HEDGEFUNDIE's excellent adventure Part II: The next journey

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
CletusCaddy
Posts: 2678
Joined: Sun Sep 12, 2021 4:23 am

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

Post by CletusCaddy »

corpgator wrote: Tue Sep 13, 2022 11:30 am
CletusCaddy wrote: Tue Sep 13, 2022 10:45 am Not sure what all the TMF gripping is all about. TMF is saving UPRO’s butt today.
TMF is down more than 2% on the day. There's nothing saving going on since cash would be better. Saving would be tmf up on a big red day like today.
The day is young my friend
unemployed_pysicist
Posts: 220
Joined: Sat Oct 09, 2021 2:32 pm
Location: Amsterdam

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

Post by unemployed_pysicist »

comeinvest wrote: Mon Sep 12, 2022 4:22 pm
Can you please explain the meaning or importance of the "value of convexity" intuitively in layman terms please? Is it significant?

I'm not sure if I would trust "surveys" regarding the yields 20-30 years from now (or regarding just about anything else that far in the future), given that "experts" can't even predict seemingly much simpler things like imminent market crashes or near-term inflation or interest rates. The long term rates estimates are largely based on short-term rate expectation between now and the time far in the future (expectations hypothesis), which means they would depend on estimates of federal reserve policy, inflation, etc., for decades. Is there really any evidence that surveys are a serious approach?

I think the yield curve itself can be used to predict (curve implied) rate changes at a specific maturity.

I can't find the Ilmainen papers or books online and not even to buy, can you please let me know where to find them?
I am also skeptical of forecasts. Nonetheless, it appears that some term structure models use them as an input. I was under the impression the forecasts were shorter term outlooks, more like "where do you see the 10 year note yield in 1 year", not necessarily 20-30 years in the future. Like you, I would rather rely on the curve itself.

Here is an example concerning the value of convexity:

In figure 9 of part 5 of Understanding the Yield Curve, Antti Ilmainen calculates the value of convexity and convexity adjusted 1 year expected returns for various notes and bonds, using the historical volatility for each tenor. The data below correspond to 1 September 1995. There are several entries, but here are examples for the 5 year par bond and a 30 year zero just to highlight how the value of convexity impacts near term expected returns at some key points on the curve.

5 year par bond:
Convexity: 0.23
Historical Vol: 1.04
Annual yield: 6.13%
Rolling yield (carry): 6.31%
Value of convexity: 0.09%
Convexity adjusted expected return: 6.4%

30 year zero:
Convexity: 8.14
Historical Vol: 0.79
Annual yield: 6.88%
Rolling yield (carry): 5.93%
Value of convexity: 2.53%
Convexity adjusted expected return: 8.46%

So in this comparison, you can see that the yield curve was downward sloping near the 30 year segment (carry is less than the yield.) The carry of the 5 year is greater than the 30 year zero. But once the convexity is accounted for, the 1 year period expected return increases by a significant amount for the 30 year zero.

The bond market (correctly) suppresses the yields at high maturities to account for the near term higher expected returns. Higher volatility enhances the performance of positively convex positions.

However, as you would expect, convexity comes at a price in the long run. The barbell-bull analysis in part 5 of Understanding the Yield Curve shows that duration matched barbells do slightly underperform bullets (though with less volatility than the bullet).
Antti Ilmainen wrote: A barbell's convexity advantage is rarely sufficient to offset the negative carry over a multi-year period.
So it is true that from a historical standpoint, convexity is a bit of a drag on returns over longer periods. I think this makes intuitive sense; the convexity buyer is paying a premium for the insurance properties of convexity, just like with buying options (volatility premium). But bond convexity does partially contribute to the return. The contribution of convexity to realized return for the barbell is not zero in Ilmainen's barbell-bullet example, it just is not enough to make up for the initial yield disadvantage.

I originally found the link to the Understanding the Yield Curve pdf in a post on stack exchange; I will see if I can find it again.
couldn't afford the h | | BUY BONDS | WEAR DIAMONDS
unemployed_pysicist
Posts: 220
Joined: Sat Oct 09, 2021 2:32 pm
Location: Amsterdam

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

Post by unemployed_pysicist »

skierincolorado wrote: Tue Sep 13, 2022 8:48 am
unemployed_pysicist wrote: Mon Sep 12, 2022 2:09 am
skierincolorado wrote: Sun Sep 11, 2022 7:44 pm
That's because long term interest rates fell dramatically over that period.
There are share classes that have been around longer. If you start in 2009 (dec 31 2008) you get a period of neae neutral rates. During this period the sharpe ratio is much higher for vglt than edv .26 vs .18.

Any period in which rates don't fall, the sharpe for vglt will be much higher. Over long time periods like 30 or 50 years this matters less (even if rates fall they are falling less per year).
Which share class? VGLSX does not seem to go back any further than VGLT. Comparing VEDTX (EDV) and TLT as another comparison gives a similar Sharpe ratio of about 0.31 for TLT and 0.3 for VEDTX from Dec 2007 to present (almost 15 years.)

Perhaps you are referring to VUSTX? VUSTX is actively managed and different holdings so not the same as VGLT. Even then the Sharpe is .34 for VUSTX vs .3 for VEDTX from Dec 2007 to present.
Again, from Dec 2007 to present long term interest rates fell 1%. Starting Dec 31 2008 produces a period of flat rates, which produces a fair comparison.

As comeinvest pointed out the carry on the 30 y is abysmal and has been very poor historically as well.

If you go back longer the sharpe on the longest maturities is terrible. Or pick a period of near flat rates. Picking relatively short periods with large interest rate decreases is going to make the longest maturities look artificially good due to the large capital gains they experience in such periods.
In the period from Jan 28 2009 to August 30 2022, DGS30 started at 3.44%, and ended at 3.23%, which is a 21 bps fall, as compared to the dates you suggested:
Dec 29 2008 to August 30 2022, where DGS30 started at 2.63% and ended at 3.23%, which is a 60 bps rise (further from 0 than 21 bps.)

The period from Jan 28 2009 to August 30 2022 gives Sharpe of .31 for VUSTX, versus .27 for VEDTX. VEDTX underperformed VUSTX for this period of ~14 years, but I would not characterize this as "terrible".

In the long run, I do not disagree that VGLT is likely to outperform EDV on a risk-adjusted basis. However, I am pushing back against the idea that EDV is somehow "terrible" or unusable for this strategy. EDV is a reasonable option for the leveraged-constrained investor who wants to take significant duration risk. VGLT and EDV are both reasonable, depending on risk appetite.
Last edited by unemployed_pysicist on Tue Sep 13, 2022 2:51 pm, edited 1 time in total.
couldn't afford the h | | BUY BONDS | WEAR DIAMONDS
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

Tamalak wrote: Tue Sep 13, 2022 10:47 am
perfectuncertainty wrote: Tue Sep 13, 2022 10:40 am
Marseille07 wrote: Tue Sep 13, 2022 9:37 am Buddy, it's already known that the long-term correlation of stocks & bonds is 0.

But here is the thing, people still flock to buy bonds if there is a big crash like March 2020. That said, the negative correlation is certainly diluted during a rising yield regime.

Basically the whole notion of a BH portfolio is called into question during a rising yield regime.
That's not quite accurate. The correlation changes between 0, + and - depending on the macro environment. It has registered as all correlations for decades. Don't count on the correlation - is my point.

Link: https://www.mfs.com/en-us/investment-pr ... ation.html

"In recent decades, negative correlations between stocks and bonds have been a tailwind for asset allocators, but negative correlations have been the historical exception, not the rule."
As I've come to understand it the HFEA perspective is this.

The "real" risk of 3x S&P is not gloomy, grinding years like this one. -50% YTD may seem like a lot, but it's only set UPRO back 18 months.. same as 1x S&P. During these "ordinary" drops, TMF will wiggle around as it pleases, hurting as often as it helps, and offering a pitiful return for its volatility on its own (my numbers say something like 2x 20 year rates)

The "real" risk of 3x S&P are lightning strike disasters like 2008 or 2020. The speed of the crash makes 3x leverage much worse and destroys underlying capital. What cuts to the bone in 1x cuts through the bone in 3x. In these scenarios UPRO gets set back, calendar-wise, much further than 1x S&P. And these are the scenarios where the Fed, and therefore TMF, will come to the rescue.
That's not my perspective. It is certainly possible both fall much more. Holding 165% stock plus 135% LTT likely has more risk than 165% stock alone both in terms of maximum drawdown and standard deviation of returns. The primary benefit of holding bonds is not the insurance. It is the expectation for positive returns that outweigh an increase in the expected standard deviation and maximum drawdown. I would expect holding bonds to increase the probability of 50% and 80% drawdowns, not decrease them. For example, holding bonds in 2022 has turned a relatively mundane 30% drawdown for 165% stocks into a 55% drawdown for 165% srock plus 135% LTT - exactly the kind of severe drawdown some thought holding bonds was supposed to help us avoid. It's important to be sanguine about the risks before embarking on this strategy. As I've said before in this thread, 135% in LTT is overtuned to the post 1982 era and it would be wise to look at a broader historical period and array of possible future outcomes to decide AA.
Last edited by skierincolorado on Tue Sep 13, 2022 3:05 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 »

unemployed_pysicist wrote: Tue Sep 13, 2022 2:29 pm
skierincolorado wrote: Tue Sep 13, 2022 8:48 am
unemployed_pysicist wrote: Mon Sep 12, 2022 2:09 am
skierincolorado wrote: Sun Sep 11, 2022 7:44 pm
That's because long term interest rates fell dramatically over that period.
There are share classes that have been around longer. If you start in 2009 (dec 31 2008) you get a period of neae neutral rates. During this period the sharpe ratio is much higher for vglt than edv .26 vs .18.

Any period in which rates don't fall, the sharpe for vglt will be much higher. Over long time periods like 30 or 50 years this matters less (even if rates fall they are falling less per year).
Which share class? VGLSX does not seem to go back any further than VGLT. Comparing VEDTX (EDV) and TLT as another comparison gives a similar Sharpe ratio of about 0.31 for TLT and 0.3 for VEDTX from Dec 2007 to present (almost 15 years.)

Perhaps you are referring to VUSTX? VUSTX is actively managed and different holdings so not the same as VGLT. Even then the Sharpe is .34 for VUSTX vs .3 for VEDTX from Dec 2007 to present.
Again, from Dec 2007 to present long term interest rates fell 1%. Starting Dec 31 2008 produces a period of flat rates, which produces a fair comparison.

