Author Topic: Shorting the shorts? Leveraged ETFs are garbage (viable exploit?)  (Read 2117 times)

Financial.Velociraptor

  • Handlebar Stache
  • *****
  • Posts: 2424
  • Age: 52
  • Location: Houston TX
  • Devour your prey raptors!
    • Living Universe Foundation
Has anyone here ever tried playing both ends against the middle on leveraged ETFs?

Here is my thesis in a nutshell: these leveraged instruments are poorly constructed and all lag what they are intended to track.  Usually badly.  "Many" fund providers offer 2x and even 3x (the legal max in US - Some European instruments are 10x the daily!) that exist in both long and inverse flavors.   Say for example 3 times the daily movement in a commodity and a 3 times the daily inverse in the same commodity.

Say you short both the long and inverse ETF in equal dollar amounts.  You are now fully hedged against what the ETFs are theoretically supposed to capture.  But you capture the inefficiency in construction both coming and going.  If you rebalanced regularly (say every 35 days to avoid wash sale), would the net drag here be greater than the cost to borrow?  Basically, the worse a product is, if it has an inverse twin, the better it is for this exploit.  It's almost a moral imperative to short these bastards.

I'm see if anyone here has experience first. The idea seems to obvious to not have been arbitraged away.  If no one can shoot it down, I think I'll try (1000/1000) pair of ETFs. 

TreeLeaf

  • Handlebar Stache
  • *****
  • Posts: 1884
Re: Shorting the shorts? Leveraged ETFs are garbage (viable exploit?)
« Reply #1 on: December 07, 2022, 02:32:43 PM »
Just thinking out loud.

Is there some way to back test this strategy and then add in the cost to borrow and see what the returns would have been historically if you followed this strategy?


MustacheAndaHalf

  • Walrus Stache
  • *******
  • Posts: 7151
  • Location: U.S. expat
Re: Shorting the shorts? Leveraged ETFs are garbage (viable exploit?)
« Reply #2 on: December 07, 2022, 02:39:43 PM »
Over on reddit someone claimed inverse leveraged ETFs always do worse than leveraged ETFs, which I found hard to believe.  When I started looking at the data, they seemed to be right, and then I had an insight: recoveries are always larger, and inverse ETFs always short the recovery.

In another thread I showed TMV (-3x TLT) losing 7% and TMF (+3x TLT) losing 5% from volatility drag since Sept 2 (TLT only lost 0.2%).  Using that data to run the suggested experiment:

$1000 short TMV, now worth $930
$1000 short TMF, now worth $950
$2000 cash versus $1880 short, profit of $120 before borrowing costs.  So that appears to be about a 6% profit, but its entirely from volatility drag.

While I wanted to claim that 1x inverse ETFs don't suffer from volatility drag, here's some evidence I might be wrong, starting on Oct 13 to now:
ARKK: $34.17 -> $34.35 is -0.5%
SARK: $67.24 -> $62.50 is -7.0%

Even if you add +0.5% to both, giving a more accurate +0% -6.5%, the difference is significant.  With put options I've noticed the time value is 3x higher for 3x leveraged ETFs.  So I wonder if borrowing costs are also 3x higher for 3x ETFs.

MustacheAndaHalf

  • Walrus Stache
  • *******
  • Posts: 7151
  • Location: U.S. expat
Re: Shorting the shorts? Leveraged ETFs are garbage (viable exploit?)
« Reply #3 on: December 07, 2022, 02:52:00 PM »
A strong warning: this next investment is so risky, you will literally lose 90% of your investment in one year if you guess wrong.  WEBS was -87% in 2020, and WEBL is -90% YTD.  If you invest in them, expect to lose everything - you were warned.

But Financial.Velociraptor isn't seeking to have a position in either - his goal is to establish equal and opposite positions that profit off volatility drag.  I looked at Dow Jones Internet (FDN) from June 16 until now:

FDN: $124.77 -> $124.36 is -0.3%
WEBS: $57.71 -> $36.92 is -35.8%
WEBL: $7.89 -> $6.02 is is -23.7%

If you shorted WEBS and WEBL equally:
$1000 cash versus short $-642 WEBS
$1000 cash versus short $-763 WEBL
$2000 cash versus short $-1405 positions, 30% profit in about 6 months

In the past I've found where Schwab and IBKR list their borrowing costs, which is the key missing piece of information.  Back in 2020, Financial.Velociraptor reminded me to check the borrowing costs of Hertz shares, and I discovered it was over 100%/year, or 2%/week.

I wonder what the borrowing costs are for WEBS, WEBL, SQQQ, TQQQ, TMF, TMV, and ARK, SARK ?

