HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

[staythecourse]

This may be the dumbest question on the thread, but I just wanted to confirm what I’ve read elsewhere on bogleheads:

Is it true that I can do an in-kind conversion of the traditional Ira (about $10k) I have in Fidelity that’s holding these ETF funds to an as-of-now-non-existent Roth also in Fidelity?

I have $5k in this project so far, and am strongly considering bumping it to $10k. Would it be better to wait to purchase the additional $5k worth before or after the conversion?

Thanks

Best to ask Fidelity. Each institution is different if they allow in kind transfers and when. Since it is a conversion you will pay taxes of course. If it was me I would have done this conversion no matter what so you can backdoor future roth ira without an issue of that 10k sitting in your traditional ira.

Good luck.

p.s. Best advice I had recent on this site is to make sure your traditional IRA is zero at the end of every year for purposes of backdoor roth ira.

[bgf]

i'll give it up to OP, this is not a harebrained strategy. here at BH we're almost irrationally conditioned to think that all active strategies will fail, but that isn't the case. the odds might be against you, and it is certaintly a tough game to play, and as a matter of logic and math all active investors cannot win - BUT its not impossible. its not even really necessarily that complicated.

im not ed thorp, jim simons, or warren buffett, and neither is OP, but its worth a shot. 15% of an already decent sized portfolio, im sure OP's human capital is solid, and 'sometimes you just gotta say, wtf.'

let it rip.

[Brassman]

I know some people have put a lot of effort into backtesting this thing, and thank you for that. I am thinking about doing the same thing and sticking a small chuck in a Roth IRA and letting it rip on this strategy.

One thing though, how does this approach compare to a 3x leveraged all weather portfolio? Given the inclusion of intermediate treasuries and gold to help during some of the other scenarios mentioned I would love to see how it stacks up. I see the back testing run during the 50-70 and 60-80 stretch was favorable, but these periods are specifically what the addition of gold and intermediate treasuries into the all weather portfolio were for.

[HEDGEFUNDIE]

I know some people have put a lot of effort into backtesting this thing, and thank you for that. I am thinking about doing the same thing and sticking a small chuck in a Roth IRA and letting it rip on this strategy.

One thing though, how does this approach compare to a 3x leveraged all weather portfolio? Given the inclusion of intermediate treasuries and gold to help during some of the other scenarios mentioned I would love to see how it stacks up. I see the back testing run during the 50-70 and 60-80 stretch was favorable, but these periods are specifically what the addition of gold and intermediate treasuries into the all weather portfolio were for.

While my strategy shares certain characteristics with the All-Weather Portfolio, I don't think All Weather is the most efficient thing to hold for the long term. Mainly because it devotes 25% of the risk budget to fight inflation - yet unexpected inflation has not been a problem for almost 40 years. To me, the smarter thing to do is to weight the long term likelihood of each "risk quadrant" to determine your asset allocation. In which case, sure, Gold and TIPS will have a role, but not 25%.

To put it another way, if you live in Florida, should your closet be 25% winter clothing?

[Coconut Cream Pie]

Following this thread with great interest, reminds me of the classic market timer thread.

If 2008 happened again, UPRO would get killed but TMF would skyrocket. [...]

It seems like this strategy puts a lot of emphasis on this "fact", but why do you think it is true? Like everything, if that was actually true and predictable, it could easily be abused to make arbitrary amounts of money.

https://www.portfoliovisualizer.com/bac ... ion2_3=100 - both UPRO and TMF are down at the same time between Aug and Oct 2016.

https://www.portfoliovisualizer.com/bac ... ion2_3=100 - both UPRO and TMF are down at the same time between Jan and Apr 2018.

What happens if this continued for half a year, or longer? Your investment would probably be wiped out. I also encourage you to look at other 3x EFTs, like NUGT (Direxion Daily Gold Miners Bull 3X ETF): https://yhoo.it/2SGmRUF. These lost 99.9% of their value over just five years. There is no recovery, whatsoever.

You're looking at 20 years. You also own a house for 1.8M according to e.g. viewtopic.php?t=242560, so you must be earning a nice salary.

If you invested your 100k into pretty much anything over the last 20 years and contributed 10k/month (which you really should, after all the savings rate is the most important), you'd have ended up with 6-8M: https://www.portfoliovisualizer.com/bac ... lBond3=100. Or with bonds for an easy 4.5M. But yeah, that'd wouldn't nearly be as cool as posting in this thread in 20 years about how you 100x your investment without contributions.

[fennewaldaj]

I don't think All Weather is the most efficient thing to hold for the long term. Mainly because it devotes 25% of the risk budget to fight inflation - yet unexpected inflation has not been a problem for almost 40 years. To me, the smarter thing to do is to weight the long term likelihood of each "risk quadrant" to determine your asset allocation. In which case, sure, Gold and TIPS will have a role, but not 25%.

To put it another way, if you live in Florida, should your closet be 25% winter clothing?

Isn't the point of the all weather portfolio to give acceptable results no matter the economic conditions. So you don't have to make any judgement on how likely inflation is. That is its not supposed to be the most efficient thing its supposed to protect you from uncertainty.

It occurred to me that combining your strategy at say the 15-20% of assets as you planned with a all weather could make a kind of sense. Things go well and you do very well in your leveraged portfolio. Things go poorly and the all weather is likely to give you an acceptable portfolio no matter what caused problems.

[hohum]

If you invested your 100k into pretty much anything over the last 20 years and contributed 10k/month (which you really should, after all the savings rate is the most important), you'd have ended up with 6-8M: https://www.portfoliovisualizer.com/bac ... lBond3=100. Or with bonds for an easy 4.5M. But yeah, that'd wouldn't nearly be as cool as posting in this thread in 20 years about how you 100x your investment without contributions.

Yeah, like the world would be a better place if nobody did anything cool.

[SVT]

I'm trying to sign up for M1 Finance online. It's making me choose 3 slices for my pie. How did you only do 2? Do you sign up with 3 and then just don't invest in the 3rd slice?

ETA: Got it. Just have to do 3 slices for signup. Once you're in, you can build a pie with only 2.

[jaj2276]

...

This is why I prefer the use of UGLD (3X gold) for 10% of my portfolio. It is also why I’ve considered the use of TYD (7-10 year treasuries) for a portion of the bond exposure. A portfolio of 35% UPRO, 55% TMF/TYD, and 10% UGLD may be more ideal from a risk parity perspective.

I too am going to add UGLD to my portfolio, although not sure how much (probably no more than 10% but maybe only 5%). Still need to mull it over for a few days. Not sure how I feel about the TYD modification to TMF, hadn't really thought about it much. I will look in to this holding as well.

EDIT 1: Going with a 45% UPRO, 50% TMF, 5% UGLD. I looked at TYD but the correlations with TMF were too high so it wouldn't add much to the portfolio other than complexity.

UGLD has at best a week positive correlation with TMF (~0.19 during the period I was looking at) and a 0 correlation with equities. Adding UGLD is in case the past doesn't predict the future (i.e. TMF does in fact go down with the ship right along UPRO). If the past does rhyme with the future, then UGLD will act as a drag on my returns but at 5% won't be too huge I hope.

