made that any returns indicated will be achieved. Certain assumptions may have been
made in this analysis which have resulted in any returns detailed herein. Transaction costs (such as commissions) are not included in the calculation of returns. Changes to the
assumptions may have a material impact on any returns detailed. Potential investors should be aware that certain legal, accounting and tax restrictions, margin requirements,
commissions and other transaction costs and changes to the assumptions set forth herein may significantly affect the economic consequences of the transactions discussed herein. The
information and analyses contained herein are not intended as tax, legal or investment advice and may not be suitable for your specific circumstances. By submitting this communication
to you, Morgan Stanley is not advising you to take any particular action based on the information, opinions or views contained herein, and acceptance of such document will be deemed
by you acceptance of these conclusions. You should consult with your own municipal, financial, accounting and legal advisors regarding the information, opinions or views contained in
this communication.
These materials may not be distributed in any jurisdiction where it is unlawful to do so. The products described in this communication may not be marketed or sold or be available for
offer or sale in a number of jurisdictions where it is unlawful to do so. This publication is disseminated in Japan by Morgan Stanley MUFG Securities Co., Ltd only to those investors
who are “Professional Investors” (tokutei toushika) as defined in the Financial Instrument Exchange Law of Japan and may not be redistributed to other types of investors without the
prior written consent of Morgan Stanley; in Hong Kong by Morgan Stanley Asia Limited; in Singapore by Morgan Stanley Asia (Singapore) Pte. and/or Morgan Stanley Asia
(Singapore) Securities Pte. Ltd., each of which is regulated by the Monetary Authority of Singapore and accepts responsibility for its contents of these materials, only to those investors
who are institutional investors, accredited investors or expert investors, as defined in the Securities and Futures Act, Chapter 289 of Singapore; in Australia by Morgan Stanley Australia
Limited A.B.N. 67 003 734 576, A.F.S.L. No. 233742 and/or Morgan Stanley Australia Securities Limited A.B.N. 55 078 652 276, A.F.S.L. No. 233741, which accept responsibility for
its contents; in Canada by Morgan Stanley Canada Limited, which has approved of, and has agreed to take responsibility for, the contents of this publication in Canada; in Spain by
Morgan Stanley, S.V., S.A., a Morgan Stanley group company, which is supervised by the Spanish Securities Markets Commission (CNMV) and states that this document has been
written and distributed in accordance with the rules of conduct applicable to financial research as established under Spanish regulations; in the United States it is directed only to
institutional clients and is distributed by Morgan Stanley & Co. Incorporated, which accepts responsibility for its contents; and in the United Kingdom it is directed only to those persons
who are eligible counterparties or professional clients and must not be acted on or relied upon by retail clients (each as defined in the UK Financial Services Authority's rules) and is
distributed in the European Union by Morgan Stanley & Co. International plc, except as provided above. Private U.K. investors should obtain the advice of their Morgan Stanley & Co.
International plc representative about the investments concerned. In Australia, this publication, and any access to it, is intended only for “wholesale clients” within the meaning of the
Australian Corporations Act. The trademarks and service marks contained herein are the property of their respective owners. Third-party data providers make no warranties or
representations of any kind relating to the accuracy, completeness, or timeliness of the data they provide and shall not have liability for any damages of any kind relating to such data.
HYPOTHETICAL PERFORMANCE RESULTS HAVE MANY INHERENT LIMITATIONS, SOME OF WHICH ARE DESCRIBED BELOW. NO REPRESENTATION IS BEING
MADE THAT ANY ACCOUNT WILL OR IS LIKELY TO ACHIEVE PROFITS OR LOSSES SIMILAR TO THOSE SHOWN. IN FACT, THERE ARE FREQUENTLY SHARP
DIFFERENCES BETWEEN HYPOTHETICAL PERFORMANCE RESULTS AND THE ACTUAL RESULTS AND THE ACTUAL RESULTS SUBSEQUENTLY ACHIEVED BY
ANY PARTICULAR TRADING PROGRAM.
ONE OF THE LIMITATIONS OF HYPOTHETICAL PERFORMANCE RESULTS IS THAT THEY ARE GENERALLY PREPARED WITH THE BENEFIT OF HINDSIGHT. IN
ADDITION, HYPOTHETICAL TRADING DOES NOT INVOLVE FINANCIAL RISK, AND NO HYPOTHETICAL TRADING RECORD CAN COMPLETELY ACCOUNT FOR
THE IMPACT OF FINANCIAL RISK IN ACTUAL TRADING. FOR EXAMPLE, THE ABILITY TO WITHSTAND LOSSES OR TO ADHERE TO A PARTICULAR TRADING
STRATEGY IN SPITE OF TRADING LOSSES ARE MATERIAL POINTS WHICH CAN ALSO ADVERSELY AFFECT ACTUAL TRADING RESULTS. THERE ARE
NUMEROUS OTHER FACTORS RELATED TO THE MARKETS IN GENERAL OR TO THE IMPLEMENTATION OF ANY SPECIFIC TRADING PROGRAM WHICH
CANNOT BE FULLY ACCOUNTED FOR IN THE PREPARATION OF HYPOTHETICAL PERFORMANCE RESULTS AND ALL OF WHICH CAN ADVERSELY AFFECT
ACTUAL TRADING RESULTS.
Any estimates, projections or predictions (including in tabular form) given in this communication are intended to be forward-looking statements. Although Morgan Stanley believes that
the expectations in such forward-looking statement are reasonable, it can give no assurance that any forward-looking statements will prove to be correct. Such estimates are subject to
actual known and unknown risks, uncertainties and other factors that could cause actual results to differ materially from those projected. These forward-looking statements speak only as
of the date of this communication. Morgan Stanley expressly disclaims any obligation or undertaking to update or revise any forward-looking statement contained herein to reflect any
change in its expectations or any change in circumstances upon which such statement is based. Prices indicated are Morgan Stanley offer prices at the close of the date indicated. Actual
transactions at these prices may not have been effected.
© 2011 Morgan Stanley. All rights reserved.
The trademarks and service marks contained herein are the property of their respective owners. Additional information on recommended securities discussed herein is available on
request. This communication or any portion hereof, may not be reprinted, resold or redistributed without the prior written consent of Morgan Stanley.
Options are not for everyone. Before engaging in the purchasing or writing of options, investors should understand the nature and extent of their rights and obligations and be aware of
the risks involved, including the risks pertaining to the business and financial condition of the issuer and the underlying stock. A secondary market may not exist for these securities. For
customers of Morgan Stanley & Co. Incorporated who are purchasing or writing exchange-traded options, your attention is called to the publication “Characteristics and Risks of
Standardized Options”. That publication, which you should have read and understood prior to investing in options, can be viewed on the Web at the following
address: http://www.optionsclearing.com/about/publications/character-risks.jsp
Clients engaging in the execution structure known as Spreading should understand that Spreading may also entail substantial commissions, because it involves at least twice the number
of contracts as a long or short position and because spreads are almost invariably closed out prior to expiration. Potential investors should be advised that the tax treatment applicable to
spread transactions should be carefully reviewed prior to entering into any transaction. Also, it should be pointed out that while the investor who engages in spread transactions may be
reducing risk, he is also reducing his profit potential. The risk/ reward ratio, hence, is an important consideration.
The risk of exercise in a spread position is the same as that in a short position. Certain investors may be able to anticipate exercise and execute a "rollover" transaction. However, should
exercise occur, it would clearly mark the end of the spread position and thereby change the risk/reward ratio. Due to early assignments of the short side of the spread, what appears to be
a limited risk spread may have more risk than initially perceived. An investor with a spread position in index options that is assigned an exercise is at risk for any adverse movement in
the current level between the time the settlement value is determined on the date when the exercise notice is filed with OCC and the time when such investor sells or exercises the long
leg of the spread. Other multiple-option strategies involving cash settled options, including combinations and straddles, present similar risk.
The trading of futures or options on futures contains inherent risks. We remind investors that these investments are subject to market risk and will fluctuate in value. The investments
discussed or recommended in this communication may be unsuitable for investors depending upon their specific investment objectives and financial position. Where an investment is
denominated in a currency other than the investor’s currency, changes in rates of exchange may have an adverse effect on the value, price of, or income derived from the investment.
Any performance data included herein represents past performance. Past performance does not guarantee future results. Investment return and principal value of an investment will
fluctuate so that an investor's shares, when redeemed, may be worth more or less than their original cost. Current performance may be lower or higher than the performance data quoted.
Please click on the following link for closing market price and net asset value (“NAV”) performance data for periods of 1, 5, and 10 years (or the life of the fund, if shorter) for funds
discussed in this communication: http://www.morganstanley.com/etf/performance/etfperformance.html.
Please be advised that Morgan Stanley’s Institutional Equity Division frequently trades as principal and may have recently traded in the securities that are the discussed in this
communication. As a result, Morgan Stanley may have a long or short position in these securities and may or may not have traded in a manner that is consistent with the above
comments prior to its dissemination.
We remind investors that these investments are subject to market risk and will fluctuate in value. Investors may obtain prospectuses for the funds described in this sales communication
from the ETF distributor. The distributor for the iShares and ProShares ETFs is SEI Investments Distribution Co. The distributor for Dow Diamonds, S&P 500, Mid Cap SPDRs SPDRs,
Nasdaq 100 Index, US Oil Fund ETF (USO), and Select Sector SPDRs is Alps Mutual Fund Services. The distributor for streetTRACKS is State Street Capital Markets, LLC. The
distributor for the SPDR Gold Trust ETF (GLD) and the KBW Regional Banking ETF is State Street Global Markets, LLC. The distributor for the Vanguard ETFs is Vanguard
Marketing Corporation. The distributor for Rydex ETFs is Rydex Distributors, Inc. The distributor for WisdomTree ETFs is ALPS Distributor, Inc. The distributor for Claymore ETFs
is Claymore Securities, Inc. The distributor for PowerShares is Invesco Aim Distributors, Inc. The distributor for ProShares is SEI Investments Distribution Co. The distributor for
iPath ETNs is Barclays Capital Inc. The distributor for Market Vectors ETFs is Van Eck Securities Corp. The distributor for Claymore ETFs is Claymore Securities Inc. The distributor
for Bear Stearns ETF is Bear Stearns & Co. Inc. The distributor for Direxion Shares and RevenueShares™ ETFs is Foreside Fund Services LLC. The distributor for ELEMENTS
SM
ETNs is Nuveen Investments/Merrill Lynch & Co. The distributor for First Trust ETFs is First Trust Portfolios L.P. The distributor for MacroShares
TM
ETFs is W.R. Hambrecht & Co.,
LLC. The distributor for RydexShares
TM
is Rydex Distributors, Inc. The distributor for UBS E-TRACS ETNs is UBS Securities LLC. The prospectuses contain more complete and
important information about the ETFs and should be read carefully before investing. We define exchange-traded funds as encompassing passively managed index-linked exchange-
traded funds, HOLDRs, and closed-end funds. This report is focused on US exchange-traded funds. These funds are registered with the SEC and trade as a single stock under SEC
exemptions. ETFs are (i) redeemable only in creation unit size aggregations and may not be individually redeemed, (ii) redeemable only through Authorized Participants, and (iii)
redeemable on an “in-kind” basis.
ETFs do not charge sales loads or other nonrecurring fees, but you may pay a brokerage commission on the purchase or sale of ETF shares. Any performance data shown in this
communication does not reflect the deduction of a brokerage commission, which would reduce the performance quoted.
Please contact your Morgan Stanley representative for the name of the name of the distributor of any funds mentioned in this sales communication that have not been included above.
Copyright © by Morgan Stanley 2011, all rights reserved.

I. A Roadmap for this Paper
Since its introduction in 1993, the VIX has moved from a seldom-used, exotic measure of equity risk to the forefront of investors’ minds as an important indicator of sentiment. The growth in the popularity of this index has led to it being rebuilt from the ground up into a new index with derivatives linked to it. Investors’ increasing concern for volatility risk and their desire to control it has led to an explosion of liquidity in VIX futures and options despite their short trading history. VIX contracts are now the most liquid way globally to trade shortterm implied volatility. Since the VIX itself is not investable, the product itself has many important intricacies that may not be obvious for the first-time user. Moreover for current users of VIX derivatives, knowing how to best leverage this product’s dynamics is crucial. We therefore provide this paper as a how-to guide on using this product for all users of VIX contracts, current and potential. Below is an outline of the content to follow:

Section III. VIX Derivatives 101
On page 8, we highlight the unique properties of VIX derivatives. Because the index is itself not tradable and has many unique properties, the pricing of VIX derivatives is very different from regular equity options. Specifically, we cover:
> How VIX derivatives account for mean-reversion

through the VIX futures term structure
> The cost of holding VIX contracts due to the shape of the

term structure
> The time sensitivity of VIX futures versus the roll cost

trade-off
> The slowness of VIX option decay versus regular options > The shape of the VIX vol surface and how VIX options

decay
> How VIX futures compare to forward variance swaps

Section IV. VIX for Equity Hedgers
We continue on page 12 with a discussion of how VIX contracts can be used as an equity hedge because of the negative correlation between volatility and equity returns. Through backtests and empirical analysis, we discuss:
> How to scale a VIX hedge for an S&P500 portfolio > How to compare the costs of VIX contracts versus

Section II. An Introduction to the VIX
Starting on page 4, we give a broad overview of the VIX and what the index truly represents. The VIX has many properties that make it unlike any other equity index. The points we discuss in this section include:
> What the VIX is and an overview of how it is priced > A history of the index and the growth of global VIX

S&P500 contracts
> Systematic performance of VIX hedges versus S&P500

look-alikes
> An analysis of the key properties of the VIX: (1) mean

hedges
> The long-term performance of VIX spread-based hedges > The potential systematic cheapness of VIX hedges > How one can use a unit mismatch between the S&P500

reversion, (2) persistent negative correlation, and (3) the convexity of its returns in a market shock
> How one gets exposure to VIX and the liquidity profile

of VIX futures and options

and VIX for downside hedges

2

Section V. VIX for Alpha-Seekers
OTC index variance swaps have long been used as a part of systematic alpha strategies. However, the events of 2008 have raised questions whether listed VIX contracts can also be used as a part of an alpha strategy On page 21, we highlight potential systematic VIX alpha strategies using the VIX term structure and the VIX option implied vol of vol. The topics we discuss include:
> How to use VIX derivatives to trade the S&P500 implied

vol term structure
> How to construct term structure trades with reduced

downside risk
> The historical richness of VIX option implied vol versus

subsequent realized vol
> The backtested performance of potential alpha trades

VI. VIX Futures ETPs
For the last major section on page 27, we review the current landscape for VIX Exchange Traded Products (“ETPs”) and how the properties of VIX futures affect their performance The points discussed are:.
> What the current landscape is for ETPs and an overview

of what is currently available
> How VIX ETPs are affected by the VIX futures rolldown > The long-term performance of VIX futures going back

15+ years
> The performance of inverse & leveraged forms of these

strategies

Section VII. Glossary of Terms
Lastly on page 30, we give a concise glossary of commonly used terms within this document and their definitions.

