HEDGEFUNDIE's excellent adventure Part II: The next journey

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Fonfo
Posts: 15
Joined: Mon Aug 02, 2021 10:55 am

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

Post by Fonfo »

Is there any etf that already replies this upro/TMF strategy or one similar? Even if it is only x2 would be interesting for me.

I am asking that because I am able to apply this strategy only in a Taxable account for now.
adamhg
Posts: 218
Joined: Sat Apr 10, 2021 8:40 am

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

Post by adamhg »

Fonfo wrote: Sat Oct 16, 2021 8:31 am Is there any etf that already replies this upro/TMF strategy or one similar? Even if it is only x2 would be interesting for me.

I am asking that because I am able to apply this strategy only in a Taxable account for now.
Psldx is 2x (100/100) with active bonds but should only be used in tax advantaged

Ntsx is 1.5x (90/60) with itt and is pretty tax efficient
Booogle
Posts: 605
Joined: Wed Nov 09, 2016 2:57 pm

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

Post by Booogle »

What makes Treasury Yields go up without the Fed increasing the Fed Funds Rate?
manlymatt83
Posts: 1286
Joined: Tue Jan 30, 2018 7:23 am

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

Post by manlymatt83 »

Two quick questions:

- For those doing this as a fixed $ of their portfolio (say, $100k, let it ride), is there a way to at least return your initial capital in the first 5 - 10 years without affecting the outcome too much? For example, something like "Every time the portfolio doubles, take 10% back until you achieve your initial $100k".

- For those doing this with regular contributions (say, $2000 every month) and quarterly rebalancing, are you splitting new contributions 55/45 between UPRO/TMF and sticking to rebalancing only on rebalancing dates, or are you buying the under balanced asset with new contributions regardless of where you are in your rebalancing cycle?
Semantics
Posts: 352
Joined: Tue Mar 10, 2020 1:42 pm

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

Post by Semantics »

Booogle wrote: Sat Oct 16, 2021 9:03 am What makes Treasury Yields go up without the Fed increasing the Fed Funds Rate?
People sell bonds and the price goes down.
Booogle
Posts: 605
Joined: Wed Nov 09, 2016 2:57 pm

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

Post by Booogle »

Semantics wrote: Sat Oct 16, 2021 10:56 am
Booogle wrote: Sat Oct 16, 2021 9:03 am What makes Treasury Yields go up without the Fed increasing the Fed Funds Rate?
People sell bonds and the price goes down.

And that makes the yield go up?
Semantics
Posts: 352
Joined: Tue Mar 10, 2020 1:42 pm

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

Post by Semantics »

Booogle wrote: Sat Oct 16, 2021 12:11 pm
Semantics wrote: Sat Oct 16, 2021 10:56 am
Booogle wrote: Sat Oct 16, 2021 9:03 am What makes Treasury Yields go up without the Fed increasing the Fed Funds Rate?
People sell bonds and the price goes down.

And that makes the yield go up?
As a percentage, yes. If a bond costs $2000 and yields $20/year in interest, then that's a 1% yield. If the price of that bond were to fall to $1000, then the $20 yield is now 2%.
User avatar
Afrofreak
Posts: 223
Joined: Sat Feb 20, 2021 2:09 pm
Location: Ontario, Canada

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

Post by Afrofreak »

manlymatt83 wrote: Sat Oct 16, 2021 9:49 am Two quick questions:

- For those doing this as a fixed $ of their portfolio (say, $100k, let it ride), is there a way to at least return your initial capital in the first 5 - 10 years without affecting the outcome too much? For example, something like "Every time the portfolio doubles, take 10% back until you achieve your initial $100k".

- For those doing this with regular contributions (say, $2000 every month) and quarterly rebalancing, are you splitting new contributions 55/45 between UPRO/TMF and sticking to rebalancing only on rebalancing dates, or are you buying the under balanced asset with new contributions regardless of where you are in your rebalancing cycle?
Regular contributor here. I buy TQQQ and TMF every 2 weeks when my pay cheque comes in, in line with the current allocation, as if those contributions had been made at the last rebalance. So, if my normal split is 70/30 but it's currently sitting at 72/28, I buy 72/28. As far as I can tell, this is the only way to adhere to rebalancing quarterly with new contributions, since both of the methods you mentioned would deviate from that plan.
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

cellis212 wrote: Sat Oct 16, 2021 8:10 am
cflannagan wrote: Fri Oct 15, 2021 6:36 pm
cellis212 wrote: Fri Oct 15, 2021 8:33 am The more correct phrase would be that equities and treasuries have "negative tail dependence." When times are good they don't really move together, but when sh*t hits the fan, they go sharply in opposite directions. That's why the strategy works so much better with treasuries than something like corporate bonds or muni bonds. In normal times, they are uncorrelated with equities, but they can crash together.

That's why it's so useful to think of the non-equity component as crash insurance and then thinking about the options in terms of:

"deductible" -- how bad do things need to be for a negative correlation. High deductible means things have to get bad before the negative correlation starts. Low deductibles are better.

"coverage limit" -- how negatively correlated are they in a crash. High coverage limit means they are really negatively correlated in very bad times. HIgher coverage limit is better.

"premium" -- expected return on investment if there is no crash. Negative premium means positive expected return, zero premium means 0 expected return, positive premium means negative expected return. Lower (ideally negative) premiums are better.

"interest rate risk" -- if interest rates go up, how does that impact your insurance value (this is a common problem in life insurance)? Note that it also means that your premiums go down by more if interest rates go down (which is one reason TMF dominates in backtests). Low interest rate risk is better.

For instance:
- TMF: High deductible, high coverage limit, positive premium, highest interest rate risk.
- TYD: High deductible, medium coverage limit, slightly higher premium than TMF, medium-low interest rate risk.
- EDV: High deductible, medium coverage limit, negative premium, medium interest rate risk.
- cash: Low deductible, low coverage limit, zero premium, zero interest rate risk.
- long short-term vol: zero deductible, highest coverage limit, very high premium, zero interest rate risk.
- long medium-term vol: low deductible, high coverage limit, high premium, zero interest rate risk.
- SPY puts: zero deductible, extremely high coverage, very high premium, zero interest rate risk.
- gold: medium deductible, medium coverage limit, medium premium, negative interest rate risk.
Maybe I'm missing some context that would assign TMF, TYD, etc those "premium" ratings - but by your definitions, positive premiums = negative expected returns, while negative premiums = positive expected returns. How is it TMF, TYD, etc earned a positive premium rating (implying negative expected returns)?

Image

Seems to be positive returns over longer periods of time, so that's why I mentioned I might be missing some context that earned TMF, TYD, etc the "positive premium" label. What am I missing here?

Interest rates were mostly flat/decreasing for that period. If you hold interest rates constant, then TYD and TMF will lose money because the cost of leverage (plus fees, daily rebalancing, etc.) is higher than the interest the bonds pay out.
Completely false. 5 yr rates from June 2009 - December 2018 *increased* and TYD returned over 100%. These are not simply 'insurance policies.' They contribute independently to the return of the portfolio in a very significant way. Especially TYD (less so TMF).

https://www.portfoliovisualizer.com/bac ... ion1_1=100
HolyGrill
Posts: 52
Joined: Fri Feb 19, 2021 5:33 pm

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

Post by HolyGrill »

skierincolorado wrote: Fri Oct 15, 2021 3:10 pm
cellis212 wrote: Fri Oct 15, 2021 8:33 am
skierincolorado wrote: Thu Oct 07, 2021 12:22 pm
kbourgu wrote: Thu Oct 07, 2021 10:52 am Can anyone present any insights into the current state of TMF/TLT? Stocks seem to be recovering but looks like the bond portion is back in a downtrend.
The whole reason this works so well is that stocks and bonds are negatively correlated (would more accurately be described as uncorrelated). For the last few weeks they were positively correlated, so it's good to see the correlation reduced again. And as DMoogle said, this is a long-term strategy, short terms things like this should be completely ignored. I wouldn't even biggen to evaluate this strategy on any time horizon less than 10 years, preferably 20.
The more correct phrase would be that equities and treasuries have "negative tail dependence." When times are good they don't really move together, but when sh*t hits the fan, they go sharply in opposite directions. That's why the strategy works so much better with treasuries than something like corporate bonds or muni bonds. In normal times, they are uncorrelated with equities, but they can crash together.

That's why it's so useful to think of the non-equity component as crash insurance and then thinking about the options in terms of:

"deductible" -- how bad do things need to be for a negative correlation. High deductible means things have to get bad before the negative correlation starts. Low deductibles are better.

"coverage limit" -- how negatively correlated are they in a crash. High coverage limit means they are really negatively correlated in very bad times. HIgher coverage limit is better.

"premium" -- expected return on investment if there is no crash. Negative premium means positive expected return, zero premium means 0 expected return, positive premium means negative expected return. Lower (ideally negative) premiums are better.

"interest rate risk" -- if interest rates go up, how does that impact your insurance value (this is a common problem in life insurance)? Note that it also means that your premiums go down by more if interest rates go down (which is one reason TMF dominates in backtests). Low interest rate risk is better.

For instance:
- TMF: High deductible, high coverage limit, positive premium, highest interest rate risk.
- TYD: High deductible, medium coverage limit, slightly higher premium than TMF, medium-low interest rate risk.
- EDV: High deductible, medium coverage limit, negative premium, medium interest rate risk.
- cash: Low deductible, low coverage limit, zero premium, zero interest rate risk.
- long short-term vol: zero deductible, highest coverage limit, very high premium, zero interest rate risk.
- long medium-term vol: low deductible, high coverage limit, high premium, zero interest rate risk.
- SPY puts: zero deductible, extremely high coverage, very high premium, zero interest rate risk.
- gold: medium deductible, medium coverage limit, medium premium, negative interest rate risk.


The main questions then are:
- How comfortable am I with "non-crash" drawdowns (like 2018)?
- How worried am I about interest rates?
- How high of a premium am I willing to pay?
This is a decent analogy. But I think more of the return is actually coming from TMF as an independent source of return than many in this thread think. And even more would come from TYD. This is why substituting ITT for LTT in a backtest dramatically increases return AND reduces the max drawdown, when done in the correct ratio.

When you say TYD has a 'medium coverage limit' that's true if we buy it in the same quantity as TMF. If we buy it in larger quantities, we increase the coverage limit. And since it has medium-low interest rate risk, we have not increased our risk.

For somebody that already has 100% of their net worth in HFEA, they can't buy more TYD. But I am finding most people in this thread do not have their entire portfolio in HFEA - only a small to medium portion. Which is why I'm suggesting people allocate a larger portion of their portfolio to a 'modified HFEA' using TYD or ITT futures.
I saw HedgeFundie himself put only 10% of his money to HEFA. I would like to hear your opinion about All-in 50/50(UPRO/TYD). Thanks.
adamhg
Posts: 218
Joined: Sat Apr 10, 2021 8:40 am

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

Post by adamhg »

cflannagan wrote: Wed Oct 13, 2021 5:58 pm
skierincolorado wrote: Wed Oct 13, 2021 4:42 pm We can see this in Simba's spreadsheet. Since 1955:

55/45 UPRO/TYD has the same CAGR and lower stdev and low max-drawdown than UPRO/TMF.

The more you look at it, the more that TMF becomes indefensible. Of course this all is only relevant for people who aren't willing/able to use futures.. because futures make the choice very easy... we can get as much bond exposure as we want at the duration we want.
What kind of difference are we talking about between 45% of TYD vs TMF, when used with 55% UPRO for returns since 1955?

