HEDGEFUNDIE's excellent adventure Part II: The next journey

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
manlymatt83
Posts: 1274
Joined: Tue Jan 30, 2018 7:23 am

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by manlymatt83 »

Lawyered_ wrote: Wed Jan 24, 2024 3:05 pm
manlymatt83 wrote: Wed Jan 24, 2024 2:49 pm
jarjarM wrote: Wed Jan 24, 2024 2:44 pm
manlymatt83 wrote: Wed Jan 24, 2024 2:12 pm Do you think MotoTrojan didn't use GOVZ in his UPRO/EDV because it wasn't around yet? Would there be any reason to not use GOVZ over EDV these days if one was concerned about the ER of TMF?
As someone who talked to MotoTrojan via pm multiple times over this and the "Part I" thread, yup, GOVZ wasn't available when he first purposed to use EDV to replace TMF due to TMF's higher volatility (and a bit higher ER).
Interesting. Would the 43/57 UPRO/EDV ratio apply when dropping in GOVZ? Or would we want 45 UPRO/55 GOVZ due to the increased volatility of GOVZ? Or could one just do 50/50 to KISS.
It just depends on how much equity risk you want to take. 50/50 is 1.5X SPY with 75% LTT equivalent so 150% SPY/75% LTTs. 45/55 is 135% SPY/82.5% LTTs.
People did 55 URPO / 45 TMF because in backtests, it seemed to match returns of 100% UPRO but with much less risk. Correct?

So 43 UPRO / 57 EDV/GOVZ was a way to decrease returns, but also decrease risk. "Smooth out the ride" as I've heard. And I assume a benefit of this is reducing carrying costs for TMF. In fact, maybe that's the main benefit?

When looking at UPRO / GOVZ, how does one measure matching UPRO / TMF as close as possible?

I guess I just don't understand what people are looking at when they pair these. 55% UPRO vs. 45% TMF for example... why was that ideal over 60% UPRO / 40% TMF? The latter would have higher returns than 100% UPRO, but maybe a larger drawdown? And that drawdown = risk?
Hydromod
Posts: 1042
Joined: Tue Mar 26, 2019 10:21 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Hydromod »

manlymatt83 wrote: Wed Jan 24, 2024 3:35 pm
Lawyered_ wrote: Wed Jan 24, 2024 3:05 pm It just depends on how much equity risk you want to take. 50/50 is 1.5X SPY with 75% LTT equivalent so 150% SPY/75% LTTs. 45/55 is 135% SPY/82.5% LTTs.
People did 55 URPO / 45 TMF because in backtests, it seemed to match returns of 100% UPRO but with much less risk. Correct?

So 43 UPRO / 57 EDV/GOVZ was a way to decrease returns, but also decrease risk. "Smooth out the ride" as I've heard. And I assume a benefit of this is reducing carrying costs for TMF. In fact, maybe that's the main benefit?

When looking at UPRO / GOVZ, how does one measure matching UPRO / TMF as close as possible?

I guess I just don't understand what people are looking at when they pair these. 55% UPRO vs. 45% TMF for example... why was that ideal over 60% UPRO / 40% TMF? The latter would have higher returns than 100% UPRO, but maybe a larger drawdown? And that drawdown = risk?
The original was 40/60 UPRO/TMF. HEDGEFUNDIE went to 55/45 UPRO/TMF because folks convinced him that TMF returns were dwindling.

When I backtested with a risk budget inverse volatility approach (see my thread for details), 55/45 worked out to assigning the risk budget as 4/1 equity/treasury. I think that the 43/57 UPRO/EDZ had a similar risk budget.

That kind of weighting works pretty well when the assets have fairly similar volatility. It will way overweight assets that have much lower volatility.

I personally prefer using a weighting strategy based on volatility (see methodology here), but I'm a geek that's fine with rebalancing fairly frequently.
manlymatt83
Posts: 1274
Joined: Tue Jan 30, 2018 7:23 am

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by manlymatt83 »

Hydromod wrote: Wed Jan 24, 2024 4:01 pm
manlymatt83 wrote: Wed Jan 24, 2024 3:35 pm
Lawyered_ wrote: Wed Jan 24, 2024 3:05 pm It just depends on how much equity risk you want to take. 50/50 is 1.5X SPY with 75% LTT equivalent so 150% SPY/75% LTTs. 45/55 is 135% SPY/82.5% LTTs.
People did 55 URPO / 45 TMF because in backtests, it seemed to match returns of 100% UPRO but with much less risk. Correct?

So 43 UPRO / 57 EDV/GOVZ was a way to decrease returns, but also decrease risk. "Smooth out the ride" as I've heard. And I assume a benefit of this is reducing carrying costs for TMF. In fact, maybe that's the main benefit?

When looking at UPRO / GOVZ, how does one measure matching UPRO / TMF as close as possible?

I guess I just don't understand what people are looking at when they pair these. 55% UPRO vs. 45% TMF for example... why was that ideal over 60% UPRO / 40% TMF? The latter would have higher returns than 100% UPRO, but maybe a larger drawdown? And that drawdown = risk?
The original was 40/60 UPRO/TMF. HEDGEFUNDIE went to 55/45 UPRO/TMF because folks convinced him that TMF returns were dwindling.

When I backtested with a risk budget inverse volatility approach (see my thread for details), 55/45 worked out to assigning the risk budget as 4/1 equity/treasury. I think that the 43/57 UPRO/EDZ had a similar risk budget.

That kind of weighting works pretty well when the assets have fairly similar volatility. It will way overweight assets that have much lower volatility.

I personally prefer using a weighting strategy based on volatility (see methodology here), but I'm a geek that's fine with rebalancing fairly frequently.
I just want a set-it-and-forget-it allocation I can rebalance quarterly, somewhere between PSLDX (100% / 100%) and 55 UPRO/45 TMF. Maybe involving GOVZ as that seems to be a "better" version of EDV.

I don't mind doing the work myself but I am unsure how to come up with the "best" allocation for UPRO / GOVZ. Is it 40/60? 53/47? 50/50? etc.

I guess if I am following the 4 / 1 rule, what would the ratio be? Or how can I find that out?
Lawyered_
Posts: 164
Joined: Wed Feb 09, 2022 10:52 am

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Lawyered_ »

manlymatt83 wrote: Wed Jan 24, 2024 4:24 pm
Hydromod wrote: Wed Jan 24, 2024 4:01 pm
manlymatt83 wrote: Wed Jan 24, 2024 3:35 pm
Lawyered_ wrote: Wed Jan 24, 2024 3:05 pm It just depends on how much equity risk you want to take. 50/50 is 1.5X SPY with 75% LTT equivalent so 150% SPY/75% LTTs. 45/55 is 135% SPY/82.5% LTTs.
People did 55 URPO / 45 TMF because in backtests, it seemed to match returns of 100% UPRO but with much less risk. Correct?

So 43 UPRO / 57 EDV/GOVZ was a way to decrease returns, but also decrease risk. "Smooth out the ride" as I've heard. And I assume a benefit of this is reducing carrying costs for TMF. In fact, maybe that's the main benefit?

When looking at UPRO / GOVZ, how does one measure matching UPRO / TMF as close as possible?

I guess I just don't understand what people are looking at when they pair these. 55% UPRO vs. 45% TMF for example... why was that ideal over 60% UPRO / 40% TMF? The latter would have higher returns than 100% UPRO, but maybe a larger drawdown? And that drawdown = risk?
The original was 40/60 UPRO/TMF. HEDGEFUNDIE went to 55/45 UPRO/TMF because folks convinced him that TMF returns were dwindling.

When I backtested with a risk budget inverse volatility approach (see my thread for details), 55/45 worked out to assigning the risk budget as 4/1 equity/treasury. I think that the 43/57 UPRO/EDZ had a similar risk budget.

That kind of weighting works pretty well when the assets have fairly similar volatility. It will way overweight assets that have much lower volatility.

I personally prefer using a weighting strategy based on volatility (see methodology here), but I'm a geek that's fine with rebalancing fairly frequently.
I just want a set-it-and-forget-it allocation I can rebalance quarterly, somewhere between PSLDX (100% / 100%) and 55 UPRO/45 TMF. Maybe involving GOVZ as that seems to be a "better" version of EDV.

I don't mind doing the work myself but I am unsure how to come up with the "best" allocation for UPRO / GOVZ. Is it 40/60? 53/47? 50/50? etc.

I guess if I am following the 4 / 1 rule, what would the ratio be? Or how can I find that out?
If you want something between PSLDX and HFEA go 40% UPRO; 60% GOVZ, which is 120/90. GOVZ does not go back far enough, but here it is with EDV. https://www.portfoliovisualizer.com/bac ... phKiFLusNn . And here it is with 50% UPRO/50% ZROZ (similar to GOVZ) back as far as ZROZ goes: https://www.portfoliovisualizer.com/bac ... BAOuY9xCQB
Hydromod
Posts: 1042
Joined: Tue Mar 26, 2019 10:21 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Hydromod »

manlymatt83 wrote: Wed Jan 24, 2024 4:24 pm I just want a set-it-and-forget-it allocation I can rebalance quarterly, somewhere between PSLDX (100% / 100%) and 55 UPRO/45 TMF. Maybe involving GOVZ as that seems to be a "better" version of EDV.

I don't mind doing the work myself but I am unsure how to come up with the "best" allocation for UPRO / GOVZ. Is it 40/60? 53/47? 50/50? etc.

I guess if I am following the 4 / 1 rule, what would the ratio be? Or how can I find that out?
In the form I refer to, risk weight is relative to variance.

The risk budget inverse volatility says that the allocation to asset i is (bi/vi)/sum(bj/vj), where b is the square root of the risk weight, v is the volatility, and there are j assets. You can use daily, monthly, or annualized volatility, just be consistent.

I'm remembering better now. I think that MojoTrojan was doing something with effective bond duration.

Anyway, PV says daily volatility for UPRO, TMF, TYD, and EDZ are 3.3%, 2.9%, 1.4%, and 4.1% since 2009 (here). GOVZ has been similar to TYD since inception.

A 4/1 risk balance with UPRO and TMF gives an average UPRO weight of (2/3.3)/[(2/3.3) + (1/2.9)] = 0.64. Or 64/36 UPRO/TMF. A pure risk parity would have 47/53 UPRO/TMF.

I ended up with lower average values in part because these weights change over time, and the time averaged values including crashes and booms are a bit different.

For comparison, 4/1 weighting with UPRO and TYD gets to an 46/54 allocation.

Hope this helps instead of overwhelming.
jarjarM
Posts: 2489
Joined: Mon Jul 16, 2018 1:21 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by jarjarM »

manlymatt83 wrote: Wed Jan 24, 2024 4:24 pm
Hydromod wrote: Wed Jan 24, 2024 4:01 pm The original was 40/60 UPRO/TMF. HEDGEFUNDIE went to 55/45 UPRO/TMF because folks convinced him that TMF returns were dwindling.

When I backtested with a risk budget inverse volatility approach (see my thread for details), 55/45 worked out to assigning the risk budget as 4/1 equity/treasury. I think that the 43/57 UPRO/EDZ had a similar risk budget.

That kind of weighting works pretty well when the assets have fairly similar volatility. It will way overweight assets that have much lower volatility.

I personally prefer using a weighting strategy based on volatility (see methodology here), but I'm a geek that's fine with rebalancing fairly frequently.
I just want a set-it-and-forget-it allocation I can rebalance quarterly, somewhere between PSLDX (100% / 100%) and 55 UPRO/45 TMF. Maybe involving GOVZ as that seems to be a "better" version of EDV.

I don't mind doing the work myself but I am unsure how to come up with the "best" allocation for UPRO / GOVZ. Is it 40/60? 53/47? 50/50? etc.

I guess if I am following the 4 / 1 rule, what would the ratio be? Or how can I find that out?
As hydromod mentioned already, the 2nd version with 55/45 is utilized because of long discussions in part I pointing out future return for TMF is muted (correctly) so HF changed his view from 40/60 (which was the best from backtest) to higher equity allocation. And yeah read thru hydromod's thread if you want to do risk budgeting. Also, one should learn from the significant drawdown the last 2 years that this is not hold and forget and really does require some active management in assigning risk budgets and making determination on overweighting an allocation depending on economic condition. Not a set and forget type of holding. Check out the riding HFEA threads for some more info.
comeinvest
Posts: 2603
Joined: Mon Mar 12, 2012 6:57 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by comeinvest »

jarjarM wrote: Wed Jan 24, 2024 5:40 pm Also, one should learn from the significant drawdown the last 2 years that this is not hold and forget and really does require some active management in assigning risk budgets and making determination on overweighting an allocation depending on economic condition. Not a set and forget type of holding. Check out the riding HFEA threads for some more info.
Very easy to parameter-fit ("predict") the past. Not so much the future. Good luck doing "active management" monitoring and evaluating "economic conditions" and probably 1000 other parameters, and tweaking and fooling with the asset allocation on a constant basis.
manlymatt83
Posts: 1274
Joined: Tue Jan 30, 2018 7:23 am

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by manlymatt83 »

Hydromod wrote: Wed Jan 24, 2024 5:35 pm
manlymatt83 wrote: Wed Jan 24, 2024 4:24 pm I just want a set-it-and-forget-it allocation I can rebalance quarterly, somewhere between PSLDX (100% / 100%) and 55 UPRO/45 TMF. Maybe involving GOVZ as that seems to be a "better" version of EDV.