3

investors have increasingly tracked implied volatility as an indication of the market’s expectation of future risk.II. we often switch from one terminology to the other. implied volatilities differ across strikes and maturities. The portfolio is weighted so that when delta-hedged has a constant dollar gamma regardless of the underlying’s path. and (3) the convexity of its returns in a market shock > Getting exposure to VIX and the liquidity profile of VIX futures and options 9. Since an option is hedged by actively managing an amount of an offsetting stock position.PSOLHG 9RODWLOLW\ 9. This methodology is largely independent of rate and dividend assumptions and is easy to quote intraday since it is based on a basket of listed options. However. distilling the implied volatilities across different strikes into a single number representing an “average” level. many misconceptions exist about the index and what it truly represents.PSOLHG 9RO 0RQWK 6 3 $70 . VIX AS A BE NCHM ARK I N DI C ATO R OF E QUI TY RIS K -DQ -DQ -DQ -DQ -DQ -DQ -DQ -DQ -DQ -DQ -DQ -DQ -DQ -DQ Source: Morgan Stanley Quantitative and Derivative Strategies A variance swap is computed using a weighted sum of out-ofthe-money calls and puts. which is closely linked to implied volatility. In this paper. 2 4 -DQ . a 1-month S&P500 variance swap Data from 2 Jan 96 through 23 Mar 11 alike indices > An analysis of the key properties of the VIX: (1) mean reversion.2 Variance swap levels are computed across listed maturities intraday and are continuously interpolated to create a constant rolling measure of 30-day risk. unlike the price of a stock. which accounts for future expectations of inflation and changes in interest rates. An Introduction to the VIX The VIX has emerged as an index that is actively tracked by multiple investor types across virtually all asset classes.. The VIX is an indication of the implied volatility of the S&P500 over the subsequent 30 calendar days using the live. A variance swap strike can be thought of as the at-the-money implied volatility plus a skew premium. The success of the VIX has led the CBOE to compute other equity volatility benchmarks such as VXV (a 3-month VIX). the largest pricing input aside from the strike is the forecast of future volatility of the underlying. This is a significant advantage of the VIX since it is quoted as a variance swap strike.cboe. The popularity of this product eventually led to the launch of other VIX-like indices globally along with VIX derivatives.pdf. In this section. higher volatility requires that the hedge be rebalanced more frequently. It later modified the methodology in 2003 to make it easier to link derivatives to it.com/micro/vix/vixwhite. This incurs higher costs and in turn increases the option premium. In fact. Options are thus a forward-looking instrument. In addition. the CBOE introduced the VIX as a benchmark of implied equity risk. (2) persistent negative correlation. VXD (DJIA VIX). As option trading has become more prevalent. intraday prices of CBOE-listed S&P500 (SPX) options. please refer to the VIX white paper which can be found at http://www. $70 . This makes it difficult to quote one single number as ‘the’ level of expected future risk.. and RVX (Russell 2000 VIX) along with other VIX-like measures on 1 In 1993.1 Exhibit 1: The historical difference between 1-month S&P500 at-themoney implied volatility and the VIX. the variance swap’s strike and the implied volatility would be equal. Computing how much risk an option is implying is a non-trivial calculation that depends on other factors such as interest rates and expected dividends over the life of the option. This makes implied volatility comparable to the yield curve. accounting for expectations of future equity risk. VXN (NASDAQ 100 VIX). if a variance swap is priced using options with the same implied volatility. we aim to cover: > What the VIX is and how it is priced > A history of the index and the growth of global VIX look- The implied volatility of an option is not a directly observable level. For more details on the VIX calculation. The terms “variance swap strike” and “implied volatility” are sometimes considered verbal substitutes for each other depending on the context (Exhibit 1). A Real-Time Measure of S&P500 Risk IM PL YI N G F U TU RE EQ UI TY RIS K TH R O UG H OP TIO NS In the Black-Scholes option pricing model.

International and EM equities typically show a positive correlation with the S&P500. after September 2008. Hence.. one of the highest sustained levels of correlation in its entire history. the VIX/S&P500 correlation was -0. SMI. the universe of single stock VIX indices was on five names: AAPL. their recent correlation to the S&P500 is similar to that of International and Emerging Market Equities. due to its negative correlation and its convexity to large negative equity returns. Nikkei 225. on the other hand.3 3 0RQWK 9. correlation was approximately 0. CAC 40.40. BEL 20. Over the past year. have had a mixed correlation history with the S&P500. V FRUUHODWLRQ LV FORVH WR We use the following indices: MSCI EAFE (International Equity). Volatility expectations tend to spike after large sell-offs but gradually creep down in a rally (Exhibit 3). the VIX more than tripled from 25 to its peak at 80 as the S&P500 dropped 40% over the same period. ranging from positive in the 1990’s to a negative relationship currently. ASX 200. Barclays Aggregate (Investment Grade Fixed Income). MSCI EM (Emerging Market Equity). TOP 40. Exhibit 3: The historical relationship between 1-month S&P500 returns <HDU 5ROOLQJ &RUUHODWLRQ and 1-month VIX returns Data from 2 Jan 90 through 23 Mar 11 . 3HUFHQW &KDQJH 0RQWK 6 3 3HUFHQW &KDQJH Source: Morgan Stanley Quantitative and Derivative Strategies This makes the VIX potentially attractive as a tail risk hedge. For example. The CBOE has also begun launching VIX-like indices on select US stocks. it is likely correlation would have been similar to recent levels. Throughout its history. the VIX tends to rise. oil (OVX). GS. AEX. equities tend to become more volatile. For example. the VIX has exhibited a strong and persistent negative correlation with S&P500 returns. As of 23 Mar 11. Euro STOXX 50.QWHUQDWLRQDO (TXLW\ WR 6 3 (PHUJLQJ 0DUNHW (TXLW\ WR 6 3 . FTSE 100. Historically. and DJ UBS Commodity Index (Aggregate Commodities). there has never been a period where the correlation has been weaker than -0. and IBM. This is consistent with investor behavior – they are more anxious to purchase protection when equities are falling than they are to sell volatility when the market is rising.QFRPH WR 6 3 $JJUHJDWH &RPPRGLWLHV WR 6 3 9. DAX. The persistent negative correlation between the VIX and the S&P500 makes the VIX a potential diversifying tool for asset allocators.gold (GVXX). and the Euro (EVZ). If S&P500 options back then had been more liquid. making them popular diversifying assets... Nifty. GOOG. Mexican Bolsa. While commodities historically have had relatively little correlation to US equities. Correlation levels were less negative prior to 2001. Other assets classes. As equity markets decline. Nifty 50. WR 6 3 9.80. one of the strongest sustained levels of negative correlation in the history of the index. as well as the correlation of other global risk assets with the S&P500. Other global options exchanges have also launched similar benchmark indices on the S&P/TSX 60. as the market anticipates increased future volatility. KOSPI 200.QYHVWPHQW *UDGH )L[HG . S&P500 options were much less liquid at the time.80. Source: Morgan Stanley Quantitative and Derivative Strategies Exhibit 2 shows the rolling 1-year correlation between the VIX and the S&P500. over the past year. Exhibit 2: 1-Year rolling correlation between daily changes in the S&P500 and other asset classes Based on weekly returns from 3 Jan 77 through 23 Mar 11 5LVN DVVHW FRUUHODWLRQV DUH FORVH WR The VIX also changes asymmetrically to moves in the S&P 500. Fixed income’s correlation has varied over time. AMZN. and HSI. 5 . L AR G E SPI KES F OR SIG NIF IC AN T E Q UI TY L OSSE S Important Properties of the VIX N EG ATI VE C O RRE L ATI O N TO E Q UI TY R E TU RNS The feature that makes VIX attractive to investors is its historically negative correlation to equity returns.

While the liquidity of VIX contracts is high relative to SPX variance swaps. which would indicate no variation in the market. VIX futures were first listed in March 2004. in practice volatility resembles interest rates. which lowers the true net vega traded. Contracts cash-settle to a special opening quote of the VIX on a unique expiration cycle.5 Trading volume 4 This expiration cycle was picked so that if a person were to hold the portfolio of options comprising the VIX they would have a perfect hedge at expiration for a 30day variance swap. after the VIX reached its all-time highs in 4Q08. Exhibit 4: The 3-month median S&P500 vega traded in VIX derivatives Getting Exposure to the VIX THE VIX I S NO T A DI RECTL Y I NVES T ABL E I N DEX Unlike equity indices. This liquidity growth is likely due to (1) increased demand for tail risk hedging. OTC SPX variance swaps trade no more than USD 10mn vega notional per day. The vega of a VIX future is simply the contract multiplier (1000).. The front-loaded nature of VIX liquidity is very persistent over time. Despite these contracts being listed out to eight months. VIX futures and options are often traded in spread positions.L O AD E D W I TH FEW L O NG -M ATU R I TY TR AD E S Trading volume in VIX contracts has grown substantially since inception. the aggregate trading volume across all listed SPX options during 4Q10 was around USD 75mm notional vega per day. This occurs 30 days prior to the next SPX option expiration.UNLI KE E Q UI TIES. it eventually will shift away from that extreme towards a longer-term average.. While there is no theoretical upper bound. This makes VIX contracts potentially the most liquid way to trade short-dated implied volatility. It is also important to note that VIX options are the only liquid vehicle for optionality on volatility. To put VIX derivative liquidity in perspective. according to our traders’ estimates. which can be replicated through a basket of stocks. The majority of VIX contract liquidity is concentrated within the first few expirations (Exhibit 5). Volatility thus has mean-reverting properties. this is on a Wednesday in the third or fourth week of the month. While volatility may be relatively low or high at some point. VIX derivatives expire on the day that is exactly 30 days prior to the next months’ SPX option expiration. The VIX is a theoretical index that is re-interpolated continuously making a replicating portfolio impossible to construct. liquidity in these contracts has increased substantially – we estimate that VIX derivatives liquidity is now comparable to that of OTC SPX variance swaps. Both contracts have monthly expirations – futures are listed up to eight months out while options are listed up to six months out. we define the vega turnover of VIX options to also be its contract multiplier (100). 2SWLRQV $SU $SU $SU $SU $SU -XO -XO -XO -XO -DQ -DQ -DQ -DQ -DQ -XO Source: Morgan Stanley Quantitative and Derivative Strategies L IQ UI DI TY IS HE AV I L Y FRO N T. The usual expiration day for SPX options is a Friday. L IQ UI DI TY I N VI X CO N TR AC TS H AS G ROW N TO RIV AL SPX V ARI ANCE SW AP LIQU I DI TY Data from 3 Jan 06 through 23 Mar 11 9. This methodology is consistent with convention on how to measure a notional options turnover. the VIX is not a directly tradable index. Nevertheless. with 30% in the second-month contract. OTC options on realized variance are available but have scarce liquidity. (2) increased trading of VIX as a diversifying asset. )XWXUHV 9. the time horizon of this liquidity is different. with the market often spending extended periods in a regime before switching to the other. increased significantly during 2009. Consistent with VIX’s 30-day tenor. only 5 6 -DQ 2FW 2FW 2FW 2FW 2FW . VIX F U TURES AND O P TI ON S EX IS T AS A S U RROG ATE 0RQWK 0HGLDQ 'DLO\ 9HJD 7UDGHG 86' 0Q The reconstitution of the VIX in 2003 allowed for futures and options contracts to be developed as a surrogate for exposure to the unreplicatable VIX.4 Typically. which makes the day 30 days prior typically a Wednesday. and (3) greater tactical usage of the product. with options following in February 2006. it is bound by zero. For consistency. which tend to float between zero and an arbitrarily high number. as shown in Exhibit 5. V IX IS M E AN REVE RTI NG Since volatility is a measure of risk. We typically observe volatility drifting between ‘low’ and ‘high’ regimes. VIX futures and options each trade over 37mn expiration vega per day (Exhibit 4). Turnover in the front-month contract accounts for 42% of the total VIX futures trading volume currently. while SPX variance trades across the term structure.

We discuss this phenomenon in more detail on page 9. depending on the volatility regime. negative skew (‘tail risk’) received a greater weighting in investors’ risk assessment. Exhibit 5: The rolling 3-month average of the percent of daily turnover each VIX future contract comprises Data from 23 Mar 07 through 23 Mar 11 VW 0RQWK 9.. we might well see further increases in the liquidity of VIX-related products. However. 6 unexpected data points observed during 2008. as “tail risk hedging” became part of investing nomenclature. The listed nature and intraday liquidity of VIX contracts has made them attractive to investors as a volatility hedge against these types of events. VIX options share the same pattern. Therefore it is difficult to infer from the data alone how likely those tail events really are. tend to become less sensitive to moves in the VIX the longer-dated they are. While these strategies are not meant to track the VIX. 7 . Exhibit 6: The perception of tail risk is likely higher than it was before 2008 after going through the credit crisis 3HUFHSWLRQ RI 5LVN 3UH &ULVLV 3HUFHSWLRQ RI 5LVN 3RVW &ULVLV Starting in January 2009. The rise of VIX futures-linked ETPs can be seen as further evidence of the broadening of the participant base. who used these contracts as short-term positioning tools. Exchange Traded Products (“ETPs”) have been launched as easy-to-access turnkey solutions to give exposure to strategies based on VIX futures. After 2008. Since VIX options are priced off VIX futures. E TPS B AS E D O N VIX FUTURES AL S O E XIS T AS A M E THO D F O R VI X EXP OS UR E 0DU 0DU 0DU 0DU 0DU Exhibit 6 shows the shift in investor risk perception pre. assets in these products are now well over $2bn. This makes longer-dated contracts less appealing to hold relative to short-dated contracts. LI Q UI D V OL V EH IC LE QG 0RQWK 9. )XWXUHV UG 0RQWK 9. a spike in the VIX does not necessarily lead to a large P&L swing for longdated VIX derivatives. . L IQ UI DI TY IS LI KEL Y TO CO N TI N UE TO GR OW AS I NVES TO RS LO O K TO OW N V OL ATI L I TY 1RY 1RY 1RY 1RY -DQ -DQ -DQ 6HS 6HS 6HS 0D\ 0D\ 0D\ 0D\ 6HS -DQ -XO -XO -XO -XO Source: Morgan Stanley Quantitative and Derivative Strategies The sensitivity of VIX futures to movements in the SPX is the likely cause for this front-loaded liquidity. they can deliver a high correlation to the VIX and exhibit many properties that are similar to those of VIX futures. We highlight the current universe of these products and the factors affecting their performances on page 27.QRZLQJ WKH LPSDFW RI WDLO HYHQWV KDV LQFUHDVHG LQYHVWRU DZDUHQHVV DQG SHUFHSWLRQ RI WKHLU OLNHOLKRRG SRVW FULVLV 0RQWK 5HWXUQ ZLWK 9RODWLOLW\ Source: Morgan Stanley Quantitative and Derivative Strategies 6 This phenomenon is studied in a field of statistics known as ‘extreme value theory. spread across 13 different products as of 23 Mar 11. some of these products exhibit significant negative carry.16% of the liquidity is in contracts beyond the first three months. Movements in the underlying cash index affect equity index futures similarly across the maturities. VIX futures. with 47% and 29% of daily turnover occurring in the first two maturities. The potential benefit of integrating volatility was shown during the credit crisis. As investors further move to integrate volatility exposure as a tail-risk hedge and for diversification. )XWXUHV WK 0RQWK 9. )XWXUHV Prior to 4Q08.. institutional and even retail market participants have increasingly used these contracts. when the VIX rose from 23 the week before Lehman to a peak of 80. VIX futures and options trading interest was mainly concentrated in the hedge fund community.and post the credit crisis. In total.’ A probability distribution may have a significant weighting towards a tail but a limited statistical sample of that distribution is likely to not have a realization from that tail. Since then. particularly when used as a tactical way to execute a volatility view. )XWXUHV 0RQWK $YHUDJH 3HUFHQW RI 'DLO\ 7XUQRYHU The Major Players and Positioning in VIX VIX DERIV ATI VE S AR E BEING I NCRE AS I NGL Y US ED B Y I NS TI TU TIO NS AS A L IS TED . The increased skew of longer dated SPX options as well as of VIX call options reflects investors’ response to the new... on the other hand. This shift in risk assessment can potentially explain the greater use of volatility products such as the VIX.

. the 6-month VIX future was over 45 vol points lower than the spot VIX. )XWXUH WK WK WK decay > How VIX futures compare to forward variance swaps Source: Morgan Stanley Quantitative and Derivative Strategies The Unique Properties of VIX Contracts C AS H + VIX FUTURES DO NO T M AKE VI X This changing shape of the term structure is necessary for a mean-reverting underlying. Shorter-dated VIX contracts are more sensitive to spot VIX. Should VIX spike but expectations are that the VIX will revert in a month’s time. VW QG > The shape of the VIX vol surface and how VIX options UG 9. VIX Derivatives 101 Derivatives on the VIX have many unique properties. VIX futures do not have this arbitrage-based pricing. At these relative extremes. Instead. because there is less time for volatility to mean-revert. It is also responsible for many of the other unique features of VIX contracts. because the VIX itself is not tradable. Exhibit 7: The VIX future term structure at relative VIX highs and lows 6HFRQG /RZHVW 9. This is because VIX futures are priced off the S&P500 implied volatility term structure. the VIX futures term structure priced in an expectation of mean reversion. which is usually upwards-sloping. accounting for the probability of such a spike. The longer the maturity of the contract. we provide an overview of the mechanics of VIX futures and options: > VIX derivatives account for volatility mean-reversion Exhibit 7 plots the VIX futures term structure on two extreme dates – the day of the VIX’s all-time high in 2008. Otherwise. This property makes VIX futures an imperfect surrogate to trade the VIX. more so than they resemble equity derivatives. Shocks to spot VIX should thus have less of an impact. while at the relative high in 2008.. +LJK 1RY -DQ OHIW D[LV through the VIX futures term structure > The shape of the volatility term structure and its impact ULJKW D[LV on the cost of holding VIX contracts > The time sensitivity of VIX futures versus roll cost trade- off > The slowness of VIX option decay versus that of regular options 9. as the VIX is not a tradable index. Six-month VIX futures generally move less Just like volatility.III. The slope reflects a volatility premium – since the VIX and implied volatility tend to spike in times of market stress. This replication arises because the underlying index can be traded using a managed basket of equities. At the relative low in 2007. the term structure of the VIX futures is upwards sloping. and a day in January 2007 when the VIX reached its second-lowest level in history. VIX F U TU RES ACCOUNT FO R M E AN REVE RSI ON The mean reversion priced into the VIX futures term structure means that different futures maturities show varying responsiveness to spot VIX. VIX F U TU RES H AV E A TIM E DEP EN DE N T BE T A TO V IX The arbitrage-based pricing of forward / futures contracts means that an investment in an equity index can be replicated through a combination of cash and futures on that index. Most of the time. one could devise a statistical arbitrage trading strategy that buys the index at its relative lows and shorts it when it is at its relative highs. which have a similarly intangible underlying. VIX futures should trade at a premium to spot VIX.. We group all historically traded VIX futures by their time to maturity and compute their betas to daily vol point changes in VIX. These contracts resemble interest rate derivatives.. In this section. Exhibit 8 plots the sensitivity of VIX futures to changes in the VIX. the 6-month VIX future was over four vol points higher than the spot VIX. VIX futures account for this property and price in expectations of mean reversion. the higher is the probability of mean reversion. VIX futures prices reflect expectations on the VIX level at expiration of the futures contract. the VIX is a mean-reverting index. 'D\ $OO 7LPH 9. 8 . a 1-month VIX futures price should remain unchanged.

