Forums - Formula For Volatility Adjusted SPY Hedge

Forums (http://www.elitetrader.com/vb/index.php)
- Trade Management (http://www.elitetrader.com/vb/forumdisplay.php?forumid=60)
-- Formula For Volatility Adjusted SPY Hedge (http://www.elitetrader.com/vb/showthread.php?threadid=239272)

Posted by jayjay121 on 03-22-12 02:17 PM:

Formula For Volatility Adjusted SPY Hedge

hi, can anyone help me?

Ill explain what Im trying to achieve.

Say Im mostly long or short a basket of 20 stocks. To protect myself how would i set up a SPY hedge?, what is the formula if any?........ideally something i can set up in Excel would be good.

When trading most days i have equal shorts and longs so pretty market neutrel, but on some days i get i can get lopsided and be mainly directional with my stocks. So on these days how can i reduce my risk with a SPY hedge?, but also how do i know how many SPY shares to buy/sell against what stocks i already have in the basket?, to make the basket as market neutrel as possible. Please help as im struggling to get my head round the formula needed to start a spreadsheet. Its the knowing the correct postion size at the time of the hedge that is getting me.

I suppose Im looking for a formula that will put me pretty dollar neutrel so the SPY hedge matches with the rest of the basket.

kind regards,

Posted by rmorse on 03-22-12 02:30 PM:

I would look at cash neutral. If your basket of stocks is long $250K, hedge it with SPY or SPY options based on that value.

EG. long portfolio value / current SPY value = hedge ratio

$250,000/139 (SPY price)=1798

So either sell 1798 shares of SPY to hedge out market risk, or 1798 deltas in SPY. There are different ways to beta adjust etc, but this is a quick and dirty way to back test your hedge. With this, your goal would be to hedge out beta risk and have alpha risk left.

Bob

__________________
Robert L. Morse
Business Development

VICTOR SECURITIES
285 Grand Avenue
Englewood, NJ 07631
rmorse@victorsecurities.com
office: 646-545-3860
www.linkedin.com/pub/robert-morse/6/8a7/617
__________________

Posted by jayjay121 on 03-22-12 03:18 PM:

hi, thanks for the reply, this simple formula looks just what ive been missing. I knew it would be something simple.

what about volatility?, does this formula keep me volatility adjusted?

kind regards

For your information im not pair trading, just sometimes i get lopsided but dont want market direct to be part of the strategy, will your formula still be valid?

Posted by Rationalize on 03-22-12 04:05 PM:

Quote from jayjay121:

hi, thanks for the reply, this simple formula looks just what ive been missing. I knew it would be something simple.

what about volatility?, does this formula keep me volatility adjusted?

kind regards

For your information im not pair trading, just sometimes i get lopsided but dont want market direct to be part of the strategy, will your formula still be valid?

It depends entirely on what you own vs the SPY basket.

If you owned the whole basket in the right weights, you'd be pretty neutral on most fronts.

Start by netting the components, to see what you have not flattened.

Posted by rmorse on 03-22-12 04:06 PM:

Quote from jayjay121:

hi, thanks for the reply, this simple formula looks just what ive been missing. I knew it would be something simple.

what about volatility?, does this formula keep me volatility adjusted?

kind regards

For your information im not pair trading, just sometimes i get lopsided but dont want market direct to be part of the strategy, will your formula still be valid?

No, not volatility adjusted. I don't recommend over hedging. In my opinion, simple is better.

Posted by Rationalize on 03-22-12 04:08 PM:

Quote from rmorse:

No, not volatility adjusted. I don't recommend over hedging. In my opinion, simple is better.

Not anything adjusted.

Need to net the deltas as step 1.

This basket isn’t that basket.

Posted by newwurldmn on 03-22-12 04:22 PM:

get beta flat. not notional flat.

Posted by Rationalize on 03-23-12 03:24 AM:

Quote from newwurldmn:

get beta flat. not notional flat.

Isn't that making a pretty big correlation assumption?

Interested in your perspective.

Posted by newwurldmn on 03-23-12 10:59 AM:

Quote from Rationalize:

Isn't that making a pretty big correlation assumption?

Interested in your perspective.

Yeah it is. It's also making a volatility assumption. But if you are net long 100k of bac clearly 100k of spy won't be a sufficient hedge. And of you are short 100k of NLy the. 100k of spy would be an over hedge.

Whatever ratio you use will be more art and science, but you need to account for the correlation an the vol somehow and beta does that kind of.

Posted by rmorse on 03-23-12 11:07 AM:

I think the issue here is simple. You've done your homework and put together a portfolio of longs and shorts. You believe the stocks your short will under perform the market. You believe the stocks your long will perform better then market. I would want to take out market risk. If SPY does not track with your portfolio, and another ETF does, use that.

If you're a billion $ fund, would I get more complicated, maybe. A simple dollar neutral strategy will help you with highly correlated large marker moves.

Bob
BTW:newwurldmn...I like it when we always have a different opinion. It's no fun to agree.