[samsdad]

I’ve edited this post to follow EfficientInvestor’s lead regarding the simulated data issue:

UPDATE (2/20/19) - The data that had previously been provided in this post was based on developing a best fit to the data of actual leveraged ETFs since their inception. Since inception of 3X ETFs in 2009, borrowing rates have been low. Therefore, this data did not take into account higher borrowing rates that occurred prior to 2009. Due to this, I have removed the data I had previously provided.

[staythecourse]

Samsdad,

Again thanks for the good work. I am NOT a big fan of backtesting, but you need some way to test any hypothesis. Even if the data is approximations it is better to have some data for it then NOT have any more data then from the time point from 1987- current which basically was the same economic environment.

Again, I think the caution is NOT in analyzing the data, but a constant reminder IF folks are going to use this approach it should be in your play money ONLY!! Play money should only be money you don't mind going to a big, fat zero. With that caveat how is this any riskier then using some random market timing strategy or screeners to select some "undervalued" stocks?

I don't know anyone (including the OP) that is suggesting using this approach for one's majority of their portfolio.

Good luck.

[jaj2276]

I now want this thread deleted (kidding!, only slightly). I understand why it was started, to hold OP to the fire. However I fear that it will make the trade crowded, a sure-fire deathknell for the success of the trade itself.

[software]

Maybe I don’t understand how these funds get their leverage, but isn’t there a non-zero (and even high) percentage chance that at some point these funds will end up at zero?

To get 3x leverage, for every share of underlying fund owned there needs to be 2 coming from leverage. If the underlying index crashes more than 33% isn’t the entire position wiped out?

[DosCommas]

It looks to me like holding this in taxable could present some fabulous tax loss harvesting opportunities due to the volatility.

UPRO and SPXL have no overlapping holdings so could be pair funds. Any suggestions for TMF? TYD perhaps, but that tracks 7-10 year treasuries instead...

I’m very hesitant to give up any Roth space to “play money”. That strikes me as against my IPS to maximize the use of my tax advantages space for my objectives. If it doesn’t go so well I’d have lost tax free space and also the ability to write it off!

[staythecourse]

It looks to me like holding this in taxable could present some fabulous tax loss harvesting opportunities due to the volatility.

UPRO and SPXL have no overlapping holdings so could be pair funds. Any suggestions for TMF? TYD perhaps, but that tracks 7-10 year treasuries instead...

I’m very hesitant to give up any Roth space to “play money”. That strikes me as against my IPS to maximize the use of my tax advantages space for my objectives. If it doesn’t go so well I’d have lost tax free space and also the ability to write it off!

I think the issue is going to be rebalancing quarterly. You will be selling gains taxed at ordinary income rates which likely eat A LARGE portion of any potential profits. I leave it to Hedge to address the question as I am sure he had thought of it himself already.

Good luck.

[michaeljc70]

If it fails, the reply is: See. We told you so.

Either way, the critic is justified and the innovator can't win.

That's one reason I like the market timer thread. He received a lot of criticism and some of it was flippant, and he never responded in kind. It takes conviction to do something different and stick with it. The high fives may not come even if your strategy works.

If it works and I have 10 million bucks I don't need kudos from some guy I don't know on the internet.

[jaj2276]

It looks to me like holding this in taxable could present some fabulous tax loss harvesting opportunities due to the volatility.

UPRO and SPXL have no overlapping holdings so could be pair funds. Any suggestions for TMF? TYD perhaps, but that tracks 7-10 year treasuries instead...

I’m very hesitant to give up any Roth space to “play money”. That strikes me as against my IPS to maximize the use of my tax advantages space for my objectives. If it doesn’t go so well I’d have lost tax free space and also the ability to write it off!

I think the issue is going to be rebalancing quarterly. You will be selling gains taxed at ordinary income rates which likely eat A LARGE portion of any potential profits. I leave it to Hedge to address the question as I am sure he had thought of it himself already.

Good luck.

The OP mentioned it in an earlier post. You would need to keep cash in the account so that rebalance in the first year would simply involve buying the low performing asset. Then in year 1+ you will likely be selling gains as LTCG and also have losses to offset those gains.

[DosCommas]

It looks to me like holding this in taxable could present some fabulous tax loss harvesting opportunities due to the volatility.

UPRO and SPXL have no overlapping holdings so could be pair funds. Any suggestions for TMF? TYD perhaps, but that tracks 7-10 year treasuries instead...

I’m very hesitant to give up any Roth space to “play money”. That strikes me as against my IPS to maximize the use of my tax advantages space for my objectives. If it doesn’t go so well I’d have lost tax free space and also the ability to write it off!

I think the issue is going to be rebalancing quarterly. You will be selling gains taxed at ordinary income rates which likely eat A LARGE portion of any potential profits. I leave it to Hedge to address the question as I am sure he had thought of it himself already.

Good luck.

The OP mentioned it in an earlier post. You would need to keep cash in the account so that rebalance in the first year would simply involve buying the low performing asset. Then in year 1+ you will likely be selling gains as LTCG and also have losses to offset those gains.

I must have missed that. Yes, that’s what I was thinking. Starting with some play money and rebalancing with new funds until hitting the target principal invested, plus aggressively tax loss harvesting the positions along with the quarterly rebalance. I guess I’ll have to run a back test to see if it is viable. Will post if I do.

[HomerJ]

I think we can all agree that the period from 1973 to 1981 pretty much sucked if you were an investor. If the data is correct, you would've been better off with a 40/60 3X mix (blue) than with any of the other options, though it is true that you would have still been negative at the end of the period, adjusted for inflation (your $10k would have been worth $5970 according to PV.) That's better than the $5100ish the red or gold options would have given you, and much better than the approx. $4700 a 100% S&P 500 (no dividends reinvested, or for any of the above with GSPC) investor would have had left in 1981.

Why are "no dividends reinvested"? That somewhat invalidates your data and your conclusions. Because in the real world, dividends matter.

I see UPRO has some dividends... Do they match the dividends from a unleveraged S&P 500 fund? Since it's a daily trading animal, I'm thinking no, but I can't tell for sure. Do you get as much in dividends in UPRO each quarter as you would from a normal S&P 500 fund?

[HomerJ]

An observation: we tend to be rather conservative investors here. So conservative, that there are some people (not pointing at anyone in particular) who will not accept past data or backtesting as a basis of any investing decision--I don't care if it's hundreds of years of data. To them, tomorrow will always have the chance of being completely different from the past. Therefore, anything from the past is irrelevant to them, ostensibly. I respectfully disagree. The human condition, especially the nature of greed and fear, are the same today as they were 69 years ago in 1950, and the same as they were 690 years ago in 1329—but that's a topic for another thread.

Like someone else mentioned. Backtesting only proves what DOESN'T work. It's a useful tool to check an investment plan and see if it blew up in the past. If it blew up in the past, then you can probably discard it.

If it didn't blow up in the past, that's a good start, but it's not enough. It's NOT an indication that it's a rock-solid plan. Because the next crash will probably be different.

EVERY investment plan that failed in the past 50 years, at first, backtested well. Sure, there are some that actually worked pretty decent going forward. Not all back-tested plans fail. But there are a ton that did fail, because tomorrow WAS different from the past.

There's good reasons to like this strategy, but don't fool yourself that back-testing proves it's safe or even low risk. Backtesting can only prove the negative, not the positive.

[siamond]

I see UPRO has some dividends... Do they match the dividends from a unleveraged S&P 500 fund? Since it's a daily trading animal, I'm thinking no, but I can't tell for sure. Do you get as much in dividends in UPRO each quarter as you would from a normal S&P 500 fund?