As comeinvest pointed out the carry on the 30 y is abysmal and has been very poor historically as well.

If you go back longer the sharpe on the longest maturities is terrible. Or pick a period of near flat rates. Picking relatively short periods with large interest rate decreases is going to make the longest maturities look artificially good due to the large capital gains they experience in such periods.
In the long run, I do not disagree that VGLT is likely to outperform EDV on a risk-adjusted basis. However, I am pushing back against the idea that EDV is somehow "terrible" or unusable for this strategy. EDV is a reasonable option for the leveraged-constrained investor who wants to take significant duration risk. VGLT and EDV are both reasonable, depending on risk appetite.
I disagree. The risk adjusted returns of EDV during periods in which interest rates do not decrease is horrendous (.18 sharpe). Inclusion of such an asset is likely to increase the risk of a portfolio far more than it increases expected returns. I'd even include unleveraged and leveraged 20 year bonds in that category (vglt and tmf) but it's not nearly as bad. Edv is potentially useful as a short term rebalancing option but I would use vglt and tmf.
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

corpgator wrote: Tue Sep 13, 2022 10:43 am
skierincolorado wrote: Tue Sep 13, 2022 9:47 am
Tamalak wrote: Tue Sep 13, 2022 9:35 am
skierincolorado wrote: Tue Sep 13, 2022 9:25 am
Tamalak wrote: Tue Sep 13, 2022 9:08 am

Only if both of the non correlated components have a good expected return. I'm unsure if TMF does. If rates never changed from here on out, what would TMF return?

Its return from 2010 thru 2019 - a period where for the start and end dates, both rates and rate expectations had to be near 0 - was 16% annualized, so maybe it is great..
Long term rates fell a lot from 2010 to present. Expected returns for tmf are not high, but are probably positive. One estimate is the term premium.
I was foolishly looking at 10 year rates when TMF is leveraged 20 year.

Let's look at 20 year rates and try to compare apples to apples:

https://www.multpl.com/20-year-treasury ... le/by-year

Jan 1 2012, rate 2.69%. Jan 1 2018, rate 2.73%. Close enough even for leverage, right?

But the CAGR of TMF from 2012 thru 2017 is only 3.8%.

Jan 1 2015, rate 2.20%. Jan 1 2022, rate 2.15%.

CAGR of TMF from from 2015 thru 2021 is only 5%.

In the long run, it doesn't look like TMF has a good return at all! So what makes it useful? Its correlation to zero, so it provides no protection to UPRO. And its return on its own is pitiful compared to its huge volatility.
I pretty much agree. 3-5% is still positive and so the expected return of hfea (165% stock 135% LtT) is probably higher than 165% stock alone. And in the future I expext tmf to return more like 3-3.5% based on the term premium near zero. Not worth the extra risk in my opinion. A modest position in ITT captures most of the potential return with less risk and similar potential for downside protection during the worst market crashes like 01 08 and 20.
I'm glad for your contributions. Because of them, I stopped rebalancing into TMF and started buying TYD. It's not much, but it has helped some. I've looked at the portfolio overall instead of bucketing. I have a large position in HFEA (and shrinking!), so it has been hard to stomach. Maybe it will recover some day...
I am glad to hear it and hope that when looked at as a whole your portfolio is not down as much as HFEA. Those have been the two biggest points I've tried to make without repeating too often. I assume that to maintain your original risk budget you've been rebalancing into the leveraged parts of your portfolio? I have been selling vti and buying upro in my 401k to maintain the same leverage as a year ago ( actually increased leverage from 1.35x to 1.55x per lifecycle investing principles).
corpgator
Posts: 217
Joined: Wed Jul 22, 2015 12:00 pm

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

Post by corpgator »

skierincolorado wrote: Tue Sep 13, 2022 2:59 pm
corpgator wrote: Tue Sep 13, 2022 10:43 am
skierincolorado wrote: Tue Sep 13, 2022 9:47 am
Tamalak wrote: Tue Sep 13, 2022 9:35 am
skierincolorado wrote: Tue Sep 13, 2022 9:25 am
Long term rates fell a lot from 2010 to present. Expected returns for tmf are not high, but are probably positive. One estimate is the term premium.
I was foolishly looking at 10 year rates when TMF is leveraged 20 year.

Let's look at 20 year rates and try to compare apples to apples:

https://www.multpl.com/20-year-treasury ... le/by-year

Jan 1 2012, rate 2.69%. Jan 1 2018, rate 2.73%. Close enough even for leverage, right?

But the CAGR of TMF from 2012 thru 2017 is only 3.8%.

Jan 1 2015, rate 2.20%. Jan 1 2022, rate 2.15%.

CAGR of TMF from from 2015 thru 2021 is only 5%.

In the long run, it doesn't look like TMF has a good return at all! So what makes it useful? Its correlation to zero, so it provides no protection to UPRO. And its return on its own is pitiful compared to its huge volatility.
I pretty much agree. 3-5% is still positive and so the expected return of hfea (165% stock 135% LtT) is probably higher than 165% stock alone. And in the future I expext tmf to return more like 3-3.5% based on the term premium near zero. Not worth the extra risk in my opinion. A modest position in ITT captures most of the potential return with less risk and similar potential for downside protection during the worst market crashes like 01 08 and 20.
I'm glad for your contributions. Because of them, I stopped rebalancing into TMF and started buying TYD. It's not much, but it has helped some. I've looked at the portfolio overall instead of bucketing. I have a large position in HFEA (and shrinking!), so it has been hard to stomach. Maybe it will recover some day...
I am glad to hear it and hope that when looked at as a whole your portfolio is not down as much as HFEA. Those have been the two biggest points I've tried to make without repeating too often. I assume that to maintain your original risk budget you've been rebalancing into the leveraged parts of your portfolio? I have been selling vti and buying upro in my 401k to maintain the same leverage as a year ago ( actually increased leverage from 1.35x to 1.55x per lifecycle investing principles).
I got a new job that pays me double and gave me access to a mega backdoor roth recently. Instead of buying 100% vti or equivalent as I always did before (aside from my forrays into HFEA where I converted roths and older rolled over 401ks), I've been buying VTI with pre-tax and UPRO with after tax since the 401k plan allows for brokerage link at fidelity and I can buy anything. It feels like lighting my money on fire though, but I'm holding strong so far.

I did add on some ibonds too. I need to find some cash somewhere to get some more after I set up trustee accounts for the wife and I. Right now I'm at 1.35/.45/-.8 if you count all my liquid assets. Strictly stocks, I'm at 1.62. Bonds/fixed are much more leveraged since I simply don't hold cash since I'm not near retirement.
FIRE55
Posts: 59
Joined: Tue Dec 15, 2020 6:09 pm

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

Post by FIRE55 »

skierincolorado wrote: Tue Sep 13, 2022 2:59 pm I assume that to maintain your original risk budget you've been rebalancing into the leveraged parts of your portfolio? I have been selling vti and buying upro in my 401k to maintain the same leverage as a year ago ( actually increased leverage from 1.35x to 1.55x per lifecycle investing principles).
Yup. Sold some VYM & JEPI & cash to buy some TQQQ & UPRO as rebalance bands fired and 52-wk lows were hit. So far I've bought some UPRO down to $35 and some TQQQ down to $25. Orders placed for $30 UPRO and $20 TQQQ.

/FIRE55
comeinvest
Posts: 2709
Joined: Mon Mar 12, 2012 6:57 pm

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

Post by comeinvest »

skierincolorado wrote: Tue Sep 13, 2022 2:59 pm I am glad to hear it and hope that when looked at as a whole your portfolio is not down as much as HFEA. Those have been the two biggest points I've tried to make without repeating too often. I assume that to maintain your original risk budget you've been rebalancing into the leveraged parts of your portfolio? I have been selling vti and buying upro in my 401k to maintain the same leverage as a year ago ( actually increased leverage from 1.35x to 1.55x per lifecycle investing principles).
You actively buy more equities even after they declined and already constitute a higher percentage of NAV from the decline? What is your current rebalancing algo and asset allocation if I may ask?
comeinvest
Posts: 2709
Joined: Mon Mar 12, 2012 6:57 pm

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

Post by comeinvest »

unemployed_pysicist wrote: Tue Sep 13, 2022 2:29 pm
skierincolorado wrote: Tue Sep 13, 2022 8:48 am
unemployed_pysicist wrote: Mon Sep 12, 2022 2:09 am
skierincolorado wrote: Sun Sep 11, 2022 7:44 pm
That's because long term interest rates fell dramatically over that period.
There are share classes that have been around longer. If you start in 2009 (dec 31 2008) you get a period of neae neutral rates. During this period the sharpe ratio is much higher for vglt than edv .26 vs .18.

Any period in which rates don't fall, the sharpe for vglt will be much higher. Over long time periods like 30 or 50 years this matters less (even if rates fall they are falling less per year).
Which share class? VGLSX does not seem to go back any further than VGLT. Comparing VEDTX (EDV) and TLT as another comparison gives a similar Sharpe ratio of about 0.31 for TLT and 0.3 for VEDTX from Dec 2007 to present (almost 15 years.)

Perhaps you are referring to VUSTX? VUSTX is actively managed and different holdings so not the same as VGLT. Even then the Sharpe is .34 for VUSTX vs .3 for VEDTX from Dec 2007 to present.
Again, from Dec 2007 to present long term interest rates fell 1%. Starting Dec 31 2008 produces a period of flat rates, which produces a fair comparison.

As comeinvest pointed out the carry on the 30 y is abysmal and has been very poor historically as well.

If you go back longer the sharpe on the longest maturities is terrible. Or pick a period of near flat rates. Picking relatively short periods with large interest rate decreases is going to make the longest maturities look artificially good due to the large capital gains they experience in such periods.
In the period from Jan 28 2009 to August 30 2022, DGS30 started at 3.44%, and ended at 3.23%, which is a 21 bps fall, as compared to the dates you suggested:
Dec 29 2008 to August 30 2022, where DGS30 started at 2.63% and ended at 3.23%, which is a 60 bps rise (further from 0 than 21 bps.)

The period from Jan 28 2009 to August 30 2022 gives Sharpe of .31 for VUSTX, versus .27 for VEDTX. VEDTX underperformed VUSTX for this period of ~14 years, but I would not characterize this as "terrible".