I don't mind doing the work myself but I am unsure how to come up with the "best" allocation for UPRO / GOVZ. Is it 40/60? 53/47? 50/50? etc.

I guess if I am following the 4 / 1 rule, what would the ratio be? Or how can I find that out?
In the form I refer to, risk weight is relative to variance.

The risk budget inverse volatility says that the allocation to asset i is (bi/vi)/sum(bj/vj), where b is the square root of the risk weight, v is the volatility, and there are j assets. You can use daily, monthly, or annualized volatility, just be consistent.

I'm remembering better now. I think that MojoTrojan was doing something with effective bond duration.

Anyway, PV says daily volatility for UPRO, TMF, TYD, and EDZ are 3.3%, 2.9%, 1.4%, and 4.1% since 2009 (here). GOVZ has been similar to TYD since inception.

A 4/1 risk balance with UPRO and TMF gives an average UPRO weight of (2/3.3)/[(2/3.3) + (1/2.9)] = 0.64. Or 64/36 UPRO/TMF. A pure risk parity would have 47/53 UPRO/TMF.

I ended up with lower average values in part because these weights change over time, and the time averaged values including crashes and booms are a bit different.

For comparison, 4/1 weighting with UPRO and TYD gets to an 46/54 allocation.

Hope this helps instead of overwhelming.
This makes sense. Thank you! Assuming the ER of GOVZ remains at 10 BP vs 6 BP for EDV, I really can’t see a reason why one would use EDV over GOVZ anymore. It basically gets you a few percentage points of free capital to allocate to UPRO if I’m not mistaken. Am I missing something?
bond93
Posts: 12
Joined: Mon Jan 22, 2024 11:38 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by bond93 »

One topic I wanted to resurrect is the conversation about entry and exit strategies.
The backtests seem to be a great example of why if contributing a lump sum, timing is everything.
But what if one is perpetually contributing, every month/quarter for years, doesn't that tip the balance in favor of
55/45 UPRO/TMF even during rough times where you will eventually be "right" as long as you keep contributing? This way one can keep the "set and forget" allocation with the knowledge there will be tough times, but continually contributing lets you enter not once, but at every point along the way. To me this is as close as you can get to "buy SPY every friday" but in a leveraged position. I would love to see a backtest of DCA perpetually from 1955-1980s to see if that changes the calculus. I remember that was a gnarly time-period for this strategy but it was based on a lump sum.

Regarding the exit point, Hedgefundie himself had a hard stop at 10M. It seems rather than keep the fund alive indefinitely to "skim" off the
cream every year, a full exit at a precise number or precise multiple of injected funds will allow you to hit your target with no risk of going back below it.(of course getting to the top is the hard part.)

Apologies if this has already been covered, hard to find in 200+ pages of talk.
comeinvest
Posts: 2603
Joined: Mon Mar 12, 2012 6:57 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by comeinvest »

bond93 wrote: Fri Jan 26, 2024 10:54 am One topic I wanted to resurrect is the conversation about entry and exit strategies.
The backtests seem to be a great example of why if contributing a lump sum, timing is everything.
But what if one is perpetually contributing, every month/quarter for years, doesn't that tip the balance in favor of
55/45 UPRO/TMF even during rough times where you will eventually be "right" as long as you keep contributing? This way one can keep the "set and forget" allocation with the knowledge there will be tough times, but continually contributing lets you enter not once, but at every point along the way. To me this is as close as you can get to "buy SPY every friday" but in a leveraged position. I would love to see a backtest of DCA perpetually from 1955-1980s to see if that changes the calculus. I remember that was a gnarly time-period for this strategy but it was based on a lump sum.

Regarding the exit point, Hedgefundie himself had a hard stop at 10M. It seems rather than keep the fund alive indefinitely to "skim" off the
cream every year, a full exit at a precise number or precise multiple of injected funds will allow you to hit your target with no risk of going back below it.(of course getting to the top is the hard part.)

Apologies if this has already been covered, hard to find in 200+ pages of talk.
Show me one rigorous study that suggests dollar cost averaging is beneficial. The perceived benefit of phasing entry points even when a lump sum is available, is usually an illusion. Maximize the time in the market and reduce your lifetime risk by adding money as soon as available.
bond93
Posts: 12
Joined: Mon Jan 22, 2024 11:38 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by bond93 »

comeinvest wrote: Fri Jan 26, 2024 2:33 pm
bond93 wrote: Fri Jan 26, 2024 10:54 am One topic I wanted to resurrect is the conversation about entry and exit strategies.
The backtests seem to be a great example of why if contributing a lump sum, timing is everything.
But what if one is perpetually contributing, every month/quarter for years, doesn't that tip the balance in favor of
55/45 UPRO/TMF even during rough times where you will eventually be "right" as long as you keep contributing? This way one can keep the "set and forget" allocation with the knowledge there will be tough times, but continually contributing lets you enter not once, but at every point along the way. To me this is as close as you can get to "buy SPY every friday" but in a leveraged position. I would love to see a backtest of DCA perpetually from 1955-1980s to see if that changes the calculus. I remember that was a gnarly time-period for this strategy but it was based on a lump sum.

Regarding the exit point, Hedgefundie himself had a hard stop at 10M. It seems rather than keep the fund alive indefinitely to "skim" off the
cream every year, a full exit at a precise number or precise multiple of injected funds will allow you to hit your target with no risk of going back below it.(of course getting to the top is the hard part.)

Apologies if this has already been covered, hard to find in 200+ pages of talk.
Show me one rigorous study that suggests dollar cost averaging is beneficial. The perceived benefit of phasing entry points even when a lump sum is available, is usually an illusion. Maximize the time in the market and reduce your lifetime risk by adding money as soon as available.
Two comments
1) As you probably already know, for most people (myself included) DCA is not necessarily about spreading out a lump sum just to reduce perceived risk. There was a Dave Ramsey speech about this where he basically says he doesn't DCA because its a great strategy, hes just doing it because thats when he has the money to make his next "lump sum"(after a paycheck). So I DCA every month or so because thats as soon as I can buy in. I might also wager it be beneficial that for people committing to this with 5-10% funny money, it's worth putting 5-10% of newly invested dollars towards it when possible rather than buy once and then try to build a parachute on the way down.
2) I agree with you that time in market is king in the majority of cases. For the standard S&P it seems one could put a lump sum down at almost any point and do alright after 10-15 years. (case in point: Bob the worlds worst market timer) but with 3x leverage and this strategy the game changes because yes at a certain point you may match/beat the market again, but that could take literally 50 years. (look at the 1955-2018 backtest) I recognize that the test was with 40/60 so if there is a 55/45 example from then that doesn't underperform for 50 years I'll happily cede my ground. I'd just personally like to be alive to enjoy my lottery :D
https://i.imgur.com/9D8QjKf.png
Last edited by bond93 on Fri Jan 26, 2024 6:56 pm, edited 5 times in total.
jarjarM
Posts: 2489
Joined: Mon Jul 16, 2018 1:21 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by jarjarM »

comeinvest wrote: Wed Jan 24, 2024 5:44 pm
jarjarM wrote: Wed Jan 24, 2024 5:40 pm Also, one should learn from the significant drawdown the last 2 years that this is not hold and forget and really does require some active management in assigning risk budgets and making determination on overweighting an allocation depending on economic condition. Not a set and forget type of holding. Check out the riding HFEA threads for some more info.
Very easy to parameter-fit ("predict") the past. Not so much the future. Good luck doing "active management" monitoring and evaluating "economic conditions" and probably 1000 other parameters, and tweaking and fooling with the asset allocation on a constant basis.
That's true, past doesn't predict the future and most ppl end up fighting the last war. I guess my point is just that it's not really a set and forget investment and requires some TLC.
comeinvest
Posts: 2603
Joined: Mon Mar 12, 2012 6:57 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by comeinvest »

jarjarM wrote: Fri Jan 26, 2024 4:40 pm
comeinvest wrote: Wed Jan 24, 2024 5:44 pm
jarjarM wrote: Wed Jan 24, 2024 5:40 pm Also, one should learn from the significant drawdown the last 2 years that this is not hold and forget and really does require some active management in assigning risk budgets and making determination on overweighting an allocation depending on economic condition. Not a set and forget type of holding. Check out the riding HFEA threads for some more info.
Very easy to parameter-fit ("predict") the past. Not so much the future. Good luck doing "active management" monitoring and evaluating "economic conditions" and probably 1000 other parameters, and tweaking and fooling with the asset allocation on a constant basis.
That's true, past doesn't predict the future and most ppl end up fighting the last war. I guess my point is just that it's not really a set and forget investment and requires some TLC.
I disagree. I believe your last two sentences contradict each other; they are the polar opposite. You just use different words.
comeinvest
Posts: 2603
Joined: Mon Mar 12, 2012 6:57 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by comeinvest »

bond93 wrote: Fri Jan 26, 2024 4:12 pm Two comments
1) As you probably already know, for most people (myself included) DCA is not necessarily about spreading out a lump sum just to reduce perceived risk. There was a Dave Ramsey speech about this where he basically says he doesn't DCA because its a great strategy, hes just doing it because thats when he has the money to make his next "lump sum"(after a paycheck). So I DCA every month or so because thats as soon as I can buy in.
I'm not understanding what you are trying to say. If you have no lump sump, the question doesn't arise. You can only invest a lump sum, and the question only arises, if you have a lump sum.
bond93 wrote: Fri Jan 26, 2024 4:12 pm I might also wager it be beneficial that for people committing to this with 5-10% funny money, it's worth putting 5-10% of newly invested dollars towards it when possible rather than buy once and then try to build a parachute on the way down.
1. "Committing to this with 5-10% funny money", i.e. an isolated bucket in dollars for a leveraged strategy, is irrationally known subpar. It is hard to argue for or against knowingly irrational decisions, where the risk-adjusted final outcome distribution is not the optimization target, but something else which I'm unable to identify.
If you rebalance in and out of this bucket, you are fine, but then why not rather define your global asset allocation including this bucket; the performance of your bucket would then also not reflect a 3x strategy.
2. Even if you did some sort of isolated bucket in dollars for a leveraged strategy, then time in the market is that much more important for a leveraged strategy. I bet it would reduce risk, not increase it.
bond93 wrote: Fri Jan 26, 2024 4:12 pm
2) I agree with you that time in market is king in the majority of cases. For the standard S&P it seems one could put a lump sum down at almost any point and do alright after 10-15 years. (case in point: Bob the worlds worst market timer) but with 3x leverage and this strategy the game changes because yes at a certain point you may match/beat the market again, but that could take literally 50 years. (look at the 1955-2018 backtest) I recognize that the test was with 40/60 so if there is a 55/45 example from then that doesn't underperform for 50 years I'll happily cede my ground. I'd just personally like to be alive to enjoy my lottery :D
https://i.imgur.com/9D8QjKf.png
If you started in ca. 1964 then you are probably in the lowest percentile of outcomes until the 1980ies.
And so what. Either you invest in a strategy with this expected return and risk, or you don't. If you do, then the longer the time in the market the better. If you think dropping parts of an available lump sum money over time is beneficial compared to dropping the lump sum once available, then show me how you would have fared dropping into the strategy between 1955 and 1964, then being fully invested in it until 1982.
From a 1955 perspective with lump sum, tell me what motivates you to put more of your money at risk between 1964 and 1982 than between 1955 and 1964. Perhaps do some reading on the lifecycle investment strategy and principles. The case of a lump sum with no further additions is a special case, but I think the principle that more time in the market reduces risk applies to both.

I have bad news for you: If you don't like the tail risk, then deleverage. Also, if you are too old, then at some point it's "game over" for starting highly leveraged strategies. I'm not sure why it's not called "lifetime" but rather "lifecycle" investing which kind of suggests that you have more than one "cycle" at your disposal; not everybody believes in reincarnation.

The majority of new comments in this thread seem to be by folks who like the returns, but not the risk of HFEA and variants. Lol. And futile if not detrimental attempts at tweaking it. It becomes repetitive.
unemployed_pysicist
Posts: 220
Joined: Sat Oct 09, 2021 2:32 pm
Location: Amsterdam

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by unemployed_pysicist »

comeinvest wrote: Sun Jan 21, 2024 4:56 am
unemployed_pysicist wrote: Fri Jan 19, 2024 4:50 am
comeinvest wrote: Thu Jan 18, 2024 5:18 pm
unemployed_pysicist wrote: Thu Jan 18, 2024 4:56 pm
Logan Roy wrote: Thu Jan 18, 2024 1:01 pm
A big problem someone flagged a few days ago was that I'm using a fixed cost of leverage. That was very lazy of me, and will invalidate these results. They suggested using the 10yr yield as the cost of leverage. That's available on Robert Shiller's page, with the S&P data.
Hi Logan,

Did your 3xETFs in your backtest have daily resetting leverage? If so, I think Fed funds is readily available and should be a good proxy, plus some additional spread to simulate the swaps, etc. Whenever I make a daily resetting leverage time series, I stitch together the following series using the FRED API:

FFHTHIGH
FFWSJHIGH
DFF

This creates a cost of daily leverage time series all the way back to 1928. The first two series are the high values of the fedfunds rate published in the Wallstreet Journal and the New York Herald Tribune, which gives the worst case scenario for the cost of leverage - the low values are also on FRED if you want to make an average or something.