No, they don't. UPRO's dividends are actually pretty small (same for its 2x equivalent, SSO). I don't think any of the simulations developed so far properly modeled dividends, which probably partly invalidates the UPRO backtesting (for which price is the main driver) and may very well totally invalidate the TMF backtesting (for which dividends/coupons are a strong driver).

EDIT: distributions were dividends, there were no capital gains distributions for UPRO (while SSO had 1.6% of it in 2006 and 2007).

Code: Select all

CAGR	26.48%	26.72%	0.19%		CAGR	17.04%	22.03%	4.26%

ER 0.92	UPRO	UPRO	UPRO			S&P 500	S&P 500	S&P 500
.	PR	TR	Distribs	.	PR	TR	Distribs
2010	35.43%	35.77%	0.25%		2010	23.12%	28.14%	4.08%
2011	-11.55%	-11.49%	0.07%		2011	-5.34%	-1.30%	4.27%
2012	46.21%	46.29%	0.06%		2012	26.55%	32.42%	4.64%
2013	118.33%	118.56%	0.10%		2013	65.88%	73.09%	4.34%
2014	37.69%	38.00%	0.23%		2014	22.48%	27.58%	4.16%
2015	-5.36%	-5.05%	0.33%		2015	-3.79%	0.35%	4.30%
2016	29.92%	30.11%	0.14%		2016	17.93%	23.21%	4.48%
2017	71.26%	71.26%	0.00%		2017	41.96%	47.75%	4.08%
2018	-25.32%	-24.91%	0.54%		2018	-14.62%	-11.21%	3.99%

I am trying to unravel this modeling issue. What is clear so far is that the 3x tracking for UPRO/SSO/S&P 500 is based on the *price* index, not on the Total Return index. Now how are dividends processed remains elusive. If anybody knows the corresponding mechanics of such leveraged fund, please speak up...

[samsdad]

I’ve edited this post to follow EfficientInvestor’s lead regarding the simulated data issue:

UPDATE (2/20/19) - The data that had previously been provided in this post was based on developing a best fit to the data of actual leveraged ETFs since their inception. Since inception of 3X ETFs in 2009, borrowing rates have been low. Therefore, this data did not take into account higher borrowing rates that occurred prior to 2009. Due to this, I have removed the data I had previously provided.

[kellykline]

Has anyone mentioned about what happened to XIV? Might happen to UPRO too. Beware.

[samsdad]

I think we can all agree that the period from 1973 to 1981 pretty much sucked if you were an investor. If the data is correct, you would've been better off with a 40/60 3X mix (blue) than with any of the other options, though it is true that you would have still been negative at the end of the period, adjusted for inflation (your $10k would have been worth $5970 according to PV.) That's better than the $5100ish the red or gold options would have given you, and much better than the approx. $4700 a 100% S&P 500 (no dividends reinvested, or for any of the above with GSPC) investor would have had left in 1981.

Why are "no dividends reinvested"? That somewhat invalidates your data and your conclusions. Because in the real world, dividends matter.

I see UPRO has some dividends... Do they match the dividends from a unleveraged S&P 500 fund? Since it's a daily trading animal, I'm thinking no, but I can't tell for sure. Do you get as much in dividends in UPRO each quarter as you would from a normal S&P 500 fund?

No dividend data is included by ^GSPC from what I'm told. So, it's sans-dividends. If you have S&P 500 data that includes dividends back to the 1950s I'd like to look at it. But, I don't know why that'd invalidate the data per se. If anything wouldn't it make it more a more conservative estimation of the returns?

[pdavi21]

The real problem here is if US markets perform over the next ten years like emerging markets did over the last ten. A 60/40 EMx3/TMF lost to EM/EDV even from the bottom of the crash. EMx3 lost to EM from the bottom of the 2009 crash. Developed Markets x3 lost from near the bottom of the crash, but TMF saves it and gives it a higher CAGR than x1 in a 60/40.

I think OP should consider diversifying as a diversified portfolio of x3 funds would've relied heavily on the outsized gains of one country (USA) and the negative correlation of long treasuries.

[HEDGEFUNDIE]

Has anyone mentioned about what happened to XIV? Might happen to UPRO too. Beware.

XIV was a 1x inverse ETN tracking the VIX. So it promised to deliver -100% of the daily movement of the VIX.

On Feb 5, 2018 the VIX went up by 116%. So guess what happened to XIV? Exactly what was supposed to happen, it got wiped out. In that sense, XIV worked perfectly.

UPRO promises 3x the daily returns of the S&P 500. So if we ever see a day when the S&P moves down -33.3%, UPRO will be wiped out as well. The point of this strategy though, is that TMF is supposed to move in the opposite direction if/when this happens.

FWIW, the largest one day drop in the S&P 500 was Black Monday 1987 when it dropped -20.4%.

[HEDGEFUNDIE]

The real problem here is if US markets perform over the next ten years like emerging markets did over the last ten. A 60/40 EMx3/TMF lost to EM/EDV even from the bottom of the crash. EMx3 lost to EM from the bottom of the 2009 crash. Developed Markets x3 lost from near the bottom of the crash, but TMF saves it and gives it a higher CAGR than x1 in a 60/40.

I think OP should consider diversifying as a diversified portfolio of x3 funds would've relied heavily on the outsized gains of one country (USA) and the negative correlation of long treasuries.

Have mentioned this previously: the higher underlying volatility of those other stock indexes poses greater risk of volatility drag and a complete wipeout of those particular 3x funds. When you're playing with 3x leverage, you want to diversify correlations, but not at the cost of unnecessarily increasing volatility.

Posting volatility drag table here again:

[HEDGEFUNDIE]

I am trying to unravel this modeling issue. What is clear so far is that the 3x tracking for UPRO/SSO/S&P 500 is based on the *price* index, not on the Total Return index. Now how are dividends processed remains elusive. If anybody knows the corresponding mechanics of such leveraged fund, please speak up...

The majority of the underlying holdings of UPRO and TMF are swaps and futures contracts, which do not yield dividends.

But both funds do hold some actual shares of stocks and bonds, probably as collateral for the derivatives. It's those holdings that yield dividends.

[samsdad]

Has anyone mentioned about what happened to XIV? Might happen to UPRO too. Beware.

XIV was a 1x inverse ETF tracking the VIX. So it promised to deliver -100% of the daily movement of the VIX.

On Feb 5, 2018 the VIX went up by 116%. So guess what happened to XIV? Exactly what was supposed to happen, it got wiped out. In that sense, the fund worked perfectly.

UPRO promises 3x the daily returns of the S&P 500. So if we ever see a day when the S&P moves down -33.3%, UPRO will be wiped out as well. The point of this strategy though, is that TMF is supposed to move in the opposite direction if/when this happens.

FWIW, the largest one day drop in the S&P 500 was Black Monday 1987 when it dropped -20.4%.

From the horse’s mouth, the New York Stock Exchange (my emphasis):

Circuit Breakers

In response to the market breaks in October 1987 and October 1989, the New York Stock Exchange instituted circuit breakers to reduce volatility and promote investor confidence. By implementing a pause in trading, investors are given time to assimilate incoming information and the ability to make informed choices during periods of high market volatility.