Financial.Velociraptor

  • Handlebar Stache
  • *****
  • Posts: 2424
  • Age: 52
  • Location: Houston TX
  • Devour your prey raptors!
    • Living Universe Foundation
Re: Shorting the shorts? Leveraged ETFs are garbage (viable exploit?)
« Reply #4 on: December 07, 2022, 03:24:34 PM »
@MustacheAndaHalf

I haven't checked the borrowing costs of the 'usual suspects' at IBKR yet.  You could theoretically dodge that with synthetic shorts but then it becomes difficult to short equal amounts and I bet the P/C parity "skew" is high (which would indicate high borrowing costs, else arbitraged away).   I do know that with leveraged instruments IBKR docks you for double or triple margin as per the underlying leverage.  So, 1000 each on two 3x long/short ETFs, would take up 6k of margin!  And maybe this has something to do with why no one seems to be doing it. 

I used to short VXX.  Borrowing rates were all over the map. During low vol, when it was steadily declining, it would be a few percent a year, under vol spikes it would hit hundreds of percent! (but usually for only a few days).  This suggests you want a commodity pair that is highly leveraged BUT slow moving.  That is probably an agricultural product.  Wheat for example can make huge moves up or down over a year but doesn't whipsaw by a dozen percent back and forth daily.  The people who are hedging ag know with quite a bit of certainty what the trend is going to be based on last year's results, weather forecast, and futures for fertilizer and feed.


MustacheAndaHalf

  • Walrus Stache
  • *******
  • Posts: 7151
  • Location: U.S. expat
Re: Shorting the shorts? Leveraged ETFs are garbage (viable exploit?)
« Reply #5 on: December 07, 2022, 03:38:18 PM »
Here are borrowing rates at Schwab right now, using their StreetSmart Edge app:

TMF 0.00% "avail",  TMF 4.25% "limited"
WEBS 14.0% "limited", WEBL 0.25% "limited"
SQQQ, TQQQ - "not available"
UPRO 0.00% "avail", and either SPXU (2.25% limited) or SPXS (1.00% limited)
SARK 0.00% "avail", TARK 3.25% "limited"
BITO 3.00% "limited, BITI 62.0% "limited"

If you want to start slow, I'd suggest UPRO / SPXS in equal amounts.  Although those are from ProShares and Direxion, I very much doubt they will have tracking error using the S&P 500 index as a benchmark.

Thanks for the idea - I have lots of idle cash at Schwab (1/3rd put options, 2/3rds cash), so I can afford to lock some up in short positions.  I'll probably go with SARK / TARK both because of volatility and the higher difficulty in not being 1:1.

MustacheAndaHalf

  • Walrus Stache
  • *******
  • Posts: 7151
  • Location: U.S. expat
Re: Shorting the shorts? Leveraged ETFs are garbage (viable exploit?)
« Reply #6 on: December 07, 2022, 03:51:37 PM »
I used to short VXX.  Borrowing rates were all over the map. During low vol, when it was steadily declining, it would be a few percent a year, under vol spikes it would hit hundreds of percent! (but usually for only a few days).
My riskiest investment right now are VIX call options.  When I predicted a surprise CPI rise in June, my VIX calls were an insanely good investment.  Which was entirely gained by luck - two wrong calculations made a correct prediction, go figure.

Since I believe things don't change, and like to have the market face punch me for that belief, I bought a tiny position in VIX calls last month, and took a nearly 100% loss.  It went from +86% to -88% in one day, so -170% one day return.  I pretty much lost that investment, which is why I keep the position size very small.  My earlier VIX profits more than paid for this attempt.

Like last time, I'm up over +70% in a few days, all of which could vanish if the CPI print falls dramatically.  So I might need to pace my selling over a few days: pre-CPI, pre-Fed, and after both.

MustacheAndaHalf

  • Walrus Stache
  • *******
  • Posts: 7151
  • Location: U.S. expat
Re: Shorting the shorts? Leveraged ETFs are garbage (viable exploit?)
« Reply #7 on: December 07, 2022, 05:44:35 PM »
I found the flaw in this idea: look at 2019 data.  Whenever UPRO doubles, it exceeds the size of your initial short position, and the overall investment is a loss.

Looking at 2019 prices from Yahoo Finance:
UPRO gained +111.4% while SPXS lost -59.1%.
UPRO $1000 -> $2114, SPXS $1000 -> $409
Short positions become worth $2523 versus the initial $2000 invested, for a loss of 26% plus borrowing fees.

Reducing leverage to 2x with SSO and SDS doesn't solve the problem.  Using the same start/end price comparisons, SSO + SDS loses 12%.  And although borrowing isn't available for TQQQ + SQQQ, shorting that pair in 2019 produced a -40% return before borrowing fees.

