Hedgie's 55/45 With Other 3x Bond ETFs

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
Topic Author
golongterm
Posts: 3
Joined: Sun Oct 04, 2020 9:50 am

Hedgie's 55/45 With Other 3x Bond ETFs

Post by golongterm »

I was thinking of possibly introducing TYD (3x mid-term) and TBT (3x short term) bond ETFs to hedgie's portfolio, just to backtest smaller/equal allocations of these and see if I could get the volatility down a bit more. Now I get it that long terms have the lowest correlation with equities of any asset in any asset class (as far as I know), and that's the idea, but have we examined the correlation of short and mid-term bonds fully? I could have sworn I saw someone advocating for this either here or on the BH reddit, stating it performed better, but unfortunately I don't have the data to go back far enough in PV to really test any allocations.

I was looking over this paper https://faculty.mccombs.utexas.edu/keit ... n-9.14.pdf which is discussing stock-bond correlations, specifically focusing on the 10-years (I know, it's a bit older, but it was the best match I could readily find). Specifically Graph 2 on page 69 (p 3 of the pdf) caught my eye, explaining that during periods of high inflation, the stock/bond correlation actually becomes quite negative. Considering the fed has given the green light to inflation at this point, which should eventually coincide with a concomitant IR increase (maybe not a crazy one, but certainly an increase from where we are now), does this then make sense to at least consider an allocation of TYD?

Again, I'd love to start backtesting, but I don't have data like UPROSIM and TMFSIM to actually start running this, and I have no idea how to actually find/put that data together. If anyone has any ideas for how to do this, and of course if anyone has any thoughts about the idea in general (I'm sure it's been discussed before), let me know.
User avatar
Uncorrelated
Posts: 1151
Joined: Sun Oct 13, 2019 3:16 pm

Re: Hedgie's 55/45 With Other 3x Bond ETFs

Post by Uncorrelated »

You can use my mean variance framework to find the best asset allocation if you have an estimate of future expected return, volatility and correlations. There is no need to have the actual data history available.

For what it's worth, I think that graph on page 3 looks like random noise. The human mind is very good at finding patterns where there are none. The best estimate for future stock-bond correlation is probably zero.

Last time I checked, short and mid-term 3x ETF's are too expensive. The problem isn't necessary that these ETF's themselves are expensive, but that you need more of them to obtain significant TERM exposure. This forces you to take a higher allocation to UPRO if you want to keep similar equity allocation, which has a hidden expense ratio of 2%. These criticisms don't necessary apply to futures-based leverage.
ljford7
Posts: 45
Joined: Tue Jul 06, 2010 8:15 pm

Re: Hedgie's 55/45 With Other 3x Bond ETFs

Post by ljford7 »

Try it for yourself:

https://www.portfoliovisualizer.com/bac ... tion4_1=20

Something else to keep in mind that people have been waving the inflation flag since 2008 and it hasn't happened. We very well could be living in a different environment from anytime in recent history or we may not be, but be careful to form your own judgement and not just follow the herd.
MotoTrojan
Posts: 10709
Joined: Wed Feb 01, 2017 8:39 pm

Re: Hedgie's 55/45 With Other 3x Bond ETFs

Post by MotoTrojan »

If you want less duration I’d use EDV; low expense, no leverage so no volatility decay, and gives a bit less than 1/2 the duration of TMF. But if you want to reduce overall volatility you also need to reduce the UPRO allocation.

When I held the HF portfolio I adjusted to 43/57 UPRO/EDV which is the same ratio of equity to bond volatility back to 1955 as 55/45 UPRO/TMF, so effectively it’s the HF portfolio but with 22% less leverage overall. Much more efficient IMHO (lower ER, lower volatility decay, more reasonable overall leverage).
000
Posts: 3345
Joined: Thu Jul 23, 2020 12:04 am