NYSE Rule 80B provides that a circuit-breaker halt for a Level 1 (7%) or Level 2 (13%) decline in the S&P 500 Index occurring after 9:30 a.m. Eastern and up to and including 3:25 p.m. Eastern, or in the case of an early scheduled close, 12:25 p.m. Eastern, would result in a trading halt in all stocks for 15 minutes. If the S&P 500 Index declines by 20%, triggering a Level 3 circuit-breaker, at any time, trading would be halted for the remainder of the day.

A Level 1 or Level 2 halt can only occur once per trading day. For example, following the reopening of trading after a Level 1 Market Decline halt, the NYSE would not halt the market again unless a Level 2 Market Decline was to occur. Likewise, following the reopening of trading after a Level 2 Market Decline, the NYSE would not halt trading again unless a Level 3 Market Decline were to occur, at which point, trading in all stocks would be halted until the primary market opens the next trading day.

Limit Up/Limit Down

On May 31, 2012, the SEC approved, on a pilot basis, a National Market System Plan, known as the Limit Up/Limit Down (“LULD”) Plan, to address extraordinary market volatility.

The Plan is designed to prevent trades in NMS Stocks from occurring outside specified price bands, which are set at a percentage level above and below the average reference price of a security over the preceding five-minute period. The percentage level is determined by a security’s designation as a Tier 1 or Tier 2 security. Tier 1 comprises all securities in the S&P 500, the Russell 1000 and select Exchange Traded Products (ETPs). Tier 2 comprises all other NMS securities, except for rights and warrants, which are specifically excluded from coverage. The Plan applies during regular trading hours of 9:30 am ET - 4:00 pm ET.

https://www.nyse.com/markets/nyse/trading-info

It appears that the United States markets and US Government have a plan in place to let cooler heads prevail overnight.

[siamond]

I am trying to unravel this modeling issue. What is clear so far is that the 3x tracking for UPRO/SSO/S&P 500 is based on the *price* index, not on the Total Return index. Now how are dividends processed remains elusive. If anybody knows the corresponding mechanics of such leveraged fund, please speak up...

The majority of the underlying holdings of UPRO and TMF are swaps and futures contracts, which do not yield dividends.

But both funds do hold some actual shares of stocks and bonds, probably as collateral for the derivatives. It's those holdings that yield dividends.

Yes, I got that. This document clarifies the matter and (rather unhelpfully) states that actual equity shares would be 10% to 80%. This doesn't tell us how those leveraged funds process dividends from the actual holdings though. They clearly reinvest *some* of it (in 2017, UPRO didn't distribute ANY dividends), while distributing a trickle (why?). How much, when and how remains to be understood.

For UPRO/SSO, I agree with samsdad that we could just do conservative price-only math, and this might be good enough for modeling (although not fully satisfying), but for TMF, the situation is clearly very different. I am focusing on UPRO/SSO/ULPIX for now, testing one possible hypothesis... This is certainly intriguing. As a side note, the ULPIX (2x fund, 1997+) trajectory against the S&P 500 is not exactly pretty (admittedly, this was a tough time period for a leveraged fund, but such is life).

[folkher0]

I am trying to unravel this modeling issue. What is clear so far is that the 3x tracking for UPRO/SSO/S&P 500 is based on the *price* index, not on the Total Return index. Now how are dividends processed remains elusive. If anybody knows the corresponding mechanics of such leveraged fund, please speak up...

The majority of the underlying holdings of UPRO and TMF are swaps and futures contracts, which do not yield dividends.

But both funds do hold some actual shares of stocks and bonds, probably as collateral for the derivatives. It's those holdings that yield dividends.

OP:

This is an intriguing strategy that has really turned a lot of heads around here. The recent questions in this thread about dividends raises a point that you addressed here but which warrants reinforcement for those of us with minimal experience with leveraged ETFs.

UPRO is not an index fund. It is a complex construction of derivatives designed to return 3x the S&P. From what I can determine reading the prospectus there is actually no guarantee that it continues to perform 3x S&P in perpetuity. In fact, from what I can tell these funds require some pretty aggressive math to bring the expected returns. Behind the scenes it’s kind of a black box and doesn’t pass the “if I don’t understand it I probably shouldn’t invest in it” test.

OP clearly gets this, but I’m not sure the casual follower will.

One of the first clues to that is that it doesn’t generate dividends like a true S&P fund does.

I’m intrigued by your presentation, and believe your back testing. But it seems that there’s more than meets the eye with these leveraged funds.

[Oicuryy]

What is clear so far is that the 3x tracking for UPRO/SSO/S&P 500 is based on the *price* index, not on the Total Return index.

Is it really that clear? The Direxion equivalent, SPXL, says it tracks the SPXT index which appears to be a total return index.
http://www.direxioninvestments.com/prod ... ull-3x-etf
https://www.marketwatch.com/investing/i ... trycode=xx

TMF tracks index IDCOT20TR. Does TR mean total return?
http://www.direxioninvestments.com/prod ... ull-3x-etf

Ron

[siamond]

What is clear so far is that the 3x tracking for UPRO/SSO/S&P 500 is based on the *price* index, not on the Total Return index.

Is it really that clear? The Direxion equivalent, SPXL, says it tracks the SPXT index which appears to be a total return index.
http://www.direxioninvestments.com/prod ... ull-3x-etf
https://www.marketwatch.com/investing/i ... trycode=xx

TMF tracks index IDCOT20TR. Does TR mean total return?
http://www.direxioninvestments.com/prod ... ull-3x-etf

Ron

Yes, I think it is for the S&P 500 funds. I downloaded the daily NAV values for SSO and for the S&P 500 index, then computed daily gains and losses. The 2x daily match is very clear (they actually do a really good job of it, except in 2007 where they screwed the pooch by losing a tiny bit every day (which ended up a lot at the end of the year!), then they did much better in the following years. I will double-check with UPRO and ULPIX, to be 100% sure. Also, I did the naive 3x 'Total Return' simulation yesterday, and it overshoots the UPRO actuals quite significantly.

Yes, for TMF, this might be different, I don't know, I'm not there yet. I don't understand how one could leverage anything else than a price, but heck, I'm learning as we go here, this is all very new to me.

[pdavi21]

The real problem here is if US markets perform over the next ten years like emerging markets did over the last ten. A 60/40 EMx3/TMF lost to EM/EDV even from the bottom of the crash. EMx3 lost to EM from the bottom of the 2009 crash. Developed Markets x3 lost from near the bottom of the crash, but TMF saves it and gives it a higher CAGR than x1 in a 60/40.

I think OP should consider diversifying as a diversified portfolio of x3 funds would've relied heavily on the outsized gains of one country (USA) and the negative correlation of long treasuries.

Have mentioned this previously: the higher underlying volatility of those other stock indexes poses greater risk of volatility drag and a complete wipeout of those particular 3x funds. When you're playing with 3x leverage, you want to diversify correlations, but not at the cost of unnecessarily increasing volatility.

Posting volatility drag table here again:

Fine, skip EM, but are you saying a Developed Markets or Total INTL or Total Global (x3 exists?) will be more volatile than a US market Fund in the next decade?

EDIT: I guess I'm trying to say, is a diversified portfolio including higher volatility assets (like small cap and INTL) is likely to have a lower volatility than just one less risky holding. Two ways to approach would be a Global x3 fund (if it exists) which could theorhetically have lower volatility amd higher risk adjusted return than US large cap, or to simply hold a basket of lower correlation assets x3, any one of which could have outsized gains and low volatility over the next decade, driving returns for the portfolio (like US large cap over the last decade).

