HEDGEFUNDIE's excellent adventure Part II: The next journey

[langlands]

EDIT: I should add - if I am understand this correctly - there is only one possible instantiation of the TV model at each ratio, versus 30 possible instantiations of HFEA.

Yes, that's right. I mean, there's a lot of possible ratios in the TV model as well so it's not like it has fewer parameters.

[Hydromod]

The backtesting I’ve done suggests that rebalancing near the Jan 1 quarters has performed noticeably better than quarterly strategies away from these dates. Maybe something to do with end of quarter actions.

Much much more likely that it's because of backtesting overfit. If end of quarter actions could have such a large (and reliable) effect, the markets would be ridiculously inefficient and we should all be doing market timing like clock-work.

Sure, it's debatable whether this is a real effect. But I would stick to it going forward regardless, out of superstition if nothing else, if I were doing quarterly rebalancing. It shouldn't hurt, at least.

Others have noticed the monthly cyclicity in bonds and equities. I would not be at all surprised if this effect may already be disappearing or have disappeared already.


The fact that different dates of rebalancing can have such a big impact on the strategy (as noticed also by many participants of this strategy who've traded it live over the past year) shows that backtesting a single "representative" of the strategy from the class of all equivalent such strategies is extremely inadequate. Basically, the variance in outcomes is massive and more independent Monte Carlo simulations are needed to get any semblance of statistical significance. Comparing single backtest results between arbitrary instantiations of different strategies is meaningless because there is probably just as much variation between instantiations of the same strategy as there is between strategies.

This point is especially obvious if one looks at some of the results from here, where I looked at different weights with monthly rebalancing. Some of the other posts look at other weighting schemes.

The figure in that post shows every 5-year sequence with 21-day rebalancing over the history of the simulated daily UPRO & TMF (starting end of 1986), looking at combinations from 30/70 to 60/40 UPROSIM/TMFSIM compared to the original 40/60 UPROSIM/TMFSIM with quarterly rebalancing.

I'd post the image again here but it seems to be too large.

The second plot on the right in the linked figure shows the cumulative distribution of all of the 5-year sequences. The squares are the mean of the distribution and the vertical bars on the bottom show the median for the distributions. The quarterly rebalance set is right in the middle of the weighted sets. The key is that the spread in 5-year outcomes is way way larger than the spread in the different weighting schemes, which implies that the time period is way way more important than the weighting scheme. In other words, "the variance in outcomes is massive."

The weighting in UPRO versus TMF didn't greatly affect the mean/median CAGR (except that the blend performed better than either UPRO or TMF in isolation), but did substantially affect the spread in 5-year CAGR. The 30/70 scheme had almost all outcomes between 0 and 30 CAGR, the 60/40 had almost all outcomes between -10 and 50 CAGR. In other words, a greater weight in UPRO gave a small increase in mean/median CAGR with a substantial increase in the spread of results. I'd expect that the effect of weighting on the spread will probably be fairly robust going forward, although the mean/median value is less robust.

[langlands]

Hi Hydromod, thanks for the data. I'll note that the choices of weightings you chose was a neighborhood of 50/50 (i.e. if you simulated 100% UPRO you'd find a pretty different average CAGR). Because of the "quadratic effect" of local optima, this gives me some confidence that 50/50 is indeed around the optimal weight. (What I mean by this is that when you find the parameters that are optimal, small deviations make to first order zero difference, which is a nice property)

I still contend (and your data seems to back it up) that quarterly rebalancing isn't really justified, although it probably isn't disastrous either.

[djeayzonne]

I have an account, but maybe I can't access your saved models.

I also can't get the numbers you're getting with HFEA 55/45. Are you adding a cashflow? Re-balancing monthly I get:
[/quote]

No added cash flow. Rebalanced quarterly.
Here are the settings

[tomphilly]

No added cash flow. Rebalanced quarterly.

Yeh, the issue was I was re-balancing monthly. Quarterly makes a big difference. I actually came across this strategy elsewhere when I first started using it last year, and there was less emphasis on a specific re-balancing frequency. The emphasis was on vigilantly maintaining the ratio to avoid getting wiped out in a sudden crash.

[Hydromod]

Hi Hydromod, thanks for the data. I'll note that the choices of weightings you chose was a neighborhood of 50/50 (i.e. if you simulated 100% UPRO you'd find a pretty different average CAGR). Because of the "quadratic effect" of local optima, this gives me some confidence that 50/50 is indeed around the optimal weight. (What I mean by this is that when you find the parameters that are optimal, small deviations make to first order zero difference, which is a nice property)

I still contend (and your data seems to back it up) that quarterly rebalancing isn't really justified, although it probably isn't disastrous either.

100% UPRO and 100% TMF are on the chart too. Both have smaller means and medians than the blends. 100% UPRO has a much wider spread than any of the blends; 100% TMF has a narrower spread.

I found that rebalancing once every day or two did give a noticeable bit of bonus in CAGR (without looking at issues of trading costs and intraday variability, which would likely add significant drag), but the bonus dropped off pretty quickly with increasing duration. Increase in duration beyond a week, up to a year, tended to increase the spread in results without strongly affecting central tendency. Quarterly rebalancing tended to depend more strongly on what part of the quarter that rebalancing occurred.