Just wanted to share with you in case that is helpful for future backtesting projects :)
The long-run average of daily money market rates will be very similar to 3-month money market rates (T-bills, Term SOFR, etc.), right?
The spread of implied financing rates of futures and options fluctuated historically. And we don't know the details of swaps; all we can assume is swaps are similar to futures, right?
Everything you said sounds about right to me. I like to use the Fedfunds rate for daily leveraged funds, because it goes back quite far (time series starts in 1928) and I presume that Fedfunds could have been used by an institution to make a daily resetting leverage fund (in theory.) I'm guessing 3 month rates are going to be fairly close, but using the fedfunds rate helps me feel like I have some level of precision with the historical 1 day/overnight rate.

As for the swaps, it is almost certain that a variable spread to Fedfunds or some other benchmark would have existed. When doing backtests I just use a static spread, which seems to be how most other people do it. Probably not entirely realistic, but that's what I do until someone devises how to make a robust (and believable) swap spread time series all the way to the 1950s or 1920s - if that's even possible.
I'm not understanding why, from a theoretical point of view, the duration of the implied financing has anything to do with the frequency of the reset (daily, monthly, etc.).
We don't know the details of the swap agreements do we? Do they specify a daily fluctuating financing charge spread based fed funds or similar daily rate, or might it be a fixed rate for 3 months for example?
I don't see why the duration of the implied financing would not depend on the frequency of the leverage reset, at least from a naïve perspective. Borrowing money for 1 day is different than borrowing money for 1 month, vs 3 months, vs 1 year, etc. I would expect a term structure to exist. In spherical cow land, the implied financing rate is just the zero-coupon yield curve of the risk-free rate (ignoring the complexities of balance sheet constraints, etc. - other real world phenomena.) Perhaps I am not understanding the point you are making.

I don't really know anything about the structure of the swap agreements used by Proshares and Direxion. I never dug into the specifics of those.

Direxion had to pay interest of 1 month LIBOR + spread for their swap contracts, according to EfficientInvestor in this post: viewtopic.php?p=4376074#p4376074
couldn't afford the h | | BUY BONDS | WEAR DIAMONDS
unemployed_pysicist
Posts: 220
Joined: Sat Oct 09, 2021 2:32 pm
Location: Amsterdam

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by unemployed_pysicist »

bond93 wrote: Fri Jan 26, 2024 10:54 am One topic I wanted to resurrect is the conversation about entry and exit strategies.
The backtests seem to be a great example of why if contributing a lump sum, timing is everything.
But what if one is perpetually contributing, every month/quarter for years, doesn't that tip the balance in favor of
55/45 UPRO/TMF even during rough times where you will eventually be "right" as long as you keep contributing? This way one can keep the "set and forget" allocation with the knowledge there will be tough times, but continually contributing lets you enter not once, but at every point along the way. To me this is as close as you can get to "buy SPY every friday" but in a leveraged position. I would love to see a backtest of DCA perpetually from 1955-1980s to see if that changes the calculus. I remember that was a gnarly time-period for this strategy but it was based on a lump sum.

Regarding the exit point, Hedgefundie himself had a hard stop at 10M. It seems rather than keep the fund alive indefinitely to "skim" off the
cream every year, a full exit at a precise number or precise multiple of injected funds will allow you to hit your target with no risk of going back below it.(of course getting to the top is the hard part.)

Apologies if this has already been covered, hard to find in 200+ pages of talk.
Perhaps not exactly what you are looking for, but I did a $100,000 lumpsum with $2,000 added monthly for a mHFEA-style portfolio here: https://bogleheads.org/forum/viewtopic. ... 9#p6716469

Note that this DCA strategy with the mHFEA-style portfolio gave about the same final result as the unlevered, $100,000 lumpsum into SP500 for the 1955-1982 period. There was no DCA into the SP500 portfolio, just an initial lumpsum of $100,000.
couldn't afford the h | | BUY BONDS | WEAR DIAMONDS
Hydromod
Posts: 1042
Joined: Tue Mar 26, 2019 10:21 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Hydromod »

bond93 wrote: Fri Jan 26, 2024 10:54 am One topic I wanted to resurrect is the conversation about entry and exit strategies.
The backtests seem to be a great example of why if contributing a lump sum, timing is everything.
But what if one is perpetually contributing, every month/quarter for years, doesn't that tip the balance in favor of
55/45 UPRO/TMF even during rough times where you will eventually be "right" as long as you keep contributing? This way one can keep the "set and forget" allocation with the knowledge there will be tough times, but continually contributing lets you enter not once, but at every point along the way. To me this is as close as you can get to "buy SPY every friday" but in a leveraged position. I would love to see a backtest of DCA perpetually from 1955-1980s to see if that changes the calculus. I remember that was a gnarly time-period for this strategy but it was based on a lump sum.

Regarding the exit point, Hedgefundie himself had a hard stop at 10M. It seems rather than keep the fund alive indefinitely to "skim" off the
cream every year, a full exit at a precise number or precise multiple of injected funds will allow you to hit your target with no risk of going back below it.(of course getting to the top is the hard part.)

Apologies if this has already been covered, hard to find in 200+ pages of talk.
I personally am not planning on exiting. I have a small portion of my portfolio sequestered off to an independent levered portfolio for logistical reasons; my remaining portfolio is in a 403b plan with a standard mix of available assets and no access to leverage, which will morph into my decumulation portfolio at retirement.

The levered portion is not straight HFEA, it's a strategy that is intended to emphasize geometric growth (see here). My goal is to let that grow for some time then gradually tithe off a portion to top up the decumulation portfolio. If it has proven to do well enough, I'll let the remainder go to potentially benefit others down the road.

So I have the viewpoint that the levered portion will be less and less mine over time.
er999
Posts: 1292
Joined: Wed Nov 05, 2008 10:00 am

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by er999 »

comeinvest wrote: Fri Jan 26, 2024 2:33 pm
bond93 wrote: Fri Jan 26, 2024 10:54 am One topic I wanted to resurrect is the conversation about entry and exit strategies.
The backtests seem to be a great example of why if contributing a lump sum, timing is everything.
But what if one is perpetually contributing, every month/quarter for years, doesn't that tip the balance in favor of
55/45 UPRO/TMF even during rough times where you will eventually be "right" as long as you keep contributing? This way one can keep the "set and forget" allocation with the knowledge there will be tough times, but continually contributing lets you enter not once, but at every point along the way. To me this is as close as you can get to "buy SPY every friday" but in a leveraged position. I would love to see a backtest of DCA perpetually from 1955-1980s to see if that changes the calculus. I remember that was a gnarly time-period for this strategy but it was based on a lump sum.

Regarding the exit point, Hedgefundie himself had a hard stop at 10M. It seems rather than keep the fund alive indefinitely to "skim" off the
cream every year, a full exit at a precise number or precise multiple of injected funds will allow you to hit your target with no risk of going back below it.(of course getting to the top is the hard part.)

Apologies if this has already been covered, hard to find in 200+ pages of talk.
Show me one rigorous study that suggests dollar cost averaging is beneficial. The perceived benefit of phasing entry points even when a lump sum is available, is usually an illusion. Maximize the time in the market and reduce your lifetime risk by adding money as soon as available.

I unfortunately lump summed $100k into HFEA end of 12/2021 at all time highs so I felt the pain of start date sensitivity. Now may be a good time to start but leveraged strategies are way more sensitive to start date then non leveraged.

On portfoliovisualizer lump sum $10k start date 1/1/22 VOO (non leveraged s&p500) would be worth $10,300, upro $7.2k, HFEA $4.6k now. That’s a big drop to climb out of.
comeinvest
Posts: 2603
Joined: Mon Mar 12, 2012 6:57 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by comeinvest »

er999 wrote: Sat Jan 27, 2024 10:21 am
comeinvest wrote: Fri Jan 26, 2024 2:33 pm
bond93 wrote: Fri Jan 26, 2024 10:54 am One topic I wanted to resurrect is the conversation about entry and exit strategies.
The backtests seem to be a great example of why if contributing a lump sum, timing is everything.
But what if one is perpetually contributing, every month/quarter for years, doesn't that tip the balance in favor of
55/45 UPRO/TMF even during rough times where you will eventually be "right" as long as you keep contributing? This way one can keep the "set and forget" allocation with the knowledge there will be tough times, but continually contributing lets you enter not once, but at every point along the way. To me this is as close as you can get to "buy SPY every friday" but in a leveraged position. I would love to see a backtest of DCA perpetually from 1955-1980s to see if that changes the calculus. I remember that was a gnarly time-period for this strategy but it was based on a lump sum.

Regarding the exit point, Hedgefundie himself had a hard stop at 10M. It seems rather than keep the fund alive indefinitely to "skim" off the
cream every year, a full exit at a precise number or precise multiple of injected funds will allow you to hit your target with no risk of going back below it.(of course getting to the top is the hard part.)

Apologies if this has already been covered, hard to find in 200+ pages of talk.
Show me one rigorous study that suggests dollar cost averaging is beneficial. The perceived benefit of phasing entry points even when a lump sum is available, is usually an illusion. Maximize the time in the market and reduce your lifetime risk by adding money as soon as available.

I unfortunately lump summed $100k into HFEA end of 12/2021 at all time highs so I felt the pain of start date sensitivity. Now may be a good time to start but leveraged strategies are way more sensitive to start date then non leveraged.

On portfoliovisualizer lump sum $10k start date 1/1/22 VOO (non leveraged s&p500) would be worth $10,300, upro $7.2k, HFEA $4.6k now. That’s a big drop to climb out of.
I don't disagree; but what I said in viewtopic.php?p=7678540#p7678540 is nevertheless correct, isn't it?
Another point of view is that "dropping" a lump-sum peu à peu expresses a decision dependency on your personal historical earnings path, which is irrational. Rational financial decisions can only have present and expected future parameters as input. Whether you inherited $1M just yesterday, or earned it over the past 30 years, shouldn't matter; only amount and timing of your expected future earnings, investment horizon, expected risk and performance of the assets, etc., should matter.
comeinvest
Posts: 2603
Joined: Mon Mar 12, 2012 6:57 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by comeinvest »

unemployed_pysicist wrote: Sat Jan 27, 2024 5:04 am
comeinvest wrote: Sun Jan 21, 2024 4:56 am
unemployed_pysicist wrote: Fri Jan 19, 2024 4:50 am
comeinvest wrote: Thu Jan 18, 2024 5:18 pm
unemployed_pysicist wrote: Thu Jan 18, 2024 4:56 pm

Hi Logan,

Did your 3xETFs in your backtest have daily resetting leverage? If so, I think Fed funds is readily available and should be a good proxy, plus some additional spread to simulate the swaps, etc. Whenever I make a daily resetting leverage time series, I stitch together the following series using the FRED API:

FFHTHIGH
FFWSJHIGH
DFF

This creates a cost of daily leverage time series all the way back to 1928. The first two series are the high values of the fedfunds rate published in the Wallstreet Journal and the New York Herald Tribune, which gives the worst case scenario for the cost of leverage - the low values are also on FRED if you want to make an average or something.

Just wanted to share with you in case that is helpful for future backtesting projects :)
The long-run average of daily money market rates will be very similar to 3-month money market rates (T-bills, Term SOFR, etc.), right?
The spread of implied financing rates of futures and options fluctuated historically. And we don't know the details of swaps; all we can assume is swaps are similar to futures, right?
Everything you said sounds about right to me. I like to use the Fedfunds rate for daily leveraged funds, because it goes back quite far (time series starts in 1928) and I presume that Fedfunds could have been used by an institution to make a daily resetting leverage fund (in theory.) I'm guessing 3 month rates are going to be fairly close, but using the fedfunds rate helps me feel like I have some level of precision with the historical 1 day/overnight rate.

As for the swaps, it is almost certain that a variable spread to Fedfunds or some other benchmark would have existed. When doing backtests I just use a static spread, which seems to be how most other people do it. Probably not entirely realistic, but that's what I do until someone devises how to make a robust (and believable) swap spread time series all the way to the 1950s or 1920s - if that's even possible.
I'm not understanding why, from a theoretical point of view, the duration of the implied financing has anything to do with the frequency of the reset (daily, monthly, etc.).
We don't know the details of the swap agreements do we? Do they specify a daily fluctuating financing charge spread based fed funds or similar daily rate, or might it be a fixed rate for 3 months for example?
I don't see why the duration of the implied financing would not depend on the frequency of the leverage reset, at least from a naïve perspective. Borrowing money for 1 day is different than borrowing money for 1 month, vs 3 months, vs 1 year, etc. I would expect a term structure to exist. In spherical cow land, the implied financing rate is just the zero-coupon yield curve of the risk-free rate (ignoring the complexities of balance sheet constraints, etc. - other real world phenomena.) Perhaps I am not understanding the point you are making.