I ran an asset class backtest with data going back to start of 1978 - nothing about that backtest - other than slightly lower drawdown - seem particularly convincing that I should switch from TMF to TYD, so maybe something major happened between 1955 to 1978? Sortino ratio was better for long-term treasuries in this case. https://www.portfoliovisualizer.com/bac ... tion3_2=45
Fwiw, thank you for posting this. I was literally about to pull the trigger when I saw this which forced me to reevaluate my earlier decision. I do still think ITT can produce superior risk adjusted returns, but you do need to leverage past 3x to get there. The earlier AA I posted could get there at a lower annual fee than hfea and a 10% lower duration risk, but it was a marginal benefit at best without using futures.

To answer skiers earlier question, why not futures? I can automate most things, including box spreads, but rolling futures takes a broker and infrastructure I'm not willing to invest in. I'm prioritizing risk adjust returns / time so to speak.
cellis212
Posts: 7
Joined: Tue Jan 19, 2021 7:01 pm

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

Post by cellis212 »

skierincolorado wrote: Sat Oct 16, 2021 7:17 pm

Completely false. 5 yr rates from June 2009 - December 2018 *increased* and TYD returned over 100%. These are not simply 'insurance policies.' They contribute independently to the return of the portfolio in a very significant way. Especially TYD (less so TMF).

https://www.portfoliovisualizer.com/bac ... ion1_1=100
10-year rates which, unlike 5-year, are actually part of TYD were basically cut in half over that time frame...
https://fred.stlouisfed.org/series/DGS10
User avatar
cflannagan
Posts: 1208
Joined: Sun Oct 21, 2007 11:44 am
Location: Working Remotely

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

Post by cflannagan »

cellis212 wrote: Sun Oct 17, 2021 5:25 am
skierincolorado wrote: Sat Oct 16, 2021 7:17 pm

Completely false. 5 yr rates from June 2009 - December 2018 *increased* and TYD returned over 100%. These are not simply 'insurance policies.' They contribute independently to the return of the portfolio in a very significant way. Especially TYD (less so TMF).

https://www.portfoliovisualizer.com/bac ... ion1_1=100
10-year rates which, unlike 5-year, are actually part of TYD were basically cut in half over that time frame...
https://fred.stlouisfed.org/series/DGS10
I was confused looking at your link - it's not the same time frame. Did you mean this time frame? https://fred.stlouisfed.org/graph/?g=HRt7
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

HolyGrill wrote: Sat Oct 16, 2021 9:47 pm
skierincolorado wrote: Fri Oct 15, 2021 3:10 pm
cellis212 wrote: Fri Oct 15, 2021 8:33 am
skierincolorado wrote: Thu Oct 07, 2021 12:22 pm
kbourgu wrote: Thu Oct 07, 2021 10:52 am Can anyone present any insights into the current state of TMF/TLT? Stocks seem to be recovering but looks like the bond portion is back in a downtrend.
The whole reason this works so well is that stocks and bonds are negatively correlated (would more accurately be described as uncorrelated). For the last few weeks they were positively correlated, so it's good to see the correlation reduced again. And as DMoogle said, this is a long-term strategy, short terms things like this should be completely ignored. I wouldn't even biggen to evaluate this strategy on any time horizon less than 10 years, preferably 20.
The more correct phrase would be that equities and treasuries have "negative tail dependence." When times are good they don't really move together, but when sh*t hits the fan, they go sharply in opposite directions. That's why the strategy works so much better with treasuries than something like corporate bonds or muni bonds. In normal times, they are uncorrelated with equities, but they can crash together.

That's why it's so useful to think of the non-equity component as crash insurance and then thinking about the options in terms of:

"deductible" -- how bad do things need to be for a negative correlation. High deductible means things have to get bad before the negative correlation starts. Low deductibles are better.

"coverage limit" -- how negatively correlated are they in a crash. High coverage limit means they are really negatively correlated in very bad times. HIgher coverage limit is better.

"premium" -- expected return on investment if there is no crash. Negative premium means positive expected return, zero premium means 0 expected return, positive premium means negative expected return. Lower (ideally negative) premiums are better.

"interest rate risk" -- if interest rates go up, how does that impact your insurance value (this is a common problem in life insurance)? Note that it also means that your premiums go down by more if interest rates go down (which is one reason TMF dominates in backtests). Low interest rate risk is better.

For instance:
- TMF: High deductible, high coverage limit, positive premium, highest interest rate risk.
- TYD: High deductible, medium coverage limit, slightly higher premium than TMF, medium-low interest rate risk.
- EDV: High deductible, medium coverage limit, negative premium, medium interest rate risk.
- cash: Low deductible, low coverage limit, zero premium, zero interest rate risk.
- long short-term vol: zero deductible, highest coverage limit, very high premium, zero interest rate risk.
- long medium-term vol: low deductible, high coverage limit, high premium, zero interest rate risk.
- SPY puts: zero deductible, extremely high coverage, very high premium, zero interest rate risk.
- gold: medium deductible, medium coverage limit, medium premium, negative interest rate risk.


The main questions then are:
- How comfortable am I with "non-crash" drawdowns (like 2018)?
- How worried am I about interest rates?
- How high of a premium am I willing to pay?
This is a decent analogy. But I think more of the return is actually coming from TMF as an independent source of return than many in this thread think. And even more would come from TYD. This is why substituting ITT for LTT in a backtest dramatically increases return AND reduces the max drawdown, when done in the correct ratio.

When you say TYD has a 'medium coverage limit' that's true if we buy it in the same quantity as TMF. If we buy it in larger quantities, we increase the coverage limit. And since it has medium-low interest rate risk, we have not increased our risk.

For somebody that already has 100% of their net worth in HFEA, they can't buy more TYD. But I am finding most people in this thread do not have their entire portfolio in HFEA - only a small to medium portion. Which is why I'm suggesting people allocate a larger portion of their portfolio to a 'modified HFEA' using TYD or ITT futures.
I saw HedgeFundie himself put only 10% of his money to HEFA. I would like to hear your opinion about All-in 50/50(UPRO/TYD). Thanks.
100% 50/50 UPRO/TYD is significantly more risk than 10% HFEA 90% VTI. Look at the total AA:

100% 50/50 UPRO/TYD: 150/150 stock/ITT

10% HFEA 90% VTI: 116.5/13.5 stock/LTT

However, the authors of "Lifecycle Investing" make a strong argument that young investors with significant future expected earnings should be heavily leveraged, and a 50/50 UPRO/TYD AA would be totally appropriate. They proved that for all cohorts since 1900, leveraging 2x into stocks while young, with a glidepath to eliminate leverage as savings increase, significantly improved retirement savings. See the bogleheads thread on the subject:

viewtopic.php?f=10&t=274390

I would personally implement with futures in order to save on fees.

My understanding is that while Hedgefundie only did HFEA on 10%, the other 90% was leveraged in other ways as well.

If 50/50 is too much risk for your personal situation (you're not in your 20s or 30s for example), one could do 45/55, or 40/60, or do it on 70% of one's net worth instead of 100%. Just look at your AA holistically.
Last edited by skierincolorado on Sun Oct 17, 2021 10:54 am, edited 2 times in total.
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

cellis212 wrote: Sun Oct 17, 2021 5:25 am
skierincolorado wrote: Sat Oct 16, 2021 7:17 pm

Completely false. 5 yr rates from June 2009 - December 2018 *increased* and TYD returned over 100%. These are not simply 'insurance policies.' They contribute independently to the return of the portfolio in a very significant way. Especially TYD (less so TMF).

https://www.portfoliovisualizer.com/bac ... ion1_1=100
10-year rates which, unlike 5-year, are actually part of TYD were basically cut in half over that time frame...
https://fred.stlouisfed.org/series/DGS10
You are right, my mistake, but the point stands even using the 10 year:

From June 31 2010 to October 31 2018, the 10 year rose from 2.96 to 3.07. The 7 year, which TYD also owns, rose from 2.4% to 2.95%.

TYD returned 61.8%.

https://www.portfoliovisualizer.com/bac ... ion1_1=100


https://fred.stlouisfed.org/series/DGS10


If we do May 31 2010 to October 2018, the 10 year fell from 3.31 to 3.07, while the 7 year rose from 2.78 to 2.95. So one fell slightly and one rose slightly.

TYD returned 71.8%:

https://www.portfoliovisualizer.com/bac ... ion1_1=100
User avatar
cflannagan
Posts: 1208
Joined: Sun Oct 21, 2007 11:44 am
Location: Working Remotely

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

Post by cflannagan »

skierincolorado wrote: Sun Oct 17, 2021 10:02 am You are right, my mistake, but the point stands even using the 10 year:

From June 31 2010 to October 31 2018, the 10 year rose from 2.96 to 3.07. The 7 year, which TYD also owns, rose from 2.4% to 2.95%.

TYD returned 61.8%.

https://www.portfoliovisualizer.com/bac ... ion1_1=100


https://fred.stlouisfed.org/series/DGS10


If we do May 31 2010 to October 2018, the 10 year fell from 3.31 to 3.07, while the 7 year rose from 2.78 to 2.95. So one fell slightly and one rose slightly.

TYD returned 71.8%:

https://www.portfoliovisualizer.com/bac ... ion1_1=100

Direct links:

June 30 2010 to October 31 2018
7-year: https://fred.stlouisfed.org/graph/?g=HRyc
10-year: https://fred.stlouisfed.org/graph/?g=HRxQ

May 31 2010 to October 31 2018
7-year: https://fred.stlouisfed.org/graph/?g=HRyl
10-year; https://fred.stlouisfed.org/graph/?g=HRyp

Thanks, good find.
User avatar
typical.investor
Posts: 5245
Joined: Mon Jun 11, 2018 3:17 am

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

Post by typical.investor »

cflannagan wrote: Wed Oct 13, 2021 5:58 pm
skierincolorado wrote: Wed Oct 13, 2021 4:42 pm We can see this in Simba's spreadsheet. Since 1955:

55/45 UPRO/TYD has the same CAGR and lower stdev and low max-drawdown than UPRO/TMF.

The more you look at it, the more that TMF becomes indefensible. Of course this all is only relevant for people who aren't willing/able to use futures.. because futures make the choice very easy... we can get as much bond exposure as we want at the duration we want.
What kind of difference are we talking about between 45% of TYD vs TMF, when used with 55% UPRO for returns since 1955?

I ran an asset class backtest with data going back to start of 1978 - nothing about that backtest - other than slightly lower drawdown - seem particularly convincing that I should switch from TMF to TYD, so maybe something major happened between 1955 to 1978?
The pre-Volker Fed (up tp 1979) set rates very differently than how has been done since then.

Also, between 1960 and 1980 I believe there were many callable bonds which could be less effective to some degree when rates drop in an economic downturn relative to a non-callable one.
User avatar
cflannagan
Posts: 1208
Joined: Sun Oct 21, 2007 11:44 am
Location: Working Remotely

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

Post by cflannagan »

typical.investor wrote: Sun Oct 17, 2021 11:30 am The pre-Volker Fed (up tp 1979) set rates very differently than how has been done since then.

Also, between 1960 and 1980 I believe there were many callable bonds which could be less effective to some degree when rates drop in an economic downturn relative to a non-callable one.
Yup that's also my understanding. If bonds were still callable today, I am skeptical that HFEA would still look appealing. If I remember correctly, Hedgefundie showed this in the charts too. TMF returns would have been medicore/non-existent (mostly flat) for long period leading up to 1982 or so, and then "suddenly" look better post-1982. Someone pointed out it was around this time bonds were no longer callable (or something like that in that vein)
cellis212
Posts: 7
Joined: Tue Jan 19, 2021 7:01 pm

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

Post by cellis212 »

skierincolorado wrote: Sun Oct 17, 2021 10:02 am
cellis212 wrote: Sun Oct 17, 2021 5:25 am
skierincolorado wrote: Sat Oct 16, 2021 7:17 pm

Completely false. 5 yr rates from June 2009 - December 2018 *increased* and TYD returned over 100%. These are not simply 'insurance policies.' They contribute independently to the return of the portfolio in a very significant way. Especially TYD (less so TMF).

https://www.portfoliovisualizer.com/bac ... ion1_1=100
10-year rates which, unlike 5-year, are actually part of TYD were basically cut in half over that time frame...
https://fred.stlouisfed.org/series/DGS10
You are right, my mistake, but the point stands even using the 10 year:

From June 31 2010 to October 31 2018, the 10 year rose from 2.96 to 3.07. The 7 year, which TYD also owns, rose from 2.4% to 2.95%.