A lot of performance difference was timing luck around a few key events.

I concluded that bands with 10 or 15 percent spread would have worked as well as anything on average, as long as rebalancing was enforced at least once a year or so, but the rebalancing strategy wasn't all that critical.

Using a criterion for derisking during bears arguably gave better performance than rebalancing strategies. But that gets into market timing, and unfortunately past market conditions are not very informative about derisking dynamics in current market conditions.

[TwoIdenticalIndexes]

After reading some of Breaking the Market's thoughts on rebalancing (only tangentially related), I made two changes on July 31st:

1. I am going to rebalance weekly. Side benefit is that this will hopefully encourage me to stop tinkering so much.
2. I am going from 55-45 to 40-15-45 (UPRO-USTL-TMF). Utilities have lower volatility than SPY and weak correlation to it. Seems like a good deal for the increase in sector-specific risk. Worked out last week, but I have noticed that it seemed like USTL had very high intraday volatility.

I am still up a little bit, but not very much. I started at essentially the worst possible time (late January), but I have to think the dates I rebalanced were unlucky, since others have mentioned being up quite a bit YTD.

[RocketShipTech]

1. I am going to rebalance weekly. Side benefit is that this will hopefully encourage me to stop tinkering so much.

Wouldn’t this have the opposite effect? You will be looking at performance every week...

[TwoIdenticalIndexes]

1. I am going to rebalance weekly. Side benefit is that this will hopefully encourage me to stop tinkering so much.

Wouldn’t this have the opposite effect? You will be looking at performance every week...

I can't help looking at my account at least every day. Definitely not healthy, but giving myself something simple to do is at least an outlet.

[Meaty]

1. I am going to rebalance weekly. Side benefit is that this will hopefully encourage me to stop tinkering so much.

Wouldn’t this have the opposite effect? You will be looking at performance every week...

I can't help looking at my account at least every day. Definitely not healthy, but giving myself something simple to do is at least an outlet.

I like looking daily too. Mostly because I’m now up 50% in 1 year

[firebirdparts]

Yeh, the issue was I was re-balancing monthly. Quarterly makes a big difference. I actually came across this strategy elsewhere when I first started using it last year, and there was less emphasis on a specific re-balancing frequency. The emphasis was on vigilantly maintaining the ratio to avoid getting wiped out in a sudden crash.

It's none of my business, but this doesn't sound right. If UPRO heads to zero, you need to vigorously avoid rebalancing. However, you don't know where it's going ahead of time. There's luck involved. If you rebalanced on January 1, April 1, and July 1, you're pretty happy today. But that's luck.

If you rebalanced daily, by the end of march you'd lost most of the entire portfolio.

[potatopancake]

Yeh, the issue was I was re-balancing monthly. Quarterly makes a big difference. I actually came across this strategy elsewhere when I first started using it last year, and there was less emphasis on a specific re-balancing frequency. The emphasis was on vigilantly maintaining the ratio to avoid getting wiped out in a sudden crash.

It's none of my business, but this doesn't sound right. If UPRO heads to zero, you need to vigorously avoid rebalancing. However, you don't know where it's going ahead of time. There's luck involved. If you rebalanced on January 1, April 1, and July 1, you're pretty happy today. But that's luck.

If you rebalanced daily, by the end of march you'd lost most of the entire portfolio.

I am not convinced that time of investment and rebalancing matters much in the long term. I compared a starting investment date with quarterly re-balancing every month in 2010 and there was not a major difference. CAGR 33. - 35.% for any given month.

Sure, if you are re-balancing daily that is a different story.

[kerstverlichting]

After reading some of Breaking the Market's thoughts on rebalancing (only tangentially related), I made two changes on July 31st:

1. I am going to rebalance weekly. Side benefit is that this will hopefully encourage me to stop tinkering so much.
2. I am going from 55-45 to 40-15-45 (UPRO-USTL-TMF). Utilities have lower volatility than SPY and weak correlation to it. Seems like a good deal for the increase in sector-specific risk. Worked out last week, but I have noticed that it seemed like USTL had very high intraday volatility.

I am still up a little bit, but not very much. I started at essentially the worst possible time (late January), but I have to think the dates I rebalanced were unlucky, since others have mentioned being up quite a bit YTD.

When I search for USTL, all I find is a liquidated Nasdaq etf, not anything related to utilities? Can you help me out, as I'm interested in what you're proposing.

edit: nvm, I see you must've meant UTSL instead

[sf1988]

Is anyone doing this within a 401k brokerage link?

My new job has this as an option and am tempted to increase my allocation (I guess I could always just do that within my Roth IRA).

Just curious if there are any issues with LETFs in 401k\ brokerage link account.

[Texanbybirth]

Is anyone doing this within a 401k brokerage link?

My new job has this as an option and am tempted to increase my allocation (I guess I could always just do that within my Roth IRA).

Just curious if there are any issues with LETFs in 401k\ brokerage link account.

I see no reason why your brokerage link wouldn't allow this. It's easy enough for you to send tickers to your rep to find out. You can certainly pack a whole lot more annual money into this strategy with a 401k. (There was discussion at some point in this thread about whether you should even contribute more or just commit a certain amount as a side bet. I believe OP had decided to not contribute any more of his [future] funds to the strategy.)