I don't really know anything about the structure of the swap agreements used by Proshares and Direxion. I never dug into the specifics of those.

Direxion had to pay interest of 1 month LIBOR + spread for their swap contracts, according to EfficientInvestor in this post: viewtopic.php?p=4376074#p4376074
I'm new to this, but my understanding is that swaps are derivative agreements that exchange two or perhaps more cash flows. They can specify anything they want to define those cash flows that determine the future settlement. How each counterparty hedges their exposure from the swap with other financial instruments, or maps their total target exposure to the underlying financial markets to the swap contracts, is their own business.
Thinking of it, I agree that due to the daily resetting leverage of the exposure, a strict "3x daily exposure" after financing cost that would be the economic equivalent of a 3x implementation with fully paid equity securities and with borrowing cash at money market rates instead of swaps, would require a swap agreement based on daily rates, because of the slight fluctuation of the implicitly borrowed amount each day. If an index is at 5,000 today, a $5,000 equity investment in the ETF has 3x exposure of $15,000 with $10,000 implicit borrowing. If the index rises to 5,100 tomorrow, the ETF's swaps rise to $15,300 and equity is ca. $5,300. The exposure is reset to 3 x $5,300 = $15,900, which requires $15,900 - $5,300 = $10,600 implicit borrowing, $600 more than the day before.
So if the ETF's swap agreements were to specify short-term rates of other than 1 day duration, then the performance of the swap agreement would not accurately match that of an underlying hedging portfolio with debit interest based on the borrowed amount on any given day, because the borrowed amounts fluctuate daily and cannot be predicted over the length of the swap agreement.
But you said yourself that Direxion pays implicit interest based on 1-month LIBOR, even though Direxion's ETFs reset daily. So if we trust them, then I guess they just live with that small inaccuracy or discrepancy to a hedging portfolio with the same exposure that they market. I guess in reality it doesn't matter because 1-month rates are very close to daily rates, and the spread of the swap's agreed implicit financing rate to whatever base rate it uses is probably by magnitudes bigger than the difference between 1-month and daily rates. I thought through this only because it struck me that you went to great lengths assembling historical daily rates. But I understand that you were able to find fed fund rates more easily than other rates anyway.
bond93
Posts: 12
Joined: Mon Jan 22, 2024 11:38 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by bond93 »

comeinvest wrote: Fri Jan 26, 2024 7:49 pm
bond93 wrote: Fri Jan 26, 2024 4:12 pm Two comments
1) As you probably already know, for most people (myself included) DCA is not necessarily about spreading out a lump sum just to reduce perceived risk. There was a Dave Ramsey speech about this where he basically says he doesn't DCA because its a great strategy, hes just doing it because thats when he has the money to make his next "lump sum"(after a paycheck). So I DCA every month or so because thats as soon as I can buy in.
I'm not understanding what you are trying to say. If you have no lump sump, the question doesn't arise. You can only invest a lump sum, and the question only arises, if you have a lump sum.
bond93 wrote: Fri Jan 26, 2024 4:12 pm I might also wager it be beneficial that for people committing to this with 5-10% funny money, it's worth putting 5-10% of newly invested dollars towards it when possible rather than buy once and then try to build a parachute on the way down.
1. "Committing to this with 5-10% funny money", i.e. an isolated bucket in dollars for a leveraged strategy, is irrationally known subpar. It is hard to argue for or against knowingly irrational decisions, where the risk-adjusted final outcome distribution is not the optimization target, but something else which I'm unable to identify.
If you rebalance in and out of this bucket, you are fine, but then why not rather define your global asset allocation including this bucket; the performance of your bucket would then also not reflect a 3x strategy.
2. Even if you did some sort of isolated bucket in dollars for a leveraged strategy, then time in the market is that much more important for a leveraged strategy. I bet it would reduce risk, not increase it.
bond93 wrote: Fri Jan 26, 2024 4:12 pm
2) I agree with you that time in market is king in the majority of cases. For the standard S&P it seems one could put a lump sum down at almost any point and do alright after 10-15 years. (case in point: Bob the worlds worst market timer) but with 3x leverage and this strategy the game changes because yes at a certain point you may match/beat the market again, but that could take literally 50 years. (look at the 1955-2018 backtest) I recognize that the test was with 40/60 so if there is a 55/45 example from then that doesn't underperform for 50 years I'll happily cede my ground. I'd just personally like to be alive to enjoy my lottery :D
https://i.imgur.com/9D8QjKf.png
If you started in ca. 1964 then you are probably in the lowest percentile of outcomes until the 1980ies.
And so what. Either you invest in a strategy with this expected return and risk, or you don't. If you do, then the longer the time in the market the better. If you think dropping parts of an available lump sum money over time is beneficial compared to dropping the lump sum once available, then show me how you would have fared dropping into the strategy between 1955 and 1964, then being fully invested in it until 1982.
From a 1955 perspective with lump sum, tell me what motivates you to put more of your money at risk between 1964 and 1982 than between 1955 and 1964. Perhaps do some reading on the lifecycle investment strategy and principles. The case of a lump sum with no further additions is a special case, but I think the principle that more time in the market reduces risk applies to both.

I have bad news for you: If you don't like the tail risk, then deleverage. Also, if you are too old, then at some point it's "game over" for starting highly leveraged strategies. I'm not sure why it's not called "lifetime" but rather "lifecycle" investing which kind of suggests that you have more than one "cycle" at your disposal; not everybody believes in reincarnation.

The majority of new comments in this thread seem to be by folks who like the returns, but not the risk of HFEA and variants. Lol. And futile if not detrimental attempts at tweaking it. It becomes repetitive.
"I'm not understanding what you are trying to say. If you have no lump sump, the question doesn't arise. You can only invest a lump sum, and the question only arises, if you have a lump sum."
-I'm simply saying that to have the best outcome or likelihood of reaping the benefits of this strategy, perpetual contribution seems superior to a one-time bet and then no more. Thats all. 100k + 1k a month is better than just 100k then nothing. This is what I already said in my original post.

"Show me one rigorous study that suggests dollar cost averaging is beneficial." Lets look at recent history. Would you rather have lump summed 100k in 99 or DCA-ed for 5 years, same question for 08. In both cases, DCA would have been better. Obviously one can't predict exactly when those occur, but if all the markers are pointing to a correction, DCA may give you better returns. Maybe if the Buffett indicator is signaling strongly overvalued, its worth not lump summing, but spreading it out a few months or years. That would have helped you in 2000 and in 2020 atleast. heck, even buffett is now sitting on a record pile of cash over a hundred billion, and even he DCA's over months when taking a position, albiet for other reasons perhaps. But shouldn't he just load it all in now according to your line of thinking?

"If you started in ca. 1964 then you are probably in the lowest percentile of outcomes until the 1980ies."
-Incorrect if you are looking at a greater than 10-year time horizon, which is probably most of us. Starting in '64 with a lump sum you'd catch up to the market 30-40 years later. Starting in '55 you'd only think you were doing fine until 10 years later when you revert back below for the aformentioned period. Yes 64 or so is when the lines diverge, but you'd catch up to the market more quickly than a 55 start date. As I detail later, starting to DCA in 64 is actually a great time to do so if your time horizon is over 10 years.

"Perhaps do some reading on the lifecycle investment strategy and principles. The case of a lump sum with no further additions is a special case, but I think the principle that more time in the market reduces risk applies to both."
-Interesting that you mention this because in the book titled "Lifecycle Investing" the authors seem to argue for diversification across time.
"If Andrew were to invest 5k in yr1, 10k in yr2, and 15k in yr3, he would have invested 30k dollar years. Andrew is much better diversified against temporal fluctuations in the stock market when he invests the total dollar years more evenly across time, here 10k in each of three years. You should think of every year of your life as a distinct investment opportunity. People make the mistake of putting 80% of their stock investments in just ten years. You are better off spreading your stock investments across several decades." Chapter 1

"The majority of new comments in this thread seem to be by folks who like the returns, but not the risk of HFEA and variants. Lol. And futile if not detrimental attempts at tweaking it. It becomes repetitive."
-If you had started in 60 or 64, you would have done better spreading out 100,000 over 10 years than as a lump sum. The first years' injections would be before the downturn threshold but all remaining injections post-64 would be buying in at low points, putting you above the market sooner. One can see this when superimposing the chart on itself and moving the HFEA blue line up to where it meets red S&P for any given year. In fact, those later injections would be back up matching and beating the S&P within 20ish years, a reasonable time horizon for 30/40 yr olds. All money invested pre-64 would not match the market until 2010. But money invested after that could match it as soon as 87. When there are clear cases where spreading out large sums proves superior, it's hard to not see that as protective when using a strategy like this...and sure maybe that mutes your overall return (not in this case), but you could reduce risk of a ridiculous drawdown and still come out ahead of the market, sooner than had you done the lump sum. I hardly see that as "detrimental" or "futile."
Also, you can just as easily make the same points without condescension.
comeinvest
Posts: 2603
Joined: Mon Mar 12, 2012 6:57 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by comeinvest »

bond93 wrote: Sun Jan 28, 2024 12:28 am
comeinvest wrote: Fri Jan 26, 2024 7:49 pm
bond93 wrote: Fri Jan 26, 2024 4:12 pm Two comments
1) As you probably already know, for most people (myself included) DCA is not necessarily about spreading out a lump sum just to reduce perceived risk. There was a Dave Ramsey speech about this where he basically says he doesn't DCA because its a great strategy, hes just doing it because thats when he has the money to make his next "lump sum"(after a paycheck). So I DCA every month or so because thats as soon as I can buy in.
I'm not understanding what you are trying to say. If you have no lump sump, the question doesn't arise. You can only invest a lump sum, and the question only arises, if you have a lump sum.
bond93 wrote: Fri Jan 26, 2024 4:12 pm I might also wager it be beneficial that for people committing to this with 5-10% funny money, it's worth putting 5-10% of newly invested dollars towards it when possible rather than buy once and then try to build a parachute on the way down.
1. "Committing to this with 5-10% funny money", i.e. an isolated bucket in dollars for a leveraged strategy, is irrationally known subpar. It is hard to argue for or against knowingly irrational decisions, where the risk-adjusted final outcome distribution is not the optimization target, but something else which I'm unable to identify.
If you rebalance in and out of this bucket, you are fine, but then why not rather define your global asset allocation including this bucket; the performance of your bucket would then also not reflect a 3x strategy.
2. Even if you did some sort of isolated bucket in dollars for a leveraged strategy, then time in the market is that much more important for a leveraged strategy. I bet it would reduce risk, not increase it.
bond93 wrote: Fri Jan 26, 2024 4:12 pm
2) I agree with you that time in market is king in the majority of cases. For the standard S&P it seems one could put a lump sum down at almost any point and do alright after 10-15 years. (case in point: Bob the worlds worst market timer) but with 3x leverage and this strategy the game changes because yes at a certain point you may match/beat the market again, but that could take literally 50 years. (look at the 1955-2018 backtest) I recognize that the test was with 40/60 so if there is a 55/45 example from then that doesn't underperform for 50 years I'll happily cede my ground. I'd just personally like to be alive to enjoy my lottery :D
https://i.imgur.com/9D8QjKf.png
If you started in ca. 1964 then you are probably in the lowest percentile of outcomes until the 1980ies.
And so what. Either you invest in a strategy with this expected return and risk, or you don't. If you do, then the longer the time in the market the better. If you think dropping parts of an available lump sum money over time is beneficial compared to dropping the lump sum once available, then show me how you would have fared dropping into the strategy between 1955 and 1964, then being fully invested in it until 1982.
From a 1955 perspective with lump sum, tell me what motivates you to put more of your money at risk between 1964 and 1982 than between 1955 and 1964. Perhaps do some reading on the lifecycle investment strategy and principles. The case of a lump sum with no further additions is a special case, but I think the principle that more time in the market reduces risk applies to both.

I have bad news for you: If you don't like the tail risk, then deleverage. Also, if you are too old, then at some point it's "game over" for starting highly leveraged strategies. I'm not sure why it's not called "lifetime" but rather "lifecycle" investing which kind of suggests that you have more than one "cycle" at your disposal; not everybody believes in reincarnation.

The majority of new comments in this thread seem to be by folks who like the returns, but not the risk of HFEA and variants. Lol. And futile if not detrimental attempts at tweaking it. It becomes repetitive.
"I'm not understanding what you are trying to say. If you have no lump sump, the question doesn't arise. You can only invest a lump sum, and the question only arises, if you have a lump sum."
-I'm simply saying that to have the best outcome or likelihood of reaping the benefits of this strategy, perpetual contribution seems superior to a one-time bet and then no more. Thats all. 100k + 1k a month is better than just 100k then nothing. This is what I already said in my original post.

"Show me one rigorous study that suggests dollar cost averaging is beneficial." Lets look at recent history. Would you rather have lump summed 100k in 99 or DCA-ed for 5 years, same question for 08. In both cases, DCA would have been better. Obviously one can't predict exactly when those occur, but if all the markers are pointing to a correction, DCA may give you better returns. Maybe if the Buffett indicator is signaling strongly overvalued, its worth not lump summing, but spreading it out a few months or years. That would have helped you in 2000 and in 2020 atleast. heck, even buffett is now sitting on a record pile of cash over a hundred billion, and even he DCA's over months when taking a position, albiet for other reasons perhaps. But shouldn't he just load it all in now according to your line of thinking?