Posted by Rationalize on 03-23-12 11:09 AM:

Quote from rmorse:

I think the issue here is simple. You've done your homework and put together a portfolio of longs and shorts. You believe the stocks your short will under perform the market. You believe the stocks your long will perform better then market. I would want to take out market risk. If SPY does not track with your portfolio, and another ETF does, use that.

If you're a billion $ fund, would I get more complicated, maybe. A simple dollar neutral strategy will help you with highly correlated large marker moves.

Bob

If an ETF tracks your portfolio, you trade the ETF.

Posted by Rationalize on 03-23-12 11:12 AM:

Quote from newwurldmn:

Yeah it is. It's also making a volatility assumption. But if you are net long 100k of bac clearly 100k of spy won't be a sufficient hedge. And of you are short 100k of NLy the. 100k of spy would be an over hedge.

Whatever ratio you use will be more art and science, but you need to account for the correlation an the vol somehow and beta does that kind of.

kind of

I'd still net the deltas first, then hedge the residual.

Issue here than annoys me is the SPY hedge may add risk, depending on the OP's basket.

Posted by newwurldmn on 03-23-12 11:22 AM:

Quote from Rationalize:

kind of

I'd still net the deltas first, then hedge the residual.

Issue here than annoys me is the SPY hedge may add risk, depending on the OP's basket.

Yeah. The hedge would be:

SUM(beta*quantity*price)/priceSPY

So you if you are long BAC and short GS in equal notional then the GS would hedge a good chunk of the BAC.

You are right about not getting a good hedge. You will have a lot of basis risk - but that's why the OP has the position in the first place. He's trying to eliminate the market risk, and he may not even do that.

Posted by Rationalize on 03-23-12 11:24 AM:

Quote from newwurldmn:

Yeah. The hedge would be:

SUM(beta*quantity*price)/priceSPY

So you if you are long BAC and short GS in equal notional then the GS would hedge a good chunk of the BAC.

You are right about not getting a good hedge. You will have a lot of basis risk - but that's why the OP has the position in the first place. He's trying to eliminate the market risk, and he may not even do that.

Agreed.

Posted by jayjay121 on 03-23-12 02:20 PM:

i realise that i will not be able to eliminate all risk. The idea is to protect myself if things go completely one side, even then i would lose and not be able to completely save the day, but difference is i want to still be in business and not wiped out, so must hedge for black swan event.

ive just backtested the simple formula and my large winner days were smaller with the hedge and the large losing days were also smaller as expected. So curve is smoother but return is less. But i take the smoother curve anyday.

Im not pair trading. I want to be able to select all stocks that fit my critria on that day without worrying too much about main market direction. This means on some days i will be too long or too short. yes these days do produce the best win days but when wrong the losses can be tough.

So with all that in mind is the first formula (total postions value/SPY price) the most ideal for my purpose?, or should i use the one above with the beta added in it?

Posted by Rationalize on 03-23-12 02:25 PM:

How is this not a pairs trade?

Posted by HurricaneUS on 03-23-12 03:11 PM:

Quote from jayjay121:

i realise that i will not be able to eliminate all risk. The idea is to protect myself if things go completely one side, even then i would lose and not be able to completely save the day, but difference is i want to still be in business and not wiped out, so must hedge for black swan event.

ive just backtested the simple formula and my large winner days were smaller with the hedge and the large losing days were also smaller as expected. So curve is smoother but return is less. But i take the smoother curve anyday.

Im not pair trading. I want to be able to select all stocks that fit my critria on that day without worrying too much about main market direction. This means on some days i will be too long or too short. yes these days do produce the best win days but when wrong the losses can be tough.

So with all that in mind is the first formula (total postions value/SPY price) the most ideal for my purpose?, or should i use the one above with the beta added in it?

what happens if one of your stocks reports an accounting irregularity that was overlooked by the auditors and the company is in fact insolvent?

in addition to market hedging you will also need company risk hedging via options on oversized positions and/or proper money management (ie never risk more than 2% of account on any one position) and/or proper diversification...

..do you see how this can get very complicated very quickly?....

Posted by jayjay121 on 03-23-12 03:34 PM:

hi, yes i see how it can become complicated so simple is best. It does not have to fit perfectly.

It is not a pair trade because pair trading takes into account sector, correlation and normally goes against another stock within the pair. Im just selecting stocks based on my critria, and this could give a very different basket each day. So im not concerned with sector or correlation when my strategy picks a stock. I am though concerned with overall market direction therefore i need the hedge if i become too long or too short within the basket.

Correct me if im wrong as i felt this is not pair trading. Im not looking for pairs specifically. It could be anything.

Also for the money management part, yes im only risking 1% of my account per trade based on average daily range.

Im not very good with math formula's hence why i ask the question. Just looking for a good formula i can use in excel that quickly protects me if needed.

kind regards,

Posted by Rationalize on 03-23-12 03:44 PM:

Quote from jayjay121:

hi, yes i see how it can become complicated so simple is best. It does not have to fit perfectly.

It is not a pair trade because pair trading takes into account sector, correlation and normally goes against another stock within the pair. Im just selecting stocks based on my critria, and this could give a very different basket each day. So im not concerned with sector or correlation when my strategy picks a stock. I am though concerned with overall market direction therefore i need the hedge if i become too long or too short within the basket.