. TH E SH O R T-E N D IS THE CO S TL IES T POI N T TO B U Y The shape of the VIX futures term structure gives insight into the holding costs of going long VIX futures. This produces a trade-off between trading the desired volatility of VIX in the front-end with higher costs associated with that position. Exhibit 9 plots the median historical term structure of VIX futures.8 to spot VIX. For example. This gives a measure of the cost of holding a VIX future per unit of exposure it gives to VIX. This marginal contribution flattens out quickly with 6-month futures adding only 0. Exhibit 10: The marginal contribution to the VIX futures term structure versus the beta of that future to spot VIX Based on daily VIX futures data from 26 Mar 04 through 23 Mar 11 0RQWK )XWXUHV $YHUDJH 0RQWK 7HUP 6WUXFWXUH 5ROO &RVW 9RO 3RLQWV 0RQWK IXWXUHV FRVW OHVV WKDQ 0RQWK 9. particularly in a low-volatility regime. as well as the marginal contribution of each futures maturity to the overall term structure. creating potential alpha opportunities. while 4 to 6-month future contracts scale much better. the lower cost per unit of spot VIX exposure of longer-dated futures has to be counterbalanced with the considerably lower liquidity in these contracts. )XWXUH 7HQRU 0 0 0 Source: Morgan Stanley Quantitative and Derivative Strategies 'D\V 8QWLO 9. )XWXUH WR 6SRW 9. Since VIX futures typically trade at a premium to spot VIX.. For example. 0RQWK )XWXUHV 0RQWK )XWXUHV WR 0RQWK )XWXUHV %HWD RI 9. This sensitivity remains low for most of the life of a VIX future until just prior to expiration. if VIX futures are priced two vol points above spot VIX. then one would expect to pay two vol points at maturity if VIX is unchanged. The tradeoff between holding costs and spikes has historically been asymmetric.6 and 0. long VIX futures holders will lose this premium most of the time. We highlight this tradeoff in more detail in Exhibit 10 where we plot these marginal contributions to the term structure versus the beta of that VIX future to spot VIX (data from Exhibit 8 and Exhibit 9). These liquidity considerations have to be taken into account when developing 9 . However. relative to the spot VIX. an exposure to a 4-month future that is scaled to give the same exposure to the spot VIX as a 1-month future would have only 40% of the holding cost of the 1-month future on average. Exhibit 8: The beta of VIX futures to changes in spot VIX as a function of time to maturity Based on daily VIX futures data from 26 Mar 04 through 23 Mar 11 Exhibit 9: The median VIX futures term structure and the steepness between each point on average Based on daily VIX futures data from 26 Mar 04 through 23 Mar 11 $YHUDJH 0DUJLQDO &RQWULEXWLRQ WR WKH 7HUP 6WUXFWXUH ULJKW 0HGLDQ +LVWRULFDO 9DOXH OHIW 9. They are compensated for this by substantial gains when volatility spikes.. %HWD WR 9.. We discuss strategies around this tradeoff starting on page 21.0 vol points higher than spot VIX while 2-month futures are 0. The median level of 1-month VIX futures is 1.68 vol points above the 1-month VIX futures.than 20% of the amount VIX does on any given day.. \ 5 /Q [ 0 0 0 9. This allows us to assess the incremental cost of holding VIX futures of longer maturity. Source: Morgan Stanley Quantitative and Derivative Strategies 1-month futures have the greatest cost per exposure ratio by far..1 vol points to the term structure.. VIX futures under 20 days to maturity typically have a beta between 0. IXWXUHV This causes front-month futures to be the most volatile while the back-month futures have only a fraction of the volatility of the VIX. )XWXUH ([SLUDWLRQ Source: Morgan Stanley Quantitative and Derivative Strategies This implies that the rolldown cost for VIX futures is greatest in the front-end while flattening out quickly beyond the 3month point.

~ ~ ª ln( F / K ) + 0. the more ATM implied vol changes per 1% change in the S&P500. The steeper the skew. Shorter-dated VIX implied volatility also reacts more to spikes in the VIX. This has no independent implied volatility and time to maturity factors.. the atthe-money strike of VIX options is determined with reference to the VIX future level for a given maturity.. Assume that the SPX fixedstrike implied vol surface is unchanged while the SPX falls. Ever since 2009. Since the VIX option implied volatility term structure is generally inverted.. rather than to the cash VIX level. This makes sense since VIX implied volatility generally rises when the VIX does (Exhibit 11). are often relatively attractive. Given mean-reverting properties of the VIX.5σ 2 º Call = FΦ « − KΦ « » » ~ ~ σ σ ¬ ¼ ¬ ¼ where F denotes the futures price. and hence a lower implied volatility. In general.and short-dated volatility shifts significantly when volatility spikes since the VIX and its 7 implied volatility both tend to rise during times of distress. However. As a consequence.PSOLHG 9RO OKV -XO -XO 0D\ 6HS 6HS -XO 6HS -DQ 0DU 0DU -DQ 0DU 1RY 1RY 1RY 0D\ 0D\ Source: Morgan Stanley Quantitative and Derivative Strategies VIX OP TIO NS DEC AY S L OW LY UNTIL EX PI RATI O N Option prices are largely functions of two inputs: (1) the moneyness of the option and (2) its non-annualized implied volatility ( ¥t). Exhibit 12: The implied vol and 130%-100% 1-month VIX option skew Based on daily data from 23 Mar 08 through 23 Mar 11 0RQWK 6NHZ 6SUHDG UKV 0RQWK 9. a longer tenor averages in less volatile periods for that VIX future.PSOLHG 9RO OKV 0RQWK $70 9. 10 0D\ 1RY 0DU 0DU . we find that VIX options implied volatility has a permanently inverted term structure.5σ 2 º ª ln( F / K ) − 0. this reflects the at-the-money level for each maturity. the spread between long. TH E VIX OP TION IM PLIE D V OL TE RM S TRUCTURE IS PE RSI S TE NTL Y I NVE RTED Source: Morgan Stanley Quantitative and Derivative Strategies The realized volatility of VIX futures declines the longerdated they are. The skew of SPX options is a measure the implied volatility of volatility. costs of out-of-the-money VIX calls are optically high relative to spot VIX. We highlight how to leverage this phenomenon on page 18. the higher the latter component is. while VIX puts appear cheap. particularly in a high-skew regime.. K the strike. the skew in the VIX implied vol surface is tilted towards calls instead of puts (Exhibit 12).PSOLHG 9RO OKV 0RQWK $70 9. Strategies that leverage this. 0DU 0D\ -XO -XO 6HS 6HS -XO 6HS -DQ -DQ 0DU 0DU -DQ -DQ 1RY 1RY 0D\ Because the VIX futures term structure is generally upwards sloping. the Black 1976 model for a VIX call option price reduces to: 7 VIX implied vol actually has a strong relationship to S&P500 skew. such as OTM call spreads. ~ If we replace the factor σ t with a single total implied volatility. a steep VIX call skew further helps to elevate the costs of VIX calls.PSOLHG 9RO OKV 9. . .strategies that leverage this systematic bias for hedging or alpha. 8 the higher the option price. VIX options tend to decay slowly until 8 Unlike S&P500 options. Exhibit 11 plots the rolling time series of ATM 1-month and 5-month VIX implied volatilities. . . Similar to the S&P500 option term structure. Since implied volatility represents the expected future realized volatility. The implied volatility of a short-dated VIX option anticipates volatility over the near term. The new ATM implied vol would have to roll up the SPX put skew to a higher level. VIX option skew has been steepening and is currently near an all-time high.. UKV VIX futures represent forward-looking expectations of future VIX levels at each maturity date. This is in contrast to the cash equities options market. which may itself be more volatile. and denotes the cumulative density of the standard normal distribution. Long-dated implied volatilities are relatively stable and significantly lower than 1-month implied volatility. as well as their spread and the VIX itself. VIX OP TIO N SKEW IS ALW AYS TI L TE D TO C AL L S Along with a generally upward sloping VIX futures term structure. which forms a more appropriate base line. σ and zero interest rates. TH E ATM VI X V OL S TR IK E IS B AS E D O N THE F U TU RES Exhibit 11: The implied vol and vol term structure of VIX options Based on daily data from 23 Mar 08 through 23 Mar 11 0RQWK 0LQXV 0RQWK 6SUHDG OKV 0RQWK $70 9.

opening up more avenues to express volatility views with precision. unlike VIX futures which have no semi-static hedge.. While the time to maturity t shrinks. We show this in Exhibit 13 where we plot the average shape of the VIX implied volatility term structure using 1.PSOLHG 9RO 7HUP 6WUXFWXUH ULJKW Variance swaps are the preferred vehicle to trade volatility globally. Exhibit 13: The average costs of ATM VIX call options versus the average shape of the VIX implied vol term structure Based on daily data from 24 Feb 06 through 23 Mar 11 volatility. VIX contracts are by far the most liquid vehicles. )XWXUH 6WULNH 3 / RI D [ )RUZDUG 9DULDQFH 6ZDS 6WULNH 7KH FRQYH[LW\ LV SDLG IRU E\ D KLJKHU VWULNH 3 / SHU RI 1RWLRQDO 9HJD 7KH YDULDQFH VZDS KDV D FRQYH[LW\ EHQHILW IRU ODUJH FKDQJHV LQ YRODWLOLW\ Source: Morgan Stanley Quantitative and Derivative Strategies VIX IS LIQ UI D FO R O NL Y NE AR . This is in contrast to OTC index variance swaps. 9 It can be shown that a variance swap can be replicated using a static portfolio of delta-hedged OTM options. the implied volatility rolls up the term structure. We assume zero interest rates. in contrast. Their liquidity is largely concentrated in the first three months.they remain the preferred vehicle to trade volatility in most markets. the expiration P&L change will be a fixed multiplier times the strike change (see Exhibit 14). . The long variance swap gives a convexity benefit for large changes in volatility.7 vol points for 1-month maturities.. These offsetting factors tend to slow the decay in option prices until a few weeks before expiration.. are standardized instruments. For every vol point change in the strike of a VIX future.to 5month ATM implied vols. they tend to be relatively illiquid with wide bid/offer spreads. For example. VIX IS THE O N L Y LI Q UI D VE HI CLE F OR VO L O P TI O NS 7LPH WR 0DWXULW\ 'D\V Source: Morgan Stanley Quantitative and Derivative Strategies VIX Contracts versus Variance Contracts COM P AR I N G THE P AYO F F O F VI X VE RS US V ARI AN C E Futures based on VIX deliver a linear payoff with respect to changes in VIX. option prices decay more slowly than in the case of a flat term structure. )XWXUH OHIW RI 9. the strike difference has averaged around 0. For forwardstarting variance swaps. This creates a convex payoff with respect to changes in volatility.close to expiration. Moreover. )XWXUH OHIW $YHUDJH &DOO 3ULFH +ROGLQJ 0 9RO &RQVWDQW $YHUDJH 9. while variance is more liquid for longer-dated views. primarily because of the ease of creation and hedging.TE RM M ATU R I TIES $YHUDJH $70 9. Since 2004. &DOO 3ULFH RI 9. the price of ATM calls drops on average from 14% for 5-month tenors to 10% for 1-month maturities. the square of VIX options are the most liquid way to obtain optionality on volatility. while variance swaps are traded across the term structure. Exhibit 14: The payoffs of a hypothetical 1-month VIX future and a hypothetical 1x1 forward starting variance swap in one month’s time 3 / RI D 0RQWK 9. This has made variance swaps relatively ease to create . 11 .9 We compare these payoffs in Exhibit 14 using a hypothetical 1-month VIX future and a 1-month variance swap that forward starts in one month (1x1 forward variance). While options on variance swaps exist in the OTC market. this implies a rapid decay just before expiration. prices would decay to around 6% for 1-month maturity. tenors and forward starting periods can be fully customized in this OTC product. Since calls that are not in the money have to converge to zero at maturity. the payoff of these products is based on variance. the payoff at the end of the forward period is quite similar to that of a VIX future. We also plot option prices assuming a flat implied volatility term structure (at the level of 5-month implied volatility). Because of the roll-up effect. depending on the market.. Maturities can exist out to ten years.. This benefit comes at the cost of a higher strike than a VIX future with the same maturity. VIX products. For short-dated volatility views. However. We use these levels to price ATM VIX calls over time as they roll down the term structure. Under a flat term structure.

shown in Exhibit 15. We then compute how many vol points each contract changed after one month – the spread of 1-month futures versus spot VIX one month later. In this section. we need more of them to overcome their lower sensitivity to VIX changes. it allows us a more 10 direct way to compute the required hedge ratios. and so on.89 2-Month VIX Contracts per $100 of S&P500 > 1.B AS E D HE DGES We then regress the 1-month S&P 500 returns on these changes in VIX futures. While this is the inverse of the typical volatility regression.63 1-Month VIX Contracts per $100 of S&P500 > 0. Based on our sample. For the entire history of VIX futures trading. )XWXUHV hedges > Long-term performance of VIX spread-based hedges > Potential systematic cheapness of VIX hedges > Exploiting the unit mismatch between the S&P500 and 0RQWK &KDQJH LQ 6 3 \ 5 [ \ 5 [ \ 5 [ 0RQWK &KDQJH LQ 9. 10 When the regression is performed in the conventional way where the change in volatility is a function of the market return. the sensitivity of a volatility-based hedge to negative returns in equities will allow us to determine optimal scaling ratios. since the performance of the equity portfolio and that of the hedge are not directly linked. Since longer-dated VIX futures are not as volatile as short-dated futures. which give us the change in the S&P 500 for a given change in the VIX futures. in contrast. Scaling a VIX trade is somewhat more complex than using regular index options. often use three to six month maturities as these tend to give the most favorable cost/benefit tradeoff. Hedges using equity index options. (2) the sensitivity of VIX futures to changes in the spot VIX. We focus on hedges using short-dated VIX contracts only. we use a historical regression model. This is driven by three factors: (1) the liquidity profile of VIX contracts. and introduce a number of backtests of the benefits of VIX-based hedging. a low beta could be computed and then inverted. In this section. we frame the set of potential strategies. 12 . we cover: > Scaling a VIX hedge into an S&P500 portfolio > Costs of VIX contracts versus S&P500 contracts > Systematic performance of VIX hedges versus S&P500 S C AL I N G A VIX POS I TI O N USI NG REG RESS IO N To compute the notional size required to scale a VIX hedge into an equity portfolio. )XWXUHV 0RQWK 9..IV. The opposite regression form mitigates this. the beta has to be inverted to compute the hedge ratio. and (3) the favorable decay characteristics. )XWXUHV 0RQWK 9. F O CUS O N SHO R T-D ATE D VIX . that of 2-month futures versus 1-month futures one month later. we compute a dataset of daily rolling 1-month. )XWXUH Source: Morgan Stanley Quantitative and Derivative Strategies VIX for downside hedges Integrating VIX into Long-Only Portfolios H E DGI N G AN E QU I TY P O RTF O LIO U SI NG VOL ATI LI TY Because of the inherent negative correlation between equity returns and volatility.07 3-Month VIX Contracts per $100 of S&P500 As expected... VIX for Equity Hedgers One of the key benefits of integrating volatility exposure in an equity portfolio is the inherent diversification benefit of the negative correlation between equity returns and volatility. which began in March 2004. We assume an S&P500 portfolio for the subsequent analysis in this section. 2month. creating a high hedge ratio and overhedging of the portfolio. Under imperfect correlation. and 3-month VIX futures prices. Exploiting this ‘tail risk hedge’ benefit requires the specification of an appropriate strategy – how and when volatility should be scaled into a portfolio. Based on our earlier analysis.. the hedge ratio increases for longer-dated contracts. there are potential diversification benefits from integrating volatility into an equity portfolio. Exhibit 15: The responsiveness of the S&P500 as a function of changes in the first three VIX futures Based on daily data from 26 Mar 04 through 23 Mar 11 3HUFHQW 0RQWK 9. we derive the following hedge ratios: > 0.