At my humble earnings level, Roth IRA provides plenty of exposure to this strategy. IF it does actually work it's trick over my timeframe (assuming nothing catastrophic I'm OUT at 59.5 yo), then the tax free withdrawals from our Roth IRA will be life changing. IF it doesn't work (for whatever reason), the amount contributed won't destroy any of our plans.

{I haven't decided whether we'll put any more cash toward the EA.}

[tomphilly]

If UPRO heads to zero, you need to vigorously avoid rebalancing. However, you don't know where it's going ahead of time. There's luck involved. If you rebalanced on January 1, April 1, and July 1, you're pretty happy today. But that's luck.

Agreed, chance is a big factor with the HFEA re-balancing dates, at least on the 10 year backtests. I wonder if the quarterly date becomes insignificant when you do it for long enough. For 2020, if you had an "unlucky re-balancing schedule" and your ratio was out of sync on Feb 15 from the ATH run up (say it was at 70% UPRO), you would've seen a staggering drawdown by mid March. I was frequently re-balancing ("buy the dip") all the way to the bottom and fully recovered by mid April - probably also just lucky there.

[locallyoptimal]

Greetings—these past few days I’ve noticed that the negative correlation between LTTs and stocks has been shaky.
As I write, TMF is down 4.2%, TQQ down 2.8% and UPRO is up 0.6%
Is this a sign that the low yields on LTTs make it no longer an attractive safe haven for the big institutions? Any thoughts on where the “risk off” money is flowing to instead?
Thanks for any insights!

[tomphilly]

As I write, TMF is down 4.2%, TQQ down 2.8% and UPRO is up 0.6%
Is this a sign that the low yields on LTTs make it no longer an attractive safe haven for the big institutions? Any thoughts on where the “risk off” money is flowing to instead?

It could be, but this is a long term strategy and you shouldn't analyze the market on a daily basis. There will be many crappy days when both are down.

That said, TQQQ was at eye-watering highs, could be getting hit by China jitters, and/or the inevitable rotation into other sectors as they finally become attractive again. Look at FAS.

I was about 10% TQQQ but have cashed out and I'm just vanilla UPRO, which still has loads of tech exposure. I have been toying with the idea of adding 5% or 10% FAS because it "looks cheap", but I also don't know [expletive removed by admin LadyGeek] about the financial sector or its outlook.

As for TMF, there's an absolute mixed bag of signals right now, but what other hedge are you going to use?

This is one reason why I'm interested in the Target Volatility model. It has less reliance on TMF during low vol periods.

[langlands]

Greetings—these past few days I’ve noticed that the negative correlation between LTTs and stocks has been shaky.
As I write, TMF is down 4.2%, TQQ down 2.8% and UPRO is up 0.6%
Is this a sign that the low yields on LTTs make it no longer an attractive safe haven for the big institutions? Any thoughts on where the “risk off” money is flowing to instead?
Thanks for any insights!

The last couple days have been "risk on," not "risk off." Look at Dow and Russell 2000.

[Gufomel]

So this is what this strategy feels like when rates rise

** I know daily movement is practically meaningless.

[guyinlaw]

As I write, TMF is down 4.2%, TQQ down 2.8% and UPRO is up 0.6%
Is this a sign that the low yields on LTTs make it no longer an attractive safe haven for the big institutions? Any thoughts on where the “risk off” money is flowing to instead?

It could be, but this is a long term strategy and you shouldn't analyze the market on a daily basis. There will be many crappy days when both are down.

This is NOT a long term strategy. As we head to the end of 40 year bond bull market, LTT will behave unexpectedly. As TMF is hitting it's roof, this strategy is getting riskier.

Even Bridgewater is abandoning LTT bonds in the All weather portfolios. UPRO/TMF is similar to leveraged all weather portfolio..

viewtopic.php?f=10&t=322574
https://www.bridgewater.com/grappling-w ... everywhere

[tomphilly]

As we head to the end of 40 year bond bull market, LTT will behave unexpectedly.

I don't disagree with you, but I still consider the framework (leveraged negatively correlated tickers) a long term strategy. TMF could be mixed with other hedges going forward, e.g BTAL, TYD. I believe many already mix the risk assets, so why not the hedge.

"Treasury Rates Can Still Hit 0%", nonetheless I share your concern.

I would be extremely thankful if a very smart person here figured out how to backtest willthrill81's 80/20 target volatility model in the 60s-80s, when rates rose. It might be a very attractive alternative going into an increasingly murky LTT environment.

[thwang99]

TMF dropping a lot recently especially today 8/11/2020 (dropped 3.3% today). Stocks were down today also. Does anyone know why? I couldn't find any news under TMF or TLT.

Unrelated note, I'm doing a 65% TQQQ 35% TMF risk parity account, today was bad cause both dropped relatively a lot. But this year it's been great.

[Meaty]

TMF dropping a lot recently especially today 8/11/2020 (dropped 3.3% today). Stocks were down today also. Does anyone know why? I couldn't find any news under TMF or TLT.

Unrelated note, I'm doing a 65% TQQQ 35% TMF risk parity account, today was bad cause both dropped relatively a lot. But this year it's been great.

