Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by skierincolorado »

LTCM wrote: Wed Sep 08, 2021 5:38 pm
EfficientInvestor wrote: Wed Sep 08, 2021 4:52 pm What I would really like to do at some point is to gather all historical yield data and develop an algorithm that determines the current optimal point on the curve. So this algorithm would be a function of roll and carry I suppose.
That would be very useful!

I liked your efficient portfolio approach listed above. What would you amend the rates to based on this discussion (if anything)?
I find the whole problem kind of too complex. You have to somehow factor in forward rate tables and do a lot of complex math. You also have to specify a date we are projecting to or the problem is intractable. This is my rough ballpark guess, using a 2 year projection, from looking at forward rate tables 2 years from now. The return of all treasuries is reduced, especially LTT, because forward rates indicate the market expects all interest rates to rise. STT and ITT do not decrease very much compared to the original values because the original values did not include return from roll. So I adjusted up from 0.8% to 1.3% first, and then down to 0.5% based on market forward rates. LTT see the biggest adjustment down because they have the most duration exposure to the market expectation of rate increases, and the smallest benefit from roll. The original value was 1.87% which I adjust down to 0.7%. Also curious what EfficientInvestor thinks

We could specify a longer projection horizon than 2 years, like 30 years. The return of the 30 year would basically be the current YTM, plus roll since we would be rolling 30s, not holding to maturity. But the return of the ITT would go up dramatically based on market forward rates > 2%. Of course having a 30 year horizon kind of defeats the purpose of a dynamic strategy that changes its allocation based on current market rates. If we're deciding what to invest in for the next 3 months or a year, we should make a 1 year projection of returns, using forward rate tables for a year from now.

https://www.portfoliovisualizer.com/eff ... ints=false
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by skierincolorado »

skierincolorado wrote: Wed Sep 08, 2021 5:48 pm
LTCM wrote: Wed Sep 08, 2021 5:38 pm
EfficientInvestor wrote: Wed Sep 08, 2021 4:52 pm What I would really like to do at some point is to gather all historical yield data and develop an algorithm that determines the current optimal point on the curve. So this algorithm would be a function of roll and carry I suppose.
That would be very useful!

I liked your efficient portfolio approach listed above. What would you amend the rates to based on this discussion (if anything)?
I find the whole problem kind of too complex. You have to somehow factor in forward rate tables and do a lot of complex math. You also have to specify a date we are projecting to or the problem is intractable. This is my rough ballpark guess, using a 2 year projection, from looking at forward rate tables 2 years from now. The return of all treasuries is reduced, especially LTT, because forward rates indicate the market expects all interest rates to rise. STT and ITT do not decrease very much compared to the original values because the original values did not include return from roll. So I adjusted up from 0.8% to 1.3% first, and then down to 0.5% based on market forward rates. LTT see the biggest adjustment down because they have the most duration exposure to the market expectation of rate increases, and the smallest benefit from roll. The original value was 1.87% which I adjust down to 0.7%. Also curious what EfficientInvestor thinks

We could specify a longer projection horizon than 2 years, like 30 years. The return of the 30 year would basically be the current YTM, plus roll since we would be rolling 30s, not holding to maturity. But the return of the ITT would go up dramatically based on market forward rates > 2%. Of course having a 30 year horizon kind of defeats the purpose of a dynamic strategy that changes its allocation based on current market rates. If we're deciding what to invest in for the next 3 months or a year, we should make a 1 year projection of returns, using forward rate tables for a year from now.

https://www.portfoliovisualizer.com/eff ... ints=false
A rough approximation that includes the return from roll is to look at the return of holding a 5 year bond for 1 year based on market forward rates:

Current 5 year rate: .8%
Market expected 5 year rate 1 year hence: ~1.05%
Current 4 year rate: .62%
Market expected 4 year interest rate 1 year hence: .87%

So if we hold a $1000 5 year bond with a .82% coupon for 1 year, at the end we will hold a 4 year bond with a YTM of .87%. The price of our bond will be $997.25 using a bond price calculator, so we lost $2.75 We will have collected $8.25 in interest. We will net $8.25 - 2.75, or $5.50. Thus our 1-yr return is 0.55%. If the 4 year rate had stayed at its current 0.62% we would have returned 1.5% instead of merely 0.55%.

We can do the same for a 30 year bond.

Current 30 year rate: 1.97%
Market expected 30 year rate 1 year hence: 2.07% (based on market expectation of 1-yr rates 9 years from now of 2.11% - indicating 30y rates over 3% - will have to increase more than .1%/year to get there).
Current 29 year rate: 1.94%
Market expected 29 year rate 1 year hence: ~2.04%

So if we hold a $1000 30 year bond with a 1.97% coupon for 1 year, at the end we will hold a 29 year bond with a YTM of 2.04%. The price of our bond will be $984.73 using a bond price calculator, so we lost $15.27. We will have collected $19.70 in interest. We will have netted $4.43, or a 0.443% return.

Rate forwards indicate that rates for both the 5 year and the 30 year are likely to increase. Even though I used a .25% rate increase for the 4 year bond, and a .1% rate increase for the 29 year bond, the expected return of the 5 year bond was still greater. And the risk is much much less.

Overall, given the market's expectation for rate increases, I'm pretty pleased that ITT still has a positive return (0.55%) even if rates increase 0.25% in the next year. On the other hand, the 30 year bond's return was barely positive (0.443%) with merely a .1% rate increase. The market does expect ITT rates to increase more than LTT rates, but both are expected to increase substantially given that the expected 1 year rate 9 years from now is 2.11% - higher than today's 30 year rate. That would suggest the 30y rate is likely over 3% 10 years from now, and will have to increase 0.1% per year to get there - which would cause returns for the next several years to be near 0%.

