Consider a two asset Himalaya under the following circumstances.

1. The option is nearing its first Observation date

2. Both the assets have similar spot levels

3. Let’s say asset A ~ 96% and asset B at ~ 100%

4. Also vols of both the assets are significantly different.

Vol A = 40%, Vol B = 20%

**Clearly the long position holder wants B to knock out since option on A is costly than option on B.**

**Keep in mind, What makes structure costly: Higher probability of B knocking out.**

When A ~ 96% and B ~ 100%

**At +1 correl **there are two possible movements,

a) **Both the spots increase:** (B moves by x, A moves by 2x), therefore B

has to move by at least 4% for A to outperform B.

b) **Both the spots decrease** (leftward movement on the spot line)

A is always the worst performer.

**At -1 correl**, again there are two possible movements ,

a) **Both move towards each other:** (B moves by -x, A moves by +2x),

therefore B has to move by at least -1.33% for A to outperform B

b) **Both move away from each other:** A is always the worst performer.

**Clearly the probability of A outperforming B is more at -1 correl and as a long position holder I don’t want A to knock out, hence structure is long correlation.**

**Reverse the spots of A and B and carry out a similar exercise to discover that structure is short correlation.**

**Hence when spot of A is less than spot of B, structure is long correlation/cross gamma and when A crosses B structure is short correlation/crossgamma.**

Hedging:

1. Buy an Outperformance digital(Call spread) on (A-B) so when A>B structure is long correlation and when spot of A is less than spot of B (outperformance is OTM) structure is short correlation. Add this cost of hedging when pricing Himalayas.

**OR**

2. Find the probability of A knocking out on the basis of vols and correlation. Sell call on A in proportion to that probability factor and do a similar exercise for B as well.

**OR**

3. Monitor the implied and realized covariance’s and re balance the delta accordingly

### Like this:

Like Loading...

*Related*

## Leave a Reply