Optimism about the economy reopening and vaccine news.

[TwoIdenticalIndexes]

As we head to the end of 40 year bond bull market, LTT will behave unexpectedly.

I don't disagree with you, but I still consider the framework (leveraged negatively correlated tickers) a long term strategy. TMF could be mixed with other hedges going forward, e.g BTAL, TYD. I believe many already mix the risk assets, so why not the hedge.

"Treasury Rates Can Still Hit 0%", nonetheless I share your concern.

I would be extremely thankful if a very smart person here figured out how to backtest willthrill81's 80/20 target volatility model in the 60s-80s, when rates rose. It might be a very attractive alternative going into an increasingly murky LTT environment.

The only asset you linked to has negative alltime returns and a 2.5% expense ratio. I'd rather buy gold, thanks...

[Semantics]

As we head to the end of 40 year bond bull market, LTT will behave unexpectedly.

I don't disagree with you, but I still consider the framework (leveraged negatively correlated tickers) a long term strategy. TMF could be mixed with other hedges going forward, e.g BTAL, TYD. I believe many already mix the risk assets, so why not the hedge.

"Treasury Rates Can Still Hit 0%", nonetheless I share your concern.

I would be extremely thankful if a very smart person here figured out how to backtest willthrill81's 80/20 target volatility model in the 60s-80s, when rates rose. It might be a very attractive alternative going into an increasingly murky LTT environment.

The only asset you linked to has negative alltime returns and a 2.5% expense ratio. I'd rather buy gold, thanks...

If you're referring to BTAL (which has 2.1% ER not 2.5%), the point of holding it isn't for its returns. The fact it's neutral, which is even in the name, is a feature - it doesn't go up, but it doesn't decay like other volatility-based hedges like VIX futures. 65/35 UPRO/BTAL gave higher CAGR and lower volatility than 55/45 UPRO/TMF from 2012-2018 with flat long-term rates. It ought to outperform by a wider margin going forward if yields don't go down. Gold did worse than both in the allocations I looked at. I wish there were a more volatile or leveraged version of BTAL though - as it is, it's hard to pair with UPRO/TQQQ because the max Sharpe ratio in backtests seems to be at around 33/67 which is too low to beat out unleveraged S&P 500.

[guyinlaw]

As we head to the end of 40 year bond bull market, LTT will behave unexpectedly.

I don't disagree with you, but I still consider the framework (leveraged negatively correlated tickers) a long term strategy. TMF could be mixed with other hedges going forward, e.g BTAL, TYD. I believe many already mix the risk assets, so why not the hedge.

"Treasury Rates Can Still Hit 0%", nonetheless I share your concern.

Rather than look for risk parity, a straight leverage is a better IMO. See the thread below.

viewtopic.php?f=10&t=274390

The paper and analysis suggests a maximum leverage of 2:1. In this case leveraged ETFs are most expensive option. Leverage through borrowing (IBRK), e-mini futures and SPY Leaps are cheaper.

TMF dropping a lot recently especially today 8/11/2020 (dropped 3.3% today). Stocks were down today also. Does anyone know why? I couldn't find any news under TMF or TLT.

10y and 30y interest rates rose today. Two possible reasons.
1. reopening optimism, inflation increase expectation
2. There is Treasury auction this week, wonder if this is manipulation or something related to this.
https://www.marketwatch.com/story/10-ye ... 2020-08-11

[PicassoSparks]

If you're referring to BTAL (which has 2.1% ER not 2.5%), the point of holding it isn't for its returns. The fact it's neutral, which is even in the name, is a feature - it doesn't go up, but it doesn't decay like other volatility-based hedges like VIX futures. 65/35 UPRO/BTAL gave higher CAGR and lower volatility than 55/45 UPRO/TMF from 2012-2018 with flat long-term rates. It ought to outperform by a wider margin going forward if yields don't go down. Gold did worse than both in the allocations I looked at. I wish there were a more volatile or leveraged version of BTAL though - as it is, it's hard to pair with UPRO/TQQQ because the max Sharpe ratio in backtests seems to be at around 33/67 which is too low to beat out unleveraged S&P 500.

A backtest that leaves out market crashes isn’t very persuasive. Sure it’s reasonable to expect rates to rise but it’s also reasonable to expect more market turmoil. The strategy needs to be robust to both and BTAL + UPRO does not do well in crashes, where UPRO + TMF does.

[Meaty]

As we head to the end of 40 year bond bull market, LTT will behave unexpectedly.

I don't disagree with you, but I still consider the framework (leveraged negatively correlated tickers) a long term strategy. TMF could be mixed with other hedges going forward, e.g BTAL, TYD. I believe many already mix the risk assets, so why not the hedge.

"Treasury Rates Can Still Hit 0%", nonetheless I share your concern.

I would be extremely thankful if a very smart person here figured out how to backtest willthrill81's 80/20 target volatility model in the 60s-80s, when rates rose. It might be a very attractive alternative going into an increasingly murky LTT environment.