Re: Hedgie's 55/45 With Other 3x Bond ETFs

Post by 000 »

golongterm wrote: Sun Oct 04, 2020 10:09 am I was thinking of possibly introducing TYD (3x mid-term) and TBT (3x short term) bond ETFs to hedgie's portfolio
That makes even less sense than leveraging long term bonds, as leverage is a negative allocation to cash and the shorter the bond duration the more cash-like it becomes.
User avatar
coingaroo
Posts: 93
Joined: Fri Apr 26, 2019 11:31 am

Re: Hedgie's 55/45 With Other 3x Bond ETFs

Post by coingaroo »

000 wrote: Sun Oct 04, 2020 10:43 pm
golongterm wrote: Sun Oct 04, 2020 10:09 am I was thinking of possibly introducing TYD (3x mid-term) and TBT (3x short term) bond ETFs to hedgie's portfolio
That makes even less sense than leveraging long term bonds, as leverage is a negative allocation to cash and the shorter the bond duration the more cash-like it becomes.
Leverage also incurs the cost of volatility drag and swap financing costs, and the shorter the duration the less vol drag you incur.

TYD has a sharpe of 0.72, compared to TMF's 0.58
000
Posts: 3345
Joined: Thu Jul 23, 2020 12:04 am

Re: Hedgie's 55/45 With Other 3x Bond ETFs

Post by 000 »

coingaroo wrote: Mon Oct 05, 2020 12:12 am Leverage also incurs the cost of volatility drag and swap financing costs, and the shorter the duration the less vol drag you incur.

TYD has a sharpe of 0.72, compared to TMF's 0.58
Not sure I follow... you would borrow money to invest in a one-month T-bill?

What is the limit of this strategy?
User avatar
Uncorrelated
Posts: 1151
Joined: Sun Oct 13, 2019 3:16 pm

Re: Hedgie's 55/45 With Other 3x Bond ETFs

Post by Uncorrelated »

ljford7 wrote: Sun Oct 04, 2020 9:32 pm Try it for yourself:

https://www.portfoliovisualizer.com/bac ... tion4_1=20

Something else to keep in mind that people have been waving the inflation flag since 2008 and it hasn't happened. We very well could be living in a different environment from anytime in recent history or we may not be, but be careful to form your own judgement and not just follow the herd.
There is no point in trying to run backtests with bonds unless your data goes back to at least the 1960's, preferably back to the 1930's. The entire period from 1980 to 2020 is completely unrepresentative of expected bond returns.

MotoTrojan wrote: Sun Oct 04, 2020 10:06 pm If you want less duration I’d use EDV; low expense, no leverage so no volatility decay, and gives a bit less than 1/2 the duration of TMF. But if you want to reduce overall volatility you also need to reduce the UPRO allocation.

When I held the HF portfolio I adjusted to 43/57 UPRO/EDV which is the same ratio of equity to bond volatility back to 1955 as 55/45 UPRO/TMF, so effectively it’s the HF portfolio but with 22% less leverage overall. Much more efficient IMHO (lower ER, lower volatility decay, more reasonable overall leverage).
The problem with EDV is that it doesn't actually have a higher sharpe ratio than TMF due to a high-beta anomaly (the longer the bond duration, the lower the sharpe ratio becomes). Practically speaking, you need twice as much EDV to obtain the same TERM exposure as with TMF. The is problematic for the same ration that TYD (3x ITT) is problematic: to obtain the same equity exposure, you need more UPRO, and UPRO is extremely expensive.

When instructed to take less risk, my mean-variance optimize either reduces the TMF allocation or trades some TMF exposure for total bond market. I have found no evidence that other bond choices make sense.
000 wrote: Sun Oct 04, 2020 10:43 pm
golongterm wrote: Sun Oct 04, 2020 10:09 am I was thinking of possibly introducing TYD (3x mid-term) and TBT (3x short term) bond ETFs to hedgie's portfolio
That makes even less sense than leveraging long term bonds, as leverage is a negative allocation to cash and the shorter the bond duration the more cash-like it becomes.
That is not quite true. Short term bonds have a much higher sharpe ratio than long term bonds, this is clearly illustrated by Frazzini (2013, pdf):

Image
The Y axis is the annual alpha after running a regression against the treasury market.

