Short ETFs vs. short an ETF?

Hi.

I have been running an interactive brokers automated trading system for while using the ETFs SPY and SH. Where SH is an unleveraged inverse to SPY.

I use market orders and I trade around 5 times a week.

Using interactive brokers what would be the advantages / disadvantages of shorting SPY instead of going long SH?

Wrt. for example bid/ask spread, order execution, fees and interest.

Is there a preferred way if I wish to take a short position on the S&P500?

-- Christian

 

Short SSO (2x S&P), but limit it to intra-day...don't hold over night as compounding will eventually remove profits.

 

My holding period is up to 2 weeks.

And I do not need the leverage of 2 times ETF.

 

Unless liquidity is a problem, it seems like doing what you're doing makes sense.

SPY's volume is around 300M/day and SH's volume is around 1.5M/day. Granted the average daily volume difference between your two ETFs is a factor of 100, but still. SH's volume is still so high as to be very liquid. Your bid/ask spread and fills should be about the same.

The one advantage that you didn't mention of being long-only is you can trade your system in an IRA (or any other non-margin cash account). You can't short stock in an IRA, but you *can* be long an inverse ETF.

You might take a second look at the ultra version of SPY. Average daily volume is 80M for SSO (2X S&P500) and 50M for SDS (2X inverse S&P500). Lots of liquidity and you just cut your position sizes in half to keep the same risk. This frees up margin so you can do other things with your capital.

 

Chvid,

What Trvlwanderer said about compounding removing the profits is not actually true,

I've thoroughly analyzed the ultrashort downward drift that appears if you look at some charts, namely SDS, SRS, SKF, looking for some kind of arbitrage opportunity. There is no intrinsic disadvantage to holding them over long periods- my conclusion was that recent volatility characteristics in the market have caused the effect. But this could change any time. Here's the math and more explanation on my blog (no advertisements or whatever):
SKF Inverse Construction and Volatility

I would actually recommend using a short ETF instead of going short an ETF because it will compound itself without you having to cover and reopen the position to maximize your profit.

Regards,
Max

 

Thanks for the replies.

Another issue: SPY has a dividend yield of 3.72% with dividends occuring every quarter. If I do short it - I suppose I would have to pay the dividend?

stevegee58 - I thought that must ETFs had market that set prices so liquidity would not be an issue? (Or is that naive.)

maxdama - the math is relevant for a leveraged short fund and it was actually therefore I used an unleveraged one. But it is nice that the math shows that people aren't getting cheated.

 

Chvid,

Unleveraged ETFs also compound on the short side- that's the big difference. And the unleveraged short ETFs still have excess leverage, in the Kelly criterion sense. They have been about 1.3x too volatile to maximize profits over the past year. If volatility decreases this will be less of a problem. Still, the effect is almost trivial for short holding periods so I wouldn't worry about it. Just interesting from a math perspective.

Regards,
Max

 

It seems like there are more and more ETF which have the same underlying market as future contracts we trade, for example SPY, SH which are the S&P500 ones. But now there are also GLD gold, FXE, Euro currency. Many markets have both long and short funds.

Maybe investors should trade back tested systems which are used to trade futures to invest their 401K before they become 101K's

 

Can we use a S&P500 or NASDAQ Futures System to trade ETFs?

 

Yes you can use system which trade SP500 futures and rework it so that you create a long only and a short only version and trade long version for example on SPY and short only version on SH.