A few pages back in the thread there was discussion about 55 / 45 UPRO / TMF from 2015 to 2018 when there was multiple 25 bps increases. CAGR was 13%. While that’s a far cry from 20%+, it’s hardly a financial ruin. I think the big risk is stagflation, rapidly rising rates, or TMF and UPRO becoming correlated. I don’t see those risks developing

[coingaroo]

I think the big risk is stagflation, rapidly rising rates, or TMF and UPRO becoming correlated. I don’t see those risks developing

All of these 3 risks can be mitigated by adding gold (UGL).

---

For those doing TQQQ / TMF like me, what's your allocation? I'm 60/40.

[Meaty]

I think the big risk is stagflation, rapidly rising rates, or TMF and UPRO becoming correlated. I don’t see those risks developing

All of these 3 risks can be mitigated by adding gold (UGL).

---

For those doing TQQQ / TMF like me, what's your allocation? I'm 60/40.

I’m running 45 UPRO, 45 TMF, and 10 TQQQ. The lack of circuit breaker for tqqq gives me pause but the above allocation (per PV) decreases max draw down, improves CAGR by 2%, and improves sharpe ratio over the 55/45 UPRO TMF version

[Semantics]

If you're referring to BTAL (which has 2.1% ER not 2.5%), the point of holding it isn't for its returns. The fact it's neutral, which is even in the name, is a feature - it doesn't go up, but it doesn't decay like other volatility-based hedges like VIX futures. 65/35 UPRO/BTAL gave higher CAGR and lower volatility than 55/45 UPRO/TMF from 2012-2018 with flat long-term rates. It ought to outperform by a wider margin going forward if yields don't go down. Gold did worse than both in the allocations I looked at. I wish there were a more volatile or leveraged version of BTAL though - as it is, it's hard to pair with UPRO/TQQQ because the max Sharpe ratio in backtests seems to be at around 33/67 which is too low to beat out unleveraged S&P 500.

A backtest that leaves out market crashes isn’t very persuasive. Sure it’s reasonable to expect rates to rise but it’s also reasonable to expect more market turmoil. The strategy needs to be robust to both and BTAL + UPRO does not do well in crashes, where UPRO + TMF does.

Feel free to backtest it yourself, I'm just posting my thoughts, not trying to persuade you of anything. Perhaps I didn't make it clear, the point of looking at the 2012-2018 period was just to examine what different assets look like under very specific conditions. I don't hold only BTAL + UPRO, just like I don't hold only TMF + UPRO. If you want to look at crashes too, over the full available history, 2012-2020, BTAL + UPRO has a Sharpe ratio of 1.18 vs. 1.21 for TMF. So, slightly worse, but as I said I suspect the former will slightly outperform over the next 8.5 years, since LTT treasury yields are lower, and I think the 30 year rate is unlikely to fall another 1.8%. TMF had 12% CAGR over that period - if it barely outperforms the BTAL version with 12% CAGR, my confidences isn't that high it will continue to do so with the currently expected 2-3%. I hope I'm wrong, I'd rather have more tools in the toolbox. On the flip side, the markets are predicting inflation going forward will be similar to the past decade, even with the Fed indicating they would like inflation to run a bit hot, so I'm not too worried about holding treasuries.

[taojaxx]

Would be great to have the OP's reaction to Bridgewater's recent article
https://www.bridgewater.com/grappling-w ... everywhere
HF, if you're lurking in the dark, just step up to the plate!

[Saaybogle]

I’m guessing TMFs poor performance is due to 30 year treasury bonds being auctioned today. Is that right? Has that alone caused rates to rise and thus TMF to drop? I see that the unemployment number also came in better than expected. Thanks.

[tomphilly]

I’m guessing TMFs poor performance is due to 30 year treasury bonds being auctioned today. Is that right? Has that alone caused rates to rise and thus TMF to drop? I see that the unemployment number also came in better than expected. Thanks.

Yields are rising, for 5 straight sessions. It obviously hurts when SPY is just hovering around an epic resistance level (the Feb all time high). The past week would have definitely played into the hands of will's 80/20 TV model, but you can't base things off a week. In any case, before this week started I decided to transition to a 65/35 TV model on one portfolio, which I may slowly adjust to 80/20.

[Semantics]

As we head to the end of 40 year bond bull market, LTT will behave unexpectedly.

I don't disagree with you, but I still consider the framework (leveraged negatively correlated tickers) a long term strategy. TMF could be mixed with other hedges going forward, e.g BTAL, TYD. I believe many already mix the risk assets, so why not the hedge.

"Treasury Rates Can Still Hit 0%", nonetheless I share your concern.

I would be extremely thankful if a very smart person here figured out how to backtest willthrill81's 80/20 target volatility model in the 60s-80s, when rates rose. It might be a very attractive alternative going into an increasingly murky LTT environment.

A few pages back in the thread there was discussion about 55 / 45 UPRO / TMF from 2015 to 2018 when there was multiple 25 bps increases. CAGR was 13%. While that’s a far cry from 20%+, it’s hardly a financial ruin. I think the big risk is stagflation, rapidly rising rates, or TMF and UPRO becoming correlated. I don’t see those risks developing

Long term treasury rates (TMF) did not go up with interest rates from 2015-2018 though. And now the yields are significantly lower than in 2015-2018. For the amount of risk this strategy entails, would a similar expected result of only beating the S&P 500 by 1-2% CAGR over the next decade be acceptable?