TYD returned 61.8%.

https://www.portfoliovisualizer.com/bac ... ion1_1=100


https://fred.stlouisfed.org/series/DGS10


If we do May 31 2010 to October 2018, the 10 year fell from 3.31 to 3.07, while the 7 year rose from 2.78 to 2.95. So one fell slightly and one rose slightly.

TYD returned 71.8%:

https://www.portfoliovisualizer.com/bac ... ion1_1=100
Looking into it more, I think that's from the daily reset impact. As discussed a bunch in this thread, if markets trend in either direction UPRO outperforms, if it is back and forth, then it underperforms. The same thing happens with the treasury funds. If interest rates trend up or down for extended periods (which they generally did during this time frame), they outperform and if rates are more volatile then they underperform.

This is still a form of interest rate risk, though much more nuanced than I originally thought. If you are willing to assume that interest rates will continue to have this trend behavior, then yeah, TYD will make you money on its own. Personally, I think that's a really good prediction and it's also a great reason to pick TYD over alternatives.
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

cellis212 wrote: Sun Oct 17, 2021 12:46 pm
skierincolorado wrote: Sun Oct 17, 2021 10:02 am
cellis212 wrote: Sun Oct 17, 2021 5:25 am
skierincolorado wrote: Sat Oct 16, 2021 7:17 pm

Completely false. 5 yr rates from June 2009 - December 2018 *increased* and TYD returned over 100%. These are not simply 'insurance policies.' They contribute independently to the return of the portfolio in a very significant way. Especially TYD (less so TMF).

https://www.portfoliovisualizer.com/bac ... ion1_1=100
10-year rates which, unlike 5-year, are actually part of TYD were basically cut in half over that time frame...
https://fred.stlouisfed.org/series/DGS10
You are right, my mistake, but the point stands even using the 10 year:

From June 31 2010 to October 31 2018, the 10 year rose from 2.96 to 3.07. The 7 year, which TYD also owns, rose from 2.4% to 2.95%.

TYD returned 61.8%.

https://www.portfoliovisualizer.com/bac ... ion1_1=100


https://fred.stlouisfed.org/series/DGS10


If we do May 31 2010 to October 2018, the 10 year fell from 3.31 to 3.07, while the 7 year rose from 2.78 to 2.95. So one fell slightly and one rose slightly.

TYD returned 71.8%:

https://www.portfoliovisualizer.com/bac ... ion1_1=100
Looking into it more, I think that's from the daily reset impact. As discussed a bunch in this thread, if markets trend in either direction UPRO outperforms, if it is back and forth, then it underperforms. The same thing happens with the treasury funds. If interest rates trend up or down for extended periods (which they generally did during this time frame), they outperform and if rates are more volatile then they underperform.

This is still a form of interest rate risk, though much more nuanced than I originally thought. If you are willing to assume that interest rates will continue to have this trend behavior, then yeah, TYD will make you money on its own. Personally, I think that's a really good prediction and it's also a great reason to pick TYD over alternatives.
It has very little to do with the daily reset. Simba's spreadsheet provides simulated returns for a 3x ITT fund since 1955, including borrowing costs. The returns are exceptional. High-volatility decay and low-vality boost can happen with any leveraged strategy (and happens with daily, quarterly, or annual leverage reset), but they are very secondary to the primary source of return - which is the large positive return of the underlying asset: ITT.

TYD owns 3x its AUM in ITT which are a constant source of return from collecting the coupon and the roll yield. Even in high interest rate volatility, these sources of return are substantial and larger than any decay from daily leverage resets. What would kill the return of TYD is the same thing that would kill the return of any unleveraged bond fund - rapidly increasing interest rates.
cellis212
Posts: 7
Joined: Tue Jan 19, 2021 7:01 pm

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

Post by cellis212 »

skierincolorado wrote: Sun Oct 17, 2021 12:55 pm
cellis212 wrote: Sun Oct 17, 2021 12:46 pm
skierincolorado wrote: Sun Oct 17, 2021 10:02 am
cellis212 wrote: Sun Oct 17, 2021 5:25 am
skierincolorado wrote: Sat Oct 16, 2021 7:17 pm

Completely false. 5 yr rates from June 2009 - December 2018 *increased* and TYD returned over 100%. These are not simply 'insurance policies.' They contribute independently to the return of the portfolio in a very significant way. Especially TYD (less so TMF).

https://www.portfoliovisualizer.com/bac ... ion1_1=100
10-year rates which, unlike 5-year, are actually part of TYD were basically cut in half over that time frame...
https://fred.stlouisfed.org/series/DGS10
You are right, my mistake, but the point stands even using the 10 year:

From June 31 2010 to October 31 2018, the 10 year rose from 2.96 to 3.07. The 7 year, which TYD also owns, rose from 2.4% to 2.95%.

TYD returned 61.8%.

https://www.portfoliovisualizer.com/bac ... ion1_1=100


https://fred.stlouisfed.org/series/DGS10


If we do May 31 2010 to October 2018, the 10 year fell from 3.31 to 3.07, while the 7 year rose from 2.78 to 2.95. So one fell slightly and one rose slightly.

TYD returned 71.8%:

https://www.portfoliovisualizer.com/bac ... ion1_1=100
Looking into it more, I think that's from the daily reset impact. As discussed a bunch in this thread, if markets trend in either direction UPRO outperforms, if it is back and forth, then it underperforms. The same thing happens with the treasury funds. If interest rates trend up or down for extended periods (which they generally did during this time frame), they outperform and if rates are more volatile then they underperform.

This is still a form of interest rate risk, though much more nuanced than I originally thought. If you are willing to assume that interest rates will continue to have this trend behavior, then yeah, TYD will make you money on its own. Personally, I think that's a really good prediction and it's also a great reason to pick TYD over alternatives.
It has very little to do with the daily reset. Simba's spreadsheet provides simulated returns for a 3x ITT fund since 1955, including borrowing costs. The returns are exceptional. High-volatility decay and low-vality boost can happen with any leveraged strategy (and happens with daily, quarterly, or annual leverage reset), but they are very secondary to the primary source of return - which is the large positive return of the underlying asset: ITT.

TYD owns 3x its AUM in ITT which are a constant source of return from collecting the coupon and the roll yield. Even in high interest rate volatility, these sources of return are substantial and larger than any decay from daily leverage resets. What would kill the return of TYD is the same thing that would kill the return of any unleveraged bond fund - rapidly increasing interest rates.

Wow yeah, didn't realize how much the roll yield brings in. My thought process for the original post in trying to abstract away from interest rate risk was ignoring the roll completely and thinking about holding the bonds to maturity. Realizing now that this was a mistake.
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

cellis212 wrote: Sun Oct 17, 2021 1:37 pm
skierincolorado wrote: Sun Oct 17, 2021 12:55 pm
cellis212 wrote: Sun Oct 17, 2021 12:46 pm
skierincolorado wrote: Sun Oct 17, 2021 10:02 am
cellis212 wrote: Sun Oct 17, 2021 5:25 am

10-year rates which, unlike 5-year, are actually part of TYD were basically cut in half over that time frame...
https://fred.stlouisfed.org/series/DGS10
You are right, my mistake, but the point stands even using the 10 year:

From June 31 2010 to October 31 2018, the 10 year rose from 2.96 to 3.07. The 7 year, which TYD also owns, rose from 2.4% to 2.95%.

TYD returned 61.8%.

https://www.portfoliovisualizer.com/bac ... ion1_1=100


https://fred.stlouisfed.org/series/DGS10


If we do May 31 2010 to October 2018, the 10 year fell from 3.31 to 3.07, while the 7 year rose from 2.78 to 2.95. So one fell slightly and one rose slightly.

TYD returned 71.8%:

https://www.portfoliovisualizer.com/bac ... ion1_1=100
Looking into it more, I think that's from the daily reset impact. As discussed a bunch in this thread, if markets trend in either direction UPRO outperforms, if it is back and forth, then it underperforms. The same thing happens with the treasury funds. If interest rates trend up or down for extended periods (which they generally did during this time frame), they outperform and if rates are more volatile then they underperform.

This is still a form of interest rate risk, though much more nuanced than I originally thought. If you are willing to assume that interest rates will continue to have this trend behavior, then yeah, TYD will make you money on its own. Personally, I think that's a really good prediction and it's also a great reason to pick TYD over alternatives.
It has very little to do with the daily reset. Simba's spreadsheet provides simulated returns for a 3x ITT fund since 1955, including borrowing costs. The returns are exceptional. High-volatility decay and low-vality boost can happen with any leveraged strategy (and happens with daily, quarterly, or annual leverage reset), but they are very secondary to the primary source of return - which is the large positive return of the underlying asset: ITT.

TYD owns 3x its AUM in ITT which are a constant source of return from collecting the coupon and the roll yield. Even in high interest rate volatility, these sources of return are substantial and larger than any decay from daily leverage resets. What would kill the return of TYD is the same thing that would kill the return of any unleveraged bond fund - rapidly increasing interest rates.

Wow yeah, didn't realize how much the roll yield brings in. My thought process for the original post in trying to abstract away from interest rate risk was ignoring the roll completely and thinking about holding the bonds to maturity. Realizing now that this was a mistake.
I mean even without roll yield, the coupon is substantial. There's not much roll-yield on EDV or TMF (rolling from 30 to 29 years is trivial), but these funds still have positive returns even during periods of flat interest rates. Even with some volatility decay they should still have positive returns. And in periods of low-volatility, they may experience low-volatility boost. You've got to remember they own positively returning assets just like UPRO. It would take a lot of volatility decay for TMF to have negative returns during a period of neutral rates, but it can happen. Alternatively, during low volatility, you would get substantially more than 3x the return of an unlevered fund like IEF.

When you own TYD or TMF you're not just exposed to the price changes of the underlying bonds, you're also collecting the coupons. Even on TYD which has a lot more roll yield than TMF, over half the historical return is from the coupon not the roll-yield.
adamhg
Posts: 218
Joined: Sat Apr 10, 2021 8:40 am

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

Post by adamhg »

cellis212 wrote: Sun Oct 17, 2021 1:37 pm
skierincolorado wrote: Sun Oct 17, 2021 12:55 pm
cellis212 wrote: Sun Oct 17, 2021 12:46 pm
skierincolorado wrote: Sun Oct 17, 2021 10:02 am
cellis212 wrote: Sun Oct 17, 2021 5:25 am

10-year rates which, unlike 5-year, are actually part of TYD were basically cut in half over that time frame...
https://fred.stlouisfed.org/series/DGS10
You are right, my mistake, but the point stands even using the 10 year:

From June 31 2010 to October 31 2018, the 10 year rose from 2.96 to 3.07. The 7 year, which TYD also owns, rose from 2.4% to 2.95%.

TYD returned 61.8%.

https://www.portfoliovisualizer.com/bac ... ion1_1=100


https://fred.stlouisfed.org/series/DGS10


If we do May 31 2010 to October 2018, the 10 year fell from 3.31 to 3.07, while the 7 year rose from 2.78 to 2.95. So one fell slightly and one rose slightly.

TYD returned 71.8%:

https://www.portfoliovisualizer.com/bac ... ion1_1=100
Looking into it more, I think that's from the daily reset impact. As discussed a bunch in this thread, if markets trend in either direction UPRO outperforms, if it is back and forth, then it underperforms. The same thing happens with the treasury funds. If interest rates trend up or down for extended periods (which they generally did during this time frame), they outperform and if rates are more volatile then they underperform.