VIX futures will capture the spike in volatility that typically coincides with an equity selloff. when the three months maturities were first listed regularly.. 6 3 )XWXUHV 0 9. ATM VIX call. )XWXUHV 4 4 4 4 4 4 4 6 3 6 3 )HE 0 9.and 3-month maturities. 6 3 )XWXUHV 0 9. 6 3 )XWXUHV )HE )HE 6 3 6 3 4 4 4 4 4 4 4 4 0 9. The futures are rolled monthly at expiration. 2. )XWXUHV 1RY 1RY $XJ $XJ 0D\ 0D\ 0D\ $XJ 1RY 0 9.. To illustrate the potential merits of VIX-based hedges. )XWXUHV 0 9.. O VE RVIEW OF VI X F U TURES HE DGI NG S TR ATE G IES hedge > Zero up-front cost strategy. We base our analysis on listed prices.. Exhibit 16 shows the historical performance of these overlays.VIX Futures as a Portfolio Hedge S YS TEM ATI C B AC K TE S TS O F COM M ON S TR ATE G IES Pros: > An intuitive and direct way to go long volatility as a We backtest a number of systematic hedging strategies using VIX contracts with rolling 1-. we consider sample backtests of various VIX futures. Exhibit 16: The historical performance of systematic VIX futures overlay strategies 4XDUWHUO\ 5HWXUQ 6WUDWHJ\ 9DOXH 6 6 6 6 $XJ 0D\ 3 3 3 3 1RY 0 9. and add a $0.. We can summarize these effects as follows: Based on daily data from 16 May 07 through 23 Mar 11 We backtest a strategy that combines a long S&P500 position with long VIX futures overlays. using 1 to 3 month maturities.. Conversely. )XWXUHV Source: Morgan Stanley Quantitative and Derivative Strategies 13 .10 bid/ask spread to account for the relative lack of liquidity in VIX contracts early in their history. )XWXUHV 0 9. 6 3 )XWXUHV 0 9. Our backtests begin in May 2007.. since these are futures contracts.. unlike long option strategies (excluding margin requirements) Cons: > Full upside and downside exposure to volatility changes > Long VIX futures is generally a negative carry strategy due to the upward-sloping shape of the term structure B AC K TE S TE D S TR ATE G Y PE RF O RM AN C E Long VIX futures strategies are the most direct way of hedging equity risk with VIX contracts. We use the hedge ratios derived in the previous section. performance would be negative if volatility were to fall sharply. Endless configurations of VIX hedges are possible. and OTM VIX call hedging strategies..

as expected. unlike futures which have no up-front premium outside of margin > The mark-to-market of VIX options may not be equal to changes in the spot VIX prior to expiration B AC K TE S TE D S TR ATE G Y PE RF O RM AN C E USI NG AT M VIX OP TIO NS Exhibit 17 shows the annualized return and risk characteristics of the different hedging strategies across the entire backtesting period. the S&P500 returned 13% on average. During 4Q08. 2Q10. RE TURN AND RIS K P ROFIL E OF F U TU RES HEDGES Exhibit 17: The return and risk profile of VIX futures hedges Based on daily data from 16 May 07 through 23 Mar 11 Return (Ann. In addition. Overall. Compared to VIX futures. and 4% per quarter.4% -4. and 3month futures overlays reduced the average returns to 3%. which hedge against both increases and decreases in volatility. 2 and 3 months to expiry. call options give similar upside volatility exposure but reduced downside exposure. whether measured by the standard deviation or by the 5% CVaR.6% -2.1% 17.9% 17. 1Q09 and 1Q08). Meanwhile. with an average quarterly outperformance of 2. with 1.23 0. VIX futures can introduce a drag on portfolio performance.0% for 2-month. compared to strategies using shorter-dated futures. respectively. The most obvious difference between the three overlays is in their long-term performance.7% -0. We start by backtesting strategies using ATM VIX call overlays. During this time.1%. Dev.4% -2. 3Q10 and 4Q10). on the other hand. such a strategy can be attractive: Pros: > Reduced downside risk in falling vol environments > Calls can reduce potential rolldown costs. this difference leads to more efficient hedge. 14 . The 3-month futures overlay strategy was the only one with a positive return over this period. As a result. is particularly beneficial for strategies using longerdated futures maturities.9% 27.03 Source: Morgan Stanley Quantitative and Derivative Strategies VIX Calls as a Portfolio Hedge O VE RVIEW OF VI X CAL L HE DG ING S TR ATE G IES VIX call options are potentially another approach to hedge downside risk in equity positions. Risk reduction was similar across all overlay strategies. Through holding higher notional exposure of 3month futures. the 3-month futures strategy outperformed the 1month strategy in 12 of 14 full quarters in our sample. the strongest S&P500 quarter (2Q09.33 -0. Exhibit 18 shows the historical performance of these overlays.All three variations of the strategies worked as effective downside hedges.5% 17. Depending on the VIX implied volatility regime (the ‘vol of vol’). 3Q09. (Ann. 2%. +16%) caused the futures overlay strategies to underperform. This is consistent with our earlier finding that per unit of spot VIX exposure. such as in the period after 4Q09.1% for 1-month. Adding a VIX futures overlay improved the average returns during these quarters to -6.6% for 3-month futures – this again demonstrates the potential benefit of using longer-dated futures. when the S&P500 dropped 22%.12 -0. Compared to VIX futures overlays.) Std. with the portfolios with overlays delivering a gain between 1% to 2%.4% 1. During the four worst S&P500 quarters (4Q08. the 3-month VIX futures strategy minimized the carry cost of the position.8% -4.3% -2. 1-.) 5% CVaR Sharpe Ratio S&P500 + S&P500 + S&P500 + 1M VIX 2M VIX 3M VIX S&P500 Future Future Future -1. the performance of the S&P500 portfolios with futures overlays ranged from a 3% loss to a 6% gain. During the four best full S&P500 quarters (2Q09. During bull markets. The strategy utilizing 3-month futures significantly outperforms the strategy using 1-month futures. 2-. -2. the payoff profile of an S&P500 position with VIX options overlays is much closer to that of the S&P500 hedged with protective puts. shorter-dated VIX futures were more expensive. A steep VIX futures term structure.6% -2. the one-sidedness of VIX options means that they only reduce risk in a single direction. depending on the VIX implied vol regime Cons: > Up-front premium payment for options exposure. the S&P500 lost -13% on average. and -1.

in contrast.4% -3.7% 15. when the S&P500 lost 22%.5% 12.1% -3. RE TURN AND RIS K P ROFIL E OF ATM C AL L HE DG ES Exhibit 19 shows the annualized return and risk characteristics of the ATM VIX call overlays.08 Source: Morgan Stanley Quantitative and Derivative Strategies 15 .or 2-month overlays are generally more attractive than for VIX futures overlays.6% 3.3% -4.4% -11. while the S&P500 gained 13%.0% 6.4% -3. while the three strategies with the VIX options overlay lost between -6% and -8%.4% -12. (Ann.6% 8.2% -21.6% -9.7% 11.5% 3. During the best S&P 500 quarter in 2Q09.8% -0.7% -6.9% 3.2% -2.7% -10.3% 0. &DOOV 0 $70 9.3% 2.0% 3.5% -4.7% -10.6% 4. On the other hand. these hedging strategies were either flat or slightly down (-4% to +1% return). in part due to the call premiums paid. the S&P500 lost on average -13%.9% 10.4% 1.9% -10. this slight outperformance comes at the cost of significantly higher premium to purchase these calls due to the (1) longer-dated nature of the options and (2) a higher hedging ratio.11 -0. This shows that scaling between the different maturities can give very similar volatility exposure.4% 27.6% -7. this represents greater upside participation during S&P500 rallies.8% 9.0% 4.3% -10.2% -1. the average returns of the 1.8% -4.2% Source: Morgan Stanley Quantitative and Derivative Strategies For example. Risk reduction was generally smaller than in the case of VIX futures overlays.4% 5.1% -3.5% -3. However.. while the hedged strategies returned between 9% and 10%.7% -5.4% -1.0% 2.9% 21. During the four best quarters for the S&P500.. the three overlay strategies gained between 7% and 8%.) 5% CVaR Sharpe Ratio S&P500 + S&P500 + S&P500 + 1M ATM 2M ATM 3M ATM S&P500 VIX Calls VIX Calls VIX Calls -1.4% 2.9% -0. the 2month strategy outperformed the 1-month strategy by 27bps a quarter while the 3-month strategy outperformed by 41bps a quarter.4% 1.6% -4.3% -0.0% 5.2% -8.2% 4. Exhibit 19: The return and risk profile of ATM VIX call hedges Based on daily data from 16 May 07 through 23 Mar 11 Return (Ann.9% 8. In the most extreme quarter (4Q08).8% -0. Dev.0% 9.12 -0.Exhibit 18: The historical performance of systematic ATM VIX call overlay strategies Based on daily data from 16 May 07 through 23 Mar 11 4XDUWHUO\ 5HWXUQ 6WUDWHJ\ 9DOXH 6 6 6 6 0D\ 3 3 3 3 0 $70 9.9% -4.0% 1.7% 11. Strategies using longer-dated VIX calls generally slightly outperform those using shorter-dated calls.6% 12.4% -2.3% 3. &DOOV 0 $70 9..7% 3Q09 4Q09 1Q10 2Q10 3Q10 4Q10 1Q11 S&P500 + 1M S&P500 + 2M S&P500 + 3M S&P500 ATM VIX Calls ATM VIX Calls ATM VIX Calls 15. &DOOV 1RY 1RY 1RY 0D\ 0D\ 0D\ $XJ $XJ $XJ 1RY )HE )HE )HE $XJ 3Q07 4Q07 1Q08 2Q08 3Q08 4Q08 1Q09 2Q09 S&P500 + 1M S&P500 + 2M S&P500 + 3M S&P500 ATM VIX Calls ATM VIX Calls ATM VIX Calls 2.0% -15.4% -11.4% -3. On average.8% 2.3% 20.9% -1.7% 4. during the four worst full quarters for the S&P500 in our sample. Compared to the overlays using VIX futures. the S&P 500 gained 16%.9% 21.7% -4.8% 7. due to the low decay properties of VIX calls.) Std.11 -0.

&DOOV 'HOWD 9. On the other hand. &DOOV 1RY 1RY $XJ $XJ 0D\ 0D\ 0D\ $XJ 1RY )HE )HE )HE $XJ 3Q07 4Q07 1Q08 2Q08 3Q08 4Q08 1Q09 2Q09 S&P500 10S&P500 20S&P500 30S&P500 Delta VIX Calls Delta VIX Calls Delta VIX Calls 2. OTM call strategies underperform ATM strategies during small declines in the S&P 500.3% -4.6% Source: Morgan Stanley Quantitative and Derivative Strategies 16 . Since the calls are OTM and thus have a lower mark-tomarket to spot VIX changes.0% -2.7% -3. This compares to an average 13% return for the S&P500. with returns ranging from 7% (30-delta calls) to 11% (10-delta calls).6% 14.9% -7.4% -7. even after the scaling of the notional exposure. and 30-delta calls.1% -9.8% 1.2% 2. OTM VIX call strategies outperformed the ATM VIX call strategies.8% -11.3% 7.0% 2. and 7%-8% returns for ATM call strategies.8% 9.6% -5.1% 11. To indicate the potential benefits of using OTM VIX calls. VIX call hedges can be configured in any combination of maturities and strikes.3% 5.6% 15. we double the notional exposure for these calls.6% 6.4% -10.9% 11.5% 4.9% 6.8% -4. In 4Q08. During the two quarters in which S&P500 returns were between 0% and 5%. with the strategies returning anywhere from 68bps (10-delta calls) to 5.1% 5. we show backtests of various 1-month OTM VIX call strategies. particularly longer-dated ATM calls .8% 2.9% 6.7% 1.4%.6% -14. Similarly. &DOOV 'HOWD 9. While this exposure in OTM VIX calls will be beneficial when volatility rises sharply. Exhibit 20 shows the results of our historical backtest.0% -12.6% -4.7% -1.6% 10.3% -5. This reflects the lower premiums for OTM VIX call options than for ATM options. A larger notional exposure in OTM VIX calls can still be cheaper than a single ATM VIX call of equal maturity.2% 12.7% 2.W H AT AB O U T O U T OF TH E M O NE Y C AL L O P TI ON S? Like hedges using regular equity puts. the ATM call strategies had returns between +80bp and -4..7% -21.9% 6.4% -3.1% -15.0% 4.1% -8.9% 12. during the four best S&P500 quarters.4% -6. Meanwhile.6% (30-delta calls).9% 0. ATM calls generally performed better than OTM calls.8% 9..1% 4. the hedges performed quite well.5% -3.8% -13.0% -11. using 10-. it would give much less protection in gradually increasing volatility scenarios.4% 3.9% 10.0% 0.1% 8. Exhibit 20: The historical performance of systematic OTM VIX call overlay strategies Based on daily data from 16 May 07 through 23 Mar 11 4XDUWHUO\ 5HWXUQ 6WUDWHJ\ 9DOXH 6 6 6 6 0D\ 3 3 3 3 1RY 'HOWD 9. the quarter with the worst S&P500 performance. depending on maturity (see Exhibit 18).3% 8.3% 7.6% 1.0% 3Q09 4Q09 1Q10 2Q10 3Q10 4Q10 1Q11 S&P500 10S&P500 20S&P500 30S&P500 Delta VIX Calls Delta VIX Calls Delta VIX Calls 15. 20-.3% 10.7% -11..

Q OLQH SHUIRUPDQFH IRU VPDOO GHFOLQHV 0RQWK 6 3 5HWXUQ 'HOWD 6 3 3XWV 'HOWD 9. versus the performance of the S&P 500. However.. Exhibit 21: The cost of 1-month SPX put hedges versus scaled 1month VIX call hedges Based on data from 16 May 07 through 23 Mar 11 &RVW RI 6FDOHG 0RQWK 9. this cheapness is based on the assumption of a constant scaling or hedging ratio. while providing nearly the same hedging power in S&P500 drawdown periods (see lower chart in Exhibit 22). During Oct08 when the S&P500 was down 17%. In practice. Over the long-run.through 50-delta options. 3XWV 'HOWD 9.. the optimal hedge ratio is variable. Moreover. We construct a set of systematic strategies that use either SPX puts or a scaled notional of VIX calls as per our previous scaling ratios. Using SPX puts does not carry that risk. the VIX call-hedged strategy performed nearly in-line with the SPX put-hedged strategy. as indicated by the R-squared of the regressions in Exhibit 15. Exhibit 21 shows the historical premiums for both. ZHUH ERWK IDOOLQJ 9. while VIX calls may seem optically cheap based on their up-front premium. ATM VIX call strikes are generally higher than spot VIX levels. this cheapness has been relatively consistent and ranges from a 55% cheaper (50-delta options) to 60% cheaper (30-delta options). VIX calls with a given delta have been persistently cheaper than SPX puts with the same delta.. This may lead to over. we value the cost of the strategies using 10. Since the term structure is generally upward-sloping.9% per month. We perform an empirical analysis of the true potential ‘cheapness’ of VIX calls versus SPX puts through a set of backtests. In months with small declines in the S&P500 (0% to -5%). &DOO +HGJH GHJUHH OLQH 4 ZKHQ 6 3 DQG 9. &DOOV 6HS -DQ 1RY 0DU &RVW RI 0RQWK 63. due to their relative cheapness. We construct sets of 1-month option overlays on the S&P500 using 10-delta to 50-delta VIX calls and SPX 17 . On each roll date. Exhibit 22: The returns of the S&P500 hedged with 30-delta S&P500 puts and VIX calls versus the S&P500 return Based on daily data from 16 May 07 through 23 Mar 11 . &DOOV 'HOWD 2SWLRQV 'HOWD 2SWLRQV 'HOWD 2SWLRQV 6WUDWHJ\ 9DOXH 6 3 6 3 6 3 -XO -XO -XO 6HS 6HS 6HS -DQ -DQ -DQ 0DU 0DU 0DU 1RY 1RY 1RY 0D\ 0D\ 0D\ 0RQWK 0RQWK -XO 0D\ 'HOWD 63.or underhedging. I N -LI NE PE RF ORM AN C E VE RS US S &P 5 0 0 P U TS Source: Morgan Stanley Quantitative and Derivative Strategies The first chart in Exhibit 22 compares the monthly performance of S&P 500 overlaid with 30-delta VIX calls and 30-delta SPX puts..VIX Calls or SPX Puts for Hedging? H IS TO RI C AL L Y. the VIX hedge outperformed the SPX put hedge by 1. In addition. 3XW +HGJH Source: Morgan Stanley Quantitative and Derivative Strategies Based on these scaling ratios.Q OLQH KHGJLQJ SRZHU LQ 2FWREHU 0RQWK 6WUDWHJ\ 5HWXUQ To compare the relative costs of using VIX calls versus SPX puts. VI X C AL L S H AV E BEE N C HE AP E R puts. VIX hedges tended to outperform SPX put hedges slightly. For very large declines.1% per month on average. an investor in these instruments also implicitly pays for the shape of the term structure. a systematic VIX call hedge would have outperformed an SPX put hedge..SPX put hedges generally outperformed VIX call hedges by 0. returns were in-line again. All options are held to expiration and then rolled into the next month’s contract. For larger declines – up to -10% . KHGJHV VOLJKWO\ RXWSHUIRUP ZKHQ WKH 6 3 LV XS . The results for other delta levels were generally comparable. the VIX call cannot be monetized. In rising market environments. Unless this gap is overcome by expiry through rising spot VIX levels. the performance of VIX call hedges outperformed the S&P500 put hedge by 32bps per month on average. During the 26 months of positive S&P500 returns in our backtest sample.