Another way to look at this, consider two investments with opposite returns.  One grows +20%/year, the other -20%/year.  After 3 years:
1.2 x 1.2 x 1.2 = 1.73
0.8 x 0.8 x 0.8 = 0.51
0.51 + 1.73 = 2.24

A large amount compounded produces a higher gain than a small losing amout taking additional losses, even though the percentages involved are the same.  In the above theoretical exercise, the result is a 12% loss.

Financial.Velociraptor

  • Handlebar Stache
  • *****
  • Posts: 2424
  • Age: 52
  • Location: Houston TX
  • Devour your prey raptors!
    • Living Universe Foundation
Re: Shorting the shorts? Leveraged ETFs are garbage (viable exploit?)
« Reply #8 on: December 07, 2022, 07:13:39 PM »
That makes a lot of sense in EMH sense.  Betting against volatility works when Vol is low, but costs you when it is "high".  I was doing some math and the concept seems to work only when prices are moving "slow".  Any big move and your long  position wipes you out over and above your short gains.

vand

  • Magnum Stache
  • ******
  • Posts: 2514
  • Location: UK
Re: Shorting the shorts? Leveraged ETFs are garbage (viable exploit?)
« Reply #9 on: December 08, 2022, 02:34:29 AM »
Why would you SHORT both sides of the trade?  In any short position your downside is unlimited while your upside is limit to the exposure.

So if you rebalance "opposing" trades to be supposedly market neutral and then the market does a moonshoot - your x3 short leverage is going to leave you hundreds of percent in loss while your x3 long leverage fund is going to go to near zero - its an assymetrically bad trade.


That said, I'm not sure that the playing double long is a slam dunk either, as the daily volatility will erode the returns on both in a flattish market.


Remember whoever is offering these derivative-based instrument probably isn't an idiot.



Financial.Velociraptor

  • Handlebar Stache
  • *****
  • Posts: 2424
  • Age: 52
  • Location: Houston TX
  • Devour your prey raptors!
    • Living Universe Foundation
Re: Shorting the shorts? Leveraged ETFs are garbage (viable exploit?)
« Reply #10 on: December 08, 2022, 01:29:41 PM »
Why would you SHORT both sides of the trade?  In any short position your downside is unlimited while your upside is limit to the exposure.

So if you rebalance "opposing" trades to be supposedly market neutral and then the market does a moonshoot - your x3 short leverage is going to leave you hundreds of percent in loss while your x3 long leverage fund is going to go to near zero - its an assymetrically bad trade.

...snip...

I was running some monte carlo style simulations last night and if daily moves are "large" you start see the benefits erode and even go negative.  I think @yand has found the crux.  The side of the bet that is "right" has a lower bound of zero while the side that is "wrong" has unlimited downside.  So, if you set a defined point in time to rebalance, you could get creamed on any major move.  If you rebalance on x%, you get killed (repeatedly) by the wash sale rule. 

Basically, this strategy has a "drag" flaw of its own!

Not going to try it.

ChpBstrd

  • Walrus Stache
  • *******
  • Posts: 7382
  • Location: A poor and backward Southern state known as minimum wage country
Re: Shorting the shorts? Leveraged ETFs are garbage (viable exploit?)
« Reply #11 on: December 09, 2022, 10:24:52 AM »
The bid-ask spread is going to be the hard part here. It is possible to be correct about the inefficiency of leveraged ETFs but to also not have an opportunity to profit because maybe the B/A is 20 cents on each leg. You can throw out some stink bids with conditional orders, but you'd have to hold a line on the profitability of the trade. That might prevent you from ever getting executed.

Another tricky problem is that a long leveraged ETF and it's short sibling will likely have much different prices. E.g. TQQQ is $21.56/share but SQQQ is $45/share. Thus, if you trade 1 option on SQQQ you're dealing with more than twice as much money and risk as if you were trading 1 option of TQQQ. So you'd have to do 1 contract of SQQQ and 2 contracts of TQQQ to be roughly arbitraged, but even then your trade is tilted 4.3% toward SQQQ because (2x21.56)<45. And these particular ETFs are close to whole multiples of each other, compared to some! To obtain full parity, you'd have to be dealing in hundreds of contracts. The markets' liquidity just isn't that deep so the B/A problem gets worse as you try to solve the parity problem.

I'd suggest looking for arbitrage opportunities between (a) liquid and illiquid options markets across non-leveraged ETFs like SPY and leveraged ETFs like UPRO, or (b) trading long index options against short long-ETF options. In either scenario, you'd be attempting to exploit the B/A spread and generally picking off retail orders from people like ourselves. There's also the possibility of doing equal-sized spreads across the different assets, but then you're doubling the number of B/A spreads you're paying AND the math gets more, not less, complicated. E.g. what does a 5% move do when you have one $10 wide bull spread on SPY and ten $4 wide short spreads on UPRO? Hell if I know!