This is still a form of interest rate risk, though much more nuanced than I originally thought. If you are willing to assume that interest rates will continue to have this trend behavior, then yeah, TYD will make you money on its own. Personally, I think that's a really good prediction and it's also a great reason to pick TYD over alternatives.
It has very little to do with the daily reset. Simba's spreadsheet provides simulated returns for a 3x ITT fund since 1955, including borrowing costs. The returns are exceptional. High-volatility decay and low-vality boost can happen with any leveraged strategy (and happens with daily, quarterly, or annual leverage reset), but they are very secondary to the primary source of return - which is the large positive return of the underlying asset: ITT.

TYD owns 3x its AUM in ITT which are a constant source of return from collecting the coupon and the roll yield. Even in high interest rate volatility, these sources of return are substantial and larger than any decay from daily leverage resets. What would kill the return of TYD is the same thing that would kill the return of any unleveraged bond fund - rapidly increasing interest rates.

Wow yeah, didn't realize how much the roll yield brings in. My thought process for the original post in trying to abstract away from interest rate risk was ignoring the roll completely and thinking about holding the bonds to maturity. Realizing now that this was a mistake.
TYD and TMF both own a sizable portion of their assets in IEF and TLT respectively. Both IEF and TLT can also lend up to 1/3 of their bonds to further juice returns over that of just the typical yields

I suspect it's one of the reasons both seem to consistently outperform peers of the same asset class. It's also why I generally try not to sim backtests with either etf unless you were planning to hold them because they aren't as pure as say the Vanguard etfs that exclude the lending language
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

adamhg wrote: Sun Oct 17, 2021 12:00 am
cflannagan wrote: Wed Oct 13, 2021 5:58 pm
skierincolorado wrote: Wed Oct 13, 2021 4:42 pm We can see this in Simba's spreadsheet. Since 1955:

55/45 UPRO/TYD has the same CAGR and lower stdev and low max-drawdown than UPRO/TMF.

The more you look at it, the more that TMF becomes indefensible. Of course this all is only relevant for people who aren't willing/able to use futures.. because futures make the choice very easy... we can get as much bond exposure as we want at the duration we want.
What kind of difference are we talking about between 45% of TYD vs TMF, when used with 55% UPRO for returns since 1955?

I ran an asset class backtest with data going back to start of 1978 - nothing about that backtest - other than slightly lower drawdown - seem particularly convincing that I should switch from TMF to TYD, so maybe something major happened between 1955 to 1978? Sortino ratio was better for long-term treasuries in this case. https://www.portfoliovisualizer.com/bac ... tion3_2=45
Fwiw, thank you for posting this. I was literally about to pull the trigger when I saw this which forced me to reevaluate my earlier decision. I do still think ITT can produce superior risk adjusted returns, but you do need to leverage past 3x to get there. The earlier AA I posted could get there at a lower annual fee than hfea and a 10% lower duration risk, but it was a marginal benefit at best without using futures.

To answer skiers earlier question, why not futures? I can automate most things, including box spreads, but rolling futures takes a broker and infrastructure I'm not willing to invest in. I'm prioritizing risk adjust returns / time so to speak.
So you already have 100% of nw in HFEA? Wow. That's some high risk tolerance!

If not just allocate a higher percentage of nw to a modified HFEA that is 45/55 UPRO/TYD or 40/60, with higher risk-adjusted returns. Forgive me if you've seen me make this point already. My post here explains the reasoning, with the backtest to prove it: viewtopic.php?p=6279416#p6279416

The other alternative, if you already have 100% of net-worth in original HFEA, is to take very slight leverage in a 45/55 or 40/60 UPRO/TYD modified HFEA. This can get you back up to your desired (very high) risk budget, with much better risk-adjusted returns. For example, 45/55 UPRO/TYD leveraged 1.1x is 148.5/181.5 stocks/ITT... which has much better risk adjusted returns than HFEA (165/135 stocks/LTT). Even a 50% drop in UPRO wouldn't get close to a margin call.

Or simply switch from 55/45 TMF to 45/55 TYD. Yeah it's less leverage, and a little less return, but it's a lot less risk. Would you rather be the blue or red here? I'd pick blue.
https://www.portfoliovisualizer.com/bac ... on4_2=-200

Here is 45/55 UPRO/TYD leveraged 1.1x vs. Original HFEA:

https://www.portfoliovisualizer.com/bac ... on4_2=-200

Or 40/60 UPRO/TYD leveraged 1.1x vs Orig HFEA

https://www.portfoliovisualizer.com/bac ... on4_2=-200
User avatar
Meaty
Posts: 854
Joined: Mon Jul 22, 2013 7:35 pm

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

Post by Meaty »

skierincolorado wrote: Sun Oct 17, 2021 5:02 pm
adamhg wrote: Sun Oct 17, 2021 12:00 am
cflannagan wrote: Wed Oct 13, 2021 5:58 pm
skierincolorado wrote: Wed Oct 13, 2021 4:42 pm We can see this in Simba's spreadsheet. Since 1955:

55/45 UPRO/TYD has the same CAGR and lower stdev and low max-drawdown than UPRO/TMF.

The more you look at it, the more that TMF becomes indefensible. Of course this all is only relevant for people who aren't willing/able to use futures.. because futures make the choice very easy... we can get as much bond exposure as we want at the duration we want.
What kind of difference are we talking about between 45% of TYD vs TMF, when used with 55% UPRO for returns since 1955?

I ran an asset class backtest with data going back to start of 1978 - nothing about that backtest - other than slightly lower drawdown - seem particularly convincing that I should switch from TMF to TYD, so maybe something major happened between 1955 to 1978? Sortino ratio was better for long-term treasuries in this case. https://www.portfoliovisualizer.com/bac ... tion3_2=45
Fwiw, thank you for posting this. I was literally about to pull the trigger when I saw this which forced me to reevaluate my earlier decision. I do still think ITT can produce superior risk adjusted returns, but you do need to leverage past 3x to get there. The earlier AA I posted could get there at a lower annual fee than hfea and a 10% lower duration risk, but it was a marginal benefit at best without using futures.

To answer skiers earlier question, why not futures? I can automate most things, including box spreads, but rolling futures takes a broker and infrastructure I'm not willing to invest in. I'm prioritizing risk adjust returns / time so to speak.
So you already have 100% of nw in HFEA? Wow. That's some high risk tolerance!

If not just allocate a higher percentage of nw to a modified HFEA that is 45/55 UPRO/TYD or 40/60, with higher risk-adjusted returns. Forgive me if you've seen me make this point already. My post here explains the reasoning, with the backtest to prove it: viewtopic.php?p=6279416#p6279416

The other alternative, if you already have 100% of net-worth in original HFEA, is to take very slight leverage in a 45/55 or 40/60 UPRO/TYD modified HFEA. This can get you back up to your desired (very high) risk budget, with much better risk-adjusted returns. For example, 45/55 UPRO/TYD leveraged 1.1x is 148.5/181.5 stocks/ITT... which has much better risk adjusted returns than HFEA (165/135 stocks/LTT). Even a 50% drop in UPRO wouldn't get close to a margin call.

Or simply switch from 55/45 TMF to 45/55 TYD. Yeah it's less leverage, and a little less return, but it's a lot less risk. Would you rather be the blue or red here? I'd pick blue.
https://www.portfoliovisualizer.com/bac ... on4_2=-200

Here is 45/55 UPRO/TYD leveraged 1.1x vs. Original HFEA:

https://www.portfoliovisualizer.com/bac ... on4_2=-200

Or 40/60 UPRO/TYD leveraged 1.1x vs Orig HFEA

https://www.portfoliovisualizer.com/bac ... on4_2=-200
This is probably a dumb question- but how is the risk much lower? HFEA has double the returns with only modest differences in sharpe ratio and max drawdown. How does the use of TYD mitigate the risk? Is it far less interest rate sensitive?
"Discipline equals Freedom" - Jocko Willink
Ramjet
Posts: 1464
Joined: Thu Feb 06, 2020 10:45 am

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

Post by Ramjet »

Deleted
Last edited by Ramjet on Mon Oct 18, 2021 10:19 am, edited 1 time in total.
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

Meaty wrote: Mon Oct 18, 2021 7:55 am
skierincolorado wrote: Sun Oct 17, 2021 5:02 pm
adamhg wrote: Sun Oct 17, 2021 12:00 am
cflannagan wrote: Wed Oct 13, 2021 5:58 pm
skierincolorado wrote: Wed Oct 13, 2021 4:42 pm We can see this in Simba's spreadsheet. Since 1955:

55/45 UPRO/TYD has the same CAGR and lower stdev and low max-drawdown than UPRO/TMF.

The more you look at it, the more that TMF becomes indefensible. Of course this all is only relevant for people who aren't willing/able to use futures.. because futures make the choice very easy... we can get as much bond exposure as we want at the duration we want.
What kind of difference are we talking about between 45% of TYD vs TMF, when used with 55% UPRO for returns since 1955?

I ran an asset class backtest with data going back to start of 1978 - nothing about that backtest - other than slightly lower drawdown - seem particularly convincing that I should switch from TMF to TYD, so maybe something major happened between 1955 to 1978? Sortino ratio was better for long-term treasuries in this case. https://www.portfoliovisualizer.com/bac ... tion3_2=45
Fwiw, thank you for posting this. I was literally about to pull the trigger when I saw this which forced me to reevaluate my earlier decision. I do still think ITT can produce superior risk adjusted returns, but you do need to leverage past 3x to get there. The earlier AA I posted could get there at a lower annual fee than hfea and a 10% lower duration risk, but it was a marginal benefit at best without using futures.

To answer skiers earlier question, why not futures? I can automate most things, including box spreads, but rolling futures takes a broker and infrastructure I'm not willing to invest in. I'm prioritizing risk adjust returns / time so to speak.
So you already have 100% of nw in HFEA? Wow. That's some high risk tolerance!

If not just allocate a higher percentage of nw to a modified HFEA that is 45/55 UPRO/TYD or 40/60, with higher risk-adjusted returns. Forgive me if you've seen me make this point already. My post here explains the reasoning, with the backtest to prove it: viewtopic.php?p=6279416#p6279416

The other alternative, if you already have 100% of net-worth in original HFEA, is to take very slight leverage in a 45/55 or 40/60 UPRO/TYD modified HFEA. This can get you back up to your desired (very high) risk budget, with much better risk-adjusted returns. For example, 45/55 UPRO/TYD leveraged 1.1x is 148.5/181.5 stocks/ITT... which has much better risk adjusted returns than HFEA (165/135 stocks/LTT). Even a 50% drop in UPRO wouldn't get close to a margin call.

Or simply switch from 55/45 TMF to 45/55 TYD. Yeah it's less leverage, and a little less return, but it's a lot less risk. Would you rather be the blue or red here? I'd pick blue.
https://www.portfoliovisualizer.com/bac ... on4_2=-200

Here is 45/55 UPRO/TYD leveraged 1.1x vs. Original HFEA:

https://www.portfoliovisualizer.com/bac ... on4_2=-200

Or 40/60 UPRO/TYD leveraged 1.1x vs Orig HFEA

https://www.portfoliovisualizer.com/bac ... on4_2=-200
This is probably a dumb question- but how is the risk much lower? HFEA has double the returns with only modest differences in sharpe ratio and max drawdown. How does the use of TYD mitigate the risk? Is it far less interest rate sensitive?
The max draw is under 54% instead of over 65%. This is because it owns less stocks. So it’s less risky because it’s less stock exposure, but it makes up for some of the lost return because tyd has higher return than tmf per dollar. I think the backrest shows it well .. I’d rather be the blue line
adamhg
Posts: 218
Joined: Sat Apr 10, 2021 8:40 am

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

Post by adamhg »

skierincolorado wrote: Mon Oct 18, 2021 8:55 am
Meaty wrote: Mon Oct 18, 2021 7:55 am
skierincolorado wrote: Sun Oct 17, 2021 5:02 pm
adamhg wrote: Sun Oct 17, 2021 12:00 am
cflannagan wrote: Wed Oct 13, 2021 5:58 pm

What kind of difference are we talking about between 45% of TYD vs TMF, when used with 55% UPRO for returns since 1955?