since VIX and S&P500 were both falling at the same time. respectively. we can also consider VIX call spread hedging strategies to target changes in volatility more precisely. The mark-to-market profile of this trade is attractive – the position only tends to lose significantly right near expiration if VIX is near the long call strike. This is because of the slow decay of VIX options. In Nov 08. our scaled VIX call hedge plus an S&P500 position returned 31%. In contrast.10 bid-ask spread for every transaction. We assume a $0. the VIX started to decline very quickly after reaching its all-time high. the higher notional VIX exposure also highlights periods of hedging mismatch. because of the sharp decline in implied volatility during this period. In May 2010. with the deltas chosen to result in approximately a 50% premium reduction. These results highlight the tactical nature of many volatilitybased hedges. The call spread strategy shows the impact of selling some upside to cheapen the hedge. Exhibit 23: The cumulative returns of a systematic 30-delta S&P500 overlays including one with amplified VIX call exposure Based on daily data from 16 May 07 through 23 Mar 11 depends on the level of the VIX as well as on the recent direction of volatility changes. the shorter-dated call should increase in value faster than the longer-dated call. While the performance in periods of rising S&P500 was generally lower (6bps outperformance over SPX put hedge per month.1% on average for the ‘regular’ VIX call hedge). full upside is given up in case of a volatility shock. Buy 2x 1-month 20- Source: Morgan Stanley Quantitative and Derivative Strategies $XJ Delta VIX call Over the longer term. performance in declining S&P 500 periods was substantially higher. 18 . The regularsized and higher notional VIX call strategy actually fell -9% and -13%. The rolldown of the term structure for the longer-dated call will typically offset some of the cost of holding the neardated call. the scaled VIX hedge was up 6% while the other two hedging strategies were down -1% to -2%. However. The upper chart in Exhibit 24 shows the results of four backtests involving commonly used VIX option hedging strategies: > Buy 1x 1-month 40-delta VIX call 6WUDWHJ\ 9DOXH 6 6 6 6 0D\ 3 3 3 3 1RY 0RQWK 0RQWK 0RQWK )HE 'HOWD 6 3 3XWV 'HOWD 9. but pushes out all maturities by one month. the performance of this scaled VIX call hedge is similar to that of the ‘regular’ VIX call hedge. However. this trade takes on other term structure risks such as (1) a persistently flat term structure and (2) parallel shifts in the term structure. During Nov 08. while the S&P500 alone was down by around -17% over this period.. &DOOV 'HOWD 9. Exhibit 23 shows the results of a hedging backtest where we scale the VIX call exposure to the same premium outlay as the SPX puts. the S&P500 was down -7% while the 30-delta S&P500 put strategy fell approximately -2%. for example.. Our benchmark is the 40-delta call strategy. Sell 1x 1-month 20([SRVXUH $XJ $XJ )HE )HE 1RY 0D\ 0D\ 1RY )HE Delta VIX call > Sell 1x 1-month 40-delta VIX call. In 4Q08. On the other hand. If volatility spikes sharply. The calendar call spread purchases a near-dated call while attempting to fund the trade by selling a longer-dated VIX call.Since the upfront premium costs of scaled 30-delta VIX calls were only about 40% of the price of 30-delta SPX puts. ‘regular’ VIX call hedges or SPX equity put hedges were up by 1-2%.B AS E D V IX HE DGE S Similar to SPX put-spread based hedging strategies. VIX hedges did not provide downside protection during this period. &DOOV $XJ 1RY 0D\ > Buy 1x 1-month 40-delta VIX calls. with the S&P500 down -8%. versus 1. even though the S&P 500 did not bottom until Mar 09. The 1x2 call spread strategy is thus approximately zero premium. The appropriateness of VIX-based hedges > Buy 1x 1-month 40-delta VIX call. and is designed to gain when the VIX is rising sharply while accepting a small loss when the VIX rises slightly. Performance of Spread-Based VIX Hedges B AC K TE S TI N G O THE R SPR E AD . sell 1x 2-month VIX call with the same strike The lower chart shows the same strategies. All strategies are rolled on the monthly VIX expiration dates.

The strategy generally decays over time. The up-front premium for this trade is generally close to zero. On the back of Middle East protests and the Japanese earthquake. in relatively flat term structure environments – like in 2007 – these strategies are generally unattractive as the decay in front-month options is much faster than that for longer-dated options. &DOO 6SUHDG 3 / $XJ $XJ $XJ $XJ )HE )HE )HE 1RY 1RY 1RY 0D\ 0D\ 0D\ 0D\ 1RY )HE 9. A V IX 1 X2 C AL L SP RE AD S N EE D TO BE AC TI V EL Y M AN AG E D TO AV O I D P O TE N TI AL L OSSES &XPXODWLYH 6WUDWHJ\ 3 / 0RQWK 0RQWK 0RQWK 0RQWK 'HOWD &DOO 'HOWD &DOO 6SUHDG 'HOWD [ &DOO 6SUHDG 'HOWD &DOHQGDU &DOO 6SUHDG The mark-to-market profile of the 1x2 call spread is greatly affected by the slow decay of VIX options. a 1x2 call spread worked better as a tail risk hedge. such as the one in force since late 2009. A trade that sells a $1 S&P500 put to finance a $1 VIX call would be selling very deep OTM S&P500 exposure to buy a more near-the-money VIX hedge for zero upfront premium. upside participation in volatility spikes was significantly reduced as well... Moreover. However. When factoring long-run holding costs.14 and created a loss. We display in Exhibit 25 the mark-to-market profile of a 20-24 1x2 call spread which was nearly zero cost on 23 Mar 11. )XWXUHV 3ULFH Source: Morgan Stanley Quantitative and Derivative Strategies The structure on a mark-to-market basis gives full upside in the case of a volatility spike. The calls decay slowly though which mitigates most of this risk. Using a Scaling Mismatch for a VIX Hedge U TI LIZ IN G TH E S C AL I N G M ISM ATC H O F I N DI CES VIX and the S&P500 are denominated in different units. VIX rose but not to crash levels. while still providing upside exposure when volatility spikes. The options settled to a March opening print of 25. this trade suffers during small volatility spikes. there is a positive probability that the 19 . upward-sloping term structures. since calendar spreads involve two separate maturities. P&L of calendar spreads tends to be more volatile than that of other strategies. If the S&P500 declines.Exhibit 24: The cumulative P&L of various systematic alternate hedging strategies using VIX calls Based on daily data from 16 May 07 through 23 Mar 11 0RQWK 0RQWK 0RQWK 0RQWK 'HOWD &DOO 'HOWD &DOO 6SUHDG 'HOWD [ &DOO 6SUHDG 'HOWD &DOHQGDU &DOO 6SUHDG &XPXODWLYH 6WUDWHJ\ 3 / On the other hand. However there is a 8 volatility point gap in the middle at expiration that could produce a loss. Therefore. This benefits performance in a falling volatility environment. making the tradeoff between cost and upside participation unattractive. When our systematic 1-month rolled in February. such as March 2011. while underperforming other hedging strategies during volatility spikes. This position should be rolled at the latest a week before expiration to avoid a loss due to rapid decay. Exhibit 25: The hypothetical future mark-to-market profile of an April VIX 20-24 1x2 calls spread as of 23 Mar 11 0RQWK 8QWLO 0DWXULW\ :HHNV 8QWLO 0DWXULW\ :HHN 8QWLO 0DWXULW\ $W 0DWXULW\ $XJ $XJ $XJ $XJ )HE )HE )HE 1RY 1RY 1RY 0D\ 0D\ 0D\ 0D\ 1RY )HE Source: Morgan Stanley Quantitative and Derivative Strategies 1 X 2 C AL L SP RE AD S D ELIVE R HI GH P AYO U TS I N CRISIS SCEN AR I O S W ITH L OW HOL DI N G CO S TS While simple VIX call spreads generally reduced the cost of the hedge. calendar call spreads can minimize holding costs due to decay. Calendar spread structures are particularly attractive during periods of steep. a VIX call option costing $1 should purchase more VIX upside than a $1 SPX put option would purchase in S&P500 downside. In such an environment. This requires the hedge to be actively managed to avoid this expiration loss. The 40-delta call worked the best in a 2008-style scenario with sharply rising volatility due to its lack of an upside cap. it traded a March 2011 20-24 1x2 call spread. but only about 8bps of the S&P500 (1/1300). $1 is about 5% (1/20) of the VIX. 0RQWK 9.

this scaling mismatch would likely cause such a strategy to be substantially unprofitable. We assume a $1 bid/offer spread for the SPX puts and compute a put strike that would have had a theoretical bid with the same price as the VIX call. S UCH AS ‘87 In our backtest. During the 1987 crash. the S&P500 never crossed the put strikes. The VIX call would be so far in the money that it would trade with very little time value – it would be equivalent to a cash + VIX futures position. THE S &P 5 00 H AS NO T CROSSED I TS P U T S TRI KE… YE T period was struck at 894. on the other hand. A crash like in 1987 would lead to an increase in the intrinsic value of the S&P 500 put of 1300* (32%-15%) = $221. mitigating the mark-to-market loss. using prices at the close of every listed VIX expiration. the trade did become unprofitable whenever the S&P500 neared the put strike. the value of the SPX put would have continued to rise more rapidly. S UDDEN M ARKE T CRAS H E S HOW EVER W ILL P RODUCE A L OSS W I TH TH IS S TR ATE GY. SI N CE V IX O P TI O NS W ERE LIS TE D IN 2 0 0 6. The S&P 500 put. For example. because the S&P500 is denominated in larger units. The mark-to-market value of the SPX put rose much faster than the value of the VIX call. As a result.5. At current S&P 500 and volatility levels. This would product a mark-to-market loss of approximately $221 $122.5. We can see this in the volatility of the P&L of our strategy in Exhibit 26. since we did not consider the time value of the options. In a 1987-style scenario. the SPX put would thus rise faster in value than the VIX call. ranging from 5% to 46% OTM. and a VIX call with about a 27. the S&P500 dropped by 32%.5 strike. As long as the S&P500 does not cross the SPX put strike. while the VIX call would have increased in intrinsic value by 150 – 27. However. the S&P500 rose 105 points on the next day. In our backtest. 1-month SPX puts with around a $1 premium were on average 15% OTM. would have a substantial short volatility exposure due to its higher time value.5 = $122. such as in the October 87 crash. Exhibit 26: The cumulative P&L of a strategy that sells 1-month S&P500 puts costing $1 and buys 1-month VIX calls costing $1 Based on daily data from 24 Feb 06 through 23 Mar 11 6WUDWHJ\ 3 / OHIW $XJ $XJ )HE )HE )HE 1RY 1RY 0D\ 0D\ 6 3 $XJ 0D\ ULJKW )HE 1RY 6 3 $XJ 0D\ 3XW 6WULNH ULJKW $XJ )HE 1RY 0D\ 1RY )HE Source: Morgan Stanley Quantitative and Derivative Strategies On a mark-to-market basis. the level of the S&P500. our strategy would lead us to short an S&P 500 put that is 15% OTM. for example during October 2008. If the S&P500 had continued to fall. the S&P500 reached 899 while the put for this 20 . while the old VIX rose from 23 to 150. We can highlight this possibility using the dynamics of the 1987 crash. If market were in a low implied volatility regime and equities were to suddenly drop.VIX call will be ITM at expiration. a 10% drop in the S&P500 corresponds to approximately 130 index points (1300 * 10%) while a 10% rise in the VIX corresponds to 2 index points (20 * 10%). and the strike of the $1 SPX put over time. For example on 10 Oct 08. an increase of 127 points.5 = $98. Exhibit 26 shows the backtest of a strategy that buys the listed 1-month VIX call with a bid closest to $1. However. an X% sell-off in the S&P500 causes a larger S&P500 point shift than the VIX point shift under an X% rise in the VIX. this strategy will yield a positive payoff in this scenario. a loss was never realized at roll time. If an equity sell-off is substantial and the S&P500 breaches the put strike. We then construct a set of 1-month theoretical SPX put prices using the volatility surface of listed options. This would tend to exacerbate the mark-to-market loss of this position. it is likely this trade would have been very unprofitable. We plot the cumulative P&L of this strategy. resulting in positive P&L in times of rapid VIX spikes. Note that this might be a conservative estimate of the mark-tomarket loss. while the strategy provided upside participation through the VIX call.

.V.PSOLHG 9RO VIX _ FutT = E t =0 [VIX T ] = E t =0 [ E t =T [ RVol 2 T −T +30 ] ] . So at time t=0. highlighting advantages as well as potential risks and drawbacks: > Using VIX derivatives to trade the S&P500 implied vol The upward-sloping term structure of implied volatility is often attributed to supply-demand dynamics. there are fewer natural sellers of volatility. Over the past five years. for example by systematically selling forward starting variance swaps. please refer to “A Tale of Two Indices. Inversions of the term structure typically only occur in distressed market scenarios. such as covered call writers or outright variance swaps sellers. and are related to the 1-month forward starting implied volatility.PSOLHG 9RO 6SUHDG 9RO SRLQWV 0RQWK 9. VIX futures are priced off the SPX implied volatility term structure. Exhibit 27 shows three sample term structures – upwards sloping. . many question whether liquid. The shape and slope of the term structure during inversions can be more extreme than when the term structure is upwards sloping – see Exhibit 28. The Journal of Derivatives. the price of a VIX future can be thought of as: $70 . 11 term structure > Constructing term structure trades with reduced Exhibit 28: The VIX futures term structure has a similar shape as the S&P500 implied vol term structure Based on daily data from 26 Mar 04 through 23 Mar 11 downside risk > VIX options implied volatility has historically traded rich versus subsequent realized vol > We backtest the performance of potential alpha trades .PSOLHG 9RO -XO -XO -XO 0DU 0DU 0DU 0DU -XO 1RY 1RY 1RY 1RY 1RY A Listed Vehicle to Arb the Term Structure VIX F U TURES VS. driven by the shape of SPX implied volatilities. it can be shown that: E t =0 [ E t =T [ RVol 2 T −T +30 ]] ≤ VIX _ FutT ≤ E t =0 [ E t =T [ RVol 2 T −T +30 ]] -DQ 0D\ 6HS 0RQWK 0RQWK 0XOWL <HDU 0 $70 9RO /RZ +LJKHVW 9RO 'D\ 'XULQJ 6RYHULHJQ 5LVN 5HIODWLRQ 7KH 'D\ 3ULRU WR /HKPDQ %DQNUXSWF\ 0RQWK <HDU which simplifies to: ForwardVolSwapT −T +30 ≤ VIX _ FutT ≤ FowardVarSwapT −T +30 For more information.PSOLHG 9RO -XO -XO -XO 0DU 0DU 0DU 1RY 1RY 0RQWK $70 63. listed VIX derivatives can complement or replace index variance swaps in these systematic alpha strategies This section provides a framework for using VIX futures & options as volatility alpha vehicles. downward sloping.. Further out the term structure. is an expectation of where the expectation of realized vol will be in the future. After the events of 2008. helping to push the term structure upwards. As a result. 11 One can express the price of the VIX at time t=T as an expectation of future realized volatility from time T to time T+30: VIX T = E t =T [ RVol 2 T −T + 30 ] . however. Volatility suppliers. and flat YIE L D F R OM UPW ARD-SLOPI NG TE RM S TRUCTURES One of the most common volatility alpha strategies is to exploit the generally upward-sloping term structure through risk premium extraction. THE SPX V OL TE RM S TRUCTURE Source: Morgan Stanley Quantitative and Derivative Strategies The SPX implied volatility term structure is generally upward sloping – the implied volatility of longer-dated options is higher than that of shorter-dated options. Exhibit 27: Three single day examples of potential term structure shapes: upward sloping. The price of a VIX future.” Carr & Wu. VIX for Alpha-Seekers Volatility has long been an alpha source for many absolute return strategies using index OTC variance swaps. .. )XWXUHV 9. The term structure of forward starting implied volatilities tends to be upwards sloping as well. the basis of VIX futures tends to track the shape of the S&P500 term structure (Exhibit 28). Using Jensen’s Inequality. the term structure was upwards sloping on 78% of trading days. fill this demand.PSOLHG 9RO 7HQRU Source: Morgan Stanley Quantitative and Derivative Strategies 21 . The premium for longer-dated tail risk increases. 0RQWK $70 63. flat and inverted. Demand for volatility comes from tactical option positions as well as downside hedging strategies. Spring 2006. Forward starting implied volatility can be traded via VIX futures.