The parity problem is real, and the position sizes needed to maintain net-neutrality would constantly fluctuate as prices deviated from their expected multiples of each other. You'd risk setting up your own inefficiencies that would dwarf the inefficiencies you're trying to arbitrage. Solve these theoretical problems and you'll be onto something!

MustacheAndaHalf

  • Walrus Stache
  • *******
  • Posts: 7151
  • Location: U.S. expat
Re: Shorting the shorts? Leveraged ETFs are garbage (viable exploit?)
« Reply #12 on: December 09, 2022, 12:00:15 PM »
The bid-ask spread is going to be the hard part here. It is possible to be correct about the inefficiency of leveraged ETFs but to also not have an opportunity to profit because maybe the B/A is 20 cents on each leg.
Did you check bid-ask spreads before you posted this?

Right now TQQQ has bid/ask of 21.35 / 21.36, and SQQQ 45.44 / 45.45.
UPRO 36.49 / 36.50   and SPXS  20.43 / 20.44
TMF  9.60 / 9.61   TMV  110.35 / 110.44

That last one was the only non $0.01 bid-ask spread I found, and it represents 0.08% of the bid price, which is a very narrow spread.  I don't understand your claim that "the bid-ask spread is going to be the hard part here".

Financial.Velociraptor

  • Handlebar Stache
  • *****
  • Posts: 2424
  • Age: 52
  • Location: Houston TX
  • Devour your prey raptors!
    • Living Universe Foundation
Re: Shorting the shorts? Leveraged ETFs are garbage (viable exploit?)
« Reply #13 on: December 09, 2022, 01:51:29 PM »
The bid-ask spread is going to be the hard part here. It is possible to be correct about the inefficiency of leveraged ETFs but to also not have an opportunity to profit because maybe the B/A is 20 cents on each leg.
Did you check bid-ask spreads before you posted this?

Right now TQQQ has bid/ask of 21.35 / 21.36, and SQQQ 45.44 / 45.45.
UPRO 36.49 / 36.50   and SPXS  20.43 / 20.44
TMF  9.60 / 9.61   TMV  110.35 / 110.44

That last one was the only non $0.01 bid-ask spread I found, and it represents 0.08% of the bid price, which is a very narrow spread.  I don't understand your claim that "the bid-ask spread is going to be the hard part here".


I think ChpBstrd is referring to the b/a on options.  Thus, why trading one "expensive" underlying and one "expensive" is problematic (e.g. lots of 100 make the LCD a huge number of shares).

You can either pay borrow fees, or pay a wide b/a spread if you go 'synthetic short'. 

Vand nailed it. There is a convexity issue that works against the strategy.

MustacheAndaHalf

  • Walrus Stache
  • *******
  • Posts: 7151
  • Location: U.S. expat
Re: Shorting the shorts? Leveraged ETFs are garbage (viable exploit?)
« Reply #14 on: December 09, 2022, 07:47:13 PM »
The bid-ask spread is going to be the hard part here. It is possible to be correct about the inefficiency of leveraged ETFs but to also not have an opportunity to profit because maybe the B/A is 20 cents on each leg.
Did you check bid-ask spreads before you posted this?

Right now TQQQ has bid/ask of 21.35 / 21.36, and SQQQ 45.44 / 45.45.
UPRO 36.49 / 36.50   and SPXS  20.43 / 20.44
TMF  9.60 / 9.61   TMV  110.35 / 110.44

That last one was the only non $0.01 bid-ask spread I found, and it represents 0.08% of the bid price, which is a very narrow spread.  I don't understand your claim that "the bid-ask spread is going to be the hard part here".
I think ChpBstrd is referring to the b/a on options.  Thus, why trading one "expensive" underlying and one "expensive" is problematic (e.g. lots of 100 make the LCD a huge number of shares).

You can either pay borrow fees, or pay a wide b/a spread if you go 'synthetic short'. 

Vand nailed it. There is a convexity issue that works against the strategy.
Oh, I 100% agree with wide bid-ask spreads on options.  The longer until expiration, and the further from the current price, the worse it gets.

But I've already gotten better at bearish investing based on this idea.  Each underlying index has two leveraged ETFs, one bullish and one bearish.  And leveraged ETFs have volatility decay, which hurts performance long term.  Surprisingly, Schwab charges almost nothing to short these ETFs.

I believe inflation will surprise the market in the next few months - possibly next week.  Instead of buying SPXS as a long position (-3x S&P 500), I have shorted UPRO!  So the usual 3x S&P 500 becomes -3x S&P 500 as a short position.  But now I have two ways to win: if the market drops, or if volatility decay kicks in.  Shorting the 3x ETF benefits from either (do not try this at home).  Unfortunately it gives up a huge benefit of leveraged ETFs: limited losses.  If you invest $1,000 in SPXS, that is the most you can lose.  If I short UPRO for $1,000 my short position can cost me $-2,000 if UPRO doubles.  Stop loss orders become a very good idea, plus watching the investment more closely.