[tomphilly]

A few pages back in the thread there was discussion about 55 / 45 UPRO / TMF from 2015 to 2018 when there was multiple 25 bps increases. CAGR was 13%. While that’s a far cry from 20%+, it’s hardly a financial ruin. I think the big risk is stagflation, rapidly rising rates, or TMF and UPRO becoming correlated. I don’t see those risks developing

Can you link a PV of 2015-2018 55/45 13% CAGR? I only get 9% CAGR with my test. Are you setting particular start and end months? Thanks

Long term treasury rates (TMF) did not go up with interest rates from 2015-2018 though. And now the yields are significantly lower than in 2015-2018. For the amount of risk this strategy entails, would a similar expected result of only beating the S&P 500 by 1-2% CAGR over the next decade be acceptable?

I get a 7% CAGR holding SPY between 2015-2018, so the difference, at least with the target volatility model at 65/35 is 5% - PV. Like you say it might not be worth the risk, but it's better than 1-2%.

[vijaym73]

Currently at 70/30 TQQQ/TMF

May move to 80/20 but most likely will wait till election is done.

Good investing

VJ

[tomphilly]

Currently at 70/30 TQQQ/TMF

May move to 80/20 but most likely will wait till election is done.

Are you quarterly re-balancing or using target volatility? Thanks

The correlation between TQQQ and TMF, at least during the pandemic, was interesting. Technology was almost a hedge. Not sure what will happen next but I assume some rotation back into sectors playing catchup. I also worry about some cataclysmic monopoly ruling in the next few years hammering TQQQ which is dominated by a few monster tickers.

[Meaty]

A few pages back in the thread there was discussion about 55 / 45 UPRO / TMF from 2015 to 2018 when there was multiple 25 bps increases. CAGR was 13%. While that’s a far cry from 20%+, it’s hardly a financial ruin. I think the big risk is stagflation, rapidly rising rates, or TMF and UPRO becoming correlated. I don’t see those risks developing

Can you link a PV of 2015-2018 55/45 13% CAGR? I only get 9% CAGR with my test. Are you setting particular start and end months? Thanks

Long term treasury rates (TMF) did not go up with interest rates from 2015-2018 though. And now the yields are significantly lower than in 2015-2018. For the amount of risk this strategy entails, would a similar expected result of only beating the S&P 500 by 1-2% CAGR over the next decade be acceptable?

I get a 7% CAGR holding SPY between 2015-2018, so the difference, at least with the target volatility model at 65/35 is 5% - PV. Like you say it might not be worth the risk, but it's better than 1-2%.

https://www.portfoliovisualizer.com/bac ... tion2_3=45

[Meaty]

As we head to the end of 40 year bond bull market, LTT will behave unexpectedly.

I don't disagree with you, but I still consider the framework (leveraged negatively correlated tickers) a long term strategy. TMF could be mixed with other hedges going forward, e.g BTAL, TYD. I believe many already mix the risk assets, so why not the hedge.

"Treasury Rates Can Still Hit 0%", nonetheless I share your concern.

I would be extremely thankful if a very smart person here figured out how to backtest willthrill81's 80/20 target volatility model in the 60s-80s, when rates rose. It might be a very attractive alternative going into an increasingly murky LTT environment.

A few pages back in the thread there was discussion about 55 / 45 UPRO / TMF from 2015 to 2018 when there was multiple 25 bps increases. CAGR was 13%. While that’s a far cry from 20%+, it’s hardly a financial ruin. I think the big risk is stagflation, rapidly rising rates, or TMF and UPRO becoming correlated. I don’t see those risks developing

Long term treasury rates (TMF) did not go up with interest rates from 2015-2018 though. And now the yields are significantly lower than in 2015-2018. For the amount of risk this strategy entails, would a similar expected result of only beating the S&P 500 by 1-2% CAGR over the next decade be acceptable?

It’s arguably more risky now than 2015, but the benchmark is 100% S&P. Do you see this strategy as riskier than 100% equities over the next 10 years? Sharpe ratio is better historically with HFEA than spy

[vijaym73]

I am strictly just quarterly rebalancing. Rebalance on first day of quarter as that’s the easiest.

Sector rotation may happen and I welcome it. Tech can’t be the only sector going up all the time. Let everything catch up and then let’s all move in tandem.

Thx

VJ

[perfectuncertainty]

I’m guessing TMFs poor performance is due to 30 year treasury bonds being auctioned today. Is that right? Has that alone caused rates to rise and thus TMF to drop? I see that the unemployment number also came in better than expected. Thanks.

Yields are rising, for 5 straight sessions. It obviously hurts when SPY is just hovering around an epic resistance level (the Feb all time high). The past week would have definitely played into the hands of will's 80/20 TV model, but you can't base things off a week. In any case, before this week started I decided to transition to a 65/35 TV model on one portfolio, which I may slowly adjust to 80/20.

https://www.marketwatch.com/story/treas ... 2020-08-13

[Semantics]

Long term treasury rates (TMF) did not go up with interest rates from 2015-2018 though. And now the yields are significantly lower than in 2015-2018. For the amount of risk this strategy entails, would a similar expected result of only beating the S&P 500 by 1-2% CAGR over the next decade be acceptable?