In the long run, I do not disagree that VGLT is likely to outperform EDV on a risk-adjusted basis. However, I am pushing back against the idea that EDV is somehow "terrible" or unusable for this strategy. EDV is a reasonable option for the leveraged-constrained investor who wants to take significant duration risk. VGLT and EDV are both reasonable, depending on risk appetite.
Not directly related to this conversation, but just as a reminder, I think equal yields at the beginning and at the end of an observation period do not guarantee a fair comparison, as I pointed out in a post in the mHFEA thread recently, different prospective yield changes may already be implied at different points on the curve, which would not necessarily be captured by holding treasuries at constant maturity or treasury futures on a rolling basis. I made that mistake before when I tried to verify the performance of certain yield curve strategies, and ended up scratching my head why the numbers don't add up.
comeinvest
Posts: 2709
Joined: Mon Mar 12, 2012 6:57 pm

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

Post by comeinvest »

unemployed_pysicist wrote: Tue Sep 13, 2022 1:53 pm I originally found the link to the Understanding the Yield Curve pdf in a post on stack exchange; I will see if I can find it again.
Thank you very much for pointing out convexity. I found the individual chapters here: https://vdoc.pub/search/Understanding%2 ... %20ilmanen
but I didn't find the entire book to download. Please let me know if you can locate a link.
unemployed_pysicist
Posts: 220
Joined: Sat Oct 09, 2021 2:32 pm
Location: Amsterdam

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

Post by unemployed_pysicist »

For anyone interested in Antti Ilmainen's Understanding the Yield Curve report, here is a link to a copy of the full document:

https://pdfcoffee.com/understanding-the ... -free.html

To the mods: please delete this post if this violates forum guidelines in any way.

Thanks
couldn't afford the h | | BUY BONDS | WEAR DIAMONDS
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

comeinvest wrote: Tue Sep 13, 2022 11:40 pm
skierincolorado wrote: Tue Sep 13, 2022 2:59 pm I am glad to hear it and hope that when looked at as a whole your portfolio is not down as much as HFEA. Those have been the two biggest points I've tried to make without repeating too often. I assume that to maintain your original risk budget you've been rebalancing into the leveraged parts of your portfolio? I have been selling vti and buying upro in my 401k to maintain the same leverage as a year ago ( actually increased leverage from 1.35x to 1.55x per lifecycle investing principles).
You actively buy more equities even after they declined and already constitute a higher percentage of NAV from the decline? What is your current rebalancing algo and asset allocation if I may ask?
When you hold upro as well as unleveraged funds, leverage decreases when the market declines. My rebalancing strategy is to maintain a fixed amount borrowed. I've increased it very slightly as I've made contributions and the market has come down and my income has increased, but the amount I have borrowed is within 15% of what it was a year ago.
Last edited by skierincolorado on Wed Sep 14, 2022 11:17 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 »

unemployed_pysicist wrote: Tue Sep 13, 2022 2:29 pm
skierincolorado wrote: Tue Sep 13, 2022 8:48 am
unemployed_pysicist wrote: Mon Sep 12, 2022 2:09 am
skierincolorado wrote: Sun Sep 11, 2022 7:44 pm
That's because long term interest rates fell dramatically over that period.
There are share classes that have been around longer. If you start in 2009 (dec 31 2008) you get a period of neae neutral rates. During this period the sharpe ratio is much higher for vglt than edv .26 vs .18.

Any period in which rates don't fall, the sharpe for vglt will be much higher. Over long time periods like 30 or 50 years this matters less (even if rates fall they are falling less per year).
Which share class? VGLSX does not seem to go back any further than VGLT. Comparing VEDTX (EDV) and TLT as another comparison gives a similar Sharpe ratio of about 0.31 for TLT and 0.3 for VEDTX from Dec 2007 to present (almost 15 years.)

Perhaps you are referring to VUSTX? VUSTX is actively managed and different holdings so not the same as VGLT. Even then the Sharpe is .34 for VUSTX vs .3 for VEDTX from Dec 2007 to present.
Again, from Dec 2007 to present long term interest rates fell 1%. Starting Dec 31 2008 produces a period of flat rates, which produces a fair comparison.

As comeinvest pointed out the carry on the 30 y is abysmal and has been very poor historically as well.

If you go back longer the sharpe on the longest maturities is terrible. Or pick a period of near flat rates. Picking relatively short periods with large interest rate decreases is going to make the longest maturities look artificially good due to the large capital gains they experience in such periods.
In the period from Jan 28 2009 to August 30 2022, DGS30 started at 3.44%, and ended at 3.23%, which is a 21 bps fall, as compared to the dates you suggested:
Dec 29 2008 to August 30 2022, where DGS30 started at 2.63% and ended at 3.23%, which is a 60 bps rise (further from 0 than 21 bps.)

The period from Jan 28 2009 to August 30 2022 gives Sharpe of .31 for VUSTX, versus .27 for VEDTX. VEDTX underperformed VUSTX for this period of ~14 years, but I would not characterize this as "terrible".

In the long run, I do not disagree that VGLT is likely to outperform EDV on a risk-adjusted basis. However, I am pushing back against the idea that EDV is somehow "terrible" or unusable for this strategy. EDV is a reasonable option for the leveraged-constrained investor who wants to take significant duration risk. VGLT and EDV are both reasonable, depending on risk appetite.
I missed the beginning part of your post. Notably in both cases the sharpe of edv was worse, the question is by how much. In your case the sharpe for vustx is about 15% better than edv. In my case it is 40% better. So the expected risk adjusted returns of vustx is likely 15-40% better.

I'd also point out, as comeinvest also pointed out, that investors expected long term rates to increase in jan 2009 as indicated by the steepness of the curve relative to a low term premium. That increase failed to materialize and so long term bonds and especially the longest term bonds have returned more than expected. If long rates had increased as expected, edv would have looked worse in comparison to vustx.

While your scenario is closer to flat rates, my scenario is closer to how investors expected long term interest rates to change. I would generally expect the sharpe of vustx to be 30 to 40% higher than edv, which makes edv uninvestable in my opinion. The sharpe of 5 to 10 year bonds I'd expect (and had been) to be even higher than 20 year. Bonds of 15 to 20 years are barely investable in my opinion and over 20 years are useless.

Also remember long term bonds are a small market that is likely distorted by pension fund demand for long dates maturities. I think it's better to stick to the broader, more competitive, and higher performing intermediate term bond market.
langlands
Posts: 1093
Joined: Wed Apr 03, 2019 10:05 pm

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

Post by langlands »

Anyone including a leveraged oil index in their portfolio (e.g. NRGU)? Seems to make more sense to me than TMF in this climate.
comeinvest
Posts: 2709
Joined: Mon Mar 12, 2012 6:57 pm

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

Post by comeinvest »

skierincolorado wrote: Wed Sep 14, 2022 7:13 pm
unemployed_pysicist wrote: Tue Sep 13, 2022 2:29 pm
skierincolorado wrote: Tue Sep 13, 2022 8:48 am
unemployed_pysicist wrote: Mon Sep 12, 2022 2:09 am
skierincolorado wrote: Sun Sep 11, 2022 7:44 pm
That's because long term interest rates fell dramatically over that period.
There are share classes that have been around longer. If you start in 2009 (dec 31 2008) you get a period of neae neutral rates. During this period the sharpe ratio is much higher for vglt than edv .26 vs .18.

Any period in which rates don't fall, the sharpe for vglt will be much higher. Over long time periods like 30 or 50 years this matters less (even if rates fall they are falling less per year).
Which share class? VGLSX does not seem to go back any further than VGLT. Comparing VEDTX (EDV) and TLT as another comparison gives a similar Sharpe ratio of about 0.31 for TLT and 0.3 for VEDTX from Dec 2007 to present (almost 15 years.)

Perhaps you are referring to VUSTX? VUSTX is actively managed and different holdings so not the same as VGLT. Even then the Sharpe is .34 for VUSTX vs .3 for VEDTX from Dec 2007 to present.
Again, from Dec 2007 to present long term interest rates fell 1%. Starting Dec 31 2008 produces a period of flat rates, which produces a fair comparison.

As comeinvest pointed out the carry on the 30 y is abysmal and has been very poor historically as well.

If you go back longer the sharpe on the longest maturities is terrible. Or pick a period of near flat rates. Picking relatively short periods with large interest rate decreases is going to make the longest maturities look artificially good due to the large capital gains they experience in such periods.
In the period from Jan 28 2009 to August 30 2022, DGS30 started at 3.44%, and ended at 3.23%, which is a 21 bps fall, as compared to the dates you suggested:
Dec 29 2008 to August 30 2022, where DGS30 started at 2.63% and ended at 3.23%, which is a 60 bps rise (further from 0 than 21 bps.)

The period from Jan 28 2009 to August 30 2022 gives Sharpe of .31 for VUSTX, versus .27 for VEDTX. VEDTX underperformed VUSTX for this period of ~14 years, but I would not characterize this as "terrible".

In the long run, I do not disagree that VGLT is likely to outperform EDV on a risk-adjusted basis. However, I am pushing back against the idea that EDV is somehow "terrible" or unusable for this strategy. EDV is a reasonable option for the leveraged-constrained investor who wants to take significant duration risk. VGLT and EDV are both reasonable, depending on risk appetite.
I missed the beginning part of your post. Notably in both cases the sharpe of edv was worse, the question is by how much. In your case the sharpe for vustx is about 15% better than edv. In my case it is 40% better. So the expected risk adjusted returns of vustx is likely 15-40% better.

I'd also point out, as comeinvest also pointed out, that investors expected long term rates to increase in jan 2009 as indicated by the steepness of the curve relative to a low term premium. That increase failed to materialize and so long term bonds and especially the longest term bonds have returned more than expected. If long rates had increased as expected, edv would have looked worse in comparison to vustx.

While your scenario is closer to flat rates, my scenario is closer to how investors expected long term interest rates to change. I would generally expect the sharpe of vustx to be 30 to 40% higher than edv, which makes edv uninvestable in my opinion. The sharpe of 5 to 10 year bonds I'd expect (and had been) to be even higher than 20 year. Bonds of 15 to 20 years are barely investable in my opinion and over 20 years are useless.