Correct me if im wrong as i felt this is not pair trading. Im not looking for pairs specifically. It could be anything.

Also for the money management part, yes im only risking 1% of my account per trade based on average daily range.

kind regards,

So, your expecting the longs to go up, and the shorts to go down in isolation. Not in relation to each other.

Which means you're expecting your selection produces absolute winners per trade, not relative winners within a paired basket.

On net here, we seem to be hedging your stock selection criteria.

Posted by jayjay121 on 03-23-12 04:27 PM:

yes correct.

so are you saying im better off without the hedge?

but what if the basket becomes too one sided?. Stocks do move on their own merit but if i feel i may need something if something major is happening with the main cash markets, dont you agree?

Posted by newwurldmn on 03-23-12 04:35 PM:

Quote from jayjay121:

yes correct.

so are you saying im better off without the hedge?

but what if the basket becomes too one sided?. Stocks do move on their own merit but if i feel i may need something if something major is happening with the main cash markets, dont you agree?

You shouldn't hedge because you are saying a stock will go up or down independent of the market. That's the view you are expressing. Or am I reading this wrong?

Posted by Rationalize on 03-23-12 04:36 PM:

Quote from jayjay121:

yes correct.

so are you saying im better off without the hedge?

but what if the basket becomes too one sided?. Stocks do move on their own merit but if i feel i may need something if something major is happening with the main cash markets, dont you agree?

If you're punting on directional signals produced from a single stock model, where every selection is made in isolation, then you profit or blow up by the efficacy of that directional selection model.

Assuming your signals are better than random, but fail occasionally, then choosing more than 1 stock is a good idea to diversify away that risk. An occasional loser should be bailed out by the vast majority winners.

This is the whole idea of a directional model, no? To pick a direction that's usually correct?

It all hinges on your signals being better than random.

Posted by jayjay121 on 03-23-12 05:03 PM:

yes my signals are better than random and proven to be so over time, so should i hedge or not?

My equity swings can be huge though, thats why i raise the question of a hedge, would it not smooth?

Posted by Rationalize on 03-23-12 05:05 PM:

Quote from jayjay121:

yes my signals are better than random and proven to be so over time, so should i hedge or not?

My equity swings can be huge though, thats why i raise the question of a hedge, would it not smooth?

Those sentences conflict.

Posted by jayjay121 on 03-23-12 05:14 PM:

not really,

i can win big if im too long but also lose big if im too long.

i can still win if im hedged but reduce big losing days if im hedged somewhat.

Posted by newwurldmn on 03-23-12 05:54 PM:

Quote from jayjay121:

not really,

i can win big if im too long but also lose big if im too long.

i can still win if im hedged but reduce big losing days if im hedged somewhat.

When the market rallies do your shorts lose money?
When the market sells off do your longs lose money?

Do you find yourself net long when the market is rallying?
Do you find youself net short when the market is selling off?

Posted by jayjay121 on 03-23-12 06:39 PM:

[QUOTE]Quote from newwurldmn:

[B]When the market rallies do your shorts lose money?

generally yes, but not always.

When the market sells off do your longs lose money?

generally yes, but not always.

Do you find yourself net long when the market is rallying?

no.

Do you find youself net short when the market is selling off?

no.

Posted by newwurldmn on 03-23-12 06:45 PM:

Quote from jayjay121:

[QUOTE]Quote from newwurldmn:

[B]When the market rallies do your shorts lose money?

generally yes, but not always.

When the market sells off do your longs lose money?

generally yes, but not always.

Do you find yourself net long when the market is rallying?

no.

Do you find youself net short when the market is selling off?

no.

You should beta hedge SPY.

Posted by jayjay121 on 03-23-12 06:52 PM:

ok thanks so the formula is?

just to be sure thanks

Posted by newwurldmn on 03-23-12 07:17 PM:

Quote from jayjay121:

ok thanks so the formula is?

just to be sure thanks

-1*Sum(Beta*quantity*price)/priceSPY

Please don't take offense to this question, but how do you build a presumably sophisticated long short system and not understand beta?

Posted by jayjay121 on 03-23-12 08:24 PM:

Quote from newwurldmn:

-1*Sum(Beta*quantity*price)/priceSPY

Please don't take offense to this question, but how do you build a presumably sophisticated long short system and not understand beta?

absolutely good question. I have self taught myself everything i know and i admit there are things im ignorant too, but im never afraid to ask or learn new things. Infact in asking is how ive learnt, now i before i posted this question i really never had a clue how to hedge, but through asking i now have new knowledge, and basically this is how ive got to this point today and also through first hand trading experience.

I knew a little about beta through and the scoring of each stock correlation and volatility.

Posted by jayjay121 on 03-23-12 08:30 PM:

sorry to sound like a complete fool but i dont understand this part of the formula.

-1*Sum

Do I understand this part correctly?:

(Beta*quantity*price)/priceSPY

beta score on FINVIZ screener x share quantity x stock price/priceSPY

many thanks you have been very helpful.

All times are GMT. The time now is 08:18 AM.