It's actually remarkable that the market expected returns of the 5 year bond and 30 year bond for the next several years are nearly identical, given there is so much more risk inherent in the 30 year bond.
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by skierincolorado »

skierincolorado wrote: Wed Sep 08, 2021 6:22 pm

A rough approximation that includes the return from roll is to look at the return of holding a 5 year bond for 1 year based on market forward rates:

Current 5 year rate: .8%
Market expected 5 year rate 1 year hence: ~1.05%
Current 4 year rate: .62%
Market expected 4 year interest rate 1 year hence: .87%

So if we hold a $1000 5 year bond with a .82% coupon for 1 year, at the end we will hold a 4 year bond with a YTM of .87%. The price of our bond will be $997.25 using a bond price calculator, so we lost $2.75 We will have collected $8.25 in interest. We will net $8.25 - 2.75, or $5.50. Thus our 1-yr return is 0.55%. If the 4 year rate had stayed at its current 0.62% we would have returned 1.5% instead of merely 0.55%.

We can do the same for a 30 year bond.

Current 30 year rate: 1.97%
Market expected 30 year rate 1 year hence: 2.07% (based on market expectation of 1-yr rates 9 years from now of 2.11% - indicating 30y rates over 3% - will have to increase more than .1%/year to get there).
Current 29 year rate: 1.94%
Market expected 29 year rate 1 year hence: ~2.04%

So if we hold a $1000 30 year bond with a 1.97% coupon for 1 year, at the end we will hold a 29 year bond with a YTM of 2.04%. The price of our bond will be $984.73 using a bond price calculator, so we lost $15.27. We will have collected $19.70 in interest. We will have netted $4.43, or a 0.443% return.

Rate forwards indicate that rates for both the 5 year and the 30 year are likely to increase. Even though I used a .25% rate increase for the 4 year bond, and a .1% rate increase for the 29 year bond, the expected return of the 5 year bond was still greater. And the risk is much much less.

Overall, given the market's expectation for rate increases, I'm pretty pleased that ITT still has a positive return (0.55%) even if rates increase 0.25% in the next year. On the other hand, the 30 year bond's return was barely positive (0.443%) with merely a .1% rate increase. The market does expect ITT rates to increase more than LTT rates, but both are expected to increase substantially given that the expected 1 year rate 9 years from now is 2.11% - higher than today's 30 year rate. That would suggest the 30y rate is likely over 3% 10 years from now, and will have to increase 0.1% per year to get there - which would cause returns for the next several years to be near 0%.

It's actually remarkable that the market expected returns of the 5 year bond and 30 year bond for the next several years are nearly identical, given there is so much more risk inherent in the 30 year bond.
An interesting and useful conclusion from this is 1) looking at forward rates and roll to project returns, the efficient portfolio does not include LTT 2) even ITT should be held in a pretty low ratio to stocks like 1:1, assuming 6% expected nominal return for stocks

If we adjust the expected return for stocks closer to the Vanguard estimate, we get back to a 1:2 ratio of stocks/ITT

For example: https://www.portfoliovisualizer.com/eff ... ints=false
Last edited by skierincolorado on Wed Sep 08, 2021 6:49 pm, edited 1 time in total.
User avatar
LTCM
Posts: 412
Joined: Wed Sep 09, 2020 3:58 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by LTCM »

Are the forward rates not already priced into the current bond prices?

If the forward rate suggests rates will go up 0.25% in the next year and they go up 0.25% in the next year would the price really fall 0.25% per year of duration?
55% VUG - 20% VEA - 20% EDV - 5% BNDX
User avatar
LTCM
Posts: 412
Joined: Wed Sep 09, 2020 3:58 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by LTCM »

https://www.investopedia.com/ask/answer ... d-rate.asp
Theoretically, the forward rate should be equal to the spot rate plus any earnings from the security (and any finance charges).
So (forward rate) - (spot rate) = roll & carry for the period b/w spot and forward?

Is it that simple?
55% VUG - 20% VEA - 20% EDV - 5% BNDX
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by skierincolorado »

LTCM wrote: Wed Sep 08, 2021 6:48 pm Are the forward rates not already priced into the current bond prices?

If the forward rate suggests rates will go up 0.25% in the next year and they go up 0.25% in the next year would the price really fall 0.25% per year of duration?
Depends what you mean by priced in. Investors buying 5 years bonds with .8% YTM know interest rates are likely to go up, so in that sense they are priced in. Investors would likely accept a lower rate if rates weren't expected to go up. But that doesn't change the fact that interest rates, across the whole yield curve, are expected to go up which will cause the 1-year return to be less than it would have been. On the other hand, roll makes 1-yr returns greater than YTM when rates don't change.

As to your question, the answer is yes. For shorter durations, this effect is mostly cancelled by roll. For longer durations, it's not. The difference between the 4-yr and 5-yr rate is .2%, so a .25% rate increase over the course of 1 year is moslty offset by the roll-down from 5 years to 4-years.