Also remember long term bonds are a small market that is likely distorted by pension fund demand for long dates maturities. I think it's better to stick to the broader, more competitive, and higher performing intermediate term bond market.
All that is convincing from a long-term strategic view, agnostic of any tactical views of the yield curve. Regarding dynamic positioning on the yield curve, last year I thought I would move on a duration-adjusted basis to the 2y and 5y from the 5y, 7y, and 10y, if and when the curve ever inverts. I abandoned that idea for the moment, as I learned that you cannot necessarily capture reversion to the mean of the geometric structure of the curve by holding treasuries or futures, as some expectations are already embedded in the curve and reflected in the roll yields. It looks like ITT are much better now than last year in comparison to LTT due to the flat curve, even though nothing can be said for sure - with the efficient market hypothesis and the expectations hypothesis one could argue that nothing changed in terms of expected returns, and that the market just expects higher short-term rates for the next 5-10 years or so than it expected last year.
comeinvest
Posts: 2709
Joined: Mon Mar 12, 2012 6:57 pm

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

Post by comeinvest »

skierincolorado wrote: Wed Sep 14, 2022 7:13 pm
unemployed_pysicist wrote: Tue Sep 13, 2022 2:29 pm
skierincolorado wrote: Tue Sep 13, 2022 8:48 am
unemployed_pysicist wrote: Mon Sep 12, 2022 2:09 am
skierincolorado wrote: Sun Sep 11, 2022 7:44 pm
That's because long term interest rates fell dramatically over that period.
There are share classes that have been around longer. If you start in 2009 (dec 31 2008) you get a period of neae neutral rates. During this period the sharpe ratio is much higher for vglt than edv .26 vs .18.

Any period in which rates don't fall, the sharpe for vglt will be much higher. Over long time periods like 30 or 50 years this matters less (even if rates fall they are falling less per year).
Which share class? VGLSX does not seem to go back any further than VGLT. Comparing VEDTX (EDV) and TLT as another comparison gives a similar Sharpe ratio of about 0.31 for TLT and 0.3 for VEDTX from Dec 2007 to present (almost 15 years.)

Perhaps you are referring to VUSTX? VUSTX is actively managed and different holdings so not the same as VGLT. Even then the Sharpe is .34 for VUSTX vs .3 for VEDTX from Dec 2007 to present.
Again, from Dec 2007 to present long term interest rates fell 1%. Starting Dec 31 2008 produces a period of flat rates, which produces a fair comparison.

As comeinvest pointed out the carry on the 30 y is abysmal and has been very poor historically as well.

If you go back longer the sharpe on the longest maturities is terrible. Or pick a period of near flat rates. Picking relatively short periods with large interest rate decreases is going to make the longest maturities look artificially good due to the large capital gains they experience in such periods.
In the period from Jan 28 2009 to August 30 2022, DGS30 started at 3.44%, and ended at 3.23%, which is a 21 bps fall, as compared to the dates you suggested:
Dec 29 2008 to August 30 2022, where DGS30 started at 2.63% and ended at 3.23%, which is a 60 bps rise (further from 0 than 21 bps.)

The period from Jan 28 2009 to August 30 2022 gives Sharpe of .31 for VUSTX, versus .27 for VEDTX. VEDTX underperformed VUSTX for this period of ~14 years, but I would not characterize this as "terrible".

In the long run, I do not disagree that VGLT is likely to outperform EDV on a risk-adjusted basis. However, I am pushing back against the idea that EDV is somehow "terrible" or unusable for this strategy. EDV is a reasonable option for the leveraged-constrained investor who wants to take significant duration risk. VGLT and EDV are both reasonable, depending on risk appetite.
I missed the beginning part of your post. Notably in both cases the sharpe of edv was worse, the question is by how much. In your case the sharpe for vustx is about 15% better than edv. In my case it is 40% better. So the expected risk adjusted returns of vustx is likely 15-40% better.

I'd also point out, as comeinvest also pointed out, that investors expected long term rates to increase in jan 2009 as indicated by the steepness of the curve relative to a low term premium. That increase failed to materialize and so long term bonds and especially the longest term bonds have returned more than expected. If long rates had increased as expected, edv would have looked worse in comparison to vustx.

While your scenario is closer to flat rates, my scenario is closer to how investors expected long term interest rates to change. I would generally expect the sharpe of vustx to be 30 to 40% higher than edv, which makes edv uninvestable in my opinion. The sharpe of 5 to 10 year bonds I'd expect (and had been) to be even higher than 20 year. Bonds of 15 to 20 years are barely investable in my opinion and over 20 years are useless.

Also remember long term bonds are a small market that is likely distorted by pension fund demand for long dates maturities. I think it's better to stick to the broader, more competitive, and higher performing intermediate term bond market.
A little anecdote: I just discovered that I earned money long the 20y short the 30y treasury futures (duration neutral), even during a time since beginning / mid April to now, when the curve flattened some more in that segment. Kind of surprising.
unemployed_pysicist
Posts: 220
Joined: Sat Oct 09, 2021 2:32 pm
Location: Amsterdam

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

Post by unemployed_pysicist »

skierincolorado wrote: Wed Sep 14, 2022 7:13 pm
unemployed_pysicist wrote: Tue Sep 13, 2022 2:29 pm
skierincolorado wrote: Tue Sep 13, 2022 8:48 am
unemployed_pysicist wrote: Mon Sep 12, 2022 2:09 am
skierincolorado wrote: Sun Sep 11, 2022 7:44 pm
That's because long term interest rates fell dramatically over that period.
There are share classes that have been around longer. If you start in 2009 (dec 31 2008) you get a period of neae neutral rates. During this period the sharpe ratio is much higher for vglt than edv .26 vs .18.

Any period in which rates don't fall, the sharpe for vglt will be much higher. Over long time periods like 30 or 50 years this matters less (even if rates fall they are falling less per year).
Which share class? VGLSX does not seem to go back any further than VGLT. Comparing VEDTX (EDV) and TLT as another comparison gives a similar Sharpe ratio of about 0.31 for TLT and 0.3 for VEDTX from Dec 2007 to present (almost 15 years.)

Perhaps you are referring to VUSTX? VUSTX is actively managed and different holdings so not the same as VGLT. Even then the Sharpe is .34 for VUSTX vs .3 for VEDTX from Dec 2007 to present.
Again, from Dec 2007 to present long term interest rates fell 1%. Starting Dec 31 2008 produces a period of flat rates, which produces a fair comparison.

As comeinvest pointed out the carry on the 30 y is abysmal and has been very poor historically as well.

If you go back longer the sharpe on the longest maturities is terrible. Or pick a period of near flat rates. Picking relatively short periods with large interest rate decreases is going to make the longest maturities look artificially good due to the large capital gains they experience in such periods.
In the period from Jan 28 2009 to August 30 2022, DGS30 started at 3.44%, and ended at 3.23%, which is a 21 bps fall, as compared to the dates you suggested:
Dec 29 2008 to August 30 2022, where DGS30 started at 2.63% and ended at 3.23%, which is a 60 bps rise (further from 0 than 21 bps.)

The period from Jan 28 2009 to August 30 2022 gives Sharpe of .31 for VUSTX, versus .27 for VEDTX. VEDTX underperformed VUSTX for this period of ~14 years, but I would not characterize this as "terrible".

In the long run, I do not disagree that VGLT is likely to outperform EDV on a risk-adjusted basis. However, I am pushing back against the idea that EDV is somehow "terrible" or unusable for this strategy. EDV is a reasonable option for the leveraged-constrained investor who wants to take significant duration risk. VGLT and EDV are both reasonable, depending on risk appetite.
I missed the beginning part of your post. Notably in both cases the sharpe of edv was worse, the question is by how much. In your case the sharpe for vustx is about 15% better than edv. In my case it is 40% better. So the expected risk adjusted returns of vustx is likely 15-40% better.

I'd also point out, as comeinvest also pointed out, that investors expected long term rates to increase in jan 2009 as indicated by the steepness of the curve relative to a low term premium. That increase failed to materialize and so long term bonds and especially the longest term bonds have returned more than expected. If long rates had increased as expected, edv would have looked worse in comparison to vustx.

While your scenario is closer to flat rates, my scenario is closer to how investors expected long term interest rates to change. I would generally expect the sharpe of vustx to be 30 to 40% higher than edv, which makes edv uninvestable in my opinion. The sharpe of 5 to 10 year bonds I'd expect (and had been) to be even higher than 20 year. Bonds of 15 to 20 years are barely investable in my opinion and over 20 years are useless.

Also remember long term bonds are a small market that is likely distorted by pension fund demand for long dates maturities. I think it's better to stick to the broader, more competitive, and higher performing intermediate term bond market.
I was editing a mistake in my post during your first response when you wrote your reply, which is why you missed it. My apologies.

You can also change the start date to November 28 2008 (1 month earlier) where DGS30 was 3.45. The Sharpes were .3 for VUSTX and .25 for VEDTX.

I would like to discuss these points:

"If long rates had increased as expected, edv would have looked worse in comparison to vustx.

While your scenario is closer to flat rates, my scenario is closer to how investors expected long term interest rates to change."

This sounds like you are saying a bond investor's default outlook should be: assume that the forward rates will be realized. Is this correct or am I misunderstanding you?
couldn't afford the h | | BUY BONDS | WEAR DIAMONDS
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

unemployed_pysicist wrote: Fri Sep 16, 2022 1:29 am
skierincolorado wrote: Wed Sep 14, 2022 7:13 pm
unemployed_pysicist wrote: Tue Sep 13, 2022 2:29 pm
skierincolorado wrote: Tue Sep 13, 2022 8:48 am
unemployed_pysicist wrote: Mon Sep 12, 2022 2:09 am

Which share class? VGLSX does not seem to go back any further than VGLT. Comparing VEDTX (EDV) and TLT as another comparison gives a similar Sharpe ratio of about 0.31 for TLT and 0.3 for VEDTX from Dec 2007 to present (almost 15 years.)

Perhaps you are referring to VUSTX? VUSTX is actively managed and different holdings so not the same as VGLT. Even then the Sharpe is .34 for VUSTX vs .3 for VEDTX from Dec 2007 to present.
Again, from Dec 2007 to present long term interest rates fell 1%. Starting Dec 31 2008 produces a period of flat rates, which produces a fair comparison.

As comeinvest pointed out the carry on the 30 y is abysmal and has been very poor historically as well.

If you go back longer the sharpe on the longest maturities is terrible. Or pick a period of near flat rates. Picking relatively short periods with large interest rate decreases is going to make the longest maturities look artificially good due to the large capital gains they experience in such periods.
In the period from Jan 28 2009 to August 30 2022, DGS30 started at 3.44%, and ended at 3.23%, which is a 21 bps fall, as compared to the dates you suggested:
Dec 29 2008 to August 30 2022, where DGS30 started at 2.63% and ended at 3.23%, which is a 60 bps rise (further from 0 than 21 bps.)