The problem with leveraging short term bonds isn't that short term bonds are bad, the problem is that you need much more of them to obtain significant TERM exposure. According to my mean-variance framework the interactions with the rest of the portfolio are such that TYD never makes any sense. But this can only be seen on the portfolio level. From the perspective of individual assets, TYD is far superior to TMF.
aristotelian
Posts: 8286
Joined: Wed Jan 11, 2017 8:05 pm

Re: Hedgie's 55/45 With Other 3x Bond ETFs

Post by aristotelian »

Why not just hold 1X long term treasuries? You would have approximately the same volatility at much lower cost. If you are going to use leverage, go big.
AnilG
Posts: 63
Joined: Wed Sep 09, 2015 1:41 am

Re: Hedgie's 55/45 With Other 3x Bond ETFs

Post by AnilG »

TBT is ProShares UltraShort 20+ Year Treasury LETF (-2x). What is the reasoning behind adding an Inverse LETF to hedgie portfolio?

TYD is 7-10 years 3x LETF and TMF is 20+ years 3x LETF.
golongterm wrote: Sun Oct 04, 2020 10:09 am I was thinking of possibly introducing TYD (3x mid-term) and TBT (3x short term) bond ETFs to hedgie's portfolio, just to backtest smaller/equal allocations of these and see if I could get the volatility down a bit more. ... ...
User avatar
Steve Reading
Posts: 2521
Joined: Fri Nov 16, 2018 10:20 pm

Re: Hedgie's 55/45 With Other 3x Bond ETFs

Post by Steve Reading »

Uncorrelated wrote: Mon Oct 05, 2020 7:00 am
MotoTrojan wrote: Sun Oct 04, 2020 10:06 pm If you want less duration I’d use EDV; low expense, no leverage so no volatility decay, and gives a bit less than 1/2 the duration of TMF. But if you want to reduce overall volatility you also need to reduce the UPRO allocation.

When I held the HF portfolio I adjusted to 43/57 UPRO/EDV which is the same ratio of equity to bond volatility back to 1955 as 55/45 UPRO/TMF, so effectively it’s the HF portfolio but with 22% less leverage overall. Much more efficient IMHO (lower ER, lower volatility decay, more reasonable overall leverage).
The problem with EDV is that it doesn't actually have a higher sharpe ratio than TMF due to a high-beta anomaly (the longer the bond duration, the lower the sharpe ratio becomes).
I knew about the BAB anomaly, but didn’t realize it was large enough to make up for the 1% expense ratio of TMF (~0.5% per leg so to speak), on top of the fact that TMF’s swaps already borrow above the RFR.
"... so high a present discounted value of wealth, it is only prudent for him to put more into common stocks compared to his present tangible wealth, borrowing if necessary" - Paul Samuelson
User avatar
Uncorrelated
Posts: 1151
Joined: Sun Oct 13, 2019 3:16 pm

Re: Hedgie's 55/45 With Other 3x Bond ETFs

Post by Uncorrelated »

Steve Reading wrote: Mon Oct 05, 2020 8:40 am
Uncorrelated wrote: Mon Oct 05, 2020 7:00 am
MotoTrojan wrote: Sun Oct 04, 2020 10:06 pm If you want less duration I’d use EDV; low expense, no leverage so no volatility decay, and gives a bit less than 1/2 the duration of TMF. But if you want to reduce overall volatility you also need to reduce the UPRO allocation.

When I held the HF portfolio I adjusted to 43/57 UPRO/EDV which is the same ratio of equity to bond volatility back to 1955 as 55/45 UPRO/TMF, so effectively it’s the HF portfolio but with 22% less leverage overall. Much more efficient IMHO (lower ER, lower volatility decay, more reasonable overall leverage).
The problem with EDV is that it doesn't actually have a higher sharpe ratio than TMF due to a high-beta anomaly (the longer the bond duration, the lower the sharpe ratio becomes).
I knew about the BAB anomaly, but didn’t realize it was large enough to make up for the 1% expense ratio of TMF (~0.5% per leg so to speak), on top of the fact that TMF’s swaps already borrow above the RFR.
Here is the average term premium, the difference between the risk free rate (here defined as the 1-month LIBOR) and the corresponding treasury rate since 1993 (the earliest FRED has data on all series):