I ran an asset class backtest with data going back to start of 1978 - nothing about that backtest - other than slightly lower drawdown - seem particularly convincing that I should switch from TMF to TYD, so maybe something major happened between 1955 to 1978? Sortino ratio was better for long-term treasuries in this case. https://www.portfoliovisualizer.com/bac ... tion3_2=45
Fwiw, thank you for posting this. I was literally about to pull the trigger when I saw this which forced me to reevaluate my earlier decision. I do still think ITT can produce superior risk adjusted returns, but you do need to leverage past 3x to get there. The earlier AA I posted could get there at a lower annual fee than hfea and a 10% lower duration risk, but it was a marginal benefit at best without using futures.

To answer skiers earlier question, why not futures? I can automate most things, including box spreads, but rolling futures takes a broker and infrastructure I'm not willing to invest in. I'm prioritizing risk adjust returns / time so to speak.
So you already have 100% of nw in HFEA? Wow. That's some high risk tolerance!

If not just allocate a higher percentage of nw to a modified HFEA that is 45/55 UPRO/TYD or 40/60, with higher risk-adjusted returns. Forgive me if you've seen me make this point already. My post here explains the reasoning, with the backtest to prove it: viewtopic.php?p=6279416#p6279416

The other alternative, if you already have 100% of net-worth in original HFEA, is to take very slight leverage in a 45/55 or 40/60 UPRO/TYD modified HFEA. This can get you back up to your desired (very high) risk budget, with much better risk-adjusted returns. For example, 45/55 UPRO/TYD leveraged 1.1x is 148.5/181.5 stocks/ITT... which has much better risk adjusted returns than HFEA (165/135 stocks/LTT). Even a 50% drop in UPRO wouldn't get close to a margin call.

Or simply switch from 55/45 TMF to 45/55 TYD. Yeah it's less leverage, and a little less return, but it's a lot less risk. Would you rather be the blue or red here? I'd pick blue.
https://www.portfoliovisualizer.com/bac ... on4_2=-200

Here is 45/55 UPRO/TYD leveraged 1.1x vs. Original HFEA:

https://www.portfoliovisualizer.com/bac ... on4_2=-200

Or 40/60 UPRO/TYD leveraged 1.1x vs Orig HFEA

https://www.portfoliovisualizer.com/bac ... on4_2=-200
This is probably a dumb question- but how is the risk much lower? HFEA has double the returns with only modest differences in sharpe ratio and max drawdown. How does the use of TYD mitigate the risk? Is it far less interest rate sensitive?
The max draw is under 54% instead of over 65%. This is because it owns less stocks. So it’s less risky because it’s less stock exposure, but it makes up for some of the lost return because tyd has higher return than tmf per dollar. I think the backrest shows it well .. I’d rather be the blue line
I appreciate your replies and generally agree with your sentiment, but I think fundamentally you're arguing for a different thing than what HFEA was originally intended to be used as even though they are structured similarly. HFEA, to me at least, is a risky bet on long term trends for both stocks and bonds, but its risky nonetheless.

Correct me if I'm wrong but your version attempts to de-risk HFEA by de-levering the the equity and both de-levering and shortening the bond duration. And your argument is that if you want to YOLO your entire account on this strategy, your version is superior. I agree, but that wasn't the point of HFEA for me at least.

This was never going to be my full account. I have unlevered passive investments, I have other levered passive investments, I have an active bot trading LETFs, we have a portion of our NW in Euros and whatever the banks over there are doing with it, we have real estate, etc etc. Our world isn't clean enough to split it out into a x:y equity/bond allocations unfortunately.

So for me, HFEA was a pile of money that I could go more aggressive on with a backstop of a much larger "safe" portion of assets to fall back on if we needed. But to get the same or comparable level of risk that I need (1:3 for a Kelly portfolio), I'd have to lever it > 3x which means either box spreads or futures:

https://www.portfoliovisualizer.com/bac ... on4_3=-560

If I had the patience / discipline to do maintain this, I'd go with "equity adjusted" version of the above ITT portfolio as a true HFEA replacement. But even the "return adjusted" version still is too levered to manage w/o futures which I (personally) cannot trade automatically. So now I've been trying to at least diversify away from LTT a little to ITT without sacrificing returns and I think Mid Caps might be the answer:

https://www.portfoliovisualizer.com/bac ... tion5_1=90

I still need to check the Kelly/CVaR/CDaR portfolios with this mix, but I'm thinking it will be something along those lines in an M1 account and set and forget for 20+ years.
Ramjet
Posts: 1464
Joined: Thu Feb 06, 2020 10:45 am

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

Post by Ramjet »

adamhg wrote: Mon Oct 18, 2021 10:09 am
skierincolorado wrote: Mon Oct 18, 2021 8:55 am
Meaty wrote: Mon Oct 18, 2021 7:55 am
skierincolorado wrote: Sun Oct 17, 2021 5:02 pm
adamhg wrote: Sun Oct 17, 2021 12:00 am

Fwiw, thank you for posting this. I was literally about to pull the trigger when I saw this which forced me to reevaluate my earlier decision. I do still think ITT can produce superior risk adjusted returns, but you do need to leverage past 3x to get there. The earlier AA I posted could get there at a lower annual fee than hfea and a 10% lower duration risk, but it was a marginal benefit at best without using futures.

To answer skiers earlier question, why not futures? I can automate most things, including box spreads, but rolling futures takes a broker and infrastructure I'm not willing to invest in. I'm prioritizing risk adjust returns / time so to speak.
So you already have 100% of nw in HFEA? Wow. That's some high risk tolerance!

If not just allocate a higher percentage of nw to a modified HFEA that is 45/55 UPRO/TYD or 40/60, with higher risk-adjusted returns. Forgive me if you've seen me make this point already. My post here explains the reasoning, with the backtest to prove it: viewtopic.php?p=6279416#p6279416

The other alternative, if you already have 100% of net-worth in original HFEA, is to take very slight leverage in a 45/55 or 40/60 UPRO/TYD modified HFEA. This can get you back up to your desired (very high) risk budget, with much better risk-adjusted returns. For example, 45/55 UPRO/TYD leveraged 1.1x is 148.5/181.5 stocks/ITT... which has much better risk adjusted returns than HFEA (165/135 stocks/LTT). Even a 50% drop in UPRO wouldn't get close to a margin call.

Or simply switch from 55/45 TMF to 45/55 TYD. Yeah it's less leverage, and a little less return, but it's a lot less risk. Would you rather be the blue or red here? I'd pick blue.
https://www.portfoliovisualizer.com/bac ... on4_2=-200

Here is 45/55 UPRO/TYD leveraged 1.1x vs. Original HFEA:

https://www.portfoliovisualizer.com/bac ... on4_2=-200

Or 40/60 UPRO/TYD leveraged 1.1x vs Orig HFEA

https://www.portfoliovisualizer.com/bac ... on4_2=-200
This is probably a dumb question- but how is the risk much lower? HFEA has double the returns with only modest differences in sharpe ratio and max drawdown. How does the use of TYD mitigate the risk? Is it far less interest rate sensitive?
The max draw is under 54% instead of over 65%. This is because it owns less stocks. So it’s less risky because it’s less stock exposure, but it makes up for some of the lost return because tyd has higher return than tmf per dollar. I think the backrest shows it well .. I’d rather be the blue line
I appreciate your replies and generally agree with your sentiment, but I think fundamentally you're arguing for a different thing than what HFEA was originally intended to be used as even though they are structured similarly. HFEA, to me at least, is a risky bet on long term trends for both stocks and bonds, but its risky nonetheless.

Correct me if I'm wrong but your version attempts to de-risk HFEA by de-levering the the equity and both de-levering and shortening the bond duration. And your argument is that if you want to YOLO your entire account on this strategy, your version is superior. I agree, but that wasn't the point of HFEA for me at least.

This was never going to be my full account. I have unlevered passive investments, I have other levered passive investments, I have an active bot trading LETFs, we have a portion of our NW in Euros and whatever the banks over there are doing with it, we have real estate, etc etc. Our world isn't clean enough to split it out into a x:y equity/bond allocations unfortunately.

So for me, HFEA was a pile of money that I could go more aggressive on with a backstop of a much larger "safe" portion of assets to fall back on if we needed. But to get the same or comparable level of risk that I need (1:3 for a Kelly portfolio), I'd have to lever it > 3x which means either box spreads or futures:

https://www.portfoliovisualizer.com/bac ... on4_3=-560

If I had the patience / discipline to do maintain this, I'd go with "equity adjusted" version of the above ITT portfolio as a true HFEA replacement. But even the "return adjusted" version still is too levered to manage w/o futures which I (personally) cannot trade automatically. So now I've been trying to at least diversify away from LTT a little to ITT without sacrificing returns and I think Mid Caps might be the answer:

https://www.portfoliovisualizer.com/bac ... tion5_1=90

I still need to check the Kelly/CVaR/CDaR portfolios with this mix, but I'm thinking it will be something along those lines in an M1 account and set and forget for 20+ years.
+ 1 million
cellis212
Posts: 7
Joined: Tue Jan 19, 2021 7:01 pm

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

Post by cellis212 »

skierincolorado wrote: Sun Oct 17, 2021 5:02 pm
adamhg wrote: Sun Oct 17, 2021 12:00 am
cflannagan wrote: Wed Oct 13, 2021 5:58 pm
skierincolorado wrote: Wed Oct 13, 2021 4:42 pm We can see this in Simba's spreadsheet. Since 1955:

55/45 UPRO/TYD has the same CAGR and lower stdev and low max-drawdown than UPRO/TMF.

The more you look at it, the more that TMF becomes indefensible. Of course this all is only relevant for people who aren't willing/able to use futures.. because futures make the choice very easy... we can get as much bond exposure as we want at the duration we want.
What kind of difference are we talking about between 45% of TYD vs TMF, when used with 55% UPRO for returns since 1955?

I ran an asset class backtest with data going back to start of 1978 - nothing about that backtest - other than slightly lower drawdown - seem particularly convincing that I should switch from TMF to TYD, so maybe something major happened between 1955 to 1978? Sortino ratio was better for long-term treasuries in this case. https://www.portfoliovisualizer.com/bac ... tion3_2=45
Fwiw, thank you for posting this. I was literally about to pull the trigger when I saw this which forced me to reevaluate my earlier decision. I do still think ITT can produce superior risk adjusted returns, but you do need to leverage past 3x to get there. The earlier AA I posted could get there at a lower annual fee than hfea and a 10% lower duration risk, but it was a marginal benefit at best without using futures.

To answer skiers earlier question, why not futures? I can automate most things, including box spreads, but rolling futures takes a broker and infrastructure I'm not willing to invest in. I'm prioritizing risk adjust returns / time so to speak.
So you already have 100% of nw in HFEA? Wow. That's some high risk tolerance!

If not just allocate a higher percentage of nw to a modified HFEA that is 45/55 UPRO/TYD or 40/60, with higher risk-adjusted returns. Forgive me if you've seen me make this point already. My post here explains the reasoning, with the backtest to prove it: viewtopic.php?p=6279416#p6279416

The other alternative, if you already have 100% of net-worth in original HFEA, is to take very slight leverage in a 45/55 or 40/60 UPRO/TYD modified HFEA. This can get you back up to your desired (very high) risk budget, with much better risk-adjusted returns. For example, 45/55 UPRO/TYD leveraged 1.1x is 148.5/181.5 stocks/ITT... which has much better risk adjusted returns than HFEA (165/135 stocks/LTT). Even a 50% drop in UPRO wouldn't get close to a margin call.