ChpBstrd

  • Walrus Stache
  • *******
  • Posts: 7382
  • Location: A poor and backward Southern state known as minimum wage country
Re: Shorting the shorts? Leveraged ETFs are garbage (viable exploit?)
« Reply #15 on: December 10, 2022, 10:59:11 AM »
The bid-ask spread is going to be the hard part here. It is possible to be correct about the inefficiency of leveraged ETFs but to also not have an opportunity to profit because maybe the B/A is 20 cents on each leg.
Did you check bid-ask spreads before you posted this?

Right now TQQQ has bid/ask of 21.35 / 21.36, and SQQQ 45.44 / 45.45.
UPRO 36.49 / 36.50   and SPXS  20.43 / 20.44
TMF  9.60 / 9.61   TMV  110.35 / 110.44

That last one was the only non $0.01 bid-ask spread I found, and it represents 0.08% of the bid price, which is a very narrow spread.  I don't understand your claim that "the bid-ask spread is going to be the hard part here".
I think ChpBstrd is referring to the b/a on options.  Thus, why trading one "expensive" underlying and one "expensive" is problematic (e.g. lots of 100 make the LCD a huge number of shares).

You can either pay borrow fees, or pay a wide b/a spread if you go 'synthetic short'. 

Vand nailed it. There is a convexity issue that works against the strategy.
Oh, I 100% agree with wide bid-ask spreads on options.  The longer until expiration, and the further from the current price, the worse it gets.

But I've already gotten better at bearish investing based on this idea.  Each underlying index has two leveraged ETFs, one bullish and one bearish.  And leveraged ETFs have volatility decay, which hurts performance long term.  Surprisingly, Schwab charges almost nothing to short these ETFs.

I believe inflation will surprise the market in the next few months - possibly next week.  Instead of buying SPXS as a long position (-3x S&P 500), I have shorted UPRO!  So the usual 3x S&P 500 becomes -3x S&P 500 as a short position.  But now I have two ways to win: if the market drops, or if volatility decay kicks in.  Shorting the 3x ETF benefits from either (do not try this at home).  Unfortunately it gives up a huge benefit of leveraged ETFs: limited losses.  If you invest $1,000 in SPXS, that is the most you can lose.  If I short UPRO for $1,000 my short position can cost me $-2,000 if UPRO doubles.  Stop loss orders become a very good idea, plus watching the investment more closely.
But at this point you're talking about a leveraged directional trade rather than an arbitrage trade. I.e. there is a way to lose big if the market goes against your thesis, as opposed to an arbitrage trade where two counter-correlated assets mostly offset each other and a small return is earned at small risk regardless of market outcome.

My thought on shorting UPRO rather than going long SPXS is that there's probably a cyclical swing like with contango and backwardization, where sometimes the drag is worse on UPRO and sometimes it's worse on SPXS. However I don't have the technical knowledge to form this hypothesis into a full theory. If nothing else, maybe there's a premium/discount to NAV that swings around between these funds.

I would be interested to know if shorting UPRO was more profitable for the same market down move than going long SPXS. It would be easy to put together a backtest which might reveal any issues across multiple calendar dates. I'd also be curious if someone could make a profit at "short $X of UPRO / long $3X of SPY" arbitrage that could beat today's 1 year treasury rates.

MustacheAndaHalf

  • Walrus Stache
  • *******
  • Posts: 7151
  • Location: U.S. expat
Re: Shorting the shorts? Leveraged ETFs are garbage (viable exploit?)
« Reply #16 on: December 10, 2022, 08:37:38 PM »
The bid-ask spread is going to be the hard part here. It is possible to be correct about the inefficiency of leveraged ETFs but to also not have an opportunity to profit because maybe the B/A is 20 cents on each leg.
Did you check bid-ask spreads before you posted this?

Right now TQQQ has bid/ask of 21.35 / 21.36, and SQQQ 45.44 / 45.45.
UPRO 36.49 / 36.50   and SPXS  20.43 / 20.44
TMF  9.60 / 9.61   TMV  110.35 / 110.44

That last one was the only non $0.01 bid-ask spread I found, and it represents 0.08% of the bid price, which is a very narrow spread.  I don't understand your claim that "the bid-ask spread is going to be the hard part here".
I think ChpBstrd is referring to the b/a on options.  Thus, why trading one "expensive" underlying and one "expensive" is problematic (e.g. lots of 100 make the LCD a huge number of shares).

You can either pay borrow fees, or pay a wide b/a spread if you go 'synthetic short'. 