I like your idea (although I won't be trying x3 funds unless the market tanks so low, I feel compelled to change my AA to 100% stocks, have no low APY debt available, and want to bet more money). I just don't see your justification for holding one country. I do see your justification for holding one type of bond.

[samsdad]

The real problem here is if US markets perform over the next ten years like emerging markets did over the last ten. A 60/40 EMx3/TMF lost to EM/EDV even from the bottom of the crash. EMx3 lost to EM from the bottom of the 2009 crash. Developed Markets x3 lost from near the bottom of the crash, but TMF saves it and gives it a higher CAGR than x1 in a 60/40.

I think OP should consider diversifying as a diversified portfolio of x3 funds would've relied heavily on the outsized gains of one country (USA) and the negative correlation of long treasuries.

Have mentioned this previously: the higher underlying volatility of those other stock indexes poses greater risk of volatility drag and a complete wipeout of those particular 3x funds. When you're playing with 3x leverage, you want to diversify correlations, but not at the cost of unnecessarily increasing volatility.

Posting volatility drag table here again:

Fine, skip EM, but are you saying a Developed Markets or Total INTL or Total Global (x3 exists?) will be more volatile than a US market Fund in the next decade?

EDIT: I guess I'm trying to say, is a diversified portfolio including higher volatility assets (like small cap and INTL) is likely to have a lower volatility than just one less risky holding. Two ways to approach would be a Global x3 fund (if it exists) which could theorhetically have lower volatility amd higher risk adjusted return than US large cap, or to simply hold a basket of lower correlation assets x3, any one of which could have outsized gains tp to high returns a low volatility over the next decade (like US large cap over the last decade).

I like your idea (although I won't be trying x3 funds unless the market tanks so low, I feel compelled to change my AA to 100% stocks, have no low APY debt available, and want to bet more money). I just don't see your justification for holding one country. I do see your justification for holding one type of bond.

Do these foreign markets have the circuit breakers and limit-down rules in place to limit daily volatility as I noted above for the US?

[hushpuppy]

I am trying to unravel this modeling issue. What is clear so far is that the 3x tracking for UPRO/SSO/S&P 500 is based on the *price* index, not on the Total Return index. Now how are dividends processed remains elusive. If anybody knows the corresponding mechanics of such leveraged fund, please speak up...

The majority of the underlying holdings of UPRO and TMF are swaps and futures contracts, which do not yield dividends.

But both funds do hold some actual shares of stocks and bonds, probably as collateral for the derivatives. It's those holdings that yield dividends.

OP:

This is an intriguing strategy that has really turned a lot of heads around here. The recent questions in this thread about dividends raises a point that you addressed here but which warrants reinforcement for those of us with minimal experience with leveraged ETFs.

UPRO is not an index fund. It is a complex construction of derivatives designed to return 3x the S&P. From what I can determine reading the prospectus there is actually no guarantee that it continues to perform 3x S&P in perpetuity. In fact, from what I can tell these funds require some pretty aggressive math to bring the expected returns. Behind the scenes it’s kind of a black box and doesn’t pass the “if I don’t understand it I probably shouldn’t invest in it” test.

OP clearly gets this, but I’m not sure the casual follower will.

One of the first clues to that is that it doesn’t generate dividends like a true S&P fund does.

I’m intrigued by your presentation, and believe your back testing. But it seems that there’s more than meets the eye with these leveraged funds.

I am very interested in this thread and wish all the best.

“if I don’t understand it I probably shouldn’t invest in it”

Frankly, I have no more ability to evaluate this strategy than my two dogs do (Both currently asleep). Therefore, I am not in the least tempted to try this for myself. That does not mean I can't appreciate the folks in this thread that are willing and able to put their ideas in front of this audience.

My thanks for an interesting thread.

Good luck,

hushpuppy

[pdavi21]

The real problem here is if US markets perform over the next ten years like emerging markets did over the last ten. A 60/40 EMx3/TMF lost to EM/EDV even from the bottom of the crash. EMx3 lost to EM from the bottom of the 2009 crash. Developed Markets x3 lost from near the bottom of the crash, but TMF saves it and gives it a higher CAGR than x1 in a 60/40.

I think OP should consider diversifying as a diversified portfolio of x3 funds would've relied heavily on the outsized gains of one country (USA) and the negative correlation of long treasuries.

Have mentioned this previously: the higher underlying volatility of those other stock indexes poses greater risk of volatility drag and a complete wipeout of those particular 3x funds. When you're playing with 3x leverage, you want to diversify correlations, but not at the cost of unnecessarily increasing volatility.

Posting volatility drag table here again:

Fine, skip EM, but are you saying a Developed Markets or Total INTL or Total Global (x3 exists?) will be more volatile than a US market Fund in the next decade?

EDIT: I guess I'm trying to say, is a diversified portfolio including higher volatility assets (like small cap and INTL) is likely to have a lower volatility than just one less risky holding. Two ways to approach would be a Global x3 fund (if it exists) which could theorhetically have lower volatility amd higher risk adjusted return than US large cap, or to simply hold a basket of lower correlation assets x3, any one of which could have outsized gains tp to high returns a low volatility over the next decade (like US large cap over the last decade).

I like your idea (although I won't be trying x3 funds unless the market tanks so low, I feel compelled to change my AA to 100% stocks, have no low APY debt available, and want to bet more money). I just don't see your justification for holding one country. I do see your justification for holding one type of bond.

Do these foreign markets have the circuit breakers and limit-down rules in place to limit daily volatility as I noted above for the US?

I don't know. Won't the market price in downside protection if it exists? That would theorhetically make US options more expensive.

[staythecourse]

Has anyone mentioned about what happened to XIV? Might happen to UPRO too. Beware.

XIV was a 1x inverse ETF tracking the VIX. So it promised to deliver -100% of the daily movement of the VIX.

On Feb 5, 2018 the VIX went up by 116%. So guess what happened to XIV? Exactly what was supposed to happen, it got wiped out. In that sense, the fund worked perfectly.

UPRO promises 3x the daily returns of the S&P 500. So if we ever see a day when the S&P moves down -33.3%, UPRO will be wiped out as well. The point of this strategy though, is that TMF is supposed to move in the opposite direction if/when this happens.

FWIW, the largest one day drop in the S&P 500 was Black Monday 1987 when it dropped -20.4%.

From the horse’s mouth, the New York Stock Exchange (my emphasis):

Circuit Breakers

In response to the market breaks in October 1987 and October 1989, the New York Stock Exchange instituted circuit breakers to reduce volatility and promote investor confidence. By implementing a pause in trading, investors are given time to assimilate incoming information and the ability to make informed choices during periods of high market volatility.

NYSE Rule 80B provides that a circuit-breaker halt for a Level 1 (7%) or Level 2 (13%) decline in the S&P 500 Index occurring after 9:30 a.m. Eastern and up to and including 3:25 p.m. Eastern, or in the case of an early scheduled close, 12:25 p.m. Eastern, would result in a trading halt in all stocks for 15 minutes. If the S&P 500 Index declines by 20%, triggering a Level 3 circuit-breaker, at any time, trading would be halted for the remainder of the day.