"If you started in ca. 1964 then you are probably in the lowest percentile of outcomes until the 1980ies."
-Incorrect if you are looking at a greater than 10-year time horizon, which is probably most of us. Starting in '64 with a lump sum you'd catch up to the market 30-40 years later. Starting in '55 you'd only think you were doing fine until 10 years later when you revert back below for the aformentioned period. Yes 64 or so is when the lines diverge, but you'd catch up to the market more quickly than a 55 start date. As I detail later, starting to DCA in 64 is actually a great time to do so if your time horizon is over 10 years.

"Perhaps do some reading on the lifecycle investment strategy and principles. The case of a lump sum with no further additions is a special case, but I think the principle that more time in the market reduces risk applies to both."
-Interesting that you mention this because in the book titled "Lifecycle Investing" the authors seem to argue for diversification across time.
"If Andrew were to invest 5k in yr1, 10k in yr2, and 15k in yr3, he would have invested 30k dollar years. Andrew is much better diversified against temporal fluctuations in the stock market when he invests the total dollar years more evenly across time, here 10k in each of three years. You should think of every year of your life as a distinct investment opportunity. People make the mistake of putting 80% of their stock investments in just ten years. You are better off spreading your stock investments across several decades." Chapter 1

"The majority of new comments in this thread seem to be by folks who like the returns, but not the risk of HFEA and variants. Lol. And futile if not detrimental attempts at tweaking it. It becomes repetitive."
-If you had started in 60 or 64, you would have done better spreading out 100,000 over 10 years than as a lump sum. The first years' injections would be before the downturn threshold but all remaining injections post-64 would be buying in at low points, putting you above the market sooner. One can see this when superimposing the chart on itself and moving the HFEA blue line up to where it meets red S&P for any given year. In fact, those later injections would be back up matching and beating the S&P within 20ish years, a reasonable time horizon for 30/40 yr olds. All money invested pre-64 would not match the market until 2010. But money invested after that could match it as soon as 87. When there are clear cases where spreading out large sums proves superior, it's hard to not see that as protective when using a strategy like this...and sure maybe that mutes your overall return (not in this case), but you could reduce risk of a ridiculous drawdown and still come out ahead of the market, sooner than had you done the lump sum. I hardly see that as "detrimental" or "futile."
Also, you can just as easily make the same points without condescension.
I didn't mean to be condescending, but to the point. My apologies.
I want to first address your paragraph where you reference the Lifecycle Investing book. Unfortunately I don't own the book; but with the example that you cited they seem to make an argument for not delaying investments. The example that you cited doesn't say how much investable money was actually available in any of those three years; but based on what I learned about Ayres' and Nalebuff's book, I assume that they would advocate investing whatever money is available as soon as it is available, if not front-loading by investing money that is not yet available, using leverage. I think that is the entire point of the book.

Regarding your other paragraphs: In your other paragraphs you are either conditioning your DCA strategy on the possibility of market timing, or you are picking specific examples on the historical timeline where delaying an investment would have been beneficial. I think Aryes/Nalebuff and HFEA are agnostic of market timing possibilities, which are an entirely different topic, and of course you can become rich very quickly if you can reliably predict markets; not to mention you can certainly improve HFEA once you introduce alpha from market timing.
Absent market timing and/or cherry-picking on the historical timeline, I think what I said is mathematically correct, and hard to factually argue against. Like I said, I think the key is my question "from a 1955 perspective with a lump sum, tell me what motivates you to put more of your money at risk between 1964 and 1982 than between 1955 and 1964." My understanding is that Ayres/Nalebuff's entire lifecycle investing is about time diversification, which means distributing the risk of negative outcomes during any specific year equally (as much as possible) across your investment horizon. I think that in the accumulation phase, this principle would imply to always invest at least the cash available as soon as it is available, if not "front-loading" with leverage to adjust for future contributions.
Like I said before with an example, DCAing by withholding cash available from investment, also doesn't make sense, as it contradicts the principle that neither your personal historical earnings timeline nor your personal historical investment performance trajectory can be an input to a rational investment process. It shouldn't matter whether you inherited $1M just yesterday, or if you accumulated $1M during the past 30 years. We can also try a proof by contradiction: If you DCA i.e. withhold x% of your then available net worth in year t1, then you would also have to withhold x% of your then available net worth in year t2 (t2 > t1), all other things being equal; which would mean you couldn't invest your entire net worth for the rest of your life. Proof by contradiction.
Please note that everything I either cited or concluded myself, is based on the assumption that I cannot time the market; and all my assertions refer to statistical ensembles with the distribution of final outcomes as optimization target, and not to specific historical time points in hindsight. Clearly, arguing something based on specific points on the historical timeline in hindsight is not productive: Whenever the market went down historically, it would have been better to delay an investment. That is trivial.
Last edited by comeinvest on Sun Jan 28, 2024 6:50 pm, edited 2 times in total.
User avatar
firebirdparts
Posts: 4331
Joined: Thu Jun 13, 2019 4:21 pm
Location: Southern Appalachia

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by firebirdparts »

Not just in this thread, but in all threads, I would advise people not to use the phrase “dollar cost averaging” to refer to any sort of payday or periodic investing. It is VERY triggering.

If you start a thread about investing on payday and use this unfortunate phrase or acronym, you may lose the concept of your whole thread and have to start a new one.

When HFEA started, it was mostly flavored by one initial investment and see how it does. I don’t think it would have been wrong to either:
1. Consider a second investment after a rising interest rate “cratering” event. This event was anticipated and well understood in the discussion if not by hedgefundie himself. Restarting did not garner much attention but its discussion in lifecycle investing should suffice here. It’s not stupid.
2. Consider regular investments through some part of your investing life. This was not discussed a whole lot in my recollection. I don’t talk in platitudes, so for me, the advantages of this, if there are any, could be explored by the user. The dynamics of these funds should be well understood by now if you’ve followed along.
This time is the same
hazlitt
Posts: 45
Joined: Sun Jan 29, 2017 9:00 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by hazlitt »

I got through the 3,000+ comments in the first HFEA thread, but haven’t plowed into this one yet. Assuming that someone was convinced to put a lull sump of UPRO/TMF at 55/45 in their Roth IRA (roughly 5% of NW) based on the thread as it existed toward then end of 2019, would it still be a good idea to get in now?

Interest rates are higher now than they were four years ago, so at least in theory there is more room for them to drop (obviously not counting on this).
Elysium
Posts: 4087
Joined: Mon Apr 02, 2007 6:22 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Elysium »

hazlitt wrote: Sun Jan 28, 2024 10:39 am I got through the 3,000+ comments in the first HFEA thread, but haven’t plowed into this one yet. Assuming that someone was convinced to put a lull sump of UPRO/TMF at 55/45 in their Roth IRA (roughly 5% of NW) based on the thread as it existed toward then end of 2019, would it still be a good idea to get in now?

Interest rates are higher now than they were four years ago, so at least in theory there is more room for them to drop (obviously not counting on this).
Since this strategy started in July 2019 to end of 2023, it is getting killed compared to plain S&P 500, with 2.5x times S/D.
here

I guess the hope here is rates reverse direction and TMF can make a come back, but if rates continue to stay higher for longer, even worse keep rising for years reversing the secular bond bull market then it will get wiped out. Any gains made from UPRO will continue to get eroded by TMF, and it will lose to plain S&P 500. Don't see that prospect discussed in the 3000+ thread I guess? I think the folks who made this strategy through elaborate back tests made a major flaw in not considering a secular bear market for bonds with rising rates for years. That said, it is a low probability and let's hope for all those invested in it here that doesn't happen. If and when rates stabilize this can crawl back to breakeven, and then may be provide some gains to justify the volatility.
Hydromod
Posts: 1042
Joined: Tue Mar 26, 2019 10:21 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Hydromod »

My understanding is that dollar cost averaging refers to the case where one has come into a sum of money and wants to invest it piece by piece over time instead of investing it immediately. In this case, I think that studies show that around 2/3 of the time one comes out ahead investing immediately (lump summing) versus spreading over time (dollar cost averaging).

If you are coming into sums of money to invest periodically (e.g., every payday), you have the choice of investing it immediately (lump summing the sequence of sums) or holding it in some low-risk form, then deploying the cumulative amount as a larger sum (lump summing or dollar cost averaging). Presumably the delay would be to provide the dry powder for a big investment at an opportune time (market time in).

In the payday series scenario, it seems that if one doesn't take the win of investing immediately (2/3 win rate) there should be some good rationale for delaying the investment. Presumably around 1/3 of the time that is a better choice.

Knowing when to switch back and forth between these two options is the trick. And remember, as the portfolio increases in size, the effect of each payday investment becomes less and less significant. If one is delaying the investment in order to redeploy at an opportune time, one has to have an increasingly large sum to make a significant difference, implying that the average delay before entering the market increases over time in order to accumulate the funds and a larger win will be needed to compensate for time out of the market.
manlymatt83
Posts: 1274
Joined: Tue Jan 30, 2018 7:23 am

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by manlymatt83 »

Decided to take some more time and read through this thread for the last year and a half. One of the things that kept getting repeated was the ideal leverage being 2X instead of 3X. Obviously, there were also discussions about whether to do this as a percentage of your overall portfolio vs an isolated portfolio.

Recently on Reddit, I saw someone who suggested that they were doing a modified version of MotoTrojan’s portfolio (43 UPRO/57 EDV) combined with the new RSSB (global stacked bonds and stocks fund).

I guess what this got them was something similar to HFEA, with less leverage, a smoother ride, and lower carrying costs. At the same time, RSSB also added a little bit of international diversification (it holds total world stock market), as well as shortened the bond duration a little bit compared to pure EDV.

Curious if this portfolio would make sense for some. It does get the leverage a little closer to 2X.
bond93
Posts: 12
Joined: Mon Jan 22, 2024 11:38 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by bond93 »

I didn't mean to be condescending, but to the point. My apologies.
I want to first address your paragraph where you reference the Lifecycle Investing book. Unfortunately I don't own the book; but with the example that you cited they seem to make an argument for not delaying investments. The example that you cited doesn't say how much investable money was actually available in any of those three years; but based on what I learned about Ayres' and Nalebuff's book, I assume that they would advocate investing whatever money is available as soon as it is available, if not front-loading by investing money that is not yet available, using leverage. I think that is the entire point of the book.

Regarding your other paragraphs: In your other paragraphs you are either conditioning your DCA strategy on the possibility of market timing, or your are picking specific examples on the historical timeline where delaying an investment would have been beneficial. I think Aryes/Nalebuff and HFEA are agnostic of market timing possibilities, which are an entirely different topic, and of course you can become rich very quickly if you can reliably predict markets; not to mention you can certainly improve HFEA once you introduce alpha from market timing.
Absent market timing and/or cherry-picking on the historical timeline, I think what I said is mathematically correct, and hard to factually argue against. Like I said, I think the key is my question "from a 1955 perspective with a lump sum, tell me what motivates you to put more of your money at risk between 1964 and 1982 than between 1955 and 1964." My understanding is that Ayres/Nalebuff's entire lifecycle investing is about time diversification, which means distributing the risk of negative outcomes during any specific year equally (as much as possible) across your investment horizon. I think that in the accumulation phase, this principle would imply to always invest at least the cash available as soon as it is available, if not "front-loading" with leverage to adjust for future contributions.
Like I said before with an example, DCAing by withholding cash available from investment, also doesn't make sense, as it contradicts the principle that neither your personal historical earnings timeline nor your personal historical investment performance trajectory can be an input to a rational investment process. It shouldn't matter whether you inherited $1M just yesterday, or if you accumulated $1M during the past 30 years. We can also try a proof by contradiction: If you DCA i.e. withhold x% of your then available net worth in year t1, then you would also have to withhold x% of your then available net worth in year t2, all other thing being equal; which would mean you couldn't invest your entire net worth for the rest of your life. Proof by contradiction.
Please note that everything I either cited or concluded myself, is based on the assumption that I cannot time the market; and all my assertions refer to statistical ensembles with the distribution of final outcomes as optimization target, and not to specific historical time points in hindsight. Clearly, arguing something based on specific points on the historical timeline in hindsight is not productive: Whenever the market went down historically, it would have been better to delay an investment. That is trivial.

Your points are taken, and I now realize you didn't mean offense. I suppose in answer to your question about the motivation to invest more in 64-82 vs 55-64 is this: "People make the mistake of putting 80% of their stock investments in just ten years. You are better off spreading your stock investments across several decades" They may be arguing for earlier stock exposure in life to accomplish this but I'd maintain that they would support the following as well: If you are 30 in 1955, if you keep investing from age 40-50 not just 30-40, this strategy, or any other, would have worked better for you especially considering your likely higher income in those later years. Similar to my point about perpetual investment rather than the onetime go. So sure, throw it all in asap when you have it, but why stop after just 10 years?