Or simply switch from 55/45 TMF to 45/55 TYD. Yeah it's less leverage, and a little less return, but it's a lot less risk. Would you rather be the blue or red here? I'd pick blue.
https://www.portfoliovisualizer.com/bac ... on4_2=-200

Here is 45/55 UPRO/TYD leveraged 1.1x vs. Original HFEA:

https://www.portfoliovisualizer.com/bac ... on4_2=-200

Or 40/60 UPRO/TYD leveraged 1.1x vs Orig HFEA

https://www.portfoliovisualizer.com/bac ... on4_2=-200
Would LEAPs on IEF be a possible alternative to ITT futures if they aren't accessible? You won't get the dividend, but that's largely baked into the option price.
adamhg
Posts: 218
Joined: Sat Apr 10, 2021 8:40 am

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

Post by adamhg »

cellis212 wrote: Mon Oct 18, 2021 10:21 am Would LEAPs on IEF be a possible alternative to ITT futures if they aren't accessible? You won't get the dividend, but that's largely baked into the option price.
Unfortunately there's no market there. Take a look at the volume and spread, and even if there was, you'll be introducing IV to the equation which is a completely different animal
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

adamhg wrote: Mon Oct 18, 2021 10:09 am
skierincolorado wrote: Mon Oct 18, 2021 8:55 am
Meaty wrote: Mon Oct 18, 2021 7:55 am
skierincolorado wrote: Sun Oct 17, 2021 5:02 pm
adamhg wrote: Sun Oct 17, 2021 12:00 am

Fwiw, thank you for posting this. I was literally about to pull the trigger when I saw this which forced me to reevaluate my earlier decision. I do still think ITT can produce superior risk adjusted returns, but you do need to leverage past 3x to get there. The earlier AA I posted could get there at a lower annual fee than hfea and a 10% lower duration risk, but it was a marginal benefit at best without using futures.

To answer skiers earlier question, why not futures? I can automate most things, including box spreads, but rolling futures takes a broker and infrastructure I'm not willing to invest in. I'm prioritizing risk adjust returns / time so to speak.
So you already have 100% of nw in HFEA? Wow. That's some high risk tolerance!

If not just allocate a higher percentage of nw to a modified HFEA that is 45/55 UPRO/TYD or 40/60, with higher risk-adjusted returns. Forgive me if you've seen me make this point already. My post here explains the reasoning, with the backtest to prove it: viewtopic.php?p=6279416#p6279416

The other alternative, if you already have 100% of net-worth in original HFEA, is to take very slight leverage in a 45/55 or 40/60 UPRO/TYD modified HFEA. This can get you back up to your desired (very high) risk budget, with much better risk-adjusted returns. For example, 45/55 UPRO/TYD leveraged 1.1x is 148.5/181.5 stocks/ITT... which has much better risk adjusted returns than HFEA (165/135 stocks/LTT). Even a 50% drop in UPRO wouldn't get close to a margin call.

Or simply switch from 55/45 TMF to 45/55 TYD. Yeah it's less leverage, and a little less return, but it's a lot less risk. Would you rather be the blue or red here? I'd pick blue.
https://www.portfoliovisualizer.com/bac ... on4_2=-200

Here is 45/55 UPRO/TYD leveraged 1.1x vs. Original HFEA:

https://www.portfoliovisualizer.com/bac ... on4_2=-200

Or 40/60 UPRO/TYD leveraged 1.1x vs Orig HFEA

https://www.portfoliovisualizer.com/bac ... on4_2=-200
This is probably a dumb question- but how is the risk much lower? HFEA has double the returns with only modest differences in sharpe ratio and max drawdown. How does the use of TYD mitigate the risk? Is it far less interest rate sensitive?
The max draw is under 54% instead of over 65%. This is because it owns less stocks. So it’s less risky because it’s less stock exposure, but it makes up for some of the lost return because tyd has higher return than tmf per dollar. I think the backrest shows it well .. I’d rather be the blue line
I appreciate your replies and generally agree with your sentiment, but I think fundamentally you're arguing for a different thing than what HFEA was originally intended to be used as even though they are structured similarly. HFEA, to me at least, is a risky bet on long term trends for both stocks and bonds, but its risky nonetheless.

Correct me if I'm wrong but your version attempts to de-risk HFEA by de-levering the the equity and both de-levering and shortening the bond duration. And your argument is that if you want to YOLO your entire account on this strategy, your version is superior. I agree, but that wasn't the point of HFEA for me at least.

This was never going to be my full account. I have unlevered passive investments, I have other levered passive investments, I have an active bot trading LETFs, we have a portion of our NW in Euros and whatever the banks over there are doing with it, we have real estate, etc etc. Our world isn't clean enough to split it out into a x:y equity/bond allocations unfortunately.

So for me, HFEA was a pile of money that I could go more aggressive on with a backstop of a much larger "safe" portion of assets to fall back on if we needed. But to get the same or comparable level of risk that I need (1:3 for a Kelly portfolio), I'd have to lever it > 3x which means either box spreads or futures:

https://www.portfoliovisualizer.com/bac ... on4_3=-560

If I had the patience / discipline to do maintain this, I'd go with "equity adjusted" version of the above ITT portfolio as a true HFEA replacement. But even the "return adjusted" version still is too levered to manage w/o futures which I (personally) cannot trade automatically. So now I've been trying to at least diversify away from LTT a little to ITT without sacrificing returns and I think Mid Caps might be the answer:

https://www.portfoliovisualizer.com/bac ... tion5_1=90

I still need to check the Kelly/CVaR/CDaR portfolios with this mix, but I'm thinking it will be something along those lines in an M1 account and set and forget for 20+ years.
As long as it is recognized this is based on an emotional/psychological segregation of risk rather than a rational holistic view of risk, return and goals and the most efficient implementation of them.

It is simply more efficient to allocate a larger portion of one's nw to a slightly less risky and much more efficient strategy. The risk/return profile for one's entire net worth is improved.

I think I may have linked to the incorrect post before. I will copy and paste the explanation here:

Instead of doing 10% of net worth in SPY plus 10% of net worth in HFEA which is 165/135 stock/LTT, do 20% of one's net worth in 135/165 stock/ITT. The latter portfolio has higher risk adjusted returns.

If we look at the 10% in HFEA and the other 10% that we are moving from (assuming) SPY to modified HFEA:

Originally 10% HFEA + 10% SPY = 132.5/67.5 stock/LTT for that 20% of your portfolio (ignoring the other 80%)

New 20% in modified HFEA = 135/165 stock/ITT (again ignoring the other 80% of your portfolio)

Let's run the PV for that 20%. The red line is 20% 'modified' HFEA. The blue line is 10% HFEA + 10% SPY. The red line has much higher CAGR and a lower max-drawdown.

https://www.portfoliovisualizer.com/bac ... on4_2=-200



Since you are saying you do not have 100% of net worth in HFEA, the rational thing to do would be to move some funds from SPY (or comparable investment) and combine into a modified HFEA. The need to take extreme risk in the HFEA account is only based on arbitrary buckets that are created psychologically. You have one bucket (10% of net worth) in HFEA. Presumably you have another bucket that is 10% (or more) of net worth in SPY (or comparable investment). If we instead draw the bucket around both of these buckets and combine them, everything about the combined bucket is improved. The max drawdown is less. The expected return is higher. And if you want to psychologically break the bucket in half, you can still do that and consider you have 10% of net worth in SPY, and 10% in extremely leveraged modified HFEA.


The other alternative, for people that insist on drawing arbitrary psychological distinctions, is to lever a modified HFEA 1.1x. This gets you right back into the YOLO territory of original HFEA, but with much less risk. Nearly the same CAGR, 51% drawdown vs 65%. Why, other than the slight complexity of leveraging UPRO/TYD 1.1x, would one prefer an allocation with a 65% max-drawdown over one with a 51% max-drawdown, given nearly identical CAGR?

https://www.portfoliovisualizer.com/bac ... on4_2=-200
Last edited by skierincolorado on Mon Oct 18, 2021 12:36 pm, edited 4 times in total.
Booogle
Posts: 605
Joined: Wed Nov 09, 2016 2:57 pm

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

Post by Booogle »

If equites go up and bond prices go down, how do any of these leveraged strategies make money?

It doesn't balance out?
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

Booogle wrote: Mon Oct 18, 2021 11:46 am If equites go up and bond prices go down, how do any of these leveraged strategies make money?

It doesn't balance out?
How does any balanced portfolio, like the traditional 60/40 retirement portfolio, make money? These strategies are just leveraged versions of those.

The main source of the return is that both assets go UP. The fact that they are negatively correlated within that upwards trend is just a (relatively minor) added benefit.
Last edited by skierincolorado on Mon Oct 18, 2021 11:49 am, edited 1 time in total.
Booogle
Posts: 605
Joined: Wed Nov 09, 2016 2:57 pm

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

Post by Booogle »

skierincolorado wrote: Mon Oct 18, 2021 11:47 am The fact that they are negatively correlated within that upwards trend is just a (relatively minor) added benefit.
Why do you want your bond prices to crash when equites go up?
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

Booogle wrote: Mon Oct 18, 2021 11:51 am
skierincolorado wrote: Mon Oct 18, 2021 11:47 am The fact that they are negatively correlated within that upwards trend is just a (relatively minor) added benefit.
Why do you want your bond prices to crash when equites go up?
I don't. But knowing bonds and stocks will both have crashes, given a choice, I would rather they occur at different times.

I'm actually more on the side of thinking this is a minor benefit. I think many in this thread overplay that benefit. But it is a benefit because it can help reduce volatility drag slightly when leveraged, and because there is a financial and psychological limit to how large of a drawdown one can tolerate.


I suppose you would prefer just 100% stocks? But I can add bonds on top of those stocks, and not increase my risk very much at all, because it's unlikely they will crash at the same time. And those bonds I've added on top have substantial positive expected return.
DMoogle
Posts: 549
Joined: Sat Oct 31, 2020 10:24 pm

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

Post by DMoogle »

Booogle wrote: Mon Oct 18, 2021 11:51 am
skierincolorado wrote: Mon Oct 18, 2021 11:47 am The fact that they are negatively correlated within that upwards trend is just a (relatively minor) added benefit.
Why do you want your bond prices to crash when equites go up?
This pops up all the time because WAY too many people mistakenly think that being negatively correlated means that one must have a positive return while the other has a negative return. That is NOT what negative correlation is; correlation should be thought of about each asset's mean return, not 0.

Example: consider two assets "A" and "B". They are perfectly negatively correlated (correlation = -1). A's expected return is 7%. What is B's expected return? If you answered -7%, you're wrong. Expected return has absolutely NOTHING to do with correlation. In this hypothetical example, I'm giving B an expected return of 5%. So if A exceeds its expected return by 1% and returns 8%, does that mean B's return will be 4%? No, because absolute volatility also has nothing to do with correlation. Plug these numbers in Excel and see what I mean:

Asset A returns over 5 years:
6%
5%
7%
9%
8%

Asset B:
5.5%
6%
5%
4%
4.5%

Correlation is perfect -1. Despite that, the expected return is positive for both assets, and volatility is independent. This concept applies to stocks and bonds; just because stocks are going up doesn't mean bonds are going down. However, abnormally high returns for one will typically result in the other asset returning abnormally low (but not necessarily negative) returns.