5 year: 0.78%
10 year: 1.33%
20 year: 1.84%
30 year: 1.89%

According to simba's backtesting spreadsheet, standard deviations of excess return are as follows (since 1980, the start of the Volcker era):
VFITX (average duration 5-6 years): 7.11%
VUSTX (average duration 18 years): 12.44%
EDV: (average duration 25 years): 28.9%

TMF has a term premium approximately of ~ 4.5% including the expense ratio, annual standard deviation around 36%. TMF usually borrows at the risk free rate.

It is clear that A) LTT and EDV have almost identical expected return, but LTT is much less risky. And B) TMF is significantly more attractive than EVD. I don't understand the appeal of EDV. All the interest in EDV must be based on faulty backtests?

For robustness I used different time periods to estimate the volatility of the various bond funds, but reached the same conclusion each time.
User avatar
Steve Reading
Posts: 2521
Joined: Fri Nov 16, 2018 10:20 pm

Re: Hedgie's 55/45 With Other 3x Bond ETFs

Post by Steve Reading »

Uncorrelated wrote: Mon Oct 05, 2020 9:24 am
It is clear that A) LTT and EDV have almost identical expected return, but LTT is much less risky. And B) TMF is significantly more attractive than EVD. I don't understand the appeal of EDV. All the interest in EDV must be based on faulty backtests?
Upon looking at the numbers, I agree that 50/50 cash/TMF and EDV have similar expected returns. Not sure about the former being less risky, since the duration is the same, I’d expect a similar volatility.

Under those conditions, I’d always pick EDV over TMF simply because expense ratios are much more important than non-arbitraged anomalies in my book. The flatness of the yield curve at the 20-30Y location might be a rational reason instead of a free-lunch. But saving on ERs always is a free lunch.

Numbers are much closer than I thought though so thanks for bringing it up :) Couldn’t fault anyone for picking TMF over EDV upon closer inspection.
"... so high a present discounted value of wealth, it is only prudent for him to put more into common stocks compared to his present tangible wealth, borrowing if necessary" - Paul Samuelson
000
Posts: 3345
Joined: Thu Jul 23, 2020 12:04 am

Re: Hedgie's 55/45 With Other 3x Bond ETFs

Post by 000 »

Uncorrelated wrote: Mon Oct 05, 2020 7:00 am
000 wrote: Sun Oct 04, 2020 10:43 pm
golongterm wrote: Sun Oct 04, 2020 10:09 am I was thinking of possibly introducing TYD (3x mid-term) and TBT (3x short term) bond ETFs to hedgie's portfolio
That makes even less sense than leveraging long term bonds, as leverage is a negative allocation to cash and the shorter the bond duration the more cash-like it becomes.
That is not quite true. Short term bonds have a much higher sharpe ratio than long term bonds, this is clearly illustrated by Frazzini (2013, pdf):

Image
The Y axis is the annual alpha after running a regression against the treasury market.

The problem with leveraging short term bonds isn't that short term bonds are bad, the problem is that you need much more of them to obtain significant TERM exposure. According to my mean-variance framework the interactions with the rest of the portfolio are such that TYD never makes any sense. But this can only be seen on the portfolio level. From the perspective of individual assets, TYD is far superior to TMF.
Right... I'm going to make borrowing to buy 3-month T-bills a core part of my investment strategy...

Somehow I think the train has come off the tracks when we're seriously discussing doing this.
NMBob
Posts: 364
Joined: Thu Apr 23, 2015 8:13 pm

Re: Hedgie's 55/45 With Other 3x Bond ETFs

Post by NMBob »

not 3x etfs but a method discussed a year ago using 2 year treasury futures

viewtopic.php?t=290919

A gentle alternative to HEDGEFUNDIE's excellent adventure with 2-Year Treasury Futures
Post Reply