such a strategy could be profitable.For example. This longer period covers three bull markets and two recessions. Based on historical rolling VIX futures data. this model has a regression R2 of 98% and we use the in-sample results to extrapolate the data using S&P500 volatilities.7. 12 Because VIX futures are strongly linked to forward starting volatility. In addition to analyzing each of these strategies individually. this ‘rolldown effect’ can be exploited by selling the VIX future.O R 4 -M ON TH S P RE AD S W OR KE D BES T VIX. we can also exploit this ‘vol-of-vol’ effect as another potential alpha source. 1 -M O N TH / 3 . and if the historical hedge ratio is accurate. allowing us to evaluate our strategies more robustly. comparing them to that of a strategy that is short the 1-month VIX futures without a hedge. Delta-hedged VIX straddles are pure plays on the ‘vol-of-vol’ effect. An inverted term structure would cause the entire strategy to work in reverse as the short 1-month future rolls up the futures curve rather than down. However. The table Selling short-dated VIX futures is the simplest way of exploiting the steepness of the upward-sloping term structure. We examine four basic types of trades as examples of pure alpha trades around the VIX: > VIX future calendar spreads > VIX put spreads > VIX calendar put spreads > Delta-hedged VIX straddles VIX future calendar spreads are the most direct way to leverage the term structure steepness. Using these historical hedge ratios of longer-dated VIX futures to 1-month futures. we find that 2. 1. we sell one 1-month VIX future while buying 2. Exhibit 10 earlier in the paper showed that the rolldown cost per unit of VIX exposure was not uniform across all maturities. 12 VIX Futures Calendar Spreads M OS T DI RE C T W AY TO TR AD E TH E TE RM S TRU C TUR E Exhibit 29 shows the performance of these VIX futures calendar spreads. 2. the strategy is exposed to sudden spikes in the VIX. We regress in-sample the VIX futures basis on (1) the spread between the equivalent ATM forward volatility level and spot 1-month implied vol and (2) the square root of time until maturity. Spikes in volatility would typically lead to a loss in this scenario. we can structure different trades that can extract this rolldown P&L while minimizing potential losses. In sample. Utilizing all the payoff profiles that combinations of VIX futures and options allow.4 times. By using some of the rolldown yield from selling a 1-month VIX future to purchase an amount of longer-dated VIX futures.to 6-month futures have a sensitivity to 1-month futures of 1. Two major risks remain: (1) parallel shifts inconsistent with these hedge ratios and (2) a sustained inversion in the term structure. We have extended the history of VIX futures from their Mar 2004 first listing back to January 1996. Through selling (delta-hedged) VIX options. For every $100 of strategy value. 22 .based alpha strategies are often based on exploiting the richness of the SPX implied volatility term structure.through 6-month futures according to our hedge ratios. In addition.1 and 2. In an upwards-sloping term structure environment. VIX options tend to trade at a premium relative to the realized volatility of VIX futures in low-risk environments. using a regression model. Successful premium extraction strategies of this type therefore require a hedging strategy that mitigates the impact of volatility spikes. The cash component earns the Effective Overnight Federal Funds rate and the strategies are rolled on the listed VIX futures expirations. we will also highlight the potential benefits of aggregating them into a single strategy. we backtested a series of systematic strategies that execute this trade monthly. If the term structure remains upwardsloping. if spot VIX remained unchanged over the lifespan of a VIX future. F O U R M ETH O DS TO GE NER ATE HE D GE D AL P H A Calendar spreads can mitigate the impact of spikes in the VIX. All futures trades assume execution at their daily closing prices. which can often be multiples of the potential risk premium gain.4. Longer-dated contracts tend to have a smaller rolldown cost than a 1-month future. respectively. the price of the VIX future has to converge towards the spot VIX as it moves towards expiry. we construct a two-factor regression model. 1. Using these ratios. we can scale a hedging strategy that would profit in steep term structure environments. the overall strategy can be effectively hedged against term structure inversions and spikes in the VIX. Put spreads can reduce the exposure to volatility spikes.9.

The remaining cash accrues the overnight Federal Funds rate. which have limited downside.separates the performance of the strategies during the backfilled period from that over the whole period. this ATM strike will be higher than the spot VIX level.D ATE D S TR ATE G IES TE N D TO W OR K BES T 8QKHGJHG 2XW RI 6DPSOH 5HVXOWV 5HWXUQ $QQ 5LVN $QQ &9D5 6KDUSH 5DWLR . as the P&L gained on the put may not be enough to cover the initial premium.15) strategy was based on hedging with 4-month VIX futures. )XWXUH 0DWXULW\ 0RQWK 0RQWK 0RQWK 0DU Similar to the VIX futures backtest. During the higher-volatility period between 2000 and 2002. We investigate selling other options to fund the VIX put premium. while the unhedged strategy required nearly two years to recover to its pre-4Q08 highs. since the longer-dated vol term structure tends to be quite flat. which includes our backfilled data.10. sell 2x an 80% 2-month put and roll monthly For every $100 of capital. sell an 80% 1-month put -DQ -DQ -DQ -DQ -DQ -DQ -DQ -DQ -DQ -DQ -DQ -DQ -DQ -DQ -DQ 0RQWK 0RQWK -DQ 0DU 0DU > Buy a 1-month ATM put. When the term structure is upwards-sloping. all strategies suffered. long ATM Source: Morgan Stanley Quantitative and Derivative Strategies During the 4Q08 vol spike. volatility and CVaR are higher for these strategies as well. when the index began one of its strongest rallies in history. The period until March 2009 was generally unprofitable for long ATM put strategies. we backtest four VIX put strategies to highlight various characteristics of these trades: > Buy a 1-month ATM put > Buy a 1-month ATM put. Over the whole period since 1996. the ATM strike is defined as the level of the corresponding VIX future. Using historical VIX futures data since 2004. The level of the VIX at expiration has to drop by more than the option premium for the strategy to break even.. Exhibit 30 shows the historical backtest of the strategies since September 2006. We would expect strategies involving VIX puts to underperform those using VIX futures in moderately upwards-sloping term structure regimes. which tends to make these volatility premium extraction strategies unprofitable. an ATM put becomes increasingly ITM as the VIX futures price approaches the VIX spot. the strategy buys a single VIX put or put spread on each roll date. the highest Sharpe Ratio (1. sell 2x an 80% 1-month put > Buy a 2-month ATM put. 23 . since longer-dated futures are a less effective hedging tool. buying VIX puts requires an up-front premium.Q 6DPSOH 5HVXOWV 5HWXUQ $QQ 5LVN $QQ &9D5 6KDUSH 5DWLR -DQ /RQJ 9. as expected – rising volatility and inverted term structures were drags on the performance. For VIX options. Most strategies outperformed the unhedged trade on both an absolute and a relative basis. However. All option transactions assume a bid-ask spread of $0. compared to shorterdated hedges. the hedged strategies were able to minimize losses and recovered relatively quickly. we can hedge using long positions in longer-dated VIX futures. Since the S&P500 low in March 2009. The implied volatility term structure oscillated from relatively flat to inverted for a sustained period. using 3-month VIX futures yielded the highest Sharpe Ratio (0. Strategies using longer-dated VIX futures as hedges yielded a higher return in general. Unlike strategies involving VIX futures.85). S HO R T. Exhibit 29: The historical performance of hedged VIX futures spread strategies and a naked short VIX futures strategy Based on daily data from 4 Jan 96 through 23 Mar 11 6KRUW 6KRUW 6KRUW 6KRUW [ [ [ [ 0 0 0 0 )XWXUH /RQJ )XWXUH /RQJ )XWXUH /RQJ )XWXUH [ 0 )XWXUHV [ 0 )XWXUHV [ 0 )XWXUHV Short-Dated VIX Ratio Put Spreads SE LLI NG THE TE RM S TRUCTURE THROUG H P U TS To mitigate the downside risk from volatility spikes that is inherent in a VIX futures selling strategy. If the term structure remains unchanged. Another strategy involves trading short-dated puts on the VIX instead.

this is equivalent to a VIX futures calendar spread. sell 1. Strategies are rolled at the monthly VIX expirations. sell 1. This was driven both by declining volatility and by a strongly upwards-sloping term structure.4x and 1. L O NGE R . The performance of ITM put spread strategies was consistently above that of the ATM put spread strategy. an investor can sell a weighted amount of a longer-dated put with the same strike as the long put. Moreover. Selling OTM VIX puts to raise partial funding for the long ATM put position can help to overcome the drag in performance of long ATM put strategies. We attribute this to the lower time value embedded in ITM puts. the roll-down benefit to performance is less strong.7x a 3-month put 0RQWK 3XWV 5HWXUQ $QQ 5LVN $QQ &9D5 6KDUSH 5DWLR 0RQWK 3XW 6SUHDG 0RQWK [ 3XW 6SUHDG 0RQWK [ 3XW 6SUHDG Source: Morgan Stanley Quantitative and Derivative Strategies Extending the maturity of the put spreads has a detrimental effect on returns and Sharpe Ratio.4x a 2-month put > Buy a 1-month 10% ITM put. Exhibit 30: The historical performance of VIX put spread strategies Based on daily data from 20 Sep 06 through 23 Mar 11 /RQJ 0RQWK $70 3XWV /RQJ 0RQWK $70 3XWV 6HOO [ 0RQWK /RQJ 0RQWK $70 3XWV 6HOO [ 0RQWK /RQJ 0RQWK $70 3XWV 6HOO [ 0RQWK 3XWV 3XWV 3XWV premium. The rolldown of the longer-dated put is less than the shorter dated put. contributing to potential losses.D ATE D . Our backtests start in May 2007.7x hedging ratios in the short puts are consistent with the ratios used for VIX futures calendar spreads in the previous sections. we set the short put strike to be the same as the long put strike.4x a 2-month put > Buy a 1-month 20% ITM put. In this trade. If the term structure remains unchanged. the long ATM put would become increasingly ITM over time as its corresponding VIX future rolls down the term structure.4x a 2-month put > Buy a 1-month 20% ITM put. Exhibit 31 shows the historical performance of the weighted calendar VIX put spread strategies. In effect. where we buy a short-dated VIX call and sell a longer-dated VIX call to reduce exposure to spikes in the VIX. this strategy buys one VIX option for every $100 of strategy value. The 1. with the remaining cash accruing Overnight Fed Funds. we did not backtest strategies using longer-dated options. Like the other backtests. The long put should increase in value faster than the short put – this difference in decay helps to pay for the long put position.10 bid-ask spread was assumed on all trades. and a $0. even in falling volatility environments. Because VIX option liquidity is concentrated in the first three months. the long put would not be ITM at maturity. In all cases. Both the 1x1 and the 1x2 put spreads showed higher returns and higher Sharpe Ratios than the long ATM put strategy in our backtest. The key risk in this strategy comes from the shape of the term structure. sell 1. sell 1. I TM C AL E N D AR PU TS W OR K BES T We consider four representative sample strategies for VIX calendar put spreads to evaluate historical benefits and risks. which reduces the net premium of the put positions that has to be overcome by the term structure Weighted VIX Calendar Put Spreads VIX F U TURES C AL E ND AR SP RE AD S THRO UG H P U TS We can introduce an additional time dimension into the put spread strategy. a breakdown in the regression relationship between changes in different volatility tenors would also produce a mishedge and potentially losses. This benefit is counterbalanced by the cost of the initial put 24 . we purchase a short-dated ATM or ITM VIX put while selling a longer-dated put with the same strike.put strategies have been profitable. Since the term structure is not as steep for longer maturities. given the shape of the term structure. Should the term structure become relatively flat or inverted. when expirations up to 3 months were listed on a consistent basis. similar to the VIX futures calendar spread we considered earlier. To help fund this. 'HF 'HF 'HF 'HF 'HF 6HS 6HS 6HS 6HS 6HS -XQ -XQ -XQ 0DU 0DU 0DU 0DU -XQ 0DU > Buy a 1-month ATM put.

the price of this call spread reduces to zero. The implied/realized volatility premium compensates for the risk of this short volatility position. We compare this strategy to one without a delta hedge. IXWXUHV UHDOL]HG YRO VSLNHV $XJ $XJ $XJ $XJ )HE )HE )HE 1RY 1RY 1RY 0D\ 0RQWK 0RQWK $70 6WUDWHJ\ 5HWXUQ $QQ 5LVN $QQ &9D5 6KDUSH 5DWLR 0D\ 0RQWK 0RQWK . we create a strip of theoretical VIX futures that have daily expirations.. The risk profile of the delta-hedged strategies is relatively similar – 5% CVaRs are nearly identical (-1.. VIX futures’ realized volatility is heavily dependent on time to maturity. Since this strategy is short more of the longer-dated puts than it owns in short-dated puts. Unlike regular equity index futures. the implied/realized VIX volatility premium is a compensation for the risk to the suppliers of this volatility. When the call strikes become further OTM.70 3XWV 6KRUW [ 0RQWK 3XWV 0RQWK . S TR O NGE R RE TU R NS TH AN S HO R T SPX S TR AD D L ES Exhibit 33 shows the backtested performance of a strategy that systematically shorts 1-month ATM VIX straddles.slide. Trades are executed at closing prices. This is driven by the same structural forces that cause SPX options to trade at a premium to realized volatility. This way for each day’s ATM 14 implied volatility level. VIX options generally trade rich – their implied volatility is typically greater than subsequent realized volatility.0% vs. 13 An easier way to think of this is to decompose these puts into positions in VIX futures and VIX calls. Again. deltahedged.. Exhibit 32: The implied – realized volatility spread in VIX options Based on daily data from 31 Dec 07 through 23 Mar 11 *HQHUDOO\ D SUHPLXP XQOHVV 9. -1. Similarly. both of the puts became deep OTM puts that converged towards a zero price. Investors are generally net demanders of optionality – for example through OTM SPX puts for hedging purposes. performance strongly picked up when the longerdated set of 3-month options was used. Being long a 2-month put and short a 1-month put is almost like being long (1) a 2-month minus 1-month futures spread and long (2) a 1-month minus 2month calendar call spread. During this rapid rise in volatility. the risk of a volatility spike is offset by the premium gain from the short put position.70 6WUDWHJ\ 1RY )HE .PSOLHG 9RO $SU $SU 'HF 'HF 'HF $SU $XJ $XJ $XJ 'HF )HE )HE )HE -XQ -XQ -XQ 2FW 2FW 2FW Source: Morgan Stanley Quantitative and Derivative Strategies Source: Morgan Stanley Quantitative and Derivative Strategies The strategy benefited greatly during the 4Q08 volatility spike. )XWXUHV 5HDOL]HG 9RO 0RQWK 9. 25 . and cash collateral accrues overnight Fed Funds. Market makers supply this optionality. as well as to a delta-hedged straddle using SPX options instead of VIX options The VIX strategy sells one VIX straddle for every $100 of index value every month.10 bid-ask spread on each option for the VIX straddle. Moreover. This is consistent with our findings from the VIX futures calendar spreads that showed stronger returns when extending to longer maturities. Exhibit 31: The historical performance of weighted calendar VIX put spread strategies Based on daily data from 16 May 07 through 23 Mar 11 /RQJ /RQJ /RQJ /RQJ 0RQWK $70 3XWV 6KRUW [ 0RQWK 3XWV 0RQWK .70 3XWV 6KRUW [ 0RQWK 3XWV 0RQWK .g.1%) while the annualized standard deviations were within 63bps of Delta-Hedged VIX Straddles TR AD I N G THE RI CHNESS O F VI X IM PLIE D V OL ATI L ITY Much like S&P500 options. corresponding VIX future. We assume a $0. $70 . the SPX strategy sells $100 of notional straddles for every $100 of strategy value. in 4Q08 or 2Q10). and a 20bps bid-ask spread for SPX options. there is a VIX future with exact 30 days until maturity.70 6WUDWHJ\ 0D\ 0RQWK 0RQWK .70 3XWV 6KRUW [ 0RQWK 3XWV 13 Unless VIX volatility is spiking (e. VIX implied volatility is priced at a premium to subsequent VIX futures realized volatility. which makes them short volatility. this produced a net premium benefit for the strategy. investors are likely net demanders of VIX options. We then compute the realized volatility of this future as it becomes increasingly volatile over time. While using ITM strikes increases potential exposure to a volatility spike. for example for VIX upside calls as tail risk hedges. Using a set of futures data. Exhibit 32 shows the 1-month ATM implied volatility of VIX options as well as the subsequent 14 realized volatility of a theoretical.PSOLHG 5HDOL]HG 0RQWK 6XEVHTXHQW 9.70 6WUDWHJ\ 0D\ 0RQWK 0RQWK .