A Level 1 or Level 2 halt can only occur once per trading day. For example, following the reopening of trading after a Level 1 Market Decline halt, the NYSE would not halt the market again unless a Level 2 Market Decline was to occur. Likewise, following the reopening of trading after a Level 2 Market Decline, the NYSE would not halt trading again unless a Level 3 Market Decline were to occur, at which point, trading in all stocks would be halted until the primary market opens the next trading day.

Limit Up/Limit Down

On May 31, 2012, the SEC approved, on a pilot basis, a National Market System Plan, known as the Limit Up/Limit Down (“LULD”) Plan, to address extraordinary market volatility.

The Plan is designed to prevent trades in NMS Stocks from occurring outside specified price bands, which are set at a percentage level above and below the average reference price of a security over the preceding five-minute period. The percentage level is determined by a security’s designation as a Tier 1 or Tier 2 security. Tier 1 comprises all securities in the S&P 500, the Russell 1000 and select Exchange Traded Products (ETPs). Tier 2 comprises all other NMS securities, except for rights and warrants, which are specifically excluded from coverage. The Plan applies during regular trading hours of 9:30 am ET - 4:00 pm ET.

https://www.nyse.com/markets/nyse/trading-info

It appears that the United States markets and US Government have a plan in place to let cooler heads prevail overnight.

I thought about this as well. The worst will be the market drops 20% (60% for the leveraged funds). I can BET if that did happen there will be a surge of active bets in the next trading day on either side of the movement (3x Sp500 or 3x LTT).

The circuit breakers again have helped the leveraged funds as they can't be wiped out in a single, crazy day.

I am still worried about the funds themselves not shutting down as i don't fully understand HOW they pay for the leverage and if that money will be asked ASAP if it hit same 50% loss in a given year or 60% loss in a given year. Good to know a 2x sp500 fund did survive 2008.

Good luck.

[samsdad]

VXUS is one of the ETPs subject to the limit-down plan FYI. See page 4. http://www.luldplan.com/plans/AppendixA ... 2_2019.pdf

[Oicuryy]

Yes, I think it is for the S&P 500 funds. I downloaded the daily NAV values for SSO and for the S&P 500 index, then computed daily gains and losses. The 2x daily match is very clear (they actually do a really good job of it, except in 2007 where they screwed the pooch by losing a tiny bit every day (which ended up a lot at the end of the year!), then they did much better in the following years. I will double-check with UPRO and ULPIX, to be 100% sure. Also, I did the naive 3x 'Total Return' simulation yesterday, and it overshoots the UPRO actuals quite significantly.

3x the price index might be a good model for 3x the total return index minus expenses. Maybe they are using their stock dividends to pay the expenses of the swap contracts.

Ron

[jaj2276]

Instead of posting pictures some more, let me just tell you what my data set says about rebalancing. Again CAUTION!

40/60 3X Initial bal:$10k

Code: Select all

1950-2018	Final Balance	CAGR	Stdev	Best Year Worst Year	Max. Drawdown	Sharpe Sortino  US Mkt Correlation
No rebalancing:   $123,159,666 	14.86% 	36.19%	124.87%	 -65.13%	-83.67% 	0.46	0.69		0.91
Annually: 	$1,723,365,187 	19.40% 	28.95%	110.13%	 -24.77%	-49.99% 	0.60	1.34		0.60
Semi: 		$1,481,610,072 	19.14% 	28.95%	118.02%	 -28.32%	-50.22% 	0.59	1.35		0.59
Quarterly:	$1,205,486,523 	18.77% 	28.88%	119.89%	 -27.70%	-50.67% 	0.58	1.31		0.61
Monthly: 	$1,070,252,880 	18.57% 	29.58%	120.28%	 -31.38%	-54.33% 	0.57	1.28		0.60
Bands: 		  $993,315,909 	18.44% 	29.31%	124.07%	 -30.42%	-54.39% 	0.57	1.26		0.61

1950-1981
No rebalancing: $162,620 	9.41% 	35.01%	110.41%	 -65.13%	-83.67% 	0.30	0.44		0.98
Annually:	$250,933 	10.96% 	20.50%	79.29%	 -24.77%	-46.34% 	0.38	0.67		0.80
Semi:		$203,021 	10.20% 	20.21%	69.54%	 -28.32%	-50.22% 	0.35	0.60		0.82
Quarterly:	$177,834 	9.73% 	20.40%	64.90%	 -27.70%	-50.67% 	0.33	0.56		0.83
Monthly:	$156,258 	9.27% 	20.59%	64.22%	 -31.38%	-54.33% 	0.31	0.52		0.83
Bands: 		$157,001 	9.29% 	20.61%	64.47%	 -30.42%	-54.39% 	0.31	0.52		0.84

1982-2018
No rebalancing: $17,512,887 	22.37% 	38.46%	101.53%	 -49.38%	-50.72% 	0.59	1.57		0.30
Annually:	$68,678,430 	26.97% 	34.34%	110.13%	 -22.43%	-49.99% 	0.74	1.81		0.53
Semi:		$72,978,304 	27.18% 	34.47%	118.02%	 -17.38%	-41.62% 	0.74	1.91		0.51
Quarterly:	$67,787,355 	26.92% 	34.26%	119.89%	 -18.04%	-43.37% 	0.74	1.86		0.52
Monthly:	$68,492,562 	26.96% 	35.25%	120.28%	 -17.04%	-45.17% 	0.73	1.84		0.52
Bands:		$63,377,287 	26.69% 	34.82%	124.07%	 -18.62%	-44.89% 	0.72	1.82		0.52

.

I don't doubt the numbers, but can you do your rebalancing bands on the time period when the things we can actually invest in (assuming we don't have the ability to do swaps and futures) are available? So maybe 6/25/2009 on (that was UPRO inception date, TMF came about a few months earlier). The reason I ask is because my rebalancing numbers show that quarterly rebal was the best (similar to the OPs). I started my portfolio in Feb of 2010 both to make sure that UPRO and TMF were around and because it was Feb when I started reading this thread.

[samsdad]

I’ve edited this post to follow EfficientInvestor’s lead regarding the simulated data issue:

UPDATE (2/20/19) - The data that had previously been provided in this post was based on developing a best fit to the data of actual leveraged ETFs since their inception. Since inception of 3X ETFs in 2009, borrowing rates have been low. Therefore, this data did not take into account higher borrowing rates that occurred prior to 2009. Due to this, I have removed the data I had previously provided.

[HEDGEFUNDIE]

Fine, skip EM, but are you saying a Developed Markets or Total INTL or Total Global (x3 exists?) will be more volatile than a US market Fund in the next decade?

EDIT: I guess I'm trying to say, is a diversified portfolio including higher volatility assets (like small cap and INTL) is likely to have a lower volatility than just one less risky holding. Two ways to approach would be a Global x3 fund (if it exists) which could theorhetically have lower volatility amd higher risk adjusted return than US large cap, or to simply hold a basket of lower correlation assets x3, any one of which could have outsized gains and low volatility over the next decade, driving returns for the portfolio (like US large cap over the last decade).

I like your idea (although I won't be trying x3 funds unless the market tanks so low, I feel compelled to change my AA to 100% stocks, have no low APY debt available, and want to bet more money). I just don't see your justification for holding one country. I do see your justification for holding one type of bond.

You're right that EAFE is not significantly more volatile than the S&P 500. Over the past 30 years EAFE annual volatility has been 17% compared to the S&P's 15%.