If rates increased by .25% for the 29-year duration bond, the roll-down is a lot less than .25% (more like .03%). So a .25% rate inrease would be catastrophic. Fortunately long-term rates aren't expected to increase .25%. More like .1%. Which is still a lot more than the roll-down, and this loss eats into the return from carry.
User avatar
LTCM
Posts: 412
Joined: Wed Sep 09, 2020 3:58 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by LTCM »

This may be useful:
https://quant.stackexchange.com/questio ... ce-of-bond
forward price = spot price - carry
carry = forward price - spot price

This seems like it would make choosing our favored bond pretty easy.
55% VUG - 20% VEA - 20% EDV - 5% BNDX
EfficientInvestor
Posts: 580
Joined: Thu Nov 01, 2018 7:02 pm
Location: Alabama

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by EfficientInvestor »

skierincolorado wrote: Wed Sep 08, 2021 6:22 pm
skierincolorado wrote: Wed Sep 08, 2021 5:48 pm
LTCM wrote: Wed Sep 08, 2021 5:38 pm
EfficientInvestor wrote: Wed Sep 08, 2021 4:52 pm What I would really like to do at some point is to gather all historical yield data and develop an algorithm that determines the current optimal point on the curve. So this algorithm would be a function of roll and carry I suppose.
That would be very useful!

I liked your efficient portfolio approach listed above. What would you amend the rates to based on this discussion (if anything)?
I find the whole problem kind of too complex. You have to somehow factor in forward rate tables and do a lot of complex math. You also have to specify a date we are projecting to or the problem is intractable. This is my rough ballpark guess, using a 2 year projection, from looking at forward rate tables 2 years from now. The return of all treasuries is reduced, especially LTT, because forward rates indicate the market expects all interest rates to rise. STT and ITT do not decrease very much compared to the original values because the original values did not include return from roll. So I adjusted up from 0.8% to 1.3% first, and then down to 0.5% based on market forward rates. LTT see the biggest adjustment down because they have the most duration exposure to the market expectation of rate increases, and the smallest benefit from roll. The original value was 1.87% which I adjust down to 0.7%. Also curious what EfficientInvestor thinks

We could specify a longer projection horizon than 2 years, like 30 years. The return of the 30 year would basically be the current YTM, plus roll since we would be rolling 30s, not holding to maturity. But the return of the ITT would go up dramatically based on market forward rates > 2%. Of course having a 30 year horizon kind of defeats the purpose of a dynamic strategy that changes its allocation based on current market rates. If we're deciding what to invest in for the next 3 months or a year, we should make a 1 year projection of returns, using forward rate tables for a year from now.

https://www.portfoliovisualizer.com/eff ... ints=false
A rough approximation that includes the return from roll is to look at the return of holding a 5 year bond for 1 year based on market forward rates:

Current 5 year rate: .8%
Market expected 5 year rate 1 year hence: ~1.05%
Current 4 year rate: .62%
Market expected 4 year interest rate 1 year hence: .87%

So if we hold a $1000 5 year bond with a .82% coupon for 1 year, at the end we will hold a 4 year bond with a YTM of .87%. The price of our bond will be $997.25 using a bond price calculator, so we lost $2.75 We will have collected $8.25 in interest. We will net $8.25 - 2.75, or $5.50. Thus our 1-yr return is 0.55%. If the 4 year rate had stayed at its current 0.62% we would have returned 1.5% instead of merely 0.55%.

We can do the same for a 30 year bond.

Current 30 year rate: 1.97%
Market expected 30 year rate 1 year hence: 2.07% (based on market expectation of 1-yr rates 9 years from now of 2.11% - indicating 30y rates over 3% - will have to increase more than .1%/year to get there).
Current 29 year rate: 1.94%
Market expected 29 year rate 1 year hence: ~2.04%

So if we hold a $1000 30 year bond with a 1.97% coupon for 1 year, at the end we will hold a 29 year bond with a YTM of 2.04%. The price of our bond will be $984.73 using a bond price calculator, so we lost $15.27. We will have collected $19.70 in interest. We will have netted $4.43, or a 0.443% return.

Rate forwards indicate that rates for both the 5 year and the 30 year are likely to increase. Even though I used a .25% rate increase for the 4 year bond, and a .1% rate increase for the 29 year bond, the expected return of the 5 year bond was still greater. And the risk is much much less.

Overall, given the market's expectation for rate increases, I'm pretty pleased that ITT still has a positive return (0.55%) even if rates increase 0.25% in the next year. On the other hand, the 30 year bond's return was barely positive (0.443%) with merely a .1% rate increase. The market does expect ITT rates to increase more than LTT rates, but both are expected to increase substantially given that the expected 1 year rate 9 years from now is 2.11% - higher than today's 30 year rate. That would suggest the 30y rate is likely over 3% 10 years from now, and will have to increase 0.1% per year to get there - which would cause returns for the next several years to be near 0%.

It's actually remarkable that the market expected returns of the 5 year bond and 30 year bond for the next several years are nearly identical, given there is so much more risk inherent in the 30 year bond.
Good stuff here. Thanks for the detailed breakdown. I'm still trying to determine what constitutes market timing and what does not. It would be interesting to see data that shows the correlation between expected future rates and actual future rates. If there is a high correlation, then maybe it is reasonable to bake those expectations into a return assumption. But if there is a low correlation, then perhaps it is appropriate to assume that rates 1 year from now will not change from where they are currently and you can bake in the roll yield to your return assumption. In the meantime, this is why I think current yield is a close enough approximation because it kind of splits the difference between the two (at least for the 5 year).
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by skierincolorado »

EfficientInvestor wrote: Thu Sep 09, 2021 10:46 am

Good stuff here. Thanks for the detailed breakdown. I'm still trying to determine what constitutes market timing and what does not. It would be interesting to see data that shows the correlation between expected future rates and actual future rates. If there is a high correlation, then maybe it is reasonable to bake those expectations into a return assumption. But if there is a low correlation, then perhaps it is appropriate to assume that rates 1 year from now will not change from where they are currently and you can bake in the roll yield to your return assumption. In the meantime, this is why I think current yield is a close enough approximation because it kind of splits the difference between the two (at least for the 5 year).
It's definitely not market timing to factor in forward rates because forward prices are fully baked into current spot prices, otherwise it would be possible to artibtrage between the two. It's also possible to synthesize forward prices from current spot prices, I think. For example, I think buying the 10-yr spot and shorting the 5-yr spot should have identical returns to buying the 5-yr forward for 5 years from now.