&DOHQGDU 3XW 'HOWD +HGJHG 9. VIX strategy returns were much stronger over the backtest period. The annualized standard deviation of the combined strategy is significantly below that of individual strategies. we have reviewed a number of different alpha strategies using VIX instruments. $70 6WUDGGOHV 'HOWD +HGJHG $XJ $XJ $XJ $XJ $XJ )HE )HE )HE )HE )HE 1RY 1RY 1RY 1RY 0D\ 0D\ 0D\ 0D\ 0D\ 1RY )HE 9.. (TXDOO\ :HLJKWHG 6SUHDGV 6SUHDGV 6WUDGGOHV $OORFDWLRQ 6KRUW 0RQWK 8QKHGJHG 9. particularly during 4Q08 and during the recovery phase that started in March 2009. Combining the strategies nevertheless provided diversification benefits resulting in a superior risk and return profile. )XWXUHV /RQJ 0RQWK $70 3XWV 6KRUW [ 0RQWK 3XWV /RQJ 0RQWK $70 3XW 6KRUW [ 0RQWK 3XWV 6KRUW 0RQWK 'HOWD +HGJHG 9. )XWXUHV /RQJ [ 0RQWK 9.. Since there is no equivalent product for the VIX. )XWXUHV &DOHQGDU 6SUHDG [ 9. as well as the implied-to-realized volatility premium. the set of suppliers of VIX volatility is likely to be much smaller. 6WUDGGOHV 5HWXUQ $QQ 5LVN $QQ &9D5 6KDUSH 5DWLR 6KRUW 0RQWK +HGJHG 9. in contrast. Aggregating VIX Options Strategies PE RF O RM AN C E O F AL L O CATI N G TO AL L S TR ATE G IES In this section... Exhibit 34: The historical performance of all VIX alpha strategies and a monthly-rebalanced strategy that allocates between them Based on daily data from 16 May 07 through 23 Mar 11 6KRUW 0RQWK 9. we allocate two dollars to the other three strategies to make the risk contributions comparable. we take the best performing strategy from each of the strategy types we considered.. which may account for VIX implied volatility staying rich. $70 6WUDGGOHV 1R 'HOWD +HGJH 6KRUW 0RQWK 9. The strategies clearly show correlation with each other. combining near-zero returns with higher risk.. To gauge the potential attractiveness of these alpha strategies.. For every dollar allocated to the VIX futures calendar spread. Exhibit 33: The historical performance of delta-hedged & unhedged short VIX straddles and delta-hedged short SPX straddles Based on daily data from 24 Feb 06 through 23 Mar 11 expiration date. are a common way of trading SPX volatility without the need to manage delta actively. 6WUDGGOHV (TXDOO\ :HLJKWHG 6WUDWHJ\ -XO -XO -XO -XO 6HS 6HS 6HS 0DU 0DU 6HS -DQ -DQ -DQ 0DU 1RY 1RY 0D\ 1RY 0D\ 0D\ 6KRUW 0RQWK 9.. VIX options were introduced to the market much more recently. 3XW 9.. particularly given our aggressive bid-ask spread assumptions. as is the CVaR. The delta-hedged VIX strategy has outperformed the SPX strategy by over 228bps per year despite its lower risk.. $70 6WUDGGOHV 'HOWD +HGJHG 6KRUW 0RQWK 63. However. Moreover. trading the richness of VIX options requires an active delta hedge.each other. the skew. 6WUDGGOHV 6KRUW 0RQWK +HGJHG 63. The strategies exploited the shape of the term structure of the VIX. SPX variance swaps. The unhedged strategy has performed poorly in comparison.. 6WUDGGOHV 5HWXUQ $QQ 5LVN $QQ &9D5 6KDUSH 5DWLR Source: Morgan Stanley Quantitative and Derivative Strategies Source: Morgan Stanley Quantitative and Derivative Strategies One potential driver for the performance differential of VIX straddles versus SPX straddles may be a greater mismatch between the supply and demand of VIX optionality. Exhibit 34 shows the historical performance of a strategy that allocates to each of those strategies on every VIX 26 0D\ 1RY .

While the other ETPs have yet to gain significant assets. Exhibit 35 provides an overview of basic characteristics of the strategies backing these products.and medium-term VIX indices.what is currently available? > Impact of VIX futures rolldown on VIX ETPs > The long-term performance of these indices going back linked to the Medium Term VIX Futures index. and focus on pure volatility products.VI. 14 ETPs were listed in the US. At the time of this publication. The remaining one is an arbitrage strategy that trades these two indices against each other to collect a risk premium. These two ETPs track the short. they do not track the spot VIX – as discussed earlier. with some products offering -1x inverse and 2x long exposure to these indices. The points we cover include: > Current ETP landscape .35% 1x 3-Jan-11 7mn 5-Month Medium Medium 0. The portfolios of VIX futures underlying these indices are designed to have a high correlation with the spot VIX. EXCHANGE TRADED NOTES VZZ XVIX UBS Long Enhanced Daily LongShort VIX VIX MidTerm ETN Futures ETN Tactical 2x Long SPVXMTR Tactical / Yield Generation VIIX TVIX XIV VIIZ TVIZ ZIV ProShares VelocityShares VIX Short.65% 1x 29-Nov-10 6mn 5-Month Medium Medium 1.89% Leveraged Short Long (Daily Reset) (Daily Reset) 2x 29-Nov-10 4mn 5-Month High High 1. The two ETPs that were listed first – the VXX and VXZ – have the majority of AUM. all other products have less than USD 100mm AUM.65% 1x 29-Nov-10 73mn 1-Month High High 1. they have increased the variety of products. We omit other listed products that contain VIX futures amongst other instruments.89% 1x 29-Jan-09 592mn 5-Month Medium Medium 0.85% 29-Nov-10 7mn 1-Month High High 0. we overview the emerging VIX ETP market and how the properties of VIX derivatives also affect their behavior.85% 1x 29-Jan-09 1.89% Leveraged Short Long Long (Daily Reset) (Daily Reset) 2x 29-Nov-10 30mn 1-Month Very High Very High 1. spot VIX does not equal cash plus VIX futures. TH E S TR ATE G IES BE HI N D VIX F U TU RES IN DI CES 15+ years > The hypothetical performance of inverse & leveraged strategies Overview of US-Listed VIX Futures ETPs CURRE N T L ANDS C AP E & AV AI L ABI L I TY OF VIX E TPS ETFs and ETNs linked to VIX futures have grown substantially since the first were listed in January 2009. VIX Futures ETPs In the previous two sections. As of 23 Mar 2011. In this section. we analyzed how the properties of VIX derivatives affected the performance of systematic VIXbased portfolio hedging and alpha extraction strategies.89% Tactical / Yield Generation 1x Short SPVXSP Short (No Reset) 1x 14-Jan-11 40mn 1-Month High High 0.85% Fee Source: Morgan Stanley Quantitative and Derivative Strategies 27 .89% Tactical / Hedging 1x Long SPVXMTR Long Suggested Usage Underlying Index Tactical Tactical 2x Long SPVXSP Tactical / Hedging 1x Long SPVXMP Tactical 2x Long SPVXMP 1x Long 1x Long SPVXTSER SPVXSP Long or Short Long Vol Exposure Leverage Launch Date 1x 3-Jan-11 Leveraged Long / Short Long Long (Daily Reset) 2x 30-Nov-10 18mn 5-Month High High 0.35% 1x 29-Nov-10 4mn 5-Month Medium Medium 0. Three products are also listed in Europe while two products are linked to equivalent indices using VSTOXX futures.89% AUM 25mn (on 23 Mar 11) Term Structure 1-Month Held Reactivity to Volatility Rolldown Exposure High High Management 0.Term Futures ETF Futures ETF Futures ETN Futures ETN Term Futures ETN Futures ETN Tactical 1x Long SPVXSP Tactical / Hedging 1x Long SPVXMP Long Tactical 1x Long SPVXSTR Long Tactical / Yield Generation 1x Short SPVXSP Short (No Reset) 1x 19-Jul-10 29mn 1-Month High High 0.89% 2x Long / 1x 1x Short 1-Dec-10 70mn Long 5m / Short 1m Low Medium 0. with 7 products linked to the Short-Term VIX Futures index and 6 Exhibit 35: US-listed VIX futures-linked volatility indices as of 23 Mar 2011 EXCHANGE TRADED FUNDS VIXY Issuer VIXM VXX XXV IVO Barclays VXZ The indices that back this new class of ETPs are based on the returns of VIX futures rather than spot VIX. However.Daily 2x VIX Daily Inverse VIX MidDaily 2x VIX Daily Inverse Term Short-Term VIX ShortTerm Mid-Term VIX MidFutures ETN Futures ETN Term ETN Futures ETN Futures ETN Term ETN Tactical / Yield Generation 1x Short SPVXSP Tactical / Yield Generation 1x Short SPVXMP Description Inverse VIX ShortVIX MidVIX ShortInverse VIX January VIX MidTerm Term Term Short-Term 2021 Short. respectively.340mn 1-Month High High 0.

Our recent report on inverse and leveraged ETFs15 highlighted that these products are best used for very short-term. an amount of the front-month contract is sold to buy the second month contract.QGH[ 0HGLXP 7HUP 9. Morgan Stanley Quantitative and Derivative Strategies. October 2010.. Each day. tactical views. To limit downside risk to the amount invested. Almost the same exact strategy is used for the medium-term VIX futures indices..The short-term futures index holds a portfolio of two VIX futures that are time-weighted to deliver a constant maturity 1month VIX future. In very volatile periods. The effect of the rolldown cost is significant – the short-term VIX futures index has lost virtually its entire value over both sample periods. )XWXUHV [ . an amount of the 4-month future is sold to buy the 7-month future. has been able to maintain its value. so would be the value of the index. Exhibit 37: Replicated performance of daily-resetting 1x inverse and 2x long strategies on the short. The directional view of the investor has to be substantial relative to the volatility of the underlying..and medium term volatility indices since their inception (bottom chart). )XWXUHV . 6HS 6HS 6HS 6HS 6HS -XQ -XQ -XQ -XQ 'HF 'HF 'HF 'HF 'HF -XQ 'HF 'HF 0DU 0DU 0DU 0DU 0DU Source: Morgan Stanley Quantitative and Derivative Strategies I NVE RSE & LEVE R AG E D VIX F U TURES I N DI CES ? The expansion of VIX futures ETPs into the inverse & leveraged space has opened up new opportunities for shortterm volatility trading. on the other hand. the compounding effects inherent in the daily resetting of leverage can cause the value of these ETFs to decay quickly. the strategy would hold both in approximately equal amounts.to 7-month futures for a time-weighted exposure of approximately 5½ months. particularly for volatile underliers. the short-term index performed much stronger than the medium-term index. Otherwise... )XWXUHV . )XWXUHV .QYHUVH 6KRUW 7HUP 9.. the rapid decay of the short-term index suggests usage for tactical volatility positioning with short expected holding periods. we also would anticipate higher rolldown costs in the short-dated index. EXP OS URE TO THE VIX F UTURES ROL LDOW N Exhibit 36: Extended short. )XWXUHV . We therefore see different use cases for these two products. 28 0DU 0DU -DQ . namely (1) less volatility than the VIX and (2) rolldown losses most of the time.QGH[ Currently. all the strategies underlying these ETPs are backed by VIX futures.QGH[ -DQ -DQ -DQ -DQ -DQ -DQ -DQ -DQ -DQ -DQ -DQ -DQ -DQ -DQ -DQ 6KRUW 7HUP 9.. For example if the first two contracts were 15 and 45 days from maturity.and medium-term VIX futures indices Based on daily data from 31 Dec 07 through 23 Mar 11 [ . which targets a part of the term structure that is typically less steep. )XWXUHV -XQ -XQ -XQ -XQ 'HF 'HF 'HF 'HF 'HF 6HS 6HS 6HS 6HS -XQ 6HS 0DU 0DU 0DU 0DU 0DU Source: Morgan Stanley Quantitative and Derivative Strategies 15 See ‘Vega Times: Inverse & Leveraged ETF Primer’.. )XWXUHV [ /RQJ 0HGLXP 7HUP 9. )XWXUHV [ /RQJ 6KRUW 7HUP 9. Exhibit 36 shows the performance of the short. They therefore exhibit the properties one would expect with VIX futures. Every day.QGH[ ([WHQGHG 0HGLXP 7HUP 9. The strategy holds 4. as these contracts become shorter-dated.and medium-term VIX futures index performance (top) and actual performance (bottom) All data through 23 Mar 11. This is the crux of many of the VIX alpha trades in Section V.QYHUVH 0HGLXP 7HUP 9. Extended performance begins on 3 Jan 96 while actual performance begins on 20 Dec 05 ([WHQGHG 6KRUW 7HUP 9. We have extended the performance in the top chart using the regression-fitted VIX futures we constructed in Section V. The term structure is particularly steep for short-dated futures in upward sloping term structure environments. While we would expect the short-dated VIX futures index to be more volatile than the medium-term index. The medium-term VIX futures index. notionals are set such that if the entire VIX futures term structure were at zero. While the medium-term index is more attractive for systematic hedging strategies.

29 . as the strategy combines the compounding effect of leveraged ETFs. The indices generally perform poorly. the rolldown decay of the index. especially the 2x long strategy based on the short-term VIX futures index.and medium-term VIX futures indices.Exhibit 37 shows the replicated performance of -1x inverse and 2x leveraged strategies on short. and the index’s high realized volatility. during 4Q08. On the other hand. This is expected. the 2x long medium-term VIX futures strategy would have produced significantly higher returns due to (1) its naturally low decay and (2) increased leverage.

The basis of VIX futures is a measure of the steepness of the futures term structure. Rolldown: With an upward-sloping term structure. a variance swap will outperform for a given change in volatility. a volatility measure would be expected to rise by 2 vol points. we expect a daily return of -3%. The deltas of a call range from 0 to 1 while the deltas of a put range from -1 to 0. It is typically quoted as the futures price minus the spot price. we often express the beta of volatility to the S&P500 as the as the change in vol points for a given percentage change in the S&P 500. CVaR: VaR is a threshold such that returns at or above that level occur with a certain probability. Delta-Hedge: To hedge out the risk to an option’s value due to movements in its underlying. Volatility is well known to be mean-reverting. It is therefore a measure of the fattailedness of a return distribution. For example. and commodities. this means the spot prices is lower than the strike while the opposite is true for calls. a beta of -2 in this case would mean for a 1% fall in the S&P500. In-the-money (ITM): An ITM option is one with a strike such that if the option expired immediately. However.VII.50 for every $1 change in the underlying.50. Glossary of Terms A Reference on Commonly Used Terms At-the-money (ATM): An ATM option is any option whose strike price is equal to the current underlying spot price. has pressure to rise when it gets to a relative low. A moneyness of 95% indicates that the strike price is 5% lower than the current spot price. Delta: The option delta represents the expected change in the value of an option for a $1 change in the underlying. it is expected to increase or decrease in value by $0. the option would have a positive payout. Mean Reversion: Mean-reversion is a phenomenon when a certain index or underlying tends to be range bound. In the case of puts. In the case of puts. Out-of-the-money (OTM): An OTM option is one with a strike such that if the option expired immediately. we can use a dynamic trading strategy known as delta-hedging. notes. By extension. For example if an option has a delta of 0. In this paper.50. the option would have no payout. Hedge ratios are set such that for a $1 change in the value of an option. a -1$ opposing change in the value of the underlying is gained. Basis: The basis of a future is the difference between its price and the underlying spot price. . this typically refers to the non-linear variance swap payoff for a linear change in volatility. Beta: The beta of an asset represents the sensitivity of its expected return to the return in another asset. The CVaR measures the expected return given that the threshold is crossed. Put Option: An option that gives an investor the right but not the obligation to sell an underlying security at a set price in the future. the hedge will have to be reset after larger changes in the underlying to keep the delta exposure at zero – hence the need for a dynamic trading strategy. Compared to a (linear) VIX future. For small changes in the underlying. the rolldown is the expected loss (in volatility points) of a long vol trade if the term structure remained unchanged until expiry of 30 . For example. Variance swaps trade at a premium to linear vol instruments to compensate for this more attractive payoff. the net position has no exposure to price movements. For example. delta also represents the number of shares of the underlying required to delta-hedge the option. a 5% CVaR of -3% means that if the 5% VaR level is crossed.50 shares of the underlying against it for a delta hedge. a 5% VaR equal to -2% means that we expect daily returns of less than 2% to occur 5% of the time. and tends to fall when it reaches a relative high. this means the spot prices is higher than the strike while the opposite is true for calls. if an investor was long a call option with a delta of 0. Call Option: An option that gives an investor the right but not the obligation to buy an underlying security at a set price in the future. Exchange Traded Product (“ETP”): An ETP is a general term that encompasses exchange traded funds. they could short sell 0. Moneyness: Moneyness is a term colloquially used to express the relative value of an option strike versus the spot. For example. Convexity: Within the volatility space.