I get a 7% CAGR holding SPY between 2015-2018, so the difference, at least with the target volatility model at 65/35 is 5% - PV. Like you say it might not be worth the risk, but it's better than 1-2%.

I was assuming a static 55/45 allocation, and also assuming that TMF will return at least a couple percent less than 2015-2018 even if rates stay this low, since yields are currently lower by more than 1%. Using a timing model seems really important in this climate, but that's where most of your 5% is coming from - your backtest does almost as well if you replace TMF with cash.

[Semantics]

As we head to the end of 40 year bond bull market, LTT will behave unexpectedly.

I don't disagree with you, but I still consider the framework (leveraged negatively correlated tickers) a long term strategy. TMF could be mixed with other hedges going forward, e.g BTAL, TYD. I believe many already mix the risk assets, so why not the hedge.

"Treasury Rates Can Still Hit 0%", nonetheless I share your concern.

I would be extremely thankful if a very smart person here figured out how to backtest willthrill81's 80/20 target volatility model in the 60s-80s, when rates rose. It might be a very attractive alternative going into an increasingly murky LTT environment.

A few pages back in the thread there was discussion about 55 / 45 UPRO / TMF from 2015 to 2018 when there was multiple 25 bps increases. CAGR was 13%. While that’s a far cry from 20%+, it’s hardly a financial ruin. I think the big risk is stagflation, rapidly rising rates, or TMF and UPRO becoming correlated. I don’t see those risks developing

Long term treasury rates (TMF) did not go up with interest rates from 2015-2018 though. And now the yields are significantly lower than in 2015-2018. For the amount of risk this strategy entails, would a similar expected result of only beating the S&P 500 by 1-2% CAGR over the next decade be acceptable?

It’s arguably more risky now than 2015, but the benchmark is 100% S&P. Do you see this strategy as riskier than 100% equities over the next 10 years? Sharpe ratio is better historically with HFEA than spy

I think it's about as risky as 165% equities, maybe a bit less because of the negative correlation (which I assume persists), but that could very well be offset and then some by volatility decay and/or rising rates. So I'm basically assuming there's a high likelihood TMF will perform no better than cash. There's a chance rates could go down more and TMF will look great, but I want something that looks good in the case that I think is more likely. At the end of the day I don't see a really compelling alternative and am hoping timing models will help to maintain strong returns even if TMF performs poorly.

Although I guess it's also the case that the spread in rates matters more than the absolute values. So even though LTT yields are down, borrowing is cheaper, so maybe TMF will do better than it did in 2015-2018 when the spread between long-term and short-term rates became quite small.

[langlands]

Long term treasury rates (TMF) did not go up with interest rates from 2015-2018 though. And now the yields are significantly lower than in 2015-2018. For the amount of risk this strategy entails, would a similar expected result of only beating the S&P 500 by 1-2% CAGR over the next decade be acceptable?

It’s arguably more risky now than 2015, but the benchmark is 100% S&P. Do you see this strategy as riskier than 100% equities over the next 10 years? Sharpe ratio is better historically with HFEA than spy

I think it's about as risky as 165% equities, maybe a bit less because of the negative correlation (which I assume persists), but that could very well be offset and then some by volatility decay and/or rising rates. So I'm basically assuming there's a high likelihood TMF will perform no better than cash. There's a chance rates could go down more and TMF will look great, but I want something that looks good in the case that I think is more likely. At the end of the day I don't see a really compelling alternative and am hoping timing models will help to maintain strong returns even if TMF performs poorly.

Although I guess it's also the case that the spread in rates matters more than the absolute values. So even though LTT yields are down, borrowing is cheaper, so maybe TMF will do better than it did in 2015-2018 when the spread between long-term and short-term rates became quite small.

Given your concerns, any reason you're not considering gold? Given the near zero yields, TMF is essentially a speculative bet for lower yields. Gold is essentially a bet for higher inflation. The difference is that gold of course has a lot more room to run.

[rockstar]

Currently at 70/30 TQQQ/TMF

May move to 80/20 but most likely will wait till election is done.

Good investing

VJ

For 5% of my portfolio, I"m 100% TQQQ. I don't get bonds at these yields. I'll sell whenever I'm up approximately 10% and buy back on the dips around 3% or higher. I do this trade as frequently as I can. I think, we're pushing toward a top for QQQ, so I'm selling my position for around a 9ish percent increase this time around.

[Semantics]

Long term treasury rates (TMF) did not go up with interest rates from 2015-2018 though. And now the yields are significantly lower than in 2015-2018. For the amount of risk this strategy entails, would a similar expected result of only beating the S&P 500 by 1-2% CAGR over the next decade be acceptable?

It’s arguably more risky now than 2015, but the benchmark is 100% S&P. Do you see this strategy as riskier than 100% equities over the next 10 years? Sharpe ratio is better historically with HFEA than spy

I think it's about as risky as 165% equities, maybe a bit less because of the negative correlation (which I assume persists), but that could very well be offset and then some by volatility decay and/or rising rates. So I'm basically assuming there's a high likelihood TMF will perform no better than cash. There's a chance rates could go down more and TMF will look great, but I want something that looks good in the case that I think is more likely. At the end of the day I don't see a really compelling alternative and am hoping timing models will help to maintain strong returns even if TMF performs poorly.