Here is a backtest of 40% UPRO vs. 20% UPRO + 20% DZK (the 3x EAFE ETF).



All I will say is, I really hope this doesn't turn into yet another "what good are international stocks" thread

EDIT: On second thought I would like to explore this further. Since 1970, EAFE has only been 0.69 correlated with the S&P 500 without being much more volatile. Theoretically it should offer a diversification benefit to the strategy without adding too much volatility risk.

[EfficientInvestor]

What is clear so far is that the 3x tracking for UPRO/SSO/S&P 500 is based on the *price* index, not on the Total Return index.

Is it really that clear? The Direxion equivalent, SPXL, says it tracks the SPXT index which appears to be a total return index.
http://www.direxioninvestments.com/prod ... ull-3x-etf
https://www.marketwatch.com/investing/i ... trycode=xx

TMF tracks index IDCOT20TR. Does TR mean total return?
http://www.direxioninvestments.com/prod ... ull-3x-etf

Ron

Yes, I think it is for the S&P 500 funds. I downloaded the daily NAV values for SSO and for the S&P 500 index, then computed daily gains and losses. The 2x daily match is very clear (they actually do a really good job of it, except in 2007 where they screwed the pooch by losing a tiny bit every day (which ended up a lot at the end of the year!), then they did much better in the following years. I will double-check with UPRO and ULPIX, to be 100% sure. Also, I did the naive 3x 'Total Return' simulation yesterday, and it overshoots the UPRO actuals quite significantly.

Yes, for TMF, this might be different, I don't know, I'm not there yet. I don't understand how one could leverage anything else than a price, but heck, I'm learning as we go here, this is all very new to me.

See the link below that will take you to Direxion's prospectus page for TMF. Under the "Target Index" section, it states that the fund tracks the ICE U.S. Treasury 20+ Year Bond Index (IDCOT20TR), "TR" standing for Total Return. Therefore, the intent is to achieve results that meet 3X the total return, not just the underlying price. How they go about meeting that total return takes a little digging.

If you click on the "Daily Holdings" tab, you will see the current holdings. If you multiply current price of $19.78 x shares outstanding listed of 5.35MM, you get a AUM of $105.8MM. Of that, you can see that $73.4MM (2/3 of fund value) is in TLT. This is where TMF's 1.48% yield currently comes from. Everything else listed in the daily holdings is either cash (Bank of NY Cash Reserve and the Goldman products) or a derivative (ISHARES BOND ETF SWAP). However, if you add up the market value in TLT and the ETF SWAPs, you get a total of $317.2MM, which is almost exactly 3X of the $105.8MM AUM. Therefore, their market exposure is 3X their AUM, which is what it is supposed to be. As for all the derivatives, I'm not going to pretend to know the exact details of all of them. However, it is my understanding that these derivatives are inclusive of dividends and not just based on the underlying price of TLT. If they weren't there would be an arbitrage possibility because you could then just go long TLT, go short the TLT swaps, and pocket the yield you get from being long TLT without any risk.

All that being said...I believe using daily data of TLT is the best way to model the performance of TMF prior to its inception (they track the same index). However, TLT hasn't been around very long either. That is why I used VUSTX instead. As shown in my article about creating the proxy funds, my TMF proxy was created using a weighted average that consisted of 94% of the adjusted daily close of VUSTX (w/ dividend) and only 6% without and then 1% for expense ratio. In other words, the actual returns of TMF are largely influenced by the dividend (total return) of the bond fund, not just the price of the index (or the price of TLT).

http://www.direxioninvestments.com/prod ... ull-3x-etf

[samsdad]

I’ve edited this post to follow EfficientInvestor’s lead regarding the simulated data issue:

UPDATE (2/20/19) - The data that had previously been provided in this post was based on developing a best fit to the data of actual leveraged ETFs since their inception. Since inception of 3X ETFs in 2009, borrowing rates have been low. Therefore, this data did not take into account higher borrowing rates that occurred prior to 2009. Due to this, I have removed the data I had previously provided.

[EfficientInvestor]

What is clear so far is that the 3x tracking for UPRO/SSO/S&P 500 is based on the *price* index, not on the Total Return index.

Is it really that clear? The Direxion equivalent, SPXL, says it tracks the SPXT index which appears to be a total return index.
http://www.direxioninvestments.com/prod ... ull-3x-etf
https://www.marketwatch.com/investing/i ... trycode=xx

TMF tracks index IDCOT20TR. Does TR mean total return?
http://www.direxioninvestments.com/prod ... ull-3x-etf

Ron

Yes, I think it is for the S&P 500 funds. I downloaded the daily NAV values for SSO and for the S&P 500 index, then computed daily gains and losses. The 2x daily match is very clear (they actually do a really good job of it, except in 2007 where they screwed the pooch by losing a tiny bit every day (which ended up a lot at the end of the year!), then they did much better in the following years. I will double-check with UPRO and ULPIX, to be 100% sure. Also, I did the naive 3x 'Total Return' simulation yesterday, and it overshoots the UPRO actuals quite significantly.

Yes, for TMF, this might be different, I don't know, I'm not there yet. I don't understand how one could leverage anything else than a price, but heck, I'm learning as we go here, this is all very new to me.

See the link below that will take you to Direxion's prospectus page for TMF. Under the "Target Index" section, it states that the fund tracks the ICE U.S. Treasury 20+ Year Bond Index (IDCOT20TR), "TR" standing for Total Return. Therefore, the intent is to achieve results that meet 3X the total return, not just the underlying price. How they go about meeting that total return takes a little digging.

If you click on the "Daily Holdings" tab, you will see the current holdings. If you multiply current price of $19.78 x shares outstanding listed of 5.35MM, you get a AUM of $105.8MM. Of that, you can see that $73.4MM (2/3 of fund value) is in TLT. This is where TMF's 1.48% yield currently comes from. Everything else listed in the daily holdings is either cash (Bank of NY Cash Reserve and the Goldman products) or a derivative (ISHARES BOND ETF SWAP). However, if you add up the market value in TLT and the ETF SWAPs, you get a total of $317.2MM, which is almost exactly 3X of the $105.8MM AUM. Therefore, their market exposure is 3X their AUM, which is what it is supposed to be. As for all the derivatives, I'm not going to pretend to know the exact details of all of them. However, it is my understanding that these derivatives are inclusive of dividends and not just based on the underlying price of TLT. If they weren't there would be an arbitrage possibility because you could then just go long TLT, go short the TLT swaps, and pocket the yield you get from being long TLT without any risk.