Although I think that future spot rates might be persistently lower than forward rates because a forward rate is inherently more risky since it's agreed upon today, not 5 years from now. I'm not sure if this is true or how to factor it in. Either way the market is forecasting much more rate increase than this could explain.

This is for exchange rates not interest rates, it finds a decent correlation, and the first sentence is basically what we are talking about:

"If forward exchange rates systematically differ from future spot rates, money can be made on
the difference: a risk-neutral agent with rational expectations can exploit the inefficiency and
hence, supposedly, the anomaly should disappear"
https://www.esm.europa.eu/sites/default/files/wp46.pdf

AKA market timing/predicting. *Not* using forward rates as your default assumption is market timing/prediction, not the other way around.

Markets consistently overestimated the pace of interest rate increases folowing 2009. However, we can't assume that the market will always be wrong in this direction based on one business cycle. If that were true, investors could just bet against forward rates and always make money and the bias would be arbitraged away. There shouldn't be a bias in the long-term.

Even if we don't fully accept forward rate prices (although I'd argue we should) we can still definitively eliminate LTT. As long as we accept that at least *some* interest rate increase is likely, maybe less than the market expects and has priced in, LTT would be eliminated from your model. For example the market expects ITT rates to rise ~.25 and LTT rates to rise ~.1 (my estimates). Even if we halve those numbers, LTT would be eliminated from the efficient frontier. Of course, I would argue that halving the market expectation is a form of market timing/prediction to be frowned upon. We'd be making a bet that rates rise less than expected. Although we wouldn't act on it, except maybe to go a little farther out on the curve (but not all the way to LTT).
User avatar
LTCM
Posts: 412
Joined: Wed Sep 09, 2020 3:58 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by LTCM »

skierincolorado wrote: Thu Sep 09, 2021 11:35 am we can still definitively eliminate LTT. As long as we accept that at least *some* interest rate increase is likely, maybe less than the market expects and has priced in, LTT would be eliminated from your model.
I'm not quite there yet. I was reading general yield curve material last night and it was saying the LTT have relatively lower yields because of convexity. A drop in interest rates raises the price of the security much more than a rise in rates decreases the price of the security. They're the least symmetrical of all the treasuries.
55% VUG - 20% VEA - 20% EDV - 5% BNDX
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by skierincolorado »

LTCM wrote: Thu Sep 09, 2021 2:14 pm
skierincolorado wrote: Thu Sep 09, 2021 11:35 am we can still definitively eliminate LTT. As long as we accept that at least *some* interest rate increase is likely, maybe less than the market expects and has priced in, LTT would be eliminated from your model.
I'm not quite there yet. I was reading general yield curve material last night and it was saying the LTT have relatively lower yields because of convexity. A drop in interest rates raises the price of the security much more than a rise in rates decreases the price of the security. They're the least symmetrical of all the treasuries.
This may be true, but it doesn't change:
1) there's no significant period where LTT were a significant part of the efficient frontier. Switching between STT and ITT produces minor changes to sharpe ratio sometimes better sometimes worse, switching in LTT produces significant cuts to sharpe ratio.

2) the market expects rates to rise which is part of the reason LTT have a decent yield right now. This has a very outsized effect on reducing the near term returns of LTT than STT or ITT, even after you consider convexity.

3) the return from roll enhances the return of ~5 year futures much more than it does for LTT (roll=0.7% carry=0.8% assuming no change in the curve, total return = 1.5%). For example, the 10-yr return of ZF contracts has been 1.68% even though ZF is a 4.5 year bond and 5 year interest rates have been well below 1.68% for most of the last 10 years. The 4.5 year rate has probably averaged barely above 1%. I got 1.4% annualized on my last ZF roll despite almost no change in rates. Try pluggin 1.4%, or even 1.2% into the efficient frontier assumptions.
https://www.spglobal.com/spdji/en/indic ... /#overview

4) beta chasing investors overpricing LTT is a real phenomenon that's unlikely to change

Any single one of these points on its own is probably enough to swing the efficient frontier from LTT to ITT. I'd also point out one can't have it both ways. Either we accept the yield curve as is, and assume no interest rate change. In this case, the return from roll adds a full 0.7% to the return of ZF and ZF will return 1.5%, while UB returns 2% with much more risk (anything above 1% for ZF swings the efficient frontier to ITT). Or one considers forward rates, as one should, and prices in some near term reduction in return for all bonds as rates rise. In that case, even the tiniest increase in long-term rates over the next few years would bring their return close to 0%. The only scenario where LTT wins are if we assume 1) ITT rates will rise 2) LTT rates don't rise at all and 3) the rate increase of ITT is so large it overcomes the 0.7% return from roll on ITT leaving us with only 0.8% of carry 4) we ignore points #1 and #4 and assume the future will be different than the past and that we can predict when and how it will be different.

In the end it simply does not make sense to believe that an asset class which has always underperformed is going to perform better now and that rates will rise substantially on ITT and not at all on LTT. It amounts to egregious market timing IMO.