The period from Jan 28 2009 to August 30 2022 gives Sharpe of .31 for VUSTX, versus .27 for VEDTX. VEDTX underperformed VUSTX for this period of ~14 years, but I would not characterize this as "terrible".

In the long run, I do not disagree that VGLT is likely to outperform EDV on a risk-adjusted basis. However, I am pushing back against the idea that EDV is somehow "terrible" or unusable for this strategy. EDV is a reasonable option for the leveraged-constrained investor who wants to take significant duration risk. VGLT and EDV are both reasonable, depending on risk appetite.
I missed the beginning part of your post. Notably in both cases the sharpe of edv was worse, the question is by how much. In your case the sharpe for vustx is about 15% better than edv. In my case it is 40% better. So the expected risk adjusted returns of vustx is likely 15-40% better.

I'd also point out, as comeinvest also pointed out, that investors expected long term rates to increase in jan 2009 as indicated by the steepness of the curve relative to a low term premium. That increase failed to materialize and so long term bonds and especially the longest term bonds have returned more than expected. If long rates had increased as expected, edv would have looked worse in comparison to vustx.

While your scenario is closer to flat rates, my scenario is closer to how investors expected long term interest rates to change. I would generally expect the sharpe of vustx to be 30 to 40% higher than edv, which makes edv uninvestable in my opinion. The sharpe of 5 to 10 year bonds I'd expect (and had been) to be even higher than 20 year. Bonds of 15 to 20 years are barely investable in my opinion and over 20 years are useless.

Also remember long term bonds are a small market that is likely distorted by pension fund demand for long dates maturities. I think it's better to stick to the broader, more competitive, and higher performing intermediate term bond market.
I was editing a mistake in my post during your first response when you wrote your reply, which is why you missed it. My apologies.

You can also change the start date to November 28 2008 (1 month earlier) where DGS30 was 3.45. The Sharpes were .3 for VUSTX and .25 for VEDTX.

I would like to discuss these points:

"If long rates had increased as expected, edv would have looked worse in comparison to vustx.

While your scenario is closer to flat rates, my scenario is closer to how investors expected long term interest rates to change."

This sounds like you are saying a bond investor's default outlook should be: assume that the forward rates will be realized. Is this correct or am I misunderstanding you?
I think forward rates include the term premium, which cannot be known precisely. But I do think it is clear from forward rates that investors expected longer rates to rise in jan 2009. The fact that that failed to materialize was to the benefit of EDV. Normally the curve would not be so steep and 30y bonds would not yield quite so much.

So I do think the Dec 31 start date is more "fair". It's possible to extend this back to the 1970s so that start dates matter less. I haven't done that for the 30y specifically, but longer durations in general have lower sharpes and I am confident from the very low carry of 30y bonds that the sharpe would be quite poor. 30 y bonds yield very little per year of duration.
comeinvest
Posts: 2709
Joined: Mon Mar 12, 2012 6:57 pm

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

Post by comeinvest »

skierincolorado wrote: Fri Sep 16, 2022 3:02 pm I think forward rates include the term premium, which cannot be known precisely. But I do think it is clear from forward rates that investors expected longer rates to rise in jan 2009. The fact that that failed to materialize was to the benefit of EDV. Normally the curve would not be so steep and 30y bonds would not yield quite so much.

So I do think the Dec 31 start date is more "fair". It's possible to extend this back to the 1970s so that start dates matter less. I haven't done that for the 30y specifically, but longer durations in general have lower sharpes and I am confident from the very low carry of 30y bonds that the sharpe would be quite poor. 30 y bonds yield very little per year of duration.
Unfortunately I'm not an interest rates scientist, but I think conceptually it can be said that any rate changes expected and therefore implied by the shape of the curve, won't translate to treasury bond returns performance, as they are already "baked into" forward rates (and thereby into the yield curve) and therefore will be reflected in the calendar rolls when you keep your maturities constant, which won't translate into investment returns. I think only current YTM, roll-down return from the slope of the curve, term premia, and unexpected rate changes translate to bond returns. Slope and term premia are somewhat related (in the long run, without slope no term premium and vice versa), although not in a direct way. I'm sure this can be all expressed with differential equations.
User avatar
hillclimber
Posts: 136
Joined: Wed Mar 10, 2021 2:32 pm

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

Post by hillclimber »

langlands wrote: Wed Sep 14, 2022 8:19 pm Anyone including a leveraged oil index in their portfolio (e.g. NRGU)? Seems to make more sense to me than TMF in this climate.
Oil generally doesn't do well in crashes, so it wouldn't be a great substitute for TMF. Also, I'd be careful with NRGU because that's an ETN.
rascott
Posts: 2957
Joined: Wed Apr 15, 2015 10:53 am

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

Post by rascott »

I've been out of this portfolio for almost 2 years now..... so not really paying attention to it.... but what kind of bloodbath has it been YTD? Down 50%?
Jags4186
Posts: 8198
Joined: Wed Jun 18, 2014 7:12 pm

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

Post by Jags4186 »

rascott wrote: Mon Sep 19, 2022 10:54 am I've been out of this portfolio for almost 2 years now..... so not really paying attention to it.... but what kind of bloodbath has it been YTD? Down 50%?
Not sure about YTD but the 1 year return is -60%. I'm back to where I started--which means I'm down vs. being 100% SP500 the last 2 years. Debating whether to cut my losses or soldier on.
ChinchillaWhiplash
Posts: 1574
Joined: Sat Jan 20, 2018 4:40 pm
Location: Land of Hypoxia

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

Post by ChinchillaWhiplash »

Question about the equity fund makeup. Would it make sense to diversify (using this word loosely) the S&P500 portion of this by using UPRO and SPXL? Looking at the holdings of these 2 funds show that they are quite different but have about the same results and ER. Would hope that both funds are pretty safe to hold in the sense that they won’t just go bust but…. Diversification is usually not a bad thing. Don’t know if the same holds true here or not. Anyone ever dig into what makes up these funds and do a comparison? Sorry if this is somewhere in this thread but it is a bit too much to dig through :shock:
Marseille07
Posts: 16054
Joined: Fri Nov 06, 2020 12:41 pm

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

Post by Marseille07 »

ChinchillaWhiplash wrote: Mon Sep 19, 2022 2:59 pm Question about the equity fund makeup. Would it make sense to diversify (using this word loosely) the S&P500 portion of this by using UPRO and SPXL? Looking at the holdings of these 2 funds show that they are quite different but have about the same results and ER. Would hope that both funds are pretty safe to hold in the sense that they won’t just go bust but…. Diversification is usually not a bad thing. Don’t know if the same holds true here or not. Anyone ever dig into what makes up these funds and do a comparison? Sorry if this is somewhere in this thread but it is a bit too much to dig through :shock:
Imo, not really. I believe they both track 3x S&P however they do it. It's interesting UPRO holds a bunch of swaps though.
Tamalak
Posts: 1988
Joined: Fri May 06, 2016 2:29 pm

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

Post by Tamalak »

On one hand fed rates being high gives this portfolio more cover (much more than it had at the beginning of the year). On the other hand it means the costs of borrowing goes up which is a drag on leveraged portfolios.. there's just no free lunch :annoyed
unemployed_pysicist
Posts: 220
Joined: Sat Oct 09, 2021 2:32 pm
Location: Amsterdam

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

Post by unemployed_pysicist »

skierincolorado wrote: Fri Sep 16, 2022 3:02 pm
unemployed_pysicist wrote: Fri Sep 16, 2022 1:29 am
skierincolorado wrote: Wed Sep 14, 2022 7:13 pm
unemployed_pysicist wrote: Tue Sep 13, 2022 2:29 pm
skierincolorado wrote: Tue Sep 13, 2022 8:48 am

Again, from Dec 2007 to present long term interest rates fell 1%. Starting Dec 31 2008 produces a period of flat rates, which produces a fair comparison.

As comeinvest pointed out the carry on the 30 y is abysmal and has been very poor historically as well.

If you go back longer the sharpe on the longest maturities is terrible. Or pick a period of near flat rates. Picking relatively short periods with large interest rate decreases is going to make the longest maturities look artificially good due to the large capital gains they experience in such periods.
In the period from Jan 28 2009 to August 30 2022, DGS30 started at 3.44%, and ended at 3.23%, which is a 21 bps fall, as compared to the dates you suggested:
Dec 29 2008 to August 30 2022, where DGS30 started at 2.63% and ended at 3.23%, which is a 60 bps rise (further from 0 than 21 bps.)

The period from Jan 28 2009 to August 30 2022 gives Sharpe of .31 for VUSTX, versus .27 for VEDTX. VEDTX underperformed VUSTX for this period of ~14 years, but I would not characterize this as "terrible".

In the long run, I do not disagree that VGLT is likely to outperform EDV on a risk-adjusted basis. However, I am pushing back against the idea that EDV is somehow "terrible" or unusable for this strategy. EDV is a reasonable option for the leveraged-constrained investor who wants to take significant duration risk. VGLT and EDV are both reasonable, depending on risk appetite.
I missed the beginning part of your post. Notably in both cases the sharpe of edv was worse, the question is by how much. In your case the sharpe for vustx is about 15% better than edv. In my case it is 40% better. So the expected risk adjusted returns of vustx is likely 15-40% better.

I'd also point out, as comeinvest also pointed out, that investors expected long term rates to increase in jan 2009 as indicated by the steepness of the curve relative to a low term premium. That increase failed to materialize and so long term bonds and especially the longest term bonds have returned more than expected. If long rates had increased as expected, edv would have looked worse in comparison to vustx.

While your scenario is closer to flat rates, my scenario is closer to how investors expected long term interest rates to change. I would generally expect the sharpe of vustx to be 30 to 40% higher than edv, which makes edv uninvestable in my opinion. The sharpe of 5 to 10 year bonds I'd expect (and had been) to be even higher than 20 year. Bonds of 15 to 20 years are barely investable in my opinion and over 20 years are useless.

Also remember long term bonds are a small market that is likely distorted by pension fund demand for long dates maturities. I think it's better to stick to the broader, more competitive, and higher performing intermediate term bond market.
I was editing a mistake in my post during your first response when you wrote your reply, which is why you missed it. My apologies.