It's also worth mentioning that stocks and bonds are, unfortunately, not perfectly negatively correlated. If they were, we could construct a risk-free portfolio, leverage infinitely, and rake infinite profit.
User avatar
cflannagan
Posts: 1208
Joined: Sun Oct 21, 2007 11:44 am
Location: Working Remotely

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

Post by cflannagan »

DMoogle wrote: Mon Oct 18, 2021 12:49 pm This pops up all the time because WAY too many people mistakenly think that being negatively correlated means that one must have a positive return while the other has a negative return. That is NOT what negative correlation is; correlation should be thought of about each asset's mean return, not 0.

Example: consider two assets "A" and "B". They are perfectly negatively correlated (correlation = -1). A's expected return is 7%. What is B's expected return? If you answered -7%, you're wrong. Expected return has absolutely NOTHING to do with correlation. In this hypothetical example, I'm giving B an expected return of 5%. So if A exceeds its expected return by 1% and returns 8%, does that mean B's return will be 4%? No, because absolute volatility also has nothing to do with correlation. Plug these numbers in Excel and see what I mean:

Asset A returns over 5 years:
6%
5%
7%
9%
8%

Asset B:
5.5%
6%
5%
4%
4.5%

Correlation is perfect -1. Despite that, the expected return is positive for both assets, and volatility is independent. This concept applies to stocks and bonds; just because stocks are going up doesn't mean bonds are going down. However, abnormally high returns for one will typically result in the other asset returning abnormally low (but not necessarily negative) returns.

It's also worth mentioning that stocks and bonds are, unfortunately, not perfectly negatively correlated. If they were, we could construct a risk-free portfolio, leverage infinitely, and rake infinite profit.
Yup. In additional to having negative correlation between UPRO and TMF (which is desired in this case as you pointed out), whenever UPRO and TMF goes in same direction together, it's going up together most of the time.
NMBob
Posts: 461
Joined: Thu Apr 23, 2015 8:13 pm

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

Post by NMBob »

wouldnt the exact opposite correlation total to zero. isnt the exact opposite the short of upro. the stock market goes up over time. the exact opposite would be something that goes down over time wouldnt it, which might not be appealing.
DMoogle
Posts: 549
Joined: Sat Oct 31, 2020 10:24 pm

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

Post by DMoogle »

NMBob wrote: Mon Oct 18, 2021 1:40 pm wouldnt the exact opposite correlation total to zero. isnt the exact opposite the short of upro. the stock market goes up over time. the exact opposite would be something that goes down over time wouldnt it, which might not be appealing.
You are correct that being short UPRO (or just long SPXU) does have -1 correlation with long UPRO, but it also has the inverse expected return. HOWEVER, just because two assets are negatively correlated does not necessarily mean that they must have the inverse expected return. It just so happens that SPXU does have a negative expected return. This concept is critical.

Study my example in my previous post closely. Two assets can be negatively correlated, yet still have a total positive expected return. Stocks and bonds are an excellent real-life example - they are negatively correlated (albeit not perfect -1) with each other, yet still both have positive expected return.
adamhg
Posts: 218
Joined: Sat Apr 10, 2021 8:40 am

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

Post by adamhg »

Booogle wrote: Mon Oct 18, 2021 11:46 am If equites go up and bond prices go down, how do any of these leveraged strategies make money?

It doesn't balance out?
The nice thing about diversification across uncorrelated assets is not only does it lower your risk but it also improves your risk adjusted returns, but can _also_ improve your nominal returns as well which is easier to conceptualize.

Here's an example comparing 100% stock, 100% bond and 75/25 stock/bond:
https://www.portfoliovisualizer.com/bac ... ol10=CASHX

Notice the 75/25 portfolio is actually outperforming the 100% stock portfolio at the end of the day. Rebalancing is the key.

Now, take a very strong risk adjusted portfolio 55/45. Notice the CAGR goes down from 11% to 10%, but stdev and max drawdown are nearly halved. Now apply 2x leverage so that your drawdown and stdev are back about what they were before, and you'll have gotten much greater returns for essentially the same "risk":
https://www.portfoliovisualizer.com/bac ... n10_3=-100

Lever it 3x and you'll find why this thread is so popular:
https://www.portfoliovisualizer.com/bac ... n10_3=-200
Last edited by adamhg on Mon Oct 18, 2021 4:39 pm, edited 5 times in total.
jarjarM
Posts: 2502
Joined: Mon Jul 16, 2018 1:21 pm

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

Post by jarjarM »

adamhg wrote: Mon Oct 18, 2021 4:29 pm
Booogle wrote: Mon Oct 18, 2021 11:46 am If equites go up and bond prices go down, how do any of these leveraged strategies make money?

It doesn't balance out?
The nice thing about diversification across uncorrelated assets is not only does it lower your risk but it also improves your risk adjusted returns, but can _also_ improve your nominal returns as well which is easier to conceptualize.

Here's an example comparing 100% stock, 100% bond and 75/25 stock/bond:
https://www.portfoliovisualizer.com/bac ... ol10=CASHX

Notice the 75/25 portfolio is actually outperforming the 100% stock portfolio at the end of the day. Rebalancing is the key.

Now, take a very strong risk adjusted portfolio 55/45. Notice the CAGR goes down from 11% to 10%, but stdev and max drawdown are nearly halved. Lever it 3x and you'll find why this thread is so popular:

https://www.portfoliovisualizer.com/bac ... n10_3=-200
+1, performance for these leveraged strategies are quite sensitive to rebalance timing/strategy due to the high volatility of the underlying components. Hence a major portion of this thread is focused on that.
darkcam
Posts: 5
Joined: Sun Feb 07, 2021 6:06 am

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

Post by darkcam »

skierincolorado wrote: Mon Oct 18, 2021 11:33 am
adamhg wrote: Mon Oct 18, 2021 10:09 am
skierincolorado wrote: Mon Oct 18, 2021 8:55 am
Meaty wrote: Mon Oct 18, 2021 7:55 am
skierincolorado wrote: Sun Oct 17, 2021 5:02 pm

So you already have 100% of nw in HFEA? Wow. That's some high risk tolerance!

If not just allocate a higher percentage of nw to a modified HFEA that is 45/55 UPRO/TYD or 40/60, with higher risk-adjusted returns. Forgive me if you've seen me make this point already. My post here explains the reasoning, with the backtest to prove it: viewtopic.php?p=6279416#p6279416

The other alternative, if you already have 100% of net-worth in original HFEA, is to take very slight leverage in a 45/55 or 40/60 UPRO/TYD modified HFEA. This can get you back up to your desired (very high) risk budget, with much better risk-adjusted returns. For example, 45/55 UPRO/TYD leveraged 1.1x is 148.5/181.5 stocks/ITT... which has much better risk adjusted returns than HFEA (165/135 stocks/LTT). Even a 50% drop in UPRO wouldn't get close to a margin call.

Or simply switch from 55/45 TMF to 45/55 TYD. Yeah it's less leverage, and a little less return, but it's a lot less risk. Would you rather be the blue or red here? I'd pick blue.
https://www.portfoliovisualizer.com/bac ... on4_2=-200

Here is 45/55 UPRO/TYD leveraged 1.1x vs. Original HFEA:

https://www.portfoliovisualizer.com/bac ... on4_2=-200

Or 40/60 UPRO/TYD leveraged 1.1x vs Orig HFEA

https://www.portfoliovisualizer.com/bac ... on4_2=-200
This is probably a dumb question- but how is the risk much lower? HFEA has double the returns with only modest differences in sharpe ratio and max drawdown. How does the use of TYD mitigate the risk? Is it far less interest rate sensitive?
The max draw is under 54% instead of over 65%. This is because it owns less stocks. So it’s less risky because it’s less stock exposure, but it makes up for some of the lost return because tyd has higher return than tmf per dollar. I think the backrest shows it well .. I’d rather be the blue line
I appreciate your replies and generally agree with your sentiment, but I think fundamentally you're arguing for a different thing than what HFEA was originally intended to be used as even though they are structured similarly. HFEA, to me at least, is a risky bet on long term trends for both stocks and bonds, but its risky nonetheless.

Correct me if I'm wrong but your version attempts to de-risk HFEA by de-levering the the equity and both de-levering and shortening the bond duration. And your argument is that if you want to YOLO your entire account on this strategy, your version is superior. I agree, but that wasn't the point of HFEA for me at least.

This was never going to be my full account. I have unlevered passive investments, I have other levered passive investments, I have an active bot trading LETFs, we have a portion of our NW in Euros and whatever the banks over there are doing with it, we have real estate, etc etc. Our world isn't clean enough to split it out into a x:y equity/bond allocations unfortunately.

So for me, HFEA was a pile of money that I could go more aggressive on with a backstop of a much larger "safe" portion of assets to fall back on if we needed. But to get the same or comparable level of risk that I need (1:3 for a Kelly portfolio), I'd have to lever it > 3x which means either box spreads or futures:

https://www.portfoliovisualizer.com/bac ... on4_3=-560

If I had the patience / discipline to do maintain this, I'd go with "equity adjusted" version of the above ITT portfolio as a true HFEA replacement. But even the "return adjusted" version still is too levered to manage w/o futures which I (personally) cannot trade automatically. So now I've been trying to at least diversify away from LTT a little to ITT without sacrificing returns and I think Mid Caps might be the answer:

https://www.portfoliovisualizer.com/bac ... tion5_1=90

I still need to check the Kelly/CVaR/CDaR portfolios with this mix, but I'm thinking it will be something along those lines in an M1 account and set and forget for 20+ years.
As long as it is recognized this is based on an emotional/psychological segregation of risk rather than a rational holistic view of risk, return and goals and the most efficient implementation of them.

It is simply more efficient to allocate a larger portion of one's nw to a slightly less risky and much more efficient strategy. The risk/return profile for one's entire net worth is improved.

I think I may have linked to the incorrect post before. I will copy and paste the explanation here:

Instead of doing 10% of net worth in SPY plus 10% of net worth in HFEA which is 165/135 stock/LTT, do 20% of one's net worth in 135/165 stock/ITT. The latter portfolio has higher risk adjusted returns.

If we look at the 10% in HFEA and the other 10% that we are moving from (assuming) SPY to modified HFEA:

Originally 10% HFEA + 10% SPY = 132.5/67.5 stock/LTT for that 20% of your portfolio (ignoring the other 80%)

New 20% in modified HFEA = 135/165 stock/ITT (again ignoring the other 80% of your portfolio)

Let's run the PV for that 20%. The red line is 20% 'modified' HFEA. The blue line is 10% HFEA + 10% SPY. The red line has much higher CAGR and a lower max-drawdown.

https://www.portfoliovisualizer.com/bac ... on4_2=-200



Since you are saying you do not have 100% of net worth in HFEA, the rational thing to do would be to move some funds from SPY (or comparable investment) and combine into a modified HFEA. The need to take extreme risk in the HFEA account is only based on arbitrary buckets that are created psychologically. You have one bucket (10% of net worth) in HFEA. Presumably you have another bucket that is 10% (or more) of net worth in SPY (or comparable investment). If we instead draw the bucket around both of these buckets and combine them, everything about the combined bucket is improved. The max drawdown is less. The expected return is higher. And if you want to psychologically break the bucket in half, you can still do that and consider you have 10% of net worth in SPY, and 10% in extremely leveraged modified HFEA.


