Variance Swaps – A Delta Problem
variance swap trading, in particular going short variance swaps (receive implied pay realized)
You mean trading variance swaps as standalone strategy not as a hedge to a portfolio of
options...
There is no optimal rolling window depends on your view of the market,
there is no typical variance notional and bid/ask spread depends on the liquidity of the market.
1/ When did the instrument began trading actively ?
2/ What is the typical bid ask spread actually, and the maximum and minimum bid ask spread
observed since they trade ?
3/ What is the minimum size (in term of vega nom) in which they usualy trade ?
1) I think 1998 first varswap (and 2005 for first Option on Var) - very active from 2003 (at
least on indices)
2) Index: SX5E, DAX, SPX <1vol NKY ~1vol and less liquid like NDX, KS200,... 2-3vols;
Equity: poor liquidity at the moment: Major stocks 3-5vols wide
3) Index: 100k-300k and Equity ~10k
Most books I have gone through mention zero delta exposure in Variance swaps. My simple
reasoning is this...
Implied Vol today is 25%
Spot Level today is 100
Next day Spot changes by a wide margin to 110
Realized Vol Becomes 27%
Pnl is actually being realized due to spot movement? Difference between realized and implied
vol...
What exactly is delta Pnl for these daily moves?
Animesh Saxena
Associate
Thought that a variance swap did not depend directly on the underlying, hence delta is zero. If
Vol does not change, and spot changes, your variance swap has the same price...
Intraday there is a delta if your variance swap is being marked based on daily movements.
If you were long the swap and mid-day the product was trading higher, then you would profit
if the underlying continued to move up and, if the underlying moved down from its intraday
price then you would be giving profits back.
I suppose in the continuous time models this isn't a problem/factor.
Well, it's exactly given by the change in the mark-to-market formula, which
decomposes into 2 terms: w(t) times the last daily realized squared return,
and (1 - w(t) times the change in the market's implied variance. Here
w(t) is the time weight at time t in the swaps lifetime.
Hi,
I am studying vol swap vs. Var swap.
To hedge a Var swap you need out-of-the-money options which are not necessary available in
an illiquid market.
Vol swap needs less out out-of-the-money options but you need to buy ATMF Volatility.
First question is, what does it mean to buy ATMF volatility? I pay x at time 0 and receive the
implied vol at time T? I pay x for an option at time 0 and P&L is (implied vol - x) at time T?
I am trying to proove that a delta hedged straddle allows someone to buy this ATMF
volatility:
Based on the daily P&L formula:
How can I get to what I want?
I read on the forum and based on B&S:
It still doesn't leed to what we want.
Another article (from JP Morgan) states that we need to buy ATM put and call, and then:
Can you please help me getting in the right direction?
Thanks a lot,
Any clues ?
If we assume a perfect B&S world...
I do understand that being long a straddle means if the vol increases enough then your P&L
will be positive.
Assuming no dividends, no interest rates,
at time 0, ATM Call = ATM Put
Therefore Straddle = 2 * ATM Call
and then? Do we hedge it to buy exactly the implied vol ? I guess yes, because we need to be
independant from the udl moves...
Thanks for your help,
Jonathan
Quote
First question is, what does it mean to buy ATMF volatility? I pay x at time 0 and receive the
implied vol at time T? I pay x for an option at time 0 and P&L is (implied vol - x) at time T?
ATMF is an option struck at where the forward price of the asset is. if if the forwrd price is
102, then strike price = 102
the equations that you show relating the premium of a call to the time tio maturty and spot is a
well-known result from the BS equation if you substitute F=K and mess around with
cumulative normal.
the best place to start with varinace swaps is the Derman paper or play around with the log
contract rather ....
I dont think you can - the p+l is path dependent.
you should definitely delta hedge it otherwise you cannot monetise the volatility.
I think you are just getting confused by what you're calling ATMF vol. As Daveangel said an
option struck ATMF is an option whose strike is ATMF, and its volatility is quoted in the
market right now.
So buying ATMF volatility simply means that you are buying the volatility of an option (or a
portfolio of options) struck ATMF.
The interesting thing with a straddle ATMF is that the delta is equal to zero (almost, not
exactly in the lognormal BS framework, and exactly equal in the normal world, but we don't
care here), so effectively your straddle looks like a pure volatility trade (true only locally of
course, as any move in the underlying will create some delta exposure).
Thanks for your answer.
Straddle has a delta = 0 at time 0 because striked at S_0.
Assume the underlying moves it does create some delta exposure (gamma is non negative).
Therefore we need to implement a delta hedging strategy (as daveangel says also to have a
path dependant strategy).
We are left with the P&L describes before.
Gamma cannot be expressed easily (Gamma Straddle = 2 * Gamma Call).
Well, at time 0, we can do something because strike is equal to the udl and then d_1 in B&S
can be simplified.
At time t, S_t is clearly different from the strike. Therefore any approximation of the premium
doesn't stand anymore.
Also I tried to integrate between 0 and t the P&L described before but
is not a Wiener process or something I know.
Even I do understand the basic idea, I cannot find any rigorous proof...
well the integral
is no more than
which in the BSM world of constant vol simplifies to
Elsewhere...
H i
My question is with regards to practical hedging of “vanilla” variance swaps.
I know that in theory on could hedge statically with a portfolio of calls and puts (Derman/Carr
papers).
However, I have been told that in practise one only buys 2-3 options around ATM and delta
hedges these options daily to “catch” the variance. This is a dynamic hedge, i.e. one changes
in option position depending on the movement of the underlying.
I am wondering if anyone has some information about this hedging strategy. Things that
would be of interest is the criteria for number of options, their strikes, the weightings and
changing the position in the portfolio depending on the underlying.
Any references/rules of thumb would be of great interest and appreciation.
Cheers,
K
Data on the Market State for Var Swaps
From what I've heard the liquidity on the single-name variance swaps market was reduced
dramatically, trading was concentrated mostly on indexes.
SG has recently suggested the "American Variance Swap", with following features:
- standard variance swap
- monthly exercise dates (for ~each expiry)
- in case of early exercise, you would receive the realised PnL only {i.e. (Vega Notional) * (%
Elapsed Time) * (RV^2 - K^2) / 2K} no matter where implied is marked
Regading the pricing, I suppose that this product is equivalent to a long position in the vanilla
variance swap + short a straddle on forward-starting var (americanised and with notional
decaying...?)
Any papers or references to this?
nope - just remember a note published a few months ago- did not find this product particularly
interesting...
Googled SG American Swaps and got this:
http://www.moritz-seibert.de/wp-content/plugins/downloads-
manager/upload/RISK_An_Aversion_To_Variance.pdf
Not sure who its written by but it pretty much answers my first question at least. Thanks a ton