Vand nailed it. There is a convexity issue that works against the strategy.
Oh, I 100% agree with wide bid-ask spreads on options.  The longer until expiration, and the further from the current price, the worse it gets.

But I've already gotten better at bearish investing based on this idea.  Each underlying index has two leveraged ETFs, one bullish and one bearish.  And leveraged ETFs have volatility decay, which hurts performance long term.  Surprisingly, Schwab charges almost nothing to short these ETFs.

I believe inflation will surprise the market in the next few months - possibly next week.  Instead of buying SPXS as a long position (-3x S&P 500), I have shorted UPRO!  So the usual 3x S&P 500 becomes -3x S&P 500 as a short position.  But now I have two ways to win: if the market drops, or if volatility decay kicks in.  Shorting the 3x ETF benefits from either (do not try this at home).  Unfortunately it gives up a huge benefit of leveraged ETFs: limited losses.  If you invest $1,000 in SPXS, that is the most you can lose.  If I short UPRO for $1,000 my short position can cost me $-2,000 if UPRO doubles.  Stop loss orders become a very good idea, plus watching the investment more closely.
But at this point you're talking about a leveraged directional trade rather than an arbitrage trade. I.e. there is a way to lose big if the market goes against your thesis, as opposed to an arbitrage trade where two counter-correlated assets mostly offset each other and a small return is earned at small risk regardless of market outcome.

My thought on shorting UPRO rather than going long SPXS is that there's probably a cyclical swing like with contango and backwardization, where sometimes the drag is worse on UPRO and sometimes it's worse on SPXS. However I don't have the technical knowledge to form this hypothesis into a full theory. If nothing else, maybe there's a premium/discount to NAV that swings around between these funds.

I would be interested to know if shorting UPRO was more profitable for the same market down move than going long SPXS. It would be easy to put together a backtest which might reveal any issues across multiple calendar dates. I'd also be curious if someone could make a profit at "short $X of UPRO / long $3X of SPY" arbitrage that could beat today's 1 year treasury rates.
I'm not sure how to fix flaws in the original idea, but I wanted to mention how I benefitted from it.  Yes, I'm making directional trades against the market.

I interpret your mention of "contango and backwardization" as meaning you should read about 3x ETFs.  Besides reading and learning about volatility drag, I've also looked at raw market data and convinced myself it explains the performance problems.

Using Yahoo Finance data, I looked at the opening price of SPY, UPRO and SPXS for the past 11 weekdays.  Dividing each day's price by the prior day gives price performance.  Consistent with those ETFs objective of +300% and -300% exposure to the S&P 500, their median leverage was -303% and +307% in the past 10 days.

Over those 10 days, SPY performance was -1.71%.  Triple that is -5.13%, yet UPRO returned -5.94% while SPXS returned 4.78%.  My favorite way to see volatility drag measures SPY at two times where it had almost no change, like 11/28 to 12/6 when it rose just +0.08%.  Yet UPRO was -0.42% and SPXS -0.90%.  This is volatility drag in action.

Extreme example, SPY -20% then +25% :
SPY 5 x -20% -> 4 x +25% = 5.0
UPRO : 5 x -60% -> 2 x +75% = 3.5
SPXS : 5 x +60% -> 8 x -75% = 2.0

During the above extreme example, holding SPXS results in a -60% drop.  Shorting UPRO results in a +30% gain (owing 3.5 on a 5.0 loan).  As I mentioned above, shorting risks unlimited losses, which can mean needing stop loss orders and careful monitoring.

You mentioned an idea, $300 SPY and short $100 of UPRO, which should have no net market exposure.  In the extreme example (15.0 SPY -> 15.0, UPRO -5.0 -> -3.5) you wind up with $40 profit from volatility drag, less the cost of shorting.

Using the past 10 days of data with this market neutral idea:
$300 SPY x -1.71% = $294.87 (loss 5.13)
shorting $100 UPRO x -5.94% = -$94.06  (gain 5.94)
Which equals a gain of $0.81 on the combined long + short position.

ChpBstrd

  • Walrus Stache
  • *******
  • Posts: 7382
  • Location: A poor and backward Southern state known as minimum wage country
Re: Shorting the shorts? Leveraged ETFs are garbage (viable exploit?)
« Reply #17 on: December 11, 2022, 10:19:09 AM »
The bid-ask spread is going to be the hard part here. It is possible to be correct about the inefficiency of leveraged ETFs but to also not have an opportunity to profit because maybe the B/A is 20 cents on each leg.
Did you check bid-ask spreads before you posted this?