You can also change the start date to November 28 2008 (1 month earlier) where DGS30 was 3.45. The Sharpes were .3 for VUSTX and .25 for VEDTX.

I would like to discuss these points:

"If long rates had increased as expected, edv would have looked worse in comparison to vustx.

While your scenario is closer to flat rates, my scenario is closer to how investors expected long term interest rates to change."

This sounds like you are saying a bond investor's default outlook should be: assume that the forward rates will be realized. Is this correct or am I misunderstanding you?
I think forward rates include the term premium, which cannot be known precisely. But I do think it is clear from forward rates that investors expected longer rates to rise in jan 2009. The fact that that failed to materialize was to the benefit of EDV. Normally the curve would not be so steep and 30y bonds would not yield quite so much.

So I do think the Dec 31 start date is more "fair". It's possible to extend this back to the 1970s so that start dates matter less. I haven't done that for the 30y specifically, but longer durations in general have lower sharpes and I am confident from the very low carry of 30y bonds that the sharpe would be quite poor. 30 y bonds yield very little per year of duration.
If the n-year forward rates had been realized, then the return of all bonds would have been the same over the n-year holding period.

As far as which scenario better predicted the path of future rates, we can look at what both scenarios would have predicted the yield curve to look like in August 2022 if we assumed that the forward rates themselves correctly forecast future rate changes (no term premium.) The yield curve goes out to 30 years, so I restrict the predicted zero yield curve to a maximum of 16 years maturity (about 14 years has elapsed between end of 2008 and August 2022.)

Here are the curves on these dates:
Image

Image

And here is the curve as of August 31 2022:
Image

Interestingly, the forward rates from your scenario are actually quite close to the front end of the yield curve that we see as of August 2022. I.e., your December 2008 scenario was better at predicting where the 0-6 year yields would end up. However, the longer end is far too low. In my scenario, the predicted zero coupon curve is closer to the longer end of the yield curve for the 10-16 year yields; my scenario was better for predicting where long term yields ended up. The predicted curves from both scenarios are significantly inverted. As far as I can tell, long term yields were "predicted" by the yield curve to fall in both scenarios.

In my backtesting, I have generally found that the highest Sharpe ratios belong to the STTs (2-3 years), then ITTs, and so on in that order for progressively longer maturities. Levering the duration with the shorter maturities involves futures contracts or margining bonds, which I presume many HFEA followers are reluctant to do. That leaves the following options for significant duration exposure: TMF (as was initially proposed), EDV/ZROZ/GOVZ, TYD, TYA, TLT, VGLT and VUSTX. VGLT (10-25 years maturity bonds) and VUSTX (15-30 years to maturity bonds) are good options if ~17 years duration is enough. TLT (20-30 years maturity bonds) outperforms VEDTX by 0.01 in Sharpe for the 2008 December to 2022 August scenario; the performance improves considerably if one ends the backtest around March 2020, which happens to be an example of where lower maturities outperform on a duration-adjusted basis (recession and sharply falling rates.)

I am skeptical of the prospects for TYD to significantly outperform EDV over a prolonged period of neutral rates. The cost of leverage and fees are a considerable impediment to its performance; it looks like TYD often loses 0.1 or more in Sharpe to its underlying index (which IEF also tracks.) Maybe a "neutral rate" backtesting opportunity will appear if rates get to August 2000 or May 2005 levels. TYD probably outperforms EDV in a sustained falling rate environment, but I have not backtested TYD earlier than what is available in portfolio visualizer yet. I am even more skeptical of TMF outperforming EDV in a neutral rate scenario.

TYA looks quite reasonable.
couldn't afford the h | | BUY BONDS | WEAR DIAMONDS
User avatar
hiddenpower
Posts: 564
Joined: Tue Nov 17, 2020 11:24 pm

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

Post by hiddenpower »

unemployed_pysicist wrote: Mon Sep 19, 2022 4:52 pm
skierincolorado wrote: Fri Sep 16, 2022 3:02 pm
unemployed_pysicist wrote: Fri Sep 16, 2022 1:29 am
skierincolorado wrote: Wed Sep 14, 2022 7:13 pm
unemployed_pysicist wrote: Tue Sep 13, 2022 2:29 pm

In the period from Jan 28 2009 to August 30 2022, DGS30 started at 3.44%, and ended at 3.23%, which is a 21 bps fall, as compared to the dates you suggested:
Dec 29 2008 to August 30 2022, where DGS30 started at 2.63% and ended at 3.23%, which is a 60 bps rise (further from 0 than 21 bps.)

The period from Jan 28 2009 to August 30 2022 gives Sharpe of .31 for VUSTX, versus .27 for VEDTX. VEDTX underperformed VUSTX for this period of ~14 years, but I would not characterize this as "terrible".

In the long run, I do not disagree that VGLT is likely to outperform EDV on a risk-adjusted basis. However, I am pushing back against the idea that EDV is somehow "terrible" or unusable for this strategy. EDV is a reasonable option for the leveraged-constrained investor who wants to take significant duration risk. VGLT and EDV are both reasonable, depending on risk appetite.
I missed the beginning part of your post. Notably in both cases the sharpe of edv was worse, the question is by how much. In your case the sharpe for vustx is about 15% better than edv. In my case it is 40% better. So the expected risk adjusted returns of vustx is likely 15-40% better.

I'd also point out, as comeinvest also pointed out, that investors expected long term rates to increase in jan 2009 as indicated by the steepness of the curve relative to a low term premium. That increase failed to materialize and so long term bonds and especially the longest term bonds have returned more than expected. If long rates had increased as expected, edv would have looked worse in comparison to vustx.

While your scenario is closer to flat rates, my scenario is closer to how investors expected long term interest rates to change. I would generally expect the sharpe of vustx to be 30 to 40% higher than edv, which makes edv uninvestable in my opinion. The sharpe of 5 to 10 year bonds I'd expect (and had been) to be even higher than 20 year. Bonds of 15 to 20 years are barely investable in my opinion and over 20 years are useless.

Also remember long term bonds are a small market that is likely distorted by pension fund demand for long dates maturities. I think it's better to stick to the broader, more competitive, and higher performing intermediate term bond market.
I was editing a mistake in my post during your first response when you wrote your reply, which is why you missed it. My apologies.

You can also change the start date to November 28 2008 (1 month earlier) where DGS30 was 3.45. The Sharpes were .3 for VUSTX and .25 for VEDTX.

I would like to discuss these points:

"If long rates had increased as expected, edv would have looked worse in comparison to vustx.

While your scenario is closer to flat rates, my scenario is closer to how investors expected long term interest rates to change."

This sounds like you are saying a bond investor's default outlook should be: assume that the forward rates will be realized. Is this correct or am I misunderstanding you?
I think forward rates include the term premium, which cannot be known precisely. But I do think it is clear from forward rates that investors expected longer rates to rise in jan 2009. The fact that that failed to materialize was to the benefit of EDV. Normally the curve would not be so steep and 30y bonds would not yield quite so much.

So I do think the Dec 31 start date is more "fair". It's possible to extend this back to the 1970s so that start dates matter less. I haven't done that for the 30y specifically, but longer durations in general have lower sharpes and I am confident from the very low carry of 30y bonds that the sharpe would be quite poor. 30 y bonds yield very little per year of duration.
If the n-year forward rates had been realized, then the return of all bonds would have been the same over the n-year holding period.

As far as which scenario better predicted the path of future rates, we can look at what both scenarios would have predicted the yield curve to look like in August 2022 if we assumed that the forward rates themselves correctly forecast future rate changes (no term premium.) The yield curve goes out to 30 years, so I restrict the predicted zero yield curve to a maximum of 16 years maturity (about 14 years has elapsed between end of 2008 and August 2022.)

Here are the curves on these dates:
Image

Image

And here is the curve as of August 31 2022:
Image

Interestingly, the forward rates from your scenario are actually quite close to the front end of the yield curve that we see as of August 2022. I.e., your December 2008 scenario was better at predicting where the 0-6 year yields would end up. However, the longer end is far too low. In my scenario, the predicted zero coupon curve is closer to the longer end of the yield curve for the 10-16 year yields; my scenario was better for predicting where long term yields ended up. The predicted curves from both scenarios are significantly inverted. As far as I can tell, long term yields were "predicted" by the yield curve to fall in both scenarios.

In my backtesting, I have generally found that the highest Sharpe ratios belong to the STTs (2-3 years), then ITTs, and so on in that order for progressively longer maturities. Levering the duration with the shorter maturities involves futures contracts or margining bonds, which I presume many HFEA followers are reluctant to do. That leaves the following options for significant duration exposure: TMF (as was initially proposed), EDV/ZROZ/GOVZ, TYD, TYA, TLT, VGLT and VUSTX. VGLT (10-25 years maturity bonds) and VUSTX (15-30 years to maturity bonds) are good options if ~17 years duration is enough. TLT (20-30 years maturity bonds) outperforms VEDTX by 0.01 in Sharpe for the 2008 December to 2022 August scenario; the performance improves considerably if one ends the backtest around March 2020, which happens to be an example of where lower maturities outperform on a duration-adjusted basis (recession and sharply falling rates.)

I am skeptical of the prospects for TYD to significantly outperform EDV over a prolonged period of neutral rates. The cost of leverage and fees are a considerable impediment to its performance; it looks like TYD often loses 0.1 or more in Sharpe to its underlying index (which IEF also tracks.) Maybe a "neutral rate" backtesting opportunity will appear if rates get to August 2000 or May 2005 levels. TYD probably outperforms EDV in a sustained falling rate environment, but I have not backtested TYD earlier than what is available in portfolio visualizer yet. I am even more skeptical of TMF outperforming EDV in a neutral rate scenario.

TYA looks quite reasonable.
TYA doesn’t seem levered enough to counteract something like a 65% UPRO allocation.
comeinvest
Posts: 2709
Joined: Mon Mar 12, 2012 6:57 pm

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

Post by comeinvest »

unemployed_pysicist wrote: Mon Sep 19, 2022 4:52 pm If the n-year forward rates had been realized, then the return of all bonds would have been the same over the n-year holding period.

As far as which scenario better predicted the path of future rates, we can look at what both scenarios would have predicted the yield curve to look like in August 2022 if we assumed that the forward rates themselves correctly forecast future rate changes (no term premium.) The yield curve goes out to 30 years, so I restrict the predicted zero yield curve to a maximum of 16 years maturity (about 14 years has elapsed between end of 2008 and August 2022.)