The other alternative, for people that insist on drawing arbitrary psychological distinctions, is to lever a modified HFEA 1.1x. This gets you right back into the YOLO territory of original HFEA, but with much less risk. Nearly the same CAGR, 51% drawdown vs 65%. Why, other than the slight complexity of leveraging UPRO/TYD 1.1x, would one prefer an allocation with a 65% max-drawdown over one with a 51% max-drawdown, given nearly identical CAGR?

https://www.portfoliovisualizer.com/bac ... on4_2=-200
This is interesting to me since that psychological aspect somewhat matters to me. Is there a ratio between UPRO/TYD that you suggest? That way I can implement that in whatever amount of leverage I prefer. Curious because a lot of your posts seem to be arguing that various different allocations of the two are all preferable.
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

darkcam wrote: Mon Oct 18, 2021 4:41 pm
skierincolorado wrote: Mon Oct 18, 2021 11:33 am
adamhg wrote: Mon Oct 18, 2021 10:09 am
skierincolorado wrote: Mon Oct 18, 2021 8:55 am
Meaty wrote: Mon Oct 18, 2021 7:55 am

This is probably a dumb question- but how is the risk much lower? HFEA has double the returns with only modest differences in sharpe ratio and max drawdown. How does the use of TYD mitigate the risk? Is it far less interest rate sensitive?
The max draw is under 54% instead of over 65%. This is because it owns less stocks. So it’s less risky because it’s less stock exposure, but it makes up for some of the lost return because tyd has higher return than tmf per dollar. I think the backrest shows it well .. I’d rather be the blue line
I appreciate your replies and generally agree with your sentiment, but I think fundamentally you're arguing for a different thing than what HFEA was originally intended to be used as even though they are structured similarly. HFEA, to me at least, is a risky bet on long term trends for both stocks and bonds, but its risky nonetheless.

Correct me if I'm wrong but your version attempts to de-risk HFEA by de-levering the the equity and both de-levering and shortening the bond duration. And your argument is that if you want to YOLO your entire account on this strategy, your version is superior. I agree, but that wasn't the point of HFEA for me at least.

This was never going to be my full account. I have unlevered passive investments, I have other levered passive investments, I have an active bot trading LETFs, we have a portion of our NW in Euros and whatever the banks over there are doing with it, we have real estate, etc etc. Our world isn't clean enough to split it out into a x:y equity/bond allocations unfortunately.

So for me, HFEA was a pile of money that I could go more aggressive on with a backstop of a much larger "safe" portion of assets to fall back on if we needed. But to get the same or comparable level of risk that I need (1:3 for a Kelly portfolio), I'd have to lever it > 3x which means either box spreads or futures:

https://www.portfoliovisualizer.com/bac ... on4_3=-560

If I had the patience / discipline to do maintain this, I'd go with "equity adjusted" version of the above ITT portfolio as a true HFEA replacement. But even the "return adjusted" version still is too levered to manage w/o futures which I (personally) cannot trade automatically. So now I've been trying to at least diversify away from LTT a little to ITT without sacrificing returns and I think Mid Caps might be the answer:

https://www.portfoliovisualizer.com/bac ... tion5_1=90

I still need to check the Kelly/CVaR/CDaR portfolios with this mix, but I'm thinking it will be something along those lines in an M1 account and set and forget for 20+ years.
As long as it is recognized this is based on an emotional/psychological segregation of risk rather than a rational holistic view of risk, return and goals and the most efficient implementation of them.

It is simply more efficient to allocate a larger portion of one's nw to a slightly less risky and much more efficient strategy. The risk/return profile for one's entire net worth is improved.

I think I may have linked to the incorrect post before. I will copy and paste the explanation here:

Instead of doing 10% of net worth in SPY plus 10% of net worth in HFEA which is 165/135 stock/LTT, do 20% of one's net worth in 135/165 stock/ITT. The latter portfolio has higher risk adjusted returns.

If we look at the 10% in HFEA and the other 10% that we are moving from (assuming) SPY to modified HFEA:

Originally 10% HFEA + 10% SPY = 132.5/67.5 stock/LTT for that 20% of your portfolio (ignoring the other 80%)

New 20% in modified HFEA = 135/165 stock/ITT (again ignoring the other 80% of your portfolio)

Let's run the PV for that 20%. The red line is 20% 'modified' HFEA. The blue line is 10% HFEA + 10% SPY. The red line has much higher CAGR and a lower max-drawdown.

https://www.portfoliovisualizer.com/bac ... on4_2=-200



Since you are saying you do not have 100% of net worth in HFEA, the rational thing to do would be to move some funds from SPY (or comparable investment) and combine into a modified HFEA. The need to take extreme risk in the HFEA account is only based on arbitrary buckets that are created psychologically. You have one bucket (10% of net worth) in HFEA. Presumably you have another bucket that is 10% (or more) of net worth in SPY (or comparable investment). If we instead draw the bucket around both of these buckets and combine them, everything about the combined bucket is improved. The max drawdown is less. The expected return is higher. And if you want to psychologically break the bucket in half, you can still do that and consider you have 10% of net worth in SPY, and 10% in extremely leveraged modified HFEA.


The other alternative, for people that insist on drawing arbitrary psychological distinctions, is to lever a modified HFEA 1.1x. This gets you right back into the YOLO territory of original HFEA, but with much less risk. Nearly the same CAGR, 51% drawdown vs 65%. Why, other than the slight complexity of leveraging UPRO/TYD 1.1x, would one prefer an allocation with a 65% max-drawdown over one with a 51% max-drawdown, given nearly identical CAGR?

https://www.portfoliovisualizer.com/bac ... on4_2=-200
This is interesting to me since that psychological aspect somewhat matters to me. Is there a ratio between UPRO/TYD that you suggest? That way I can implement that in whatever amount of leverage I prefer. Curious because a lot of your posts seem to be arguing that various different allocations of the two are all preferable.
40/60 or 45/55. There are arguments for both, and both are clearly better than HFEA. Can leverage 1.1x if desired in Reg T if desired to get back to the stratospheric levels of risk/YOLO-ness of original HFEA. Or combine two of your psychological buckets into one bigger better bucket. I personally prefer to be heavier on stocks (for several reasons not related to sharpe ratio), so 45/55 is most similar to the ratio I do with futures. 40/60 would be most similar to HFEA in terms of the stock vs bond duration ratio, and have the higher sharpe ratio, at least since 1982.
investor.was.here
Posts: 88
Joined: Thu Oct 08, 2020 2:52 am

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

Post by investor.was.here »

skierincolorado wrote: Mon Oct 18, 2021 11:33 am The other alternative, for people that insist on drawing arbitrary psychological distinctions, is to lever a modified HFEA 1.1x.
I thought about this too. It'd work for me but I also want to set this up for my kids, in custodial accounts. I don't think we have access to margin or derivatives in them.

To summarize, here's the proposals:
  • SIC: 45/55 UPRO/TYD
  • AHG: 20/30/30/20 UPRO/MIDU/TMF/TYD.
  • HFEA: 55/45 UPRO/TMF
The data for AHG looks great to me. My kids could actually be looking at these numbers when they're my age.

I checked with M1Finance about TYA and, unfortunately, they won't add it.
Last edited by investor.was.here on Tue Oct 19, 2021 1:25 pm, edited 1 time in total.
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

investor.was.here wrote: Mon Oct 18, 2021 5:09 pm
skierincolorado wrote: Mon Oct 18, 2021 11:33 am The other alternative, for people that insist on drawing arbitrary psychological distinctions, is to lever a modified HFEA 1.1x.
I thought about this too. It'd work for me but I also want to set this up for my kids, in custodial accounts. I don't think we have access to margin or derivatives in them.

To summarize, here's the proposals:
  • SIC: 45/55 UPRO/TYD
  • AHG: 20/30/30/20 UPRO/MIDU/TMF/TYD.
  • HFEA: 55/45 UPRO/TMF
The data for AHG looks great to me. My kids could actually be looking at these numbers when they're my age.

I checked with M1Finance about TYA and, unfortunately, they won't add it.
Interesting use case. This could be one of the few cases where original HFEA actually makes sense. I can think of a couple reasons it might not though:

1) If you plan to give more money to the kids in the future, and this money is invested in something more conservative, like SPY. The efficient thing to do would be to think of it as one bucket, for your kids, and invest all of it in the most efficient way possible. Instead of 50k of HFEA (in a custodial account) and 50k of SPY (in your accounts but earmarked for them), it would be far suprior to do SIC for both 50k buckets.

2) If #1 does not apply, and this was therefore all of the money I ever intended to give directly, I would probably invest it a bit more conservatively than HFEA. Probably similar to SIC, which has the added benefit of being more efficient. But risk tolerance is a personal decision for you (on behalf of your kids).

Many people intend to leave an inheritance beyond whatever happens to be leftover if they die younger than expected. If you feel that that's the case, #1 probably applies. Even though I don't have kids yet, it's one of the reasons I invest more aggressively than I might otherwise. Arguably assuming I have kids, my time horizon is 80+ years not 50 years because I will likely have a large amount invested even in my 70s 80s and 90s (if I'm lucky enough to live that long).
investor.was.here
Posts: 88
Joined: Thu Oct 08, 2020 2:52 am

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

Post by investor.was.here »

skierincolorado wrote: Mon Oct 18, 2021 5:37 pm 1) If you plan to give more money to the kids in the future, and this money is invested in something more conservative, like SPY. The efficient thing to do would be to think of it as one bucket, for your kids, and invest all of it in the most efficient way possible. Instead of 50k of HFEA (in a custodial account) and 50k of SPY (in your accounts but earmarked for them), it would be far suprior to do SIC for both 50k buckets.
I'm on board with the efficiency of thinking holistically but not the practicality of it. As such, I've been mostly a "balanced fund as reserves" kind of investor. Don't really think about time horizons, just have faith that in the long run the opportunity costs of having money outside the market exceeds the unwanted risk of sub-optimal investment decisions in the short run.

I'm struggling to figure out what my new reality will be. Thinking in buckets still, I like the idea of keeping some balanced fund around, in case none of this stuff plays out as expected. But, for the rest, why do I need safety? We've only got 25yrs until retirement, we're well insured, and we live off income.

An alternative I'm toying with the Ginger Ale portfolio. It looks more diversified. I'm afraid that Dalio is going to end up being right, and we're going to be needing international and other inflation hedges. In this case, our leveraged funds, as is, may perform horribly. Or am I mistaken?
Fonfo
Posts: 15
Joined: Mon Aug 02, 2021 10:55 am

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

Post by Fonfo »

adamhg wrote: Sat Oct 16, 2021 8:48 am
Fonfo wrote: Sat Oct 16, 2021 8:31 am Is there any etf that already replies this upro/TMF strategy or one similar? Even if it is only x2 would be interesting for me.

I am asking that because I am able to apply this strategy only in a Taxable account for now.
Psldx is 2x (100/100) with active bonds but should only be used in tax advantaged

Ntsx is 1.5x (90/60) with itt and is pretty tax efficient
OK Thanks! Hard to believe that no one has come up with a etf doing hedgi strategy.... I think the time will come.
haranoth
Posts: 151
Joined: Wed Apr 06, 2016 12:07 pm

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

Post by haranoth »

Those who are quarterly rebalancing, are you following quarterly rebalancing dates that of S&p 500?
For reference, the rebalance dates for s&p 500 is shown here
http://modernir.com/wp-content/uploads/ ... -Final.pdf
Or, do you pick a specific date/day every 3 months?
Wondering what's the cadence for the quarterly folks
Thanks
Hydromod
Posts: 1048
Joined: Tue Mar 26, 2019 10:21 pm

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

Post by Hydromod »

haranoth wrote: Tue Oct 19, 2021 3:53 pm Those who are quarterly rebalancing, are you following quarterly rebalancing dates that of S&p 500?
For reference, the rebalance dates for s&p 500 is shown here
http://modernir.com/wp-content/uploads/ ... -Final.pdf
Or, do you pick a specific date/day every 3 months?
Wondering what's the cadence for the quarterly folks
Thanks
Backtesting suggests that rebalancing within a few days of the turn of the calendar quarters has been especially good and rebalancing near the middle of the quarters has been especially bad when rebalancing quarterly. Similarly, rebalancing within a few days of the turn of the month has been especially good and rebalancing in the middle of the month worse when rebalancing monthly.

That linked document suggest that the historical pattern may be related to a bunch of activity aimed at truing up portfolios occurs at the end of each month and options stuff occurs in the middle of each month, so perhaps the historical patterns will continue.
Post Reply