Of course nothing is a guarantee, it's entirely possible that LTT rates don't rise at all and ITT rates rise significantly. I just wouldn't bet on it.
User avatar
LTCM
Posts: 412
Joined: Wed Sep 09, 2020 3:58 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by LTCM »

skierincolorado wrote: Thu Sep 09, 2021 2:35 pm
1) there's no significant period where LTT were a significant part of the efficient frontier.
There is though (depending on your definition of significant, but its well above ITT):
Efficient Frontier Results (Jan 1978 - Aug 2021)

10% LTT is a pretty significant part of the efficient portfolio considering the volatility/risk involved compared to STT/ITT.
55% VUG - 20% VEA - 20% EDV - 5% BNDX
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by skierincolorado »

LTCM wrote: Thu Sep 09, 2021 3:15 pm
skierincolorado wrote: Thu Sep 09, 2021 2:35 pm
1) there's no significant period where LTT were a significant part of the efficient frontier.
There is though (depending on your definition of significant, but its well above ITT):
Efficient Frontier Results (Jan 1978 - Aug 2021)

10% LTT is a pretty significant part of the efficient portfolio considering the volatility/risk involved compared to STT/ITT.
Depends on your definition of significant, also if you eliminate the LTT, the sharpe ratio is essentially unchanged at .691 (vs .695) so it's not actually making a meaningful contribution. In yours it is using a small amount of LTT (10%) to balance with STT. So when you eliminate LTT, it just uses ITT and you get the same sharpe ratio.

If someone wanted to do the portfolio you've linked, that would be very reasonable. 10% is not much. Personally I find it easier to just do 200-300% in ITT rather than 400% in STT and 50% in LTT. You can keep adding asset classes and the efficient frontier will take tiny bits of each of them for negligible improvements in sharpe ratio.

stock+STT vs stocks+ITT vs stock+LTT... the sharpe ratio for the former two is much better than the latter. so if you had to only pick one, it's clearly STT or ITT. If you want to get more compicated, adding a little LTT gives you a tiny (theoretical) benefit. Adding gold gives you just as much benefit
User avatar
LTCM
Posts: 412
Joined: Wed Sep 09, 2020 3:58 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by LTCM »

I can see the appeal of 1:3 Stocks/ITT. The simplicity is great. It appeals to my round numbers obsession which helps maintains rebalance discipline. Secondly, when working with relatively low amounts for futures you don't want to be dealing with too many securities.
55% VUG - 20% VEA - 20% EDV - 5% BNDX
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by skierincolorado »

LTCM wrote: Thu Sep 09, 2021 4:34 pm I can see the appeal of 1:3 Stocks/ITT. The simplicity is great. It appeals to my round numbers obsession which helps maintains rebalance discipline. Secondly, when working with relatively low amounts for futures you don't want to be dealing with too many securities.
For the EF since 1978 I get 3:7, plus there's that post of mine on page 1 giving several arguments for 1:2, even if a hair less "efficient". You disproved one of the arguments (that borrowing cost would effect the ratio - thank you!), but I think the other arguments still hold.
User avatar
LTCM
Posts: 412
Joined: Wed Sep 09, 2020 3:58 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by LTCM »

skierincolorado wrote: Thu Sep 09, 2021 4:47 pm
LTCM wrote: Thu Sep 09, 2021 4:34 pm I can see the appeal of 1:3 Stocks/ITT. The simplicity is great. It appeals to my round numbers obsession which helps maintains rebalance discipline. Secondly, when working with relatively low amounts for futures you don't want to be dealing with too many securities.
For the EF since 1978 I get 3:7, plus there's that post of mine on page 1 giving several arguments for 1:2, even if a hair less "efficient". You disproved one of the arguments (that borrowing cost would effect the ratio - thank you!), but I think the other arguments still hold.
The main thing pushing me to a higher stock holding right now is the predicted efficient frontier that efficientinvestor talked about. I feel like the market (and the bond market in particular) just tells you what to do. If bond yield is low risk on. Bond yield is high risk off. We just need one of you to work on those implied yields from the forward prices!

This goes up to 30 years forward: https://data.nasdaq.com/data/FED/SVENF- ... rate-curve
This explains how to derive implied rates from forwards or forwards from current rates/prices etc. It seems they're all derivatives of the standard yield curve but the math is way beyond me: https://www.federalreserve.gov/pubs/fed ... 628pap.pdf

Right now I'm basically still getting 1:1 stocks/LTT:
https://www.portfoliovisualizer.com/eff ... ints=false
55% VUG - 20% VEA - 20% EDV - 5% BNDX
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by skierincolorado »

LTCM wrote: Thu Sep 09, 2021 6:02 pm
skierincolorado wrote: Thu Sep 09, 2021 4:47 pm
LTCM wrote: Thu Sep 09, 2021 4:34 pm I can see the appeal of 1:3 Stocks/ITT. The simplicity is great. It appeals to my round numbers obsession which helps maintains rebalance discipline. Secondly, when working with relatively low amounts for futures you don't want to be dealing with too many securities.
For the EF since 1978 I get 3:7, plus there's that post of mine on page 1 giving several arguments for 1:2, even if a hair less "efficient". You disproved one of the arguments (that borrowing cost would effect the ratio - thank you!), but I think the other arguments still hold.
The main thing pushing me to a higher stock holding right now is the predicted efficient frontier that efficientinvestor talked about. I feel like the market (and the bond market in particular) just tells you what to do. If bond yield is low risk on. Bond yield is high risk off. We just need one of you to work on those implied yields from the forward prices!