Right now TQQQ has bid/ask of 21.35 / 21.36, and SQQQ 45.44 / 45.45.
UPRO 36.49 / 36.50   and SPXS  20.43 / 20.44
TMF  9.60 / 9.61   TMV  110.35 / 110.44

That last one was the only non $0.01 bid-ask spread I found, and it represents 0.08% of the bid price, which is a very narrow spread.  I don't understand your claim that "the bid-ask spread is going to be the hard part here".
I think ChpBstrd is referring to the b/a on options.  Thus, why trading one "expensive" underlying and one "expensive" is problematic (e.g. lots of 100 make the LCD a huge number of shares).

You can either pay borrow fees, or pay a wide b/a spread if you go 'synthetic short'. 

Vand nailed it. There is a convexity issue that works against the strategy.
Oh, I 100% agree with wide bid-ask spreads on options.  The longer until expiration, and the further from the current price, the worse it gets.

But I've already gotten better at bearish investing based on this idea.  Each underlying index has two leveraged ETFs, one bullish and one bearish.  And leveraged ETFs have volatility decay, which hurts performance long term.  Surprisingly, Schwab charges almost nothing to short these ETFs.

I believe inflation will surprise the market in the next few months - possibly next week.  Instead of buying SPXS as a long position (-3x S&P 500), I have shorted UPRO!  So the usual 3x S&P 500 becomes -3x S&P 500 as a short position.  But now I have two ways to win: if the market drops, or if volatility decay kicks in.  Shorting the 3x ETF benefits from either (do not try this at home).  Unfortunately it gives up a huge benefit of leveraged ETFs: limited losses.  If you invest $1,000 in SPXS, that is the most you can lose.  If I short UPRO for $1,000 my short position can cost me $-2,000 if UPRO doubles.  Stop loss orders become a very good idea, plus watching the investment more closely.
But at this point you're talking about a leveraged directional trade rather than an arbitrage trade. I.e. there is a way to lose big if the market goes against your thesis, as opposed to an arbitrage trade where two counter-correlated assets mostly offset each other and a small return is earned at small risk regardless of market outcome.

My thought on shorting UPRO rather than going long SPXS is that there's probably a cyclical swing like with contango and backwardization, where sometimes the drag is worse on UPRO and sometimes it's worse on SPXS. However I don't have the technical knowledge to form this hypothesis into a full theory. If nothing else, maybe there's a premium/discount to NAV that swings around between these funds.

I would be interested to know if shorting UPRO was more profitable for the same market down move than going long SPXS. It would be easy to put together a backtest which might reveal any issues across multiple calendar dates. I'd also be curious if someone could make a profit at "short $X of UPRO / long $3X of SPY" arbitrage that could beat today's 1 year treasury rates.
I'm not sure how to fix flaws in the original idea, but I wanted to mention how I benefitted from it.  Yes, I'm making directional trades against the market.

I interpret your mention of "contango and backwardization" as meaning you should read about 3x ETFs.  Besides reading and learning about volatility drag, I've also looked at raw market data and convinced myself it explains the performance problems.

Using Yahoo Finance data, I looked at the opening price of SPY, UPRO and SPXS for the past 11 weekdays.  Dividing each day's price by the prior day gives price performance.  Consistent with those ETFs objective of +300% and -300% exposure to the S&P 500, their median leverage was -303% and +307% in the past 10 days.

Over those 10 days, SPY performance was -1.71%.  Triple that is -5.13%, yet UPRO returned -5.94% while SPXS returned 4.78%.  My favorite way to see volatility drag measures SPY at two times where it had almost no change, like 11/28 to 12/6 when it rose just +0.08%.  Yet UPRO was -0.42% and SPXS -0.90%.  This is volatility drag in action.

Extreme example, SPY -20% then +25% :
SPY 5 x -20% -> 4 x +25% = 5.0
UPRO : 5 x -60% -> 2 x +75% = 3.5
SPXS : 5 x +60% -> 8 x -75% = 2.0

During the above extreme example, holding SPXS results in a -60% drop.  Shorting UPRO results in a +30% gain (owing 3.5 on a 5.0 loan).  As I mentioned above, shorting risks unlimited losses, which can mean needing stop loss orders and careful monitoring.

You mentioned an idea, $300 SPY and short $100 of UPRO, which should have no net market exposure.  In the extreme example (15.0 SPY -> 15.0, UPRO -5.0 -> -3.5) you wind up with $40 profit from volatility drag, less the cost of shorting.

Using the past 10 days of data with this market neutral idea:
$300 SPY x -1.71% = $294.87 (loss 5.13)
shorting $100 UPRO x -5.94% = -$94.06  (gain 5.94)
Which equals a gain of $0.81 on the combined long + short position.
Interesting research, suggesting a very high level of under-performance in recent days!