Here are the curves on these dates:
...
...
And here is the curve as of August 31 2022:
...
Interestingly, the forward rates from your scenario are actually quite close to the front end of the yield curve that we see as of August 2022. I.e., your December 2008 scenario was better at predicting where the 0-6 year yields would end up. However, the longer end is far too low. In my scenario, the predicted zero coupon curve is closer to the longer end of the yield curve for the 10-16 year yields; my scenario was better for predicting where long term yields ended up. The predicted curves from both scenarios are significantly inverted. As far as I can tell, long term yields were "predicted" by the yield curve to fall in both scenarios.

In my backtesting, I have generally found that the highest Sharpe ratios belong to the STTs (2-3 years), then ITTs, and so on in that order for progressively longer maturities. Levering the duration with the shorter maturities involves futures contracts or margining bonds, which I presume many HFEA followers are reluctant to do. That leaves the following options for significant duration exposure: TMF (as was initially proposed), EDV/ZROZ/GOVZ, TYD, TYA, TLT, VGLT and VUSTX. VGLT (10-25 years maturity bonds) and VUSTX (15-30 years to maturity bonds) are good options if ~17 years duration is enough. TLT (20-30 years maturity bonds) outperforms VEDTX by 0.01 in Sharpe for the 2008 December to 2022 August scenario; the performance improves considerably if one ends the backtest around March 2020, which happens to be an example of where lower maturities outperform on a duration-adjusted basis (recession and sharply falling rates.)

I am skeptical of the prospects for TYD to significantly outperform EDV over a prolonged period of neutral rates. The cost of leverage and fees are a considerable impediment to its performance; it looks like TYD often loses 0.1 or more in Sharpe to its underlying index (which IEF also tracks.) Maybe a "neutral rate" backtesting opportunity will appear if rates get to August 2000 or May 2005 levels. TYD probably outperforms EDV in a sustained falling rate environment, but I have not backtested TYD earlier than what is available in portfolio visualizer yet. I am even more skeptical of TMF outperforming EDV in a neutral rate scenario.

TYA looks quite reasonable.
I don't understand everything you posted, but very interesting stuff, thanks for sharing! Would you mind sharing your backtesting of STTs, and which data you used for this backtesting? Maybe this fits more into the mHFEA thread.

"cost of leverage" (of TYD) - isn't its cost of leverage the usual spread to treasuries of the implied financing cost of futures and options?

What is your definition of "neutral rate" in your last paragraph, how is "neural" relevant to a backtesting period, and how do you determine a "fair" backtesting period? Is a fair backtesting period a period where no unexpected rate changes occurred, i.e. where the expected rate changes (maybe adjusted for an estimate of the term premium) were realized, rather than a period with equal rates at the beginning and end of the period? Sorry if the answer to my question is already implied in your post. I'm trying to understand the yield curve mathematics.

Would you agree that the historical carry as per the Simplify chart that I posted here viewtopic.php?p=6333645&hilit=simplify#p6333645 is a "fair" backtest with good predictive capability of long-term treasury returns per maturity (for those who have no opinion on the future path of interest rates), as it basically (in my understanding) is agnostic to expected and unexpected yield changes yield curve changes, i.e. backed out the effect of yield curve changes? ("Carry" by definition assumes a static yield curve.)
Tamalak
Posts: 1988
Joined: Fri May 06, 2016 2:29 pm

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

Post by Tamalak »

Isn't the cost of borrowing for these 3x funds short term?

Wouldn't the best time to run this strategy be when short term rates are much less than long term ones, rather than the inversion we have now?
CletusCaddy
Posts: 2678
Joined: Sun Sep 12, 2021 4:23 am

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

Post by CletusCaddy »

Tamalak wrote: Tue Sep 20, 2022 9:32 am Isn't the cost of borrowing for these 3x funds short term?

Wouldn't the best time to run this strategy be when short term rates are much less than long term ones, rather than the inversion we have now?
What inversion?

As of yesterday:
1 month Treasury: 2.6%
20 year Treasury: 3.8%
Tamalak
Posts: 1988
Joined: Fri May 06, 2016 2:29 pm

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

Post by Tamalak »

CletusCaddy wrote: Tue Sep 20, 2022 9:35 am
Tamalak wrote: Tue Sep 20, 2022 9:32 am Isn't the cost of borrowing for these 3x funds short term?

Wouldn't the best time to run this strategy be when short term rates are much less than long term ones, rather than the inversion we have now?
What inversion?

As of yesterday:
1 month Treasury: 2.6%
20 year Treasury: 3.8%
Is the cost of borrowing closest to 1mo? I guess it's not an inversion there, but the gap could/should be a lot wider.

Of course if the treasury yield curve is steep, that's a signal that the economy is going to stay healthy and so stocks will likely be highly valued by then.. drat.
CletusCaddy
Posts: 2678
Joined: Sun Sep 12, 2021 4:23 am

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

Post by CletusCaddy »

Tamalak wrote: Tue Sep 20, 2022 9:39 am
CletusCaddy wrote: Tue Sep 20, 2022 9:35 am
Tamalak wrote: Tue Sep 20, 2022 9:32 am Isn't the cost of borrowing for these 3x funds short term?

Wouldn't the best time to run this strategy be when short term rates are much less than long term ones, rather than the inversion we have now?
What inversion?

As of yesterday:
1 month Treasury: 2.6%
20 year Treasury: 3.8%
Is the cost of borrowing closest to 1mo? I guess it's not an inversion there, but the gap could/should be a lot wider.

Of course if the treasury yield curve is steep, that's a signal that the economy is going to stay healthy and so stocks will likely be highly valued by then.. drat.
Just checked TMF’s holdings, borrowing cost is 1 month LIBOR plus a spread which in January was ~0.2%.

1 month LIBOR is 3.0% today so borrowing cost might be something like 3.2%
comeinvest
Posts: 2709
Joined: Mon Mar 12, 2012 6:57 pm

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

Post by comeinvest »

unemployed_pysicist wrote: Mon Sep 19, 2022 4:52 pm
skierincolorado wrote: Fri Sep 16, 2022 3:02 pm
unemployed_pysicist wrote: Fri Sep 16, 2022 1:29 am
skierincolorado wrote: Wed Sep 14, 2022 7:13 pm
unemployed_pysicist wrote: Tue Sep 13, 2022 2:29 pm

In the period from Jan 28 2009 to August 30 2022, DGS30 started at 3.44%, and ended at 3.23%, which is a 21 bps fall, as compared to the dates you suggested:
Dec 29 2008 to August 30 2022, where DGS30 started at 2.63% and ended at 3.23%, which is a 60 bps rise (further from 0 than 21 bps.)

The period from Jan 28 2009 to August 30 2022 gives Sharpe of .31 for VUSTX, versus .27 for VEDTX. VEDTX underperformed VUSTX for this period of ~14 years, but I would not characterize this as "terrible".

In the long run, I do not disagree that VGLT is likely to outperform EDV on a risk-adjusted basis. However, I am pushing back against the idea that EDV is somehow "terrible" or unusable for this strategy. EDV is a reasonable option for the leveraged-constrained investor who wants to take significant duration risk. VGLT and EDV are both reasonable, depending on risk appetite.
I missed the beginning part of your post. Notably in both cases the sharpe of edv was worse, the question is by how much. In your case the sharpe for vustx is about 15% better than edv. In my case it is 40% better. So the expected risk adjusted returns of vustx is likely 15-40% better.

I'd also point out, as comeinvest also pointed out, that investors expected long term rates to increase in jan 2009 as indicated by the steepness of the curve relative to a low term premium. That increase failed to materialize and so long term bonds and especially the longest term bonds have returned more than expected. If long rates had increased as expected, edv would have looked worse in comparison to vustx.

While your scenario is closer to flat rates, my scenario is closer to how investors expected long term interest rates to change. I would generally expect the sharpe of vustx to be 30 to 40% higher than edv, which makes edv uninvestable in my opinion. The sharpe of 5 to 10 year bonds I'd expect (and had been) to be even higher than 20 year. Bonds of 15 to 20 years are barely investable in my opinion and over 20 years are useless.

Also remember long term bonds are a small market that is likely distorted by pension fund demand for long dates maturities. I think it's better to stick to the broader, more competitive, and higher performing intermediate term bond market.
I was editing a mistake in my post during your first response when you wrote your reply, which is why you missed it. My apologies.

You can also change the start date to November 28 2008 (1 month earlier) where DGS30 was 3.45. The Sharpes were .3 for VUSTX and .25 for VEDTX.

I would like to discuss these points:

"If long rates had increased as expected, edv would have looked worse in comparison to vustx.

While your scenario is closer to flat rates, my scenario is closer to how investors expected long term interest rates to change."

This sounds like you are saying a bond investor's default outlook should be: assume that the forward rates will be realized. Is this correct or am I misunderstanding you?
I think forward rates include the term premium, which cannot be known precisely. But I do think it is clear from forward rates that investors expected longer rates to rise in jan 2009. The fact that that failed to materialize was to the benefit of EDV. Normally the curve would not be so steep and 30y bonds would not yield quite so much.

So I do think the Dec 31 start date is more "fair". It's possible to extend this back to the 1970s so that start dates matter less. I haven't done that for the 30y specifically, but longer durations in general have lower sharpes and I am confident from the very low carry of 30y bonds that the sharpe would be quite poor. 30 y bonds yield very little per year of duration.
If the n-year forward rates had been realized, then the return of all bonds would have been the same over the n-year holding period.

As far as which scenario better predicted the path of future rates, we can look at what both scenarios would have predicted the yield curve to look like in August 2022 if we assumed that the forward rates themselves correctly forecast future rate changes (no term premium.) The yield curve goes out to 30 years, so I restrict the predicted zero yield curve to a maximum of 16 years maturity (about 14 years has elapsed between end of 2008 and August 2022.)

Here are the curves on these dates:
Image

Image

And here is the curve as of August 31 2022:
Image

Interestingly, the forward rates from your scenario are actually quite close to the front end of the yield curve that we see as of August 2022. I.e., your December 2008 scenario was better at predicting where the 0-6 year yields would end up. However, the longer end is far too low. In my scenario, the predicted zero coupon curve is closer to the longer end of the yield curve for the 10-16 year yields; my scenario was better for predicting where long term yields ended up. The predicted curves from both scenarios are significantly inverted. As far as I can tell, long term yields were "predicted" by the yield curve to fall in both scenarios.