This goes up to 30 years forward: https://data.nasdaq.com/data/FED/SVENF- ... rate-curve
This explains how to derive implied rates from forwards or forwards from current rates/prices etc. It seems they're all derivatives of the standard yield curve but the math is way beyond me: https://www.federalreserve.gov/pubs/fed ... 628pap.pdf

Right now I'm basically still getting 1:1 stocks/LTT:
https://www.portfoliovisualizer.com/eff ... ints=false
You used 0.55 for ITT and 1.97 for LTT. That assumes that ITT rates rise .25% per year, while LTT rates only go up by .02%. If you assume rates for both don't change at all, you should use 1.5% for ITT and 2.2% for LTT. If you assume rates rise as the market expects using forward rates, you should use ~.5% for both. If you split the difference you should use 1% for ITT, and 1.35% for LTT. Any of these would be at least a reasonable starting point, and would all strongly exclude LTT from the efficient frontier.
User avatar
LTCM
Posts: 412
Joined: Wed Sep 09, 2020 3:58 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by LTCM »

I'm using data from the fed page:
https://www.treasury.gov/resource-cente ... data=yield
It shows 0.79 ITT and 1.9 LTT
I'm not comfortable switching in the forward affected prices until I understand that a bit better.

https://www.portfoliovisualizer.com/eff ... ints=false
55% VUG - 20% VEA - 20% EDV - 5% BNDX
User avatar
LTCM
Posts: 412
Joined: Wed Sep 09, 2020 3:58 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by LTCM »

So...I'm not 100% sure on this. Math is not my strong point...

There are 2 data sources for forwards:
US Treasury instantaneous forward rate curve
1 year forward rate curve

The 1 year treasury 1 year forward is 0.37
The average of the instantaneous forward 1+2 is 0.37

I think the instantaneous forward shows the per year interest rate for up to 30 years. So I think it should average out to what the 30 year treasury will be in 1 year. Because holding the 30 year for 30 years should be the same as holding a succession of 1 years for 30 years right?

the average of the instantaneous forwards for 30 years is 2.15 - so that's either the 30 year in 1 year or a 29 in 1 year I think. I'm not certain which (though they'll almost certainly be the same anyway).

The average of the IF 1+2 is 0.37 which I think is the 1 year treasury in 1 year.
The average of the IF 1+2+3 is 0.596 which I think is the 2 year treasury in 1 year.
The average of the IF 1+2+3+4+5+6 is 0.975 which I think is the 5 year treasury in 1 year.
The average of the IF 1+2+3...29+30 is 2.15 which I think is the 30 year treasury in 1 year.

Does this seem right?

I just checked with the 1year 4 year forwards and 9 years forward and its really close...
The average of IF 4+5 is 1.51 and the 1YF04 is "supposed" to be 1.52
The average of IF 9+10 is 2.13 and the 1YF09 is "supposed" to be 2.14

So I feel like this is probably close enough for our purposes?
55% VUG - 20% VEA - 20% EDV - 5% BNDX
User avatar
LTCM
Posts: 412
Joined: Wed Sep 09, 2020 3:58 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by LTCM »

Following on from that expected 1 year returns...

2YR = $1.25 (interest) -$2.80 (capital) = -1.55%
5YR = $7.50 (interest) -$6.00 (capital) = 0.15%
30YR = $20 (interest) -$50.00 (capital) = -3.00%

Is that right? It's pretty ugly!
55% VUG - 20% VEA - 20% EDV - 5% BNDX
billb
Posts: 275
Joined: Fri Jun 12, 2009 10:46 am
Location: Kennesaw, GA

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by billb »

Apologies if this is too pedestrian, but much of what's being said here is going over my head and I'm just trying to extract if this is a method that I could adopt, so I have some very basic questions. I'm currently using the 3x ETF version and I'm up just over 25% since starting this in April. I've heard skier talk about futures being more efficient and ITTs achieving a better result than LTT and I'd like explore this in my own account. Questions are:

1. If I have 100K to use on this strategy, is that enough?
2. To keep it simple, what does a 50/50 strategy look like in term of # of contracts for S&P and /ZF?
3. If I only use 100K, is it possible to easily rebalance if things go to 60/40 or higher. It feels like the notional of the futures contracts make that difficult in a 100K allocation. Probably reasonable to do w/ micros, but I don't see /ZF in a micro version.

Thanks for your patience.
User avatar
Bentonkb
Posts: 83
Joined: Sun Mar 07, 2021 5:47 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by Bentonkb »

billb wrote: Fri Sep 10, 2021 8:22 am Apologies if this is too pedestrian, but much of what's being said here is going over my head and I'm just trying to extract if this is a method that I could adopt, so I have some very basic questions. I'm currently using the 3x ETF version and I'm up just over 25% since starting this in April. I've heard skier talk about futures being more efficient and ITTs achieving a better result than LTT and I'd like explore this in my own account. Questions are:

1. If I have 100K to use on this strategy, is that enough?
2. To keep it simple, what does a 50/50 strategy look like in term of # of contracts for S&P and /ZF?
3. If I only use 100K, is it possible to easily rebalance if things go to 60/40 or higher. It feels like the notional of the futures contracts make that difficult in a 100K allocation. Probably reasonable to do w/ micros, but I don't see /ZF in a micro version.

Thanks for your patience.
Skier did post an example portfolio for a 100k investment. If you use ZF contracts the big contract size isn't a big problem because you have to use a big allocation to treasuries to balance out the volatility of your stocks. 50/50 seems like it will be heavily tilted toward stocks if your bonds only have a 4.5 year effective duration. 50/50 is more suited to longer duration treasuries.