A few years ago, we were theorizing in a similar way about VIX futures funds like VXX, VIXY, UVXY, and SVXY (back in the good ole days when some of them had real leverage). A lot of money was lost before we learned that contango and backwardization conditions come and go in VIX futures markets. Contango and backwardization would swing back and forth with market sentiment, oscillating sometimes on a scale of months or years. Volatility drag would unexpectedly become volatility boost and wipe out arbitrage positions that had been doing well for weeks or months!

To pick up such a signal, one would have to look at a lot more than 11 days worth of data. The pitfall was to look at a brief dataset and conclude whichever condition was observed was a permanent condition. IIRC, at one point I had a couple of years of information on a spreadsheet and the outcome of net simulated trades was an oscillating pattern of gains and losses.

Of course, in that case, volatility itself was a contributing factor in VIX futures contango/backwardization, and I never did figure out an exploitable relationship. Stock indices are maybe more straightforward in the sense we're not confounding variables.

While we're coming up with ideas, I'm intrigued by the idea of continuously rolled spread trades. For example, enter OTM bear call spreads on both SPXS and UPRO, earning a credit on each side and having a market risk profile similar to a short iron condor on the S&P500. As the market moves one direction or another, roll the spread closer to the money on the unchallenged or farther-OTM side, earning another credit. One could repeat this process forever by rolling out the duration too. In the end, you'll have "lost" both legs of the trade by ratcheting closer and closer to the current price and perhaps overlapping each other. That's a limited loss equal to the width of both spreads. The goal would be to collect more premiums than the loss, which seems almost inevitable if enough rolls into longer durations are done at the same spread amount. The point of doing this on both SPXS and UPRO instead of just one would be to offset some market beta, capturing large amounts of time decay from two highly volatile instruments, and to capture the combined inefficiency of both in a single "condor".

If you win such a "synthetic condor" half the time (e.g. average 50% gain), and recover all or most of your spread width the other half of the time (e.g. average -30% loss), then you have a net-winning strategy that could be repeated long-term. Wins and losses in this strategy might follow the contours of the contango/backwardization cycle, but by holding one's maximum risk constant, by running the trade long enough to span over the cycles, and by partially hedging market beta by playing both sides against the middle, this might be the closest we can get to harvesting the pure inefficiency of these funds. Still thinking through the implications.


MustacheAndaHalf

  • Walrus Stache
  • *******
  • Posts: 7151
  • Location: U.S. expat
Re: Shorting the shorts? Leveraged ETFs are garbage (viable exploit?)
« Reply #18 on: December 11, 2022, 05:45:19 PM »
Although its called "volatility drag", when markets rise straight up there's an opposite effect: UPRO can outperform 3x SPY, like in 2019 (SPY +31% , UPRO +102%).  Since I didn't want to type more than 11 days of data for 3 different ETFs, that limited my measurement.  But you can see volatility drag by comparing the past month and YTD performance of SPY vs UPRO:
SPY     +8.51% last month , -13.30% YTD
UPRO  +24.68% last month , -47.77% YTD


The idea of matched bear spreads has a flaw: they don't protect each other.  If the market crashes, your SPXS calls could get assigned and you lose the spread.  But in a quick recovery, UPRO can cover the spread and you lose there, too.  In the calculations I did, you collect a premium equal to 20-25% of the spread, so you have to be right more than 75% of the time for it to work.

Maybe you can give examples of the condor prices for UPRO and SPXS, as I'm seeing much smaller payouts in the spreads.

MustacheAndaHalf

  • Walrus Stache
  • *******
  • Posts: 7151
  • Location: U.S. expat
Re: Shorting the shorts? Leveraged ETFs are garbage (viable exploit?)
« Reply #19 on: December 15, 2022, 01:41:53 PM »
When profitting off volatility drag, you need a gain then a loss (or the reverse).  If you rebalance in between, you eliminate volatility drag, and the strategy won't work.  But if you let it run too far, compounding sends the two positions farther apart, and could cause losses.  So there's a balancing act with rebalancing.

I'm running three experiments.  The worst was shorting both SSO and SDS.  In theory, the -2x S&P 500 ETF will have more drag, but despite big moves I saw very little gap between them.  Since I'm paying margin rates to short, I closed those positions.  Next up is (3 * FDN - WEBL), which has mostly moved in a straight line so far, so I'll need more time to see.  A bad early sign is WEBL losing less than expected, which is a risk I knew about but thought was more rare.  Finally, (1.5 VIXY - UVXY), which has a tiny profit.  I increased its size and rebalanced, so now UVXY and WEBL are short the same dollar amount.

Note with these long-short speculations, both positions can grow or decline - but their gains and losses are nearly equal.  So $3,000 FDN and $-1,000 WEBL might decline to $2,700 FDN and $-700 WEBL, which is the same net position.  I also need to look at loss vs gain compounding to see if that explains WEBL vs FDN performance.