In my backtesting, I have generally found that the highest Sharpe ratios belong to the STTs (2-3 years), then ITTs, and so on in that order for progressively longer maturities. Levering the duration with the shorter maturities involves futures contracts or margining bonds, which I presume many HFEA followers are reluctant to do. That leaves the following options for significant duration exposure: TMF (as was initially proposed), EDV/ZROZ/GOVZ, TYD, TYA, TLT, VGLT and VUSTX. VGLT (10-25 years maturity bonds) and VUSTX (15-30 years to maturity bonds) are good options if ~17 years duration is enough. TLT (20-30 years maturity bonds) outperforms VEDTX by 0.01 in Sharpe for the 2008 December to 2022 August scenario; the performance improves considerably if one ends the backtest around March 2020, which happens to be an example of where lower maturities outperform on a duration-adjusted basis (recession and sharply falling rates.)

I am skeptical of the prospects for TYD to significantly outperform EDV over a prolonged period of neutral rates. The cost of leverage and fees are a considerable impediment to its performance; it looks like TYD often loses 0.1 or more in Sharpe to its underlying index (which IEF also tracks.) Maybe a "neutral rate" backtesting opportunity will appear if rates get to August 2000 or May 2005 levels. TYD probably outperforms EDV in a sustained falling rate environment, but I have not backtested TYD earlier than what is available in portfolio visualizer yet. I am even more skeptical of TMF outperforming EDV in a neutral rate scenario.

TYA looks quite reasonable.
Instructive web sites on forward curves:
https://www.chathamfinancial.com/techno ... ard-curves
https://www.chathamfinancial.com/insigh ... ward-curve

We can see that forward rates typically extrapolate the recent past in a way.
The realized rate movements typically exacerbate expected movements.
Funny, ca. 2016 the forward rates finally "gave up" after ca. 8 years, thinking they might as well "catch up with reality" and become more flat - precisely when the 2017-2019 rate hikes more or less realized the forward curve that was dominant between 2008 and 2015 (not 2016). Lol.

Please someone correct me if I'm wrong, but my interpretation of the historical chart (the second chart) is that bonds did exceptionally well between 2008 and 2021, as realized rates mostly undercut expected rates, resulting in gains from unexpected yield changes. This effect is mean-reverting by nature, as it cannot last forever.

Image

Image
er999
Posts: 1343
Joined: Wed Nov 05, 2008 10:00 am

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

Post by er999 »

9/21/22 the negative correlation showed up — upro down 5.33% and tmf up 4.64%.
bgf
Posts: 2085
Joined: Fri Nov 10, 2017 8:35 am

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

Post by bgf »

er999 wrote: Wed Sep 21, 2022 6:54 pm 9/21/22 the negative correlation showed up — upro down 5.33% and tmf up 4.64%.
Yes, and that UPRO whipsaw around Powell’s speech was something else.
“TE OCCIDERE POSSUNT SED TE EDERE NON POSSUNT NEFAS EST"
Lawyered_
Posts: 176
Joined: Wed Feb 09, 2022 10:52 am

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

Post by Lawyered_ »

er999 wrote: Wed Sep 21, 2022 6:54 pm 9/21/22 the negative correlation showed up — upro down 5.33% and tmf up 4.64%.
Not going to last. The negative correlation cannot stand up to the rising rates.
kbourgu
Posts: 23
Joined: Tue Sep 28, 2021 10:05 am

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

Post by kbourgu »

Can someone explain to me why this strategy isn't dead? Down about 60-70% since starting out.
Jags4186
Posts: 8198
Joined: Wed Jun 18, 2014 7:12 pm

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

Post by Jags4186 »

kbourgu wrote: Thu Sep 22, 2022 9:13 am Can someone explain to me why this strategy isn't dead? Down about 60-70% since starting out.
It is dead. I basically got out with what I went in with.
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

comeinvest wrote: Tue Sep 20, 2022 5:49 pm
Funny, ca. 2016 the forward rates finally "gave up" after ca. 8 years, thinking they might as well "catch up with reality" and become more flat - precisely when the 2017-2019 rate hikes more or less realized the forward curve that was dominant between 2008 and 2015 (not 2016). Lol.
Probably not a coincidence.. those low forward rates may have helped stimulate the economy finally. May both be related to qe3
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

kbourgu wrote: Thu Sep 22, 2022 9:13 am Can someone explain to me why this strategy isn't dead? Down about 60-70% since starting out.
Owning bonds and stocks is dead?

There are some valid criticisms 1) the bonds are too long in duration 2) too much bonds 3) too much leverage for all but the youngest and riskiest investors to use as their whole portfolio 4) using a highly leveraged strategy as a small part of one's portfolio is less efficient than leveraging the whole portfolio slightly, unless you rebalance between each strategy of the portfolio and view it as a cohesive whole.

But the basic principles of diversification and leverage are valid and not going anywhere.
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

kbourgu wrote: Thu Sep 22, 2022 9:13 am Can someone explain to me why this strategy isn't dead? Down about 60-70% since starting out.
Owning bonds and stocks is dead?

There are some valid criticisms 1) the bonds are too long in duration 2) too much bonds 3) too much leverage for all but the youngest and riskiest investors to use as their whole portfolio 4) using a highly leveraged strategy as a small part of one's portfolio is less efficient than leveraging the whole portfolio slightly, unless you rebalance between each strategy of the portfolio and view it as a cohesive whole. Maybe a fifth criticism: owning bonds and stocks during massive quantitative easing. This wasn't an issue during previous rounds of qe, but seems to have been problematic this time. But that gets into market timing and I am not at all convinced the decline in bonds was foreseeable, and even less so for stocks.

But the basic principles of diversification and leverage are valid and not going anywhere.
Last edited by skierincolorado on Thu Sep 22, 2022 12:29 pm, edited 1 time in total.
bgf
Posts: 2085
Joined: Fri Nov 10, 2017 8:35 am

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

Post by bgf »

lawyeredCLO wrote: Thu Sep 22, 2022 8:00 am
er999 wrote: Wed Sep 21, 2022 6:54 pm 9/21/22 the negative correlation showed up — upro down 5.33% and tmf up 4.64%.
Not going to last. The negative correlation cannot stand up to the rising rates.
maybe, maybe not.

i do think that we are unlikely to experience the same type of underperformance over the last year moving forward for the next year. the size of move necessary for that to happen would lead to very high long term rates.

we can either stay with a grossly inverted yield curve, short term rates can drop relative to long term, or long term rates can rise relative to short term. of those potential outcomes, i think one is more likely than the other two.

this kind of thing happens in markets all the time. it won't last forever, and rates will go down again. i think its a question of whether investors in this strategy are able to hold on and accept the gut punches along the way. i'll definitely admit that its getting harder and harder to do.
“TE OCCIDERE POSSUNT SED TE EDERE NON POSSUNT NEFAS EST"
Football2408
Posts: 26
Joined: Mon Sep 05, 2022 12:31 pm

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

Post by Football2408 »

If we are in a stagnant or slow increasing rate environment, negative correlation will return IMO. The large jump environment seems to still be here for a bit longer.
corpgator
Posts: 217
Joined: Wed Jul 22, 2015 12:00 pm

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

Post by corpgator »

The only reason I'm holding tmf is in case we have a real short term system shock, but I kick myself after every rate hike. It happens every time: flat or bump on rate day then dump the next. It mightcome up a little in the next few days, but I'm leaning towards dropping it and waiting until the hikes are over, but if I keep I'm definitely not rebalancing into it until after the hikes stop.
lkar
Posts: 379
Joined: Sat May 04, 2019 4:02 pm

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

Post by lkar »

rascott wrote: Mon Sep 19, 2022 10:54 am I've been out of this portfolio for almost 2 years now..... so not really paying attention to it.... but what kind of bloodbath has it been YTD? Down 50%?
I wish I'd never seen this thread.

I remember when this originally was posted there was a bit of angst whether it was a worthwhile thread given its inconsistency with Bogle principles. I think it has been a valuable thread and I've learned an awful lot about what I own, how bond markets work, etc. But I unduly discounted those concerns. I'm staying the course, and will continue to do so. It is a tiny portion of my portfolio, happily.

But the truth is that it does not match my temperament or philosophy and I did it any way out of FOMO. It's a good lesson for me to realize who I am, and a not too expensive one. But it really is a testament to how misleading everything can seem in a stable bull market.
Hydromod
Posts: 1051
Joined: Tue Mar 26, 2019 10:21 pm

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

Post by Hydromod »

er999 wrote: Wed Sep 21, 2022 6:54 pm 9/21/22 the negative correlation showed up — upro down 5.33% and tmf up 4.64%.
Daily correlation has persisted as negative to slightly positive.

Monthly correlation has gone strongly positive.
er999
Posts: 1343
Joined: Wed Nov 05, 2008 10:00 am

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

Post by er999 »

I am still sticking with it as it’s a small percentage of my portfolio (particularly now with the huge losses) but fear that tmf will go to zero (or close enough to it before they reverse split)

I was reminded of a podcast that said the Renaissance Medallion fund got their billions from thousands of rich dentists losing their money with poor individual stock picks. I do feel that some sophisticated trader is on the other side of the TMF trade has sucked away millions in bogleheads funds who are in HFEA. I’m not a dentist but fall into the same camp as a doctor with some money to blow on potentially stupid trades and being financially unsophisticated. Definitely the leveraged fund providers have made out with their fees and since they have both sides of the trade with long and short funds.

Anyway this thread is good for real time sentiment for these leveraged strategies.
nineth_dwarf
Posts: 13
Joined: Mon Feb 07, 2022 1:56 pm

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

Post by nineth_dwarf »

er999 wrote: Thu Sep 22, 2022 12:45 pm I was reminded of a podcast that said the Renaissance Medallion fund got their billions from thousands of rich dentists losing their money with poor individual stock picks. I do feel that some sophisticated trader is on the other side of the TMF trade has sucked away millions in bogleheads funds who are in HFEA.
BND has >80B assets under management
https://ycharts.com/companies/BND/total ... management
and they are numerous bond funds like BND with more billions under management.

TMF has <400M
https://ycharts.com/companies/TMF/total ... management

Bond funds have made individuals lose lots of net worth since Jan 2022, but TMF is a drop in the bucket.
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

Honestly surprised anybody would have a change of heart based on the performance... hfea drew down over 50% in 1974 1982 2001 and 2009. I guess seeing is believing
Post Reply