Term Structure (Futures): The futures term structure refers to the strip of futures prices for different maturities. A delta-hedged straddle can be used for pure implied volatility exposure. This replication property has made variance swaps the most popular way to trade volatility globally. The Sharpe Ratio is a well-defined when returns are normally distributed. unlike a variance swap. For example. Sharpe Ratio: The Sharpe Ratio is a measure of the relative attractiveness of an investing strategy. For example if 1-month VIX futures were 3 vol points higher than spot VIX. Skew: Skew measures how much higher OTM implied volatilities are relative to an ATM implied volatility. The shape of the term structure is often quoted as the basis. Straddle: An option strategy that goes long both a put and a call. This is unlike a volatility swap which cannot be directly replicated. a vega of $100 on a variance swap means we would expect to gain $200 if realized volatility is 2 vol points higher than the variance swap strike. Variance swaps have attractive replication properties – they can be semi-statically hedged using a strip of options. This gives a market-implied view on where the price of the underlying will be at maturity. This product is much less common in the US because it cannot be directly replicated. 31 . The vega of a variance swap represents the expected P&L swing per 1 vol point change from the swap’s strike. while the opposite is true for VIX options. A vega of $1 means if an option’s implied vol goes from 15% to 16%. the option would increase in value by $1. Volatility Swap: A long volatility swap is an OTC contract that receives the future realized volatility of an underlying in exchange for a pre-set strike. Term Structure (Volatility): The implied vol term structure is the strip of ATM implied volatilities for different maturities. The strategy benefits if the underlying either goes up or down.the trade. One of the most common and simple measures of skew is the different of the implied vols of a 90% moneyness option and a 110% moneyness option. except that the realized volatility does not begin calculation until some point later on in the future. this measure fails to fully measure the strategy’s true downside risk. A forward starting variance swap is similar. Typically. However. by comparing the annualized excess return of an asset / strategy over a risk-free rate with the strategy’s annualized volatility. Vega: The vega of an option represents the expected dollar change in the price of an option per point change in its implied volatility. for heavy-tailed returns. but loses if the underlying stays close to its current value. generally put implied vols are higher than call implied vols. Variance Swap: A variance swap is a contract whose long position receives the square of the difference between the realized volatility over the life of the variance and a certain strike. For regular index options. the S&P500 implied term structure is upward sloping while the VIX option implied vol term structure is inverted. we would be expected to lose 3 vol points in rolldown if spot VIX were unchanged at expiration. The shape of the term structure is a visual indicator or the cost of holding onto a volatility position due to rolldown. Other indicators such as CVaR or the Sortino Ratio aim to measure this risk.

in Singapore by Morgan Stanley Asia (Singapore) Pte. Morgan Stanley on its own behalf and on behalf of its affiliates disclaims any and all liability relating to these materials. Past performance is not indicative of future returns. or as a primary basis for any particular plan investment decision. Morgan Stanley Capital Group Inc. IN ADDITION. in the United States it is directed only to institutional clients and is distributed by Morgan Stanley & Co. The trademarks and service marks contained herein are the property of their respective owners. Morgan Stanley may indirectly benefit from increases in investments in hedge funds.V. In Australia. and any access to it. Morgan Stanley is not advising you to take any particular action based on the information. it can give no assurance that any forward-looking statements will prove to be correct. completeness. the material contained herein has not been based on a consideration of any individual client circumstances and as such should not be considered to be a personal recommendation. . No. legal or investment advice and may not be suitable for your specific circumstances. AND NO HYPOTHETICAL TRADING RECORD CAN COMPLETELY ACCOUNT FOR THE IMPACT OF FINANCIAL RISK IN ACTUAL TRADING. but is not a product of Morgan Stanley's Equity Research or Fixed Income Research Departments. Morgan Stanley MUFG Securities Co. Morgan Stanley & Co. and/or Morgan Stanley Asia Limited (together with their affiliates. and do not constitute.B.N. No representation or warranty is made that any returns indicated will be achieved. which accept responsibility for its contents. 233742 and/or Morgan Stanley Australia Securities Limited A. This material is not (and should not be construed to be) investment advice (as defined under ERISA or similar concepts under applicable law) from Morgan Stanley with respect to an employee benefit plan or to any person acting as a Fiduciary for an employee benefit plan. THERE ARE NUMEROUS OTHER FACTORS RELATED TO THE MARKETS IN GENERAL OR TO THE IMPLEMENTATION OF ANY SPECIFIC TRADING PROGRAM WHICH CANNOT BE FULLY ACCOUNTED FOR IN THE PREPARATION OF HYPOTHETICAL PERFORMANCE RESULTS AND ALL OF WHICH CAN ADVERSELY AFFECT ACTUAL TRADING RESULTS. HYPOTHETICAL TRADING DOES NOT INVOLVE FINANCIAL RISK. Morgan Stanley expressly disclaims any obligation or undertaking to update or revise any forward-looking statement contained herein to reflect any change in its expectations or any change in circumstances upon which such statement is based. Transaction costs (such as commissions) are not included in the calculation of returns. changes in rates of exchange may have an adverse effect on the value. this publication. Private U.Important Information and Qualifications The information provided herein was prepared by sales.K. investors should obtain the advice of their Morgan Stanley & Co. Potential investors should be aware that certain legal. 55 078 652 276. You should consult with your own municipal. International plc representative about the investments concerned. 233741. in Canada by Morgan Stanley Canada Limited.A. The investments discussed or recommended in this communication may be unsuitable for investors depending upon their specific investment objectives and financial position. and/or Morgan Stanley Asia (Singapore) Securities Pte. Any estimates. Certain assumptions may have been made in this analysis which have resulted in any returns detailed herein. which has approved of.F. Where an investment is denominated in a currency other than the investor’s currency. No.F. THERE ARE FREQUENTLY SHARP DIFFERENCES BETWEEN HYPOTHETICAL PERFORMANCE RESULTS AND THE ACTUAL RESULTS AND THE ACTUAL RESULTS SUBSEQUENTLY ACHIEVED BY ANY PARTICULAR TRADING PROGRAM. NO REPRESENTATION IS BEING MADE THAT ANY ACCOUNT WILL OR IS LIKELY TO ACHIEVE PROFITS OR LOSSES SIMILAR TO THOSE SHOWN. is intended only for “wholesale clients” within the meaning of the Australian Corporations Act. and they may change without notice. We are not commenting on the fundamentals of any companies mentioned. 67 003 734 576. Incorporated. These forward-looking statements speak only as of the date of this communication. uncertainties and other factors that could cause actual results to differ materially from those projected. without limitation.com. Changes to the assumptions may have a material impact on any returns detailed. margin requirements.L. which accepts responsibility for its contents. SOME OF WHICH ARE DESCRIBED BELOW. which is supervised by the Spanish Securities Markets Commission (CNMV) and states that this document has been written and distributed in accordance with the rules of conduct applicable to financial research as established under Spanish regulations.L. Although Morgan Stanley believes that the expectations in such forward-looking statement are reasonable. This publication is disseminated in Japan by Morgan Stanley MUFG Securities Co. trading. Third-party data providers make no warranties or representations of any kind relating to the accuracy. or timeliness of the data they provide and shall not have liability for any damages of any kind relating to such data. The products described in this communication may not be marketed or sold or be available for offer or sale in a number of jurisdictions where it is unlawful to do so. as defined in the Securities and Futures Act. Incorporated. Prices indicated are Morgan Stanley offer prices at the close of the date indicated. No representation is given with respect to their accuracy or completeness.. S. these materials. hereinafter “Morgan Stanley”). All rights reserved. financial. Chapter 289 of Singapore. projections or predictions (including in tabular form) given in this communication are intended to be forward-looking statements. ONE OF THE LIMITATIONS OF HYPOTHETICAL PERFORMANCE RESULTS IS THAT THEY ARE GENERALLY PREPARED WITH THE BENEFIT OF HINDSIGHT. Morgan Stanley and others associated with it may make markets or specialize in.. THE ABILITY TO WITHSTAND LOSSES OR TO ADHERE TO A PARTICULAR TRADING STRATEGY IN SPITE OF TRADING LOSSES ARE MATERIAL POINTS WHICH CAN ALSO ADVERSELY AFFECT ACTUAL TRADING RESULTS. Ltd. a Morgan Stanley group company. Unless stated otherwise. These materials have been based upon information generally available to the public from sources believed to be reliable. and acceptance of such document will be deemed by you acceptance of these conclusions. see https://secure. the contents of this publication in Canada.N. advice.. International PLC. For additional information. commissions and other transaction costs and changes to the assumptions set forth herein may significantly affect the economic consequences of the transactions discussed herein. By submitting this communication to you. Ltd only to those investors who are “Professional Investors” (tokutei toushika) as defined in the Financial Instrument Exchange Law of Japan and may not be redistributed to other types of investors without the prior written consent of Morgan Stanley. each of which is regulated by the Monetary Authority of Singapore and accepts responsibility for its contents of these materials.B. These materials may not be distributed in any jurisdiction where it is unlawful to do so. IN FACT. or omissions from..S. including. including within the meaning of Section 975 of the Dodd-Frank Wall Street Reform and Consumer Protection Act. or income derived from the investment. accredited investors or expert investors. HYPOTHETICAL PERFORMANCE RESULTS HAVE MANY INHERENT LIMITATIONS. opinions or views contained in this communication. This communication is a marketing communication and is not a research report.S. in Australia by Morgan Stanley Australia Limited A. in Hong Kong by Morgan Stanley Asia Limited. FOR EXAMPLE. Morgan Stanley is not acting as a municipal advisor and the opinions or views contained herein are not intended to be. brokerage or other services for those companies and may enter into transactions with them. The information and analyses contained herein are not intended as tax.ms. International plc. have or may in the future enter into principal or proprietary positions (long or short) in and effect transactions in securities of companies or trading strategies mentioned or described herein and may also perform or seek to perform investment banking. We remind investors that these investments are subject to market risk and will fluctuate in value. and has agreed to take responsibility for. Ltd.. Such estimates are subject to actual known and unknown risks. price of. accounting and legal advisors regarding the information. research reports and important disclosures. The performance data quoted represents past performance. S. opinions or views contained herein. A. accounting and tax restrictions. Morgan Stanley acts as “prime broker” and lender for a number of hedge funds. © 2011 Morgan Stanley. only to those investors who are institutional investors. or other non-research personnel of one of the following: Morgan Stanley & Co. in Spain by Morgan Stanley. Unless indicated. and in the United Kingdom it is directed only to those persons who are eligible counterparties or professional clients and must not be acted on or relied upon by retail clients (each as defined in the UK Financial Services Authority's rules) and is distributed in the European Union by Morgan Stanley & Co. We may at any time modify or liquidate all or a portion of such positions and we are under no obligation to contact you to disclose any such intention to modify or liquidate or any such modification or liquidation. A. except as provided above. all views expressed herein are the views of the author’s and may differ from or conflict with those of the Morgan Stanley Equity Research or Fixed Income Research Departments or others in the Firm. As a result. any express or implied representations or warranties for statements or errors contained in. Actual transactions at these prices may not have been effected. though it may refer to a Morgan Stanley research report or the views of a Morgan Stanley research analyst. The information provided herein has been prepared solely for informational purposes and is not an offer to buy or sell or a solicitation of an offer to buy or sell the securities or instruments mentioned or to participate in any particular trading strategy.

The distributor for streetTRACKS is State Street Capital Markets. Inc. Potential investors should be advised that the tax treatment applicable to spread transactions should be carefully reviewed prior to entering into any transaction. and closed-end funds. Please be advised that Morgan Stanley’s Institutional Equity Division frequently trades as principal and may have recently traded in the securities that are the discussed in this communication. Certain investors may be able to anticipate exercise and execute a "rollover" transaction. ETFs do not charge sales loads or other nonrecurring fees. is an important consideration. The distributor for PowerShares is Invesco Aim Distributors. should exercise occur. Nasdaq 100 Index. The distributor for Dow Diamonds. Current performance may be lower or higher than the performance data quoted.P. including combinations and straddles. Other multiple-option strategies involving cash settled options.com/about/publications/character-risks. The distributor for the SPDR Gold Trust ETF (GLD) and the KBW Regional Banking ETF is State Street Global Markets. This report is focused on US exchange-traded funds. The distributor for ProShares is SEI Investments Distribution Co. Inc. The distributor for First Trust ETFs is First Trust Portfolios L.com/etf/performance/etfperformance. which you should have read and understood prior to investing in options. The distributor for Direxion Shares and RevenueShares™ ETFs is Foreside Fund Services LLC. HOLDRs. present similar risk. Investment return and principal value of an investment will fluctuate so that an investor's shares. The distributor for UBS E-TRACS ETNs is UBS Securities LLC. The distributor for the iShares and ProShares ETFs is SEI Investments Distribution Co. The distributor for MacroSharesTM ETFs is W. Also. Any performance data shown in this communication does not reflect the deduction of a brokerage commission. may be worth more or less than their original cost. The distributor for Bear Stearns ETF is Bear Stearns & Co. The distributor for Market Vectors ETFs is Van Eck Securities Corp. Inc. ETFs are (i) redeemable only in creation unit size aggregations and may not be individually redeemed. We remind investors that these investments are subject to market risk and will fluctuate in value. We remind investors that these investments are subject to market risk and will fluctuate in value. which would reduce the performance quoted. because it involves at least twice the number of contracts as a long or short position and because spreads are almost invariably closed out prior to expiration. © 2011 Morgan Stanley. The distributor for Claymore ETFs is Claymore Securities. and (iii) redeemable on an “in-kind” basis. LLC. . Inc. he is also reducing his profit potential. your attention is called to the publication “Characteristics and Risks of Standardized Options”. Morgan Stanley may have a long or short position in these securities and may or may not have traded in a manner that is consistent with the above comments prior to its dissemination.morganstanley. The risk/ reward ratio. The distributor for the Vanguard ETFs is Vanguard Marketing Corporation. The distributor for RydexSharesTM is Rydex Distributors. Hambrecht & Co. investors should understand the nature and extent of their rights and obligations and be aware of the risks involved. if shorter) for funds discussed in this communication: http://www. it should be pointed out that while the investor who engages in spread transactions may be reducing risk.optionsclearing. This communication or any portion hereof. Past performance does not guarantee future results. All rights reserved. Copyright © by Morgan Stanley 2011. Any performance data included herein represents past performance. when redeemed.. For customers of Morgan Stanley & Co. An investor with a spread position in index options that is assigned an exercise is at risk for any adverse movement in the current level between the time the settlement value is determined on the date when the exercise notice is filed with OCC and the time when such investor sells or exercises the long leg of the spread. Additional information on recommended securities discussed herein is available on request. The risk of exercise in a spread position is the same as that in a short position. LLC. changes in rates of exchange may have an adverse effect on the value. it would clearly mark the end of the spread position and thereby change the risk/reward ratio. Where an investment is denominated in a currency other than the investor’s currency. all rights reserved.The trademarks and service marks contained herein are the property of their respective owners. 5. and Select Sector SPDRs is Alps Mutual Fund Services. Please contact your Morgan Stanley representative for the name of the name of the distributor of any funds mentioned in this sales communication that have not been included above. The prospectuses contain more complete and important information about the ETFs and should be read carefully before investing. can be viewed on the Web at the following address: http://www. The distributor for Rydex ETFs is Rydex Distributors.jsp Clients engaging in the execution structure known as Spreading should understand that Spreading may also entail substantial commissions. S&P 500. what appears to be a limited risk spread may have more risk than initially perceived. LLC. As a result. Options are not for everyone. Inc. Before engaging in the purchasing or writing of options. The distributor for ELEMENTSSM ETNs is Nuveen Investments/Merrill Lynch & Co. (ii) redeemable only through Authorized Participants. or income derived from the investment. The trading of futures or options on futures contains inherent risks. Please click on the following link for closing market price and net asset value (“NAV”) performance data for periods of 1. US Oil Fund ETF (USO). but you may pay a brokerage commission on the purchase or sale of ETF shares. hence. resold or redistributed without the prior written consent of Morgan Stanley. The distributor for WisdomTree ETFs is ALPS Distributor.html.R. The distributor for Claymore ETFs is Claymore Securities Inc. These funds are registered with the SEC and trade as a single stock under SEC exemptions. Mid Cap SPDRs SPDRs. Due to early assignments of the short side of the spread. We define exchange-traded funds as encompassing passively managed index-linked exchangetraded funds. price of. may not be reprinted. The investments discussed or recommended in this communication may be unsuitable for investors depending upon their specific investment objectives and financial position. The distributor for iPath ETNs is Barclays Capital Inc. Incorporated who are purchasing or writing exchange-traded options. and 10 years (or the life of the fund. However. Investors may obtain prospectuses for the funds described in this sales communication from the ETF distributor. That publication. Inc. including the risks pertaining to the business and financial condition of the issuer and the underlying stock. A secondary market may not exist for these securities.

Sign up to vote on this title
UsefulNot useful