In response to the other commenters RE alternatives/modifications, the main litmus test to me at least is surviving the 1955 period while still beating the market over a 20+ year time horizon instead of 40-50 years. If there are backtests showing variations on HFEA that accomplish that I'd appreciate directions to them. Searching these threads seems a bit like finding a needle in a haystack at times.

For example: atleast from 1987 onward, 77/23 UPRO/CASH with quarterly rebalancing still outperforms S&P and only goes below it for 3-4 years, it may have also been safer during the 55-75 period.

EDIT: After reading more of the lifecycle book, it actually seems the authors DO take a one-foot-in one foot out stance on DCA. They use the shiller PE ratio as part of their formula and advise adjusting the target Samuelson share based on the current PE. They go so far as to say "when the P/E ratio goes above 27.7 our number crunching suggests that people should completely stop investing in stocks. When the market is this overpriced, the expected future stock returns dont justify the risk."
moootz
Posts: 1
Joined: Thu Feb 01, 2024 9:36 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by moootz »

manlymatt83 wrote: Wed Jan 24, 2024 5:53 pm
Hydromod wrote: Wed Jan 24, 2024 5:35 pm
manlymatt83 wrote: Wed Jan 24, 2024 4:24 pm I just want a set-it-and-forget-it allocation I can rebalance quarterly, somewhere between PSLDX (100% / 100%) and 55 UPRO/45 TMF. Maybe involving GOVZ as that seems to be a "better" version of EDV.

I don't mind doing the work myself but I am unsure how to come up with the "best" allocation for UPRO / GOVZ. Is it 40/60? 53/47? 50/50? etc.

I guess if I am following the 4 / 1 rule, what would the ratio be? Or how can I find that out?
In the form I refer to, risk weight is relative to variance.

The risk budget inverse volatility says that the allocation to asset i is (bi/vi)/sum(bj/vj), where b is the square root of the risk weight, v is the volatility, and there are j assets. You can use daily, monthly, or annualized volatility, just be consistent.

I'm remembering better now. I think that MojoTrojan was doing something with effective bond duration.

Anyway, PV says daily volatility for UPRO, TMF, TYD, and EDZ are 3.3%, 2.9%, 1.4%, and 4.1% since 2009 (here). GOVZ has been similar to TYD since inception.

A 4/1 risk balance with UPRO and TMF gives an average UPRO weight of (2/3.3)/[(2/3.3) + (1/2.9)] = 0.64. Or 64/36 UPRO/TMF. A pure risk parity would have 47/53 UPRO/TMF.

I ended up with lower average values in part because these weights change over time, and the time averaged values including crashes and booms are a bit different.

For comparison, 4/1 weighting with UPRO and TYD gets to an 46/54 allocation.

Hope this helps instead of overwhelming.
This makes sense. Thank you! Assuming the ER of GOVZ remains at 10 BP vs 6 BP for EDV, I really can’t see a reason why one would use EDV over GOVZ anymore. It basically gets you a few percentage points of free capital to allocate to UPRO if I’m not mistaken. Am I missing something?
Good question. It just sounds "too easy" of a decision now.
I have the UPRO/EDV combo going in a Roth IRA and am clicks away of selling all EDV and buying GOVZ.
manlymatt83
Posts: 1274
Joined: Tue Jan 30, 2018 7:23 am

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by manlymatt83 »

moootz wrote: Thu Feb 01, 2024 9:45 pm
manlymatt83 wrote: Wed Jan 24, 2024 5:53 pm
Hydromod wrote: Wed Jan 24, 2024 5:35 pm
manlymatt83 wrote: Wed Jan 24, 2024 4:24 pm I just want a set-it-and-forget-it allocation I can rebalance quarterly, somewhere between PSLDX (100% / 100%) and 55 UPRO/45 TMF. Maybe involving GOVZ as that seems to be a "better" version of EDV.

I don't mind doing the work myself but I am unsure how to come up with the "best" allocation for UPRO / GOVZ. Is it 40/60? 53/47? 50/50? etc.

I guess if I am following the 4 / 1 rule, what would the ratio be? Or how can I find that out?
In the form I refer to, risk weight is relative to variance.

The risk budget inverse volatility says that the allocation to asset i is (bi/vi)/sum(bj/vj), where b is the square root of the risk weight, v is the volatility, and there are j assets. You can use daily, monthly, or annualized volatility, just be consistent.

I'm remembering better now. I think that MojoTrojan was doing something with effective bond duration.

Anyway, PV says daily volatility for UPRO, TMF, TYD, and EDZ are 3.3%, 2.9%, 1.4%, and 4.1% since 2009 (here). GOVZ has been similar to TYD since inception.

A 4/1 risk balance with UPRO and TMF gives an average UPRO weight of (2/3.3)/[(2/3.3) + (1/2.9)] = 0.64. Or 64/36 UPRO/TMF. A pure risk parity would have 47/53 UPRO/TMF.

I ended up with lower average values in part because these weights change over time, and the time averaged values including crashes and booms are a bit different.

For comparison, 4/1 weighting with UPRO and TYD gets to an 46/54 allocation.

Hope this helps instead of overwhelming.
This makes sense. Thank you! Assuming the ER of GOVZ remains at 10 BP vs 6 BP for EDV, I really can’t see a reason why one would use EDV over GOVZ anymore. It basically gets you a few percentage points of free capital to allocate to UPRO if I’m not mistaken. Am I missing something?
Good question. It just sounds "too easy" of a decision now.
I have the UPRO/EDV combo going in a Roth IRA and am clicks away of selling all EDV and buying GOVZ.
Curious what you end up doing!
comeinvest
Posts: 2603
Joined: Mon Mar 12, 2012 6:57 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by comeinvest »

moootz wrote: Thu Feb 01, 2024 9:45 pm
manlymatt83 wrote: Wed Jan 24, 2024 5:53 pm
Hydromod wrote: Wed Jan 24, 2024 5:35 pm
manlymatt83 wrote: Wed Jan 24, 2024 4:24 pm I just want a set-it-and-forget-it allocation I can rebalance quarterly, somewhere between PSLDX (100% / 100%) and 55 UPRO/45 TMF. Maybe involving GOVZ as that seems to be a "better" version of EDV.

I don't mind doing the work myself but I am unsure how to come up with the "best" allocation for UPRO / GOVZ. Is it 40/60? 53/47? 50/50? etc.

I guess if I am following the 4 / 1 rule, what would the ratio be? Or how can I find that out?
In the form I refer to, risk weight is relative to variance.

The risk budget inverse volatility says that the allocation to asset i is (bi/vi)/sum(bj/vj), where b is the square root of the risk weight, v is the volatility, and there are j assets. You can use daily, monthly, or annualized volatility, just be consistent.

I'm remembering better now. I think that MojoTrojan was doing something with effective bond duration.

Anyway, PV says daily volatility for UPRO, TMF, TYD, and EDZ are 3.3%, 2.9%, 1.4%, and 4.1% since 2009 (here). GOVZ has been similar to TYD since inception.

A 4/1 risk balance with UPRO and TMF gives an average UPRO weight of (2/3.3)/[(2/3.3) + (1/2.9)] = 0.64. Or 64/36 UPRO/TMF. A pure risk parity would have 47/53 UPRO/TMF.

I ended up with lower average values in part because these weights change over time, and the time averaged values including crashes and booms are a bit different.

For comparison, 4/1 weighting with UPRO and TYD gets to an 46/54 allocation.

Hope this helps instead of overwhelming.
This makes sense. Thank you! Assuming the ER of GOVZ remains at 10 BP vs 6 BP for EDV, I really can’t see a reason why one would use EDV over GOVZ anymore. It basically gets you a few percentage points of free capital to allocate to UPRO if I’m not mistaken. Am I missing something?
Good question. It just sounds "too easy" of a decision now.
I have the UPRO/EDV combo going in a Roth IRA and am clicks away of selling all EDV and buying GOVZ.
I posted this in another thread: The slope between 20y and 30y has historically been rather negative - almost the entire history, except ca. 5 years during ZIRP. See the chart with the historical yields of 20y and 30y treasuries since 1986 in this post: viewtopic.php?p=7470297#p7470297
I have not found a way to do a meaningful long-term backtest comparison of long-term treasuries vs very long-term treasuries, because the performance charts as well as the Sharpe ratios are dominated by the effect of yield changes, as opposed to yield curve carry in the long run; but I would be concerned about getting inferior returns in the long-run from lower bond yields, and/or negative or smaller roll-down returns from negative (or less positive) slope, when going out too far on the yield curve; not to mention the higher duration risk. Yield curve carry controls term premia and therefore returns in the long run. In other words, I see a possibility of a combination of lower returns and higher risk, which doesn't sound like an appealing tradeoff ;)
It looks like 30y treasuries were only available since 1986. One might argue that the lower 30y rates were an artifact of generally decreasing rates, as the majority of this time interest rates were generally decreasing, which might have led to lower long-term rates as the market was anticipating decreasing short-term rates. But this argument doesn't hold, because the yield curve generally had a positive slope between 0 and 20y, despite generally decreasing rates. The yield curve slope between 20y and 30y has also been negative during the recent rising rates environment.
kimagical
Posts: 2
Joined: Sun Feb 04, 2024 10:56 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by kimagical »

From my read over, it seems hedgefundie bet that 2 things:

1) inflation was a solved problem because the fed would raise interest rates ahead in time to control inflation, causing a drop in TMF that would be compensated with a boom in UPRO that would occur because it was also assumed that inflation would only occur as a consequence of a booming economy causing increased demand for goods, raising prices.

Clearly this assumption was false as supply chain disruptions from covid and ukraine also caused rising prices of commodities and energy, causing inflation of every product that needed those commodities (ie. Everything). This type of inflation easily correlates with poor stock performance. It also seems that the cash injection from the Fed contributed, but I'm not sure how since cash injection in previous years didn't hurt this strategy. Or maybe they did, I'm not sure what caused the 50% drawdown in ~2010 according to OP's backtest. Anyway, the stagflation of 2022 caused a 75% drawdown that, assuming an optimistic 20% CAGR going forward, will take 7 years to return just to your original investment. This may not sound too bad if you're holding for the long run, unless stagflation occurs again too often.

Stagflation can be hedged against with DBMF (it's just an index of the top 10 managed futures hedge funds), but I think in the long run it will eat into the gains too much and the cost will be too high. Maybe could add 10% DBMF for 45 UPRO 45 TMF 10 DBMF, rebalanced at the start of every quarter. Backtesting this seems rather difficult though. There was testing somewhere of another managed future similarly uncorrelated, KMLM, and iirc it killed the gains so much you might as well have forgotten all about HFEA and went with 100% VOO.

2) Hedgefundie also bet that the underlyings of UPRO and TMF wouldn't mosey around sideways too much and kill UPRO and TMF due to volatility decay. An example where a 3x leveraged fund died due to underlying volatility includes LABU, which is down 87% in 5 years despite the underlying biotech index being up 7%.

I think the underlyings of UPRO and TMF are less volatile than this due to being massive and UPRO being diversified, but with the future being uncertain, I'm not sure I can bet against these volatilities increasing in the future. I can imagine a case of the Fed , continual supply chain disruptions due to further global conflicts, deglobalization, or increasing climate change disasters, introducing volatility previous never seen before in bonds. On the upside, the fed rate is 5.5% right now, so there's plenty of room to fall later for TMF to perform another 30 year bond bull run.

So overall, I think the question is do you think supply chains will continually see disruption? If so hedging with DBMF would be your bet. If supply chains will go back to pre-pandemic stages, then adding DBMF would just kill your gains to 100% VOO levels. Finally, either way, 50 UPRO 50 TMF, or 33 UPRO 33 TMF 33 DBMF would go bust in all the "right" conditions if volatility of stocks and bonds increases due to the simple whims of people increasing volatility. Last time people bet on low volatility they made tons of money on XIV 7 years ago until suddenly overnight they lost everything.

Personally, as a young adult with income to come and will not be needing my current savings, I'm leaning on putting everything in PSLDX as a less volatility reliant 2x version of HEFA in my tax free savings. A factor in this decision is to be able to think about it less as I won't have to do any management myself whatsoever with automatic dividend reinvestment. As my career has nothing to do with the stock market, I want to be able think about my portfolio as little as possible, and going on HEFA would cause me to think about the state of all these economic factors too much as part of management.
rockstar
Posts: 6208
Joined: Mon Feb 03, 2020 5:51 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by rockstar »

kimagical wrote: Thu Feb 08, 2024 12:00 pm From my read over, it seems hedgefundie bet that 2 things:

1) inflation was a solved problem because the fed would raise interest rates ahead in time to control inflation, causing a drop in TMF that would be compensated with a boom in UPRO that would occur because it was also assumed that inflation would only occur as a consequence of a booming economy causing increased demand for goods, raising prices.

Clearly this assumption was false as supply chain disruptions from covid and ukraine also caused rising prices of commodities and energy, causing inflation of every product that needed those commodities (ie. Everything). This type of inflation easily correlates with poor stock performance. It also seems that the cash injection from the Fed contributed, but I'm not sure how since cash injection in previous years didn't hurt this strategy. Or maybe they did, I'm not sure what caused the 50% drawdown in ~2010 according to OP's backtest. Anyway, the stagflation of 2022 caused a 75% drawdown that, assuming an optimistic 20% CAGR going forward, will take 7 years to return just to your original investment. This may not sound too bad if you're holding for the long run, unless stagflation occurs again too often.

