IVolatility Egar Service

IV_Trader · Dec 1, 2005

Quote from IV_Trader:

yep , but actually I will make money if market goes up or down as long as all (or most) components will go to the same direction (collect IV difference from basket's calls and Index puts). Can happens in non-reporting month.

so far , so good. Almost all (27) Dow stocks are up.

mysticman · Dec 1, 2005

That's nice but it would help if we knew your position. Did you really buy the stocks and sell ITM calls, and if so why did you not just sell puts? How far ITM(calls) or OTM(puts) are your options? Did you do this for a credit or debit?

IV_Trader · Dec 1, 2005

Quote from mysticman:

That's nice but it would help if we knew your position. Did you really buy the stocks and sell ITM calls, and if so why did you not just sell puts? How far OTM are your options? Did you do this for a credit or debit?

I stated my position before:
1. Bought Dow stocks
2. Sold slightly ITM components call with ratio < 1 vs. stocks
3. Bought ATM DIA puts for full hedge of portfolio.
A lot of room for adjustments here.
BTW , regular EGAR position (sell Index , buy basket) looks like an absolute disaster so far , about 5.5 minus (and down 8% vs. benchmark/market).
Good luck all.

Profitaker · Dec 2, 2005

IV_Trader

Was it necessary to 100% replicate the index ? Could you not have chosen a broadly representative basket of stocks from various sectors ?

Did you buy the stocks in their correct weightings ?

How did you balance the long DIA Puts ? Was it by $ delta or what other method ?

I cannot understand why you chose to go long stock and short stock calls rather than simply short stock puts. Is it me, or have I missed something here ?

Profitaker · Dec 5, 2005

Quote from sle:

If you want to be really correct, you should use separate correlation for every pair, but it is not too hard to derive the same formula for the index to component correlation (the correlation of the two components to each other would be a weighted ratio of the two correlations to the index).

..............Corr to Index...........Index Weight
Stock A.......0.60..........................0.10
Stock B.......0.70..........................0.05

What would be the correlation between stock A/B ?

TIA.

Profitaker · Dec 6, 2005

I've looked at it from all angles and can't see how it's done.

Anyone know ?

sle · Dec 6, 2005

Quote from Profitaker:

..............Corr to Index...........Index Weight
Stock A.......0.60..........................0.10
Stock B.......0.70..........................0.05

What would be the correlation between stock A/B ?

TIA.

yes. if you do it on a spreadsheet, make sure it is the correlation of log change: ln px(today)/px(yesterday)

Profitaker · Dec 6, 2005

Quote from sle:

yes. if you do it on a spreadsheet, make sure it is the correlation of log change: ln px(today)/px(yesterday)

Still can't see how it could be calculated by using the weighted ratio of LN price change correlations rather than stock price correlations - correlation values of LN price changes bear no resemblance to the actual correlation value of 2 x stocks.

Anyway, that's probably irrelevant, as I understood from your initial post that you could derive a stock/stock correlation by using stock/index corr / index weighting, i.e. using just the info below;

..............Corr to Index...........Index Weight
Stock A.......0.60..........................0.10
Stock B.......0.70..........................0.05

Stock to index corr is readily available on the net so I was hoping to avoid having to manually compare every stock/stock corr in the index.

I'm working on the FTSE100 which would need 4950 different correlation values.

Any shortcuts / tips etc ?

Thanks again.

Profitaker · Dec 11, 2005

Quote from riskarb:

1:1 correlations will produce a pure arbitrage trading dispersion, but you'll be producing a lock of a -100bp or worse.

Isn't that a contradiction ? If you have an implied correlation of +1 or more and replicate the mother index 100% with the components, you cannot make any intrinsic loss. Where the implied correlation is more than +1 the trade would be opened for a credit and that $ value would actually be the arb, where the worst case scenario is the +1 correlation holding. Of course if the stocks then dispersed, Christmas has come early.

Anyone ever opened a +disperion with 100% replication for a credit ?

Also, at what implied correlation value would you think about running a reverse dispersion trade ?

Thanks.

riskarb · Dec 11, 2005

Quote from Profitaker:

Isn't that a contradiction ? Thanks.

Slippage, loss of edge in execution.