There is a new micro treasury futures contract. It has only been trading for a few weeks now. I've looked into it some. The contracts are a lot smaller, but I think it would be a pain to use. The contract is quoted in terms of yield rather than price, so you would need to short the micro to make a position that is equivalent to the standard futures contract. I had a hard time figuring out the right ratio between the standard contracts and the micro. It seems like you would probably want to use the DV01 for both products.

I don't recommend being the guinea pig for the experiment with micro treasuries.

edit:
For a Roth IRA with a 100k balance 125/200 might look like this:
-70k VTI
-10k VB or VBR
-20k cash collateral for futures. check broker for requirements. IB would require 10k I think. And I'd hold 10k extra as buffer.
- 2 MES contracts (S&P500) (~45k)
- 2 ZF contracts (~250k)

You could skip the 10k VB/VBR and just do 80k VTI and it would make very little difference.
viewtopic.php?f=10&t=288192&start=9800
billb
Posts: 275
Joined: Fri Jun 12, 2009 10:46 am
Location: Kennesaw, GA

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by billb »

The 50/50 allocation was just meant to keep it simple in hypothetical land. I'll attempt to locate skier's example setup. I'm looking for a 'model' portfolio that I can see and then analyze the pieces to see how it all fits together. If I had 100K, /ZF notional is 100K. So in my simple 50/50 world, if I'm going for 3x leverage, I would need a contract and a half? This is where I get lost.
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by skierincolorado »

LTCM wrote: Fri Sep 10, 2021 12:55 am I'm using data from the fed page:
https://www.treasury.gov/resource-cente ... data=yield
It shows 0.79 ITT and 1.9 LTT
I'm not comfortable switching in the forward affected prices until I understand that a bit better.

https://www.portfoliovisualizer.com/eff ... ints=false
I'm not 100% sure how to factor in forward rates yet either, but we should definitely factor in roll since that is known and easily calculable. Factoring in roll gives 1.5% and 2.2% for ITT and LTT respectively. Assuming ITT rates rise .25% and LTT rates rise .1% brings them down to .55% and .45% respectively. It's this latter assumption I'm not 100% confident in, but seems like a very reasonable estimate given forward rates.
User avatar
Bentonkb
Posts: 83
Joined: Sun Mar 07, 2021 5:47 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by Bentonkb »

billb wrote: Fri Sep 10, 2021 8:48 am The 50/50 allocation was just meant to keep it simple in hypothetical land. I'll attempt to locate skier's example setup. I'm looking for a 'model' portfolio that I can see and then analyze the pieces to see how it all fits together. If I had 100K, /ZF notional is 100K. So in my simple 50/50 world, if I'm going for 3x leverage, I would need a contract and a half? This is where I get lost.
https://www.cmegroup.com/tools-informat ... ytics.html

The par value is $100k. The invoice price if the contract expired today would be 123-19.25 * $1000 * TCF = $98,869. A lot of people use the settlement price * $1000 = $123,600 as the notional contract value, but I'm not convinced that is the right way to model it.
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by skierincolorado »

billb wrote: Fri Sep 10, 2021 8:48 am The 50/50 allocation was just meant to keep it simple in hypothetical land. I'll attempt to locate skier's example setup. I'm looking for a 'model' portfolio that I can see and then analyze the pieces to see how it all fits together. If I had 100K, /ZF notional is 100K. So in my simple 50/50 world, if I'm going for 3x leverage, I would need a contract and a half? This is where I get lost.
Yes if you had 100k, with a 50/50 AA leveraged 3x to 150/150, you would want 1.5 ZFs. Which isn't possible of course, but there are other ways to own 150k of bonds.

1) You could buy 50k of VGIT (ITT ETF) and 1 ZF. This gets you to 150k of bonds, but you still need 150k of stock. You have 50k of cash remaining. You could buy 6 MES contracts (each expose 22.5k of S&P500). That gets you to 135k of stock. Then buy 15k of VXUS or VBR. You'd have 35k of cash remaining as collateral for 1 ZF and 6 MES. Each ZF has a collateral requirement of 800 in brokerage, 1600 in IRA. Each MES is 1500 in taxable, 3000 in IRA. Assuming this is in an IRA the collateral requirement would be 19,800. You would hold 35k in cash to meet this requirement, with $15.2k to spare. Or you could sell all the VXUS/VBR (15k) and buy a 7th MES. This would give you 50k of cash and the requirement would go up by 3k to 22.8k (19.8k +3k). Now you have 50k of cash with a 22.8k requirement, leaving 27.2k to spare in case the market goes down.

2) You could buy a ZN instead of a ZF. The average duration of ZN is 7 years instead of 4.5. This makes it approximately 1.5x riskier, and also approximately 1.5x higher return/interest. So it's functionally very similar to holding 1 ZF, and is still "ITT". So you have 100k of ITT w/ 7 year duration now, but functionally 150k of ITT w/ 4.5 year duration. Now you need 150k of stock. I'd buy 60k of VXUS/VBR and 4 MES (90k). Leaving you with 40k of cash collateral for 4 MES and 1 ZN.

I would prefer method #2.

The goal is simply to own 150k of stock and 150k of ITT (converted to 5-yr duration). There's a lot of ways to do this. The higher your NW is the easier it gets (although doing across multiple accounts as I do gets tricky), although it can really be done with any account size.