Stagflation can be hedged against with DBMF (it's just an index of the top 10 managed futures hedge funds), but I think in the long run it will eat into the gains too much and the cost will be too high. Maybe could add 10% DBMF for 45 UPRO 45 TMF 10 DBMF, rebalanced at the start of every quarter. Backtesting this seems rather difficult though. There was testing somewhere of another managed future similarly uncorrelated, KMLM, and iirc it killed the gains so much you might as well have forgotten all about HFEA and went with 100% VOO.

2) Hedgefundie also bet that the underlyings of UPRO and TMF wouldn't mosey around sideways too much and kill UPRO and TMF due to volatility decay. An example where a 3x leveraged fund died due to underlying volatility includes LABU, which is down 87% in 5 years despite the underlying biotech index being up 7%.

I think the underlyings of UPRO and TMF are less volatile than this due to being massive and UPRO being diversified, but with the future being uncertain, I'm not sure I can bet against these volatilities increasing in the future. I can imagine a case of the Fed , continual supply chain disruptions due to further global conflicts, deglobalization, or increasing climate change disasters, introducing volatility previous never seen before in bonds. On the upside, the fed rate is 5.5% right now, so there's plenty of room to fall later for TMF to perform another 30 year bond bull run.

So overall, I think the question is do you think supply chains will continually see disruption? If so hedging with DBMF would be your bet. If supply chains will go back to pre-pandemic stages, then adding DBMF would just kill your gains to 100% VOO levels. Finally, either way, 50 UPRO 50 TMF, or 33 UPRO 33 TMF 33 DBMF would go bust in all the "right" conditions if volatility of stocks and bonds increases due to the simple whims of people increasing volatility. Last time people bet on low volatility they made tons of money on XIV 7 years ago until suddenly overnight they lost everything.

Personally, as a young adult with income to come and will not be needing my current savings, I'm leaning on putting everything in PSLDX as a less volatility reliant 2x version of HEFA in my tax free savings. A factor in this decision is to be able to think about it less as I won't have to do any management myself whatsoever with automatic dividend reinvestment. As my career has nothing to do with the stock market, I want to be able think about my portfolio as little as possible, and going on HEFA would cause me to think about the state of all these economic factors too much as part of management.
Lots of leverage in bond funds is bad when rates go up really fast.
manlymatt83
Posts: 1274
Joined: Tue Jan 30, 2018 7:23 am

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by manlymatt83 »

kimagical wrote: Thu Feb 08, 2024 12:00 pm Personally, as a young adult with income to come and will not be needing my current savings, I'm leaning on putting everything in PSLDX as a less volatility reliant 2x version of HEFA in my tax free savings. A factor in this decision is to be able to think about it less as I won't have to do any management myself whatsoever with automatic dividend reinvestment. As my career has nothing to do with the stock market, I want to be able think about my portfolio as little as possible, and going on HEFA would cause me to think about the state of all these economic factors too much as part of management.
I used to be 55 UPRO/45 TMF but recently switched to 43 UPRO / 57 EDV. PSLDX is great, but I can't hold it at my broker. The one thing PSLDX has is corporate bonds, so more closely tracks 35 UPRO, 20 TMF, 45 BLV for an ETF version. If I change anything in the future, I may go 35% UPRO and 65% EDV to get closer to PSLDX leverage, but for now am happy.
Lawyered_
Posts: 164
Joined: Wed Feb 09, 2022 10:52 am

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Lawyered_ »

manlymatt83 wrote: Thu Feb 08, 2024 12:15 pm
kimagical wrote: Thu Feb 08, 2024 12:00 pm Personally, as a young adult with income to come and will not be needing my current savings, I'm leaning on putting everything in PSLDX as a less volatility reliant 2x version of HEFA in my tax free savings. A factor in this decision is to be able to think about it less as I won't have to do any management myself whatsoever with automatic dividend reinvestment. As my career has nothing to do with the stock market, I want to be able think about my portfolio as little as possible, and going on HEFA would cause me to think about the state of all these economic factors too much as part of management.
I used to be 55 UPRO/45 TMF but recently switched to 43 UPRO / 57 EDV. PSLDX is great, but I can't hold it at my broker. The one thing PSLDX has is corporate bonds, so more closely tracks 35 UPRO, 20 TMF, 45 BLV for an ETF version. If I change anything in the future, I may go 35% UPRO and 65% EDV to get closer to PSLDX leverage, but for now am happy.
What made you decide to use EDV over ZROZ or GOVZ?
manlymatt83
Posts: 1274
Joined: Tue Jan 30, 2018 7:23 am

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by manlymatt83 »

Lawyered_ wrote: Fri Feb 09, 2024 9:05 am
manlymatt83 wrote: Thu Feb 08, 2024 12:15 pm I used to be 55 UPRO/45 TMF but recently switched to 43 UPRO / 57 EDV. PSLDX is great, but I can't hold it at my broker. The one thing PSLDX has is corporate bonds, so more closely tracks 35 UPRO, 20 TMF, 45 BLV for an ETF version. If I change anything in the future, I may go 35% UPRO and 65% EDV to get closer to PSLDX leverage, but for now am happy.
What made you decide to use EDV over ZROZ or GOVZ?
Honestly, I originally went with GOVZ because of the slightly longer duration, but after some googling / reddit posts, found that EDV is:

- Talked about more often, especially in the context of an alternative to HFEA. I wanted to be able to compare apples to apples 5 years from now when people are talking about HFEA and performance again :)
- Higher volume and lower spreads, easier to rebalance
- Lower expense ratio

I'd be open to GOVZ or ZROZ, but the question is, would the proportions still be 43 UPRO / 57 GOVZ? Or would they adjust a bit since GOVZ is more volatile... such as 45 UPRO / 55 GOVZ? Same with ZROZ.
User avatar
hiddenpower
Posts: 540
Joined: Tue Nov 17, 2020 11:24 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by hiddenpower »

Next time the fed says their raising rates and inflation is high.. I'm definitely going to continue this strategy.
User avatar
firebirdparts
Posts: 4331
Joined: Thu Jun 13, 2019 4:21 pm
Location: Southern Appalachia

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by firebirdparts »

hiddenpower wrote: Sat Feb 10, 2024 11:26 am Next time the fed says their raising rates and inflation is high.. I'm definitely going to continue this strategy.
Honestly I never thought I'd see that clear of a warning for that long. But I'm sure glad we got it.
This time is the same
User avatar
hiddenpower
Posts: 540
Joined: Tue Nov 17, 2020 11:24 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by hiddenpower »

firebirdparts wrote: Sat Feb 10, 2024 12:01 pm
hiddenpower wrote: Sat Feb 10, 2024 11:26 am Next time the fed says their raising rates and inflation is high.. I'm definitely going to continue this strategy.
Honestly I never thought I'd see that clear of a warning for that long. But I'm sure glad we got it.
Did you sell out of everything? If only. I didn't have that much conviction. I'd sell out of LETFs though if such a warning came about again. Poor risk reward..
manlymatt83
Posts: 1274
Joined: Tue Jan 30, 2018 7:23 am

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by manlymatt83 »

In reading this thread from start to finish, one thing I often see mentioned is that some think doing this in isolation (for example... $100k in HFEA, the rest in VT/BND) is a bad strategy. And that something like this should be an overall part of your portfolio (10% of a 1 million dollar portfolio for example, to get at $100k).

In those cases, how often are you rebalancing out of UPRO/TMF into your other 90%?

For example, this portfolio:

5.5% UPRO
4.5% TMF
81% VT
9% BNDW

While you're rebalancing between UPRO and TMF quarterly, would you also rebalance into VT/BNDW quarterly? And is it fair to say you'd never rebalance FROM VT/BNDW *into* UPRO/TMF, only the other way around?

Or is the expectation that one is coming up with an amount of leverage they're comfortable with ... 1.25x for example. And then always maintaining that.
Hydromod
Posts: 1042
Joined: Tue Mar 26, 2019 10:21 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Hydromod »

I think the argument is that the way to handle this as an overall part of the portfolio is to simply set the overall allocations (including leverage), then selecting the minimum number of LETFs needed to maintain that allocation. So you might only need one of UPRO or TMF, for example, depending on your overall leverage level.

That's a little different than having something like 20/80 HFEA/normie and rebalancing to maintain both the HFEA and normie sleeves, but effectively the same with a little smaller total ER.

That is different from allocating 20/80 percent of inflows to separate HFEA and normie sleeves and letting them both go as they will. One would expect that the HFEA sleeve will gradually deviate from 20 percent (hopefully to the up side), and the overall portfolio will have a larger dispersion of outcomes.
director84
Posts: 48
Joined: Thu Dec 23, 2010 9:59 am

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by director84 »

manlymatt83 wrote: Sat Feb 10, 2024 9:14 pm In reading this thread from start to finish, one thing I often see mentioned is that some think doing this in isolation (for example... $100k in HFEA, the rest in VT/BND) is a bad strategy. And that something like this should be an overall part of your portfolio (10% of a 1 million dollar portfolio for example, to get at $100k).

In those cases, how often are you rebalancing out of UPRO/TMF into your other 90%?

For example, this portfolio:

5.5% UPRO
4.5% TMF
81% VT
9% BNDW

While you're rebalancing between UPRO and TMF quarterly, would you also rebalance into VT/BNDW quarterly? And is it fair to say you'd never rebalance FROM VT/BNDW *into* UPRO/TMF, only the other way around?

Or is the expectation that one is coming up with an amount of leverage they're comfortable with ... 1.25x for example. And then always maintaining that.
Using your example, your overall portfolio is 97.5% stock/9% ITT/13.5% LTT. LTT has been shown to be a worse investment than ITT, and you could achieve similar returns with much lower max drawdown by doing something like 105% stock/105% ITT.
https://www.portfoliovisualizer.com/bac ... Iw0bR1ozeC

Futures are a better way to leverage, but if you want to stick with LETFs for simplicity, you could do something like 30 UPRO / 15 VTI / 25 TYD / 30 VGIT.


(The examples I used only show US stocks/bonds so you'd need to adjust a bit to include your international allocation, but the principals are the same.)

To answer your question about rebalancing, quarterly seems the most popular. Annually is probably fine, too. Probably don't want to do any more or less frequently than that, though.

Given you were willing to put in the effort to read this whole thread, I'd recommend reading the Modified HFEA thread, also. There is a lot of discussion on setting leverage targets for your portfolio as a whole and how to adjust in accordance with lifecycle investing principals.
User avatar
firebirdparts
Posts: 4331
Joined: Thu Jun 13, 2019 4:21 pm
Location: Southern Appalachia

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by firebirdparts »

hiddenpower wrote: Sat Feb 10, 2024 1:43 pm
firebirdparts wrote: Sat Feb 10, 2024 12:01 pm
hiddenpower wrote: Sat Feb 10, 2024 11:26 am Next time the fed says their raising rates and inflation is high.. I'm definitely going to continue this strategy.
Honestly I never thought I'd see that clear of a warning for that long. But I'm sure glad we got it.
Did you sell out of everything? If only. I didn't have that much conviction. I'd sell out of LETFs though if such a warning came about again. Poor risk reward..
Just TMF. My problem always is buying back in too soon.
This time is the same
manlymatt83
Posts: 1274
Joined: Tue Jan 30, 2018 7:23 am

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by manlymatt83 »

firebirdparts wrote: Mon Feb 12, 2024 9:12 am
hiddenpower wrote: Sat Feb 10, 2024 1:43 pm
firebirdparts wrote: Sat Feb 10, 2024 12:01 pm
hiddenpower wrote: Sat Feb 10, 2024 11:26 am Next time the fed says their raising rates and inflation is high.. I'm definitely going to continue this strategy.
Honestly I never thought I'd see that clear of a warning for that long. But I'm sure glad we got it.
Did you sell out of everything? If only. I didn't have that much conviction. I'd sell out of LETFs though if such a warning came about again. Poor risk reward..
Just TMF. My problem always is buying back in too soon.
So are you 100% UPRO?
bond93
Posts: 12
Joined: Mon Jan 22, 2024 11:38 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by bond93 »

Curious to know what people think of doing a Ray Dalio style all weather portfolio modified towards HFEA.
I have been experimenting with a small amount of funds, a few grand, with this allocation:
55 UPRO
20 TMF
10 TYD 7-10yr trsrs.
7.5 UGL 2x Gold
7.5 UTSL 3x Utilities

In all the backtesting one can do with the historical data for each fund, it beats HFEA in 90% of cases with smaller drawdowns . It seems to me that the main issue with HFEA is not just the huge drag that TMF can occur, but the lack of diversification overall. The stocks can still drive great returns, but this safety net is more diverse with a larger number of uncorrelated assets. Looking for someone to shoot holes in this approach. Hard to gather historical data for the last three beyond 5-10 years.