Although I guess it's also the case that the spread in rates matters more than the absolute values. So even though LTT yields are down, borrowing is cheaper, so maybe TMF will do better than it did in 2015-2018 when the spread between long-term and short-term rates became quite small.

Given your concerns, any reason you're not considering gold? Given the near zero yields, TMF is essentially a speculative bet for lower yields. Gold is essentially a bet for higher inflation. The difference is that gold of course has a lot more room to run.

Gold has almost zero historical return and high volatility - meaning a fund with daily leveraged resets like UGLD has a high chance of dragging the portfolio down due to volatility decay. Additionally, if I want the volatility to work in my favor, it ought to have a significant negative correlation with equities. Gold has a neutral correlation. It also has a mild positive correlation with treasuries (makes sense - when yields get too low people move some money into gold). Going forward, if LTT fall due to higher inflation expectations, gold will then decline as people rebalance from gold into treasuries to reap the higher yield.

I am not opposed to holding gold, but only either as a minority component of a portfolio and not as a replacement for treasuries in HFEA, or in a timing model where it's possible to dial up exposure while treasuries are falling, and then dial it back down once they've stabilized. Given the volatility of UPRO/TQQQ, I think using ballast that doesn't have a strong negative correlation would be a bit too uncomfortable for me.

Would love to hear more if anyone is successfully using it in a variant of HFEA though.

[coingaroo]

Bit of a different topic.

Anyone know what financing rates TQQQ and TMF are currently paying in this environment?

Especially interested in TMF. Want to quantify the "borrow short; lend long" effect.

EDIT: The only things I could find is the annual reports for last year.

It's hard to tell what the spread is, because you don't see when these swaps originated, but it looks like TMF is paying a spread of ~~0.36% above the 1 month LIBOR rate. As the LIBOR rate approaches zero, it's unknown what the spreads are (I suspect it's likely a lil bit higher).



Using a 0.40% spread guess on top of the current LIBOR, seems like the net effect of TMF's borrow short; lend long is 1.94% of yield per year.

If only the mgmt fee could be lower

[langlands]

Long term treasury rates (TMF) did not go up with interest rates from 2015-2018 though. And now the yields are significantly lower than in 2015-2018. For the amount of risk this strategy entails, would a similar expected result of only beating the S&P 500 by 1-2% CAGR over the next decade be acceptable?

It’s arguably more risky now than 2015, but the benchmark is 100% S&P. Do you see this strategy as riskier than 100% equities over the next 10 years? Sharpe ratio is better historically with HFEA than spy

I think it's about as risky as 165% equities, maybe a bit less because of the negative correlation (which I assume persists), but that could very well be offset and then some by volatility decay and/or rising rates. So I'm basically assuming there's a high likelihood TMF will perform no better than cash. There's a chance rates could go down more and TMF will look great, but I want something that looks good in the case that I think is more likely. At the end of the day I don't see a really compelling alternative and am hoping timing models will help to maintain strong returns even if TMF performs poorly.

Although I guess it's also the case that the spread in rates matters more than the absolute values. So even though LTT yields are down, borrowing is cheaper, so maybe TMF will do better than it did in 2015-2018 when the spread between long-term and short-term rates became quite small.

Given your concerns, any reason you're not considering gold? Given the near zero yields, TMF is essentially a speculative bet for lower yields. Gold is essentially a bet for higher inflation. The difference is that gold of course has a lot more room to run.

Gold has almost zero historical return and high volatility - meaning a fund with daily leveraged resets like UGLD has a high chance of dragging the portfolio down due to volatility decay. Additionally, if I want the volatility to work in my favor, it ought to have a significant negative correlation with equities. Gold has a neutral correlation. It also has a mild positive correlation with treasuries (makes sense - when yields get too low people move some money into gold). Going forward, if LTT fall due to higher inflation expectations, gold will then decline as people rebalance from gold into treasuries to reap the higher yield.

I am not opposed to holding gold, but only either as a minority component of a portfolio and not as a replacement for treasuries in HFEA, or in a timing model where it's possible to dial up exposure while treasuries are falling, and then dial it back down once they've stabilized. Given the volatility of UPRO/TQQQ, I think using ballast that doesn't have a strong negative correlation would be a bit too uncomfortable for me.

You really think gold will decline with increasing inflation expectations? I think you have it backwards. LTT will fall as you say, and that is exactly because people will rebalance from LTT into gold.

I view stocks and bonds very differently. Stocks go through fairly short term cycles of high P/E and low P/E and the average person experiences a couple of each in their lifetime. Because of this, you can essentially take the long term view with stocks and be reasonably confident that you'll overall get more or less the historical average stock return. Bonds go through much longer cycles and boomers for instance have pretty much gone an entire lifetime (in the US) of secularly decreasing interest rates. There really is a limit to how low the rates can go. It makes sense to say that stocks tend to always go up in boom bust cycles so "stay the course" and reap the equity premium. I don't think it makes sense to say interest rates tend to always go down so "stay the course" and reap the "interest rates always decrease" premium. That will blow up. It's just some people are lucky enough to live an entire lifetime without seeing it.