All that being said...I believe using daily data of TLT is the best way to model the performance of TMF prior to its inception (they track the same index). However, TLT hasn't been around very long either. That is why I used VUSTX instead. As shown in my article about creating the proxy funds, my TMF proxy was created using a weighted average that consisted of 94% of the adjusted daily close of VUSTX (w/ dividend) and only 6% without and then 1% for expense ratio. In other words, the actual returns of TMF are largely influenced by the dividend (total return) of the bond fund, not just the price of the index (or the price of TLT).

http://www.direxioninvestments.com/prod ... ull-3x-etf

Now for UPRO or SPXL (Direxion's 3X S&P 500 fund). I will stick with SPXL since we are now used to navigating Direxion's prospectus page for TMF. See the link below for the SPXL prospectus page. The Target Index is SPXT which is also a total return index. As with TMF, SPXL has a large portion of their AUM in an unleveraged ETF that tracks the same index. In this case, they have what appears to be over 70% of AUM in IVV. The rest is in cash and S&P 500 index swaps. Total market value exposure in the ETF and index swaps add up to 3X the AUM. The IVV holdings produce a dividend and the index swaps are assumed to be inclusive of dividend payments due to the arbitrage opportunity that would otherwise occur (unless I'm not understanding something there). All that said, the actual return of UPRO or SPXL is somewhere in between the price return and total return due to fees or other expenses or slippages involved in operating the funds.

http://www.direxioninvestments.com/prod ... ull-3x-etf

[samsdad]

Here's UPROs portfolio internals according to M*, at least the top stuff:

[EfficientInvestor]

What is clear so far is that the 3x tracking for UPRO/SSO/S&P 500 is based on the *price* index, not on the Total Return index.

Is it really that clear? The Direxion equivalent, SPXL, says it tracks the SPXT index which appears to be a total return index.
http://www.direxioninvestments.com/prod ... ull-3x-etf
https://www.marketwatch.com/investing/i ... trycode=xx

TMF tracks index IDCOT20TR. Does TR mean total return?
http://www.direxioninvestments.com/prod ... ull-3x-etf

Ron

Yes, I think it is for the S&P 500 funds. I downloaded the daily NAV values for SSO and for the S&P 500 index, then computed daily gains and losses. The 2x daily match is very clear (they actually do a really good job of it, except in 2007 where they screwed the pooch by losing a tiny bit every day (which ended up a lot at the end of the year!), then they did much better in the following years. I will double-check with UPRO and ULPIX, to be 100% sure. Also, I did the naive 3x 'Total Return' simulation yesterday, and it overshoots the UPRO actuals quite significantly.

Yes, for TMF, this might be different, I don't know, I'm not there yet. I don't understand how one could leverage anything else than a price, but heck, I'm learning as we go here, this is all very new to me.

See the link below that will take you to Direxion's prospectus page for TMF. Under the "Target Index" section, it states that the fund tracks the ICE U.S. Treasury 20+ Year Bond Index (IDCOT20TR), "TR" standing for Total Return. Therefore, the intent is to achieve results that meet 3X the total return, not just the underlying price. How they go about meeting that total return takes a little digging.

If you click on the "Daily Holdings" tab, you will see the current holdings. If you multiply current price of $19.78 x shares outstanding listed of 5.35MM, you get a AUM of $105.8MM. Of that, you can see that $73.4MM (2/3 of fund value) is in TLT. This is where TMF's 1.48% yield currently comes from. Everything else listed in the daily holdings is either cash (Bank of NY Cash Reserve and the Goldman products) or a derivative (ISHARES BOND ETF SWAP). However, if you add up the market value in TLT and the ETF SWAPs, you get a total of $317.2MM, which is almost exactly 3X of the $105.8MM AUM. Therefore, their market exposure is 3X their AUM, which is what it is supposed to be. As for all the derivatives, I'm not going to pretend to know the exact details of all of them. However, it is my understanding that these derivatives are inclusive of dividends and not just based on the underlying price of TLT. If they weren't there would be an arbitrage possibility because you could then just go long TLT, go short the TLT swaps, and pocket the yield you get from being long TLT without any risk.

All that being said...I believe using daily data of TLT is the best way to model the performance of TMF prior to its inception (they track the same index). However, TLT hasn't been around very long either. That is why I used VUSTX instead. As shown in my article about creating the proxy funds, my TMF proxy was created using a weighted average that consisted of 94% of the adjusted daily close of VUSTX (w/ dividend) and only 6% without and then 1% for expense ratio. In other words, the actual returns of TMF are largely influenced by the dividend (total return) of the bond fund, not just the price of the index (or the price of TLT).

http://www.direxioninvestments.com/prod ... ull-3x-etf

Now for UPRO or SPXL (Direxion's 3X S&P 500 fund). I will stick with SPXL since we are now used to navigating Direxion's prospectus page for TMF. See the link below for the SPXL prospectus page. The Target Index is SPXT which is also a total return index. As with TMF, SPXL has a large portion of their AUM in an unleveraged ETF that tracks the same index. In this case, they have what appears to be over 70% of AUM in IVV. The rest is in cash and S&P 500 index swaps. Total market value exposure in the ETF and index swaps add up to 3X the AUM. The IVV holdings produce a dividend and the index swaps are assumed to be inclusive of dividend payments due to the arbitrage opportunity that would otherwise occur (unless I'm not understanding something there). All that said, the actual return of UPRO or SPXL is somewhere in between the price return and total return due to fees or other expenses or slippages involved in operating the funds.

http://www.direxioninvestments.com/prod ... ull-3x-etf

A few other thoughts...I'll use SPXL as an example. It appears that almost 1/3 of current AUM is sitting in cash or cash equivalents. This makes me consider 2 things. First is that they are earning a rate of return on that cash (current rate around 2%), which is a good thing. Second is that they have cash on hand in the event of a margin call associated with the swaps. So for those that have been asking about what happens when there are big dips in the index and they get a margin call, I think this is most of the answer. They either move some AUM from cash to the swaps, or they sell some shares of the underlying ETFs and move that over to cover the required margin.

One more thing to note...futures margin is different than broker margin. They are not borrowing money to use futures the way you may borrow money from your broker in your margin account. However, there is a requirement that they put up a certain amount of collateral (aka, margin) in order to show an act of good faith towards their counter-party in the contract. The leverage that comes from futures/swaps is not from borrowing. It comes from the fact that you can get much greater exposure to the index with much less money because the margin requirement is only 5% of the amount of market value you are actually exposed to.

[Nickel & Dime]

Hmm, wonder if TMF 3x treasury dividends are exempt from state tax...

[samsdad]

From the UPRO annual report page:

For the year ended May 31, 2018, the Fund had a total return of 35.82%1. For the same period, the Index had a total return of 14.38% ² and a volatility of 12.39%. For the period, the Fund had an average daily volume of 5,042,725 shares and an average daily statistical correlation of over 0.99 to three times that of the daily perform- ance of the Index.³

The Fund takes positions in securities and/or derivatives that, in combination, should have similar daily return characteristics as three times the daily return of the Index. The Index is a measure of large-cap U.S. stock market performance. It is a float-adjusted, market capitalization-weighted index of 500 U.S. operating companies and real estate investment trusts selected through a process that factors in criteria such as liquidity, price, market capitalization and financial viability.

During the year ended May 31, 2018, the Fund invested in swap agreements and futures contracts in order to gain leveraged exposure to the Index. These derivatives generally tracked the performance of their underlying index and the Fund was generally negatively impacted by financing rates associated with swap agreements. The Fund entered into swap agreements with counterparties that the Fund’s advisor determined to be major, global financial institutions. If a counterparty becomes insolvent or otherwise fails to perform on its obligations, the value of investments in the Fund may decline. The Fund has sought to mitigate this risk by generally requiring counterparties for the Fund to post collateral for the benefit of the Fund, marked to market daily, in an amount approximately equal to the amount the counterparty owed the Fund, subject to certain minimum thresholds.

http://www.proshares.com/media/document ... 9832890959

Finally found some more info on page 123-125ish; 80% is common stock from the S&P 500 (page 123), the rest is short-term investments, repurchase agreements, and futures contracts (pages 124-125).