On another note, there's been a fair bit of talk about the conference board LEI Index. It apparently has a perfect track record of signaling big market downturns and it's currently in danger territory.
Last edited by bond93 on Mon Feb 12, 2024 9:49 am, edited 1 time in total.
Lawyered_
Posts: 164
Joined: Wed Feb 09, 2022 10:52 am

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Lawyered_ »

bond93 wrote: Mon Feb 12, 2024 9:41 am
director84 wrote: Sun Feb 11, 2024 4:15 pm
manlymatt83 wrote: Sat Feb 10, 2024 9:14 pm In reading this thread from start to finish, one thing I often see mentioned is that some think doing this in isolation (for example... $100k in HFEA, the rest in VT/BND) is a bad strategy. And that something like this should be an overall part of your portfolio (10% of a 1 million dollar portfolio for example, to get at $100k).

In those cases, how often are you rebalancing out of UPRO/TMF into your other 90%?

For example, this portfolio:

5.5% UPRO
4.5% TMF
81% VT
9% BNDW

While you're rebalancing between UPRO and TMF quarterly, would you also rebalance into VT/BNDW quarterly? And is it fair to say you'd never rebalance FROM VT/BNDW *into* UPRO/TMF, only the other way around?

Or is the expectation that one is coming up with an amount of leverage they're comfortable with ... 1.25x for example. And then always maintaining that.
Using your example, your overall portfolio is 97.5% stock/9% ITT/13.5% LTT. LTT has been shown to be a worse investment than ITT, and you could achieve similar returns with much lower max drawdown by doing something like 105% stock/105% ITT.
https://www.portfoliovisualizer.com/bac ... Iw0bR1ozeC

Futures are a better way to leverage, but if you want to stick with LETFs for simplicity, you could do something like 30 UPRO / 15 VTI / 25 TYD / 30 VGIT.


(The examples I used only show US stocks/bonds so you'd need to adjust a bit to include your international allocation, but the principals are the same.)

To answer your question about rebalancing, quarterly seems the most popular. Annually is probably fine, too. Probably don't want to do any more or less frequently than that, though.

Given you were willing to put in the effort to read this whole thread, I'd recommend reading the Modified HFEA thread, also. There is a lot of discussion on setting leverage targets for your portfolio as a whole and how to adjust in accordance with lifecycle investing principals.
Curious to know what people think of doing a Ray Dalio style all weather portfolio modified towards HFEA.
I have been experimenting with a small amount of funds, a few grand, with this allocation:
55 UPRO
20 TMF
10 TYD 7-10yr trsrs.
7.5 UGL 2x Gold
7.5 UTSL 3x Utilities

In all the backtesting one can do with the historical data for each fund, it beats HFEA in 90% of cases with smaller drawdowns . It seems to me that the main issue with HFEA is not just the huge drag that TMF can occur, but the lack of diversification overall. The stocks can still drive great returns, but this safety net is more diverse with a larger number of uncorrelated assets. Looking for someone to shoot holes in this approach.
Not enough ballast for too much expense. Why not just tone down the leverage and run something like 20% UPRO; 20% AVUV; 60% VXUS?
bond93
Posts: 12
Joined: Mon Jan 22, 2024 11:38 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by bond93 »

Lawyered_ wrote: Mon Feb 12, 2024 9:45 am
bond93 wrote: Mon Feb 12, 2024 9:41 am
director84 wrote: Sun Feb 11, 2024 4:15 pm
manlymatt83 wrote: Sat Feb 10, 2024 9:14 pm In reading this thread from start to finish, one thing I often see mentioned is that some think doing this in isolation (for example... $100k in HFEA, the rest in VT/BND) is a bad strategy. And that something like this should be an overall part of your portfolio (10% of a 1 million dollar portfolio for example, to get at $100k).

In those cases, how often are you rebalancing out of UPRO/TMF into your other 90%?

For example, this portfolio:

5.5% UPRO
4.5% TMF
81% VT
9% BNDW

While you're rebalancing between UPRO and TMF quarterly, would you also rebalance into VT/BNDW quarterly? And is it fair to say you'd never rebalance FROM VT/BNDW *into* UPRO/TMF, only the other way around?

Or is the expectation that one is coming up with an amount of leverage they're comfortable with ... 1.25x for example. And then always maintaining that.
Using your example, your overall portfolio is 97.5% stock/9% ITT/13.5% LTT. LTT has been shown to be a worse investment than ITT, and you could achieve similar returns with much lower max drawdown by doing something like 105% stock/105% ITT.
https://www.portfoliovisualizer.com/bac ... Iw0bR1ozeC

Futures are a better way to leverage, but if you want to stick with LETFs for simplicity, you could do something like 30 UPRO / 15 VTI / 25 TYD / 30 VGIT.


(The examples I used only show US stocks/bonds so you'd need to adjust a bit to include your international allocation, but the principals are the same.)

To answer your question about rebalancing, quarterly seems the most popular. Annually is probably fine, too. Probably don't want to do any more or less frequently than that, though.

Given you were willing to put in the effort to read this whole thread, I'd recommend reading the Modified HFEA thread, also. There is a lot of discussion on setting leverage targets for your portfolio as a whole and how to adjust in accordance with lifecycle investing principals.
Curious to know what people think of doing a Ray Dalio style all weather portfolio modified towards HFEA.
I have been experimenting with a small amount of funds, a few grand, with this allocation:
55 UPRO
20 TMF
10 TYD 7-10yr trsrs.
7.5 UGL 2x Gold
7.5 UTSL 3x Utilities

In all the backtesting one can do with the historical data for each fund, it beats HFEA in 90% of cases with smaller drawdowns . It seems to me that the main issue with HFEA is not just the huge drag that TMF can occur, but the lack of diversification overall. The stocks can still drive great returns, but this safety net is more diverse with a larger number of uncorrelated assets. Looking for someone to shoot holes in this approach.
Not enough ballast for too much expense. Why not just tone down the leverage and run something like 20% UPRO; 20% AVUV; 60% VXUS?
The expense ratio is lower than that of original HFEA I believe, but you get more variety of safety-net. That must be an improvement no?
Yes thats a smoother ride in the last 4 years, but there just doesn't seem to be enough data to compare the two portfolios over a meaningful time period... AVUV is just 4 years old? Substituting it with VBR for a longer lookback shows significant underperformance vs S+P and the Dalio-fied HFEA. Also, 60% is quite heavy on international for me. Seriously considering hiring someone to put together datasets for all of these funds pre-respective-inception year.
User avatar
firebirdparts
Posts: 4331
Joined: Thu Jun 13, 2019 4:21 pm
Location: Southern Appalachia

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by firebirdparts »

manlymatt83 wrote: Mon Feb 12, 2024 9:14 am
firebirdparts wrote: Mon Feb 12, 2024 9:12 am
hiddenpower wrote: Sat Feb 10, 2024 1:43 pm
firebirdparts wrote: Sat Feb 10, 2024 12:01 pm
hiddenpower wrote: Sat Feb 10, 2024 11:26 am Next time the fed says their raising rates and inflation is high.. I'm definitely going to continue this strategy.
Honestly I never thought I'd see that clear of a warning for that long. But I'm sure glad we got it.
Did you sell out of everything? If only. I didn't have that much conviction. I'd sell out of LETFs though if such a warning came about again. Poor risk reward..
Just TMF. My problem always is buying back in too soon.
So are you 100% UPRO?
Gosh, no. The warning was years ago. I am pretty enthusiastic about long term bonds now.
This time is the same
producer
Posts: 5
Joined: Wed Dec 28, 2022 12:30 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by producer »

Hi all, this was originally a separate thread but didn't get much attention and someone suggested that I post here instead so here goes:

For investors comfortable with leveraged ETFs and who are seeking to implement a 2x large cap leverage strategy (or some variation of it) in a taxable account, what are your thoughts on 100% SSO vs. 50/50 VOO/UPRO? I have read previous threads on this and am aware that 50/50 VOO/UPRO has produced slightly higher returns over the long run, at least in part due to the lower effective expense ratio. For example, for an investor with a $1M portfolio, the 50/50 VOO/UPRO would be estimated to cost about $4500 less annually in expenses relative to 100% SSO (a pretty small figure relatively IMO but not negligible).

But my primary concern is this: could there be a situation, such as a catastrophic market crash where there are multiple consecutive daily market drops of 20% or so (say 3-4 days for example), where the 50% UPRO holding drops so close to zero that the fund closes or liquidates (and so leaving no way to recoup the losses), whereas SSO drops substantially but survives? I realize this is very unlikely but am wondering if the excess return of the 50/50 VOO/UPRO over 100% SSO is actually worth the risk.
User avatar
OohLaLa
Posts: 456
Joined: Tue Mar 09, 2021 7:26 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by OohLaLa »

producer wrote: Tue Feb 13, 2024 5:54 pm Hi all, this was originally a separate thread but didn't get much attention and someone suggested that I post here instead so here goes:

For investors comfortable with leveraged ETFs and who are seeking to implement a 2x large cap leverage strategy (or some variation of it) in a taxable account, what are your thoughts on 100% SSO vs. 50/50 VOO/UPRO? I have read previous threads on this and am aware that 50/50 VOO/UPRO has produced slightly higher returns over the long run, at least in part due to the lower effective expense ratio. For example, for an investor with a $1M portfolio, the 50/50 VOO/UPRO would be estimated to cost about $4500 less annually in expenses relative to 100% SSO (a pretty small figure relatively IMO but not negligible).

But my primary concern is this: could there be a situation, such as a catastrophic market crash where there are multiple consecutive daily market drops of 20% or so (say 3-4 days for example), where the 50% UPRO holding drops so close to zero that the fund closes or liquidates (and so leaving no way to recoup the losses), whereas SSO drops substantially but survives? I realize this is very unlikely but am wondering if the excess return of the 50/50 VOO/UPRO over 100% SSO is actually worth the risk.
Hello, I will call the two approaches "100" and "5050", for sanity's sake. Also, sorry if anything ends up being ultra obvious for you, as I don't know what you know. Might be useful for some other readers, as well. :D

From your comments, I imagine you already ran the following backtest to get an idea: https://www.portfoliovisualizer.com/bac ... eIkpdegIUU

Disclaimer: I know that I am just showing one starting point with that test and the differences will vary as you change start and end points. If you change the start point to any other month of 2010, for example, you naturally see varying results; sometimes "better", sometimes "worse". Also, PV only gives you end-of-month closings as data points, so fluctuations, as felt by any actual, living investor, are not exactly reflected.

There was quite a bit of talk about this, throughout the threads, and the consensus indeed was that, in theory, 5050 wins out. Currently, you can squeeze out approx. a 0.40-0.41% saving in expenses with 5050. When you look at the actual difference in key metrics (+0.21% CAGR, +0.12% Standard Deviation and -0.23 Max Drawdown over approx. 14 years, with Feb. 2010 start), you see the difference is not as great as a simple comparison of expense ratios would lead one to believe. Half of those savings seem to evaporate.

My semi-informed 2 cents about all this:
-5050 loses some efficiency to 100, when you see the loss in theoretical savings. This is probably due to the different leverage ETFs showing different velocity (surges and drawdowns, "decay"), mixed with rebalancing timing.

-Is the 0.2% really worth it, in your personal case? 5050 will theoretically require more rebalancing, bringing potentially higher fees and taxation. Add in additional bookkeeping and other messing around with more ETFs, as little as it is.

-A smaller risk that gets diversified with 5050 is organisational risk, as you have two different investment management companies instead of one. There are of course protections against various issues such a company might cause for investors, but the risk still exists and should be kept in mind.

-I wouldn't worry about a scenario where either UPRO or SSO get shuttered during a severe crash. My gut feeling is that UPRO is more at risk, but both are based on an extremely popular index/ market, that sees huge volume of traffic. The idea that loss of assets through underlying S&P depreciation, combined with investor outflows would be so great that Proshares would be unable to profitably service the fund, seems a bit far-fetched. Add in the fact that 3x are now grandfathered-in funds, that will not be recreated barring regulatory changes. The examples that often get floated are of relatively low-volume but highly-leveraged funds of very volative underlying products (reverse VIX, crude oil futures). IMO, the risks are in completely different realms.

-In catastrophic situations, the proposed safety behind 5050 is the unleveraged ETF. It's true that the 3x will drop more drastically than the lone 2x, but half of your invested assets are in the 1x. Looking at a simple one-day example of a 20% drop in SPY:
5050 scenario: (50$ - 50 * 0.20) + (50$ - 50 * 0.60) = 40 + 20 = 60
100 scenario: 100$ - 100 * 0.40 = 60

Two-day scenario, with another 20% drop:
5050 scenario: (40 - 40 * 0.20) + (20 - 20 * 0.60) = 32 + 8 = 40
100 scenario: 60$ - 60 * 0.40 = 36

Three-day scenario, with another 20% drop:
5050 scenario: (32 - 32 * 0.20) + (8 - 8 * 0.60) = 25.6 + 3.2 = 28.8
100 scenario: 36$ - 36 * 0.40 = 21.6

With 3 days of back-to-back 20% drops, you are still alive, but more battered if you're in team 100. Now, is this bound to happen? I don't think it's very realistic, but it really comes down to personal risk tolerance whether you go with one or the other. One has to keep in mind that the opposite is true: with 100, you will get chunkier surges. With increased assets, nominal values after future, big drawdowns might not be much different, keeping in mind past phases of outperformance for 100.
Post Reply