Also I would say a 1:1 ratio of stock to ITT is far too low. The ITT will have some effect on the portfolio, but not much. Historically a 3:7 ratio was best since 1978. I do 1:2 to keep it simple and because I don't want to take a ton of risk in bonds while everyone else owns stocks. It's bad to lose money when everyone else is making money.
Last edited by skierincolorado on Fri Sep 10, 2021 10:05 am, edited 1 time in total.
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by skierincolorado »

Bentonkb wrote: Fri Sep 10, 2021 8:42 am
billb wrote: Fri Sep 10, 2021 8:22 am Apologies if this is too pedestrian, but much of what's being said here is going over my head and I'm just trying to extract if this is a method that I could adopt, so I have some very basic questions. I'm currently using the 3x ETF version and I'm up just over 25% since starting this in April. I've heard skier talk about futures being more efficient and ITTs achieving a better result than LTT and I'd like explore this in my own account. Questions are:

1. If I have 100K to use on this strategy, is that enough?
2. To keep it simple, what does a 50/50 strategy look like in term of # of contracts for S&P and /ZF?
3. If I only use 100K, is it possible to easily rebalance if things go to 60/40 or higher. It feels like the notional of the futures contracts make that difficult in a 100K allocation. Probably reasonable to do w/ micros, but I don't see /ZF in a micro version.

Thanks for your patience.
Skier did post an example portfolio for a 100k investment. If you use ZF contracts the big contract size isn't a big problem because you have to use a big allocation to treasuries to balance out the volatility of your stocks. 50/50 seems like it will be heavily tilted toward stocks if your bonds only have a 4.5 year effective duration. 50/50 is more suited to longer duration treasuries.

There is a new micro treasury futures contract. It has only been trading for a few weeks now. I've looked into it some. The contracts are a lot smaller, but I think it would be a pain to use. The contract is quoted in terms of yield rather than price, so you would need to short the micro to make a position that is equivalent to the standard futures contract. I had a hard time figuring out the right ratio between the standard contracts and the micro. It seems like you would probably want to use the DV01 for both products.

I don't recommend being the guinea pig for the experiment with micro treasuries.

edit:
For a Roth IRA with a 100k balance 125/200 might look like this:
-70k VTI
-10k VB or VBR
-20k cash collateral for futures. check broker for requirements. IB would require 10k I think. And I'd hold 10k extra as buffer.
- 2 MES contracts (S&P500) (~45k)
- 2 ZF contracts (~250k)

You could skip the 10k VB/VBR and just do 80k VTI and it would make very little difference.
viewtopic.php?f=10&t=288192&start=9800
I've looked into micros too. The volume seems a bit thin and I don't want to be the guinea pig either, but are you sure that you short them to profit from interest rates? I know they are quoted in terms of yield, but just because they are quoted like that doesn't really mean that's what the actual value is. I assume they are just quoting them that way to make them easier to understand, but that you make money by buying them not shorting them (unless betting on an interest rate increase or doing some kind of hedging).
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by skierincolorado »

LTCM wrote: Fri Sep 10, 2021 5:07 am So...I'm not 100% sure on this. Math is not my strong point...

There are 2 data sources for forwards:
US Treasury instantaneous forward rate curve
1 year forward rate curve

The 1 year treasury 1 year forward is 0.37
The average of the instantaneous forward 1+2 is 0.37

I think the instantaneous forward shows the per year interest rate for up to 30 years. So I think it should average out to what the 30 year treasury will be in 1 year. Because holding the 30 year for 30 years should be the same as holding a succession of 1 years for 30 years right?

the average of the instantaneous forwards for 30 years is 2.15 - so that's either the 30 year in 1 year or a 29 in 1 year I think. I'm not certain which (though they'll almost certainly be the same anyway).

The average of the IF 1+2 is 0.37 which I think is the 1 year treasury in 1 year.
The average of the IF 1+2+3 is 0.596 which I think is the 2 year treasury in 1 year.
The average of the IF 1+2+3+4+5+6 is 0.975 which I think is the 5 year treasury in 1 year.
The average of the IF 1+2+3...29+30 is 2.15 which I think is the 30 year treasury in 1 year.

Does this seem right?

I just checked with the 1year 4 year forwards and 9 years forward and its really close...
The average of IF 4+5 is 1.51 and the 1YF04 is "supposed" to be 1.52
The average of IF 9+10 is 2.13 and the 1YF09 is "supposed" to be 2.14

So I feel like this is probably close enough for our purposes?
This isn't quite right. For one, long-term rates are not merely the average of short term rates over the time horizon. There's also the term premium. This was an approximate quote from the paper you posted yesterday

"Long term rates are the average of expected short term rates plus a term premium"

The current 30 year rate should already reflect that.. not just the 30 year rate 1 year from now. So since 1 year rates are clearly supposed to average over 2% for the next 30 years, the term premium on the 30 year is currently negative. The term premium 1 year from now could also be negative. But at some point it will likely become positive, and rates on the 30 year would exceed the 2.14% you calculate, possibly going to 3 or 4% depending on how large the term premium gets.

It's also not exactly clear to me whether forward rates also include a term premium within themselves.

What is clear is that rates are supposed to rise across the whole curve. Which is usually not a good time to own LTT historically because rising rates impact LTT more. From the forward rates, it looks to me like ITT rates should rise ~.25%/yr and LTT rates should rise ~.1% per year, eventually settling at 2.1% short term rates, 2.6% ITT rates, and 3% LTT rates. If we use these estimates, ITT will return .55% in the next year including roll, and LTT rates will return .45% including roll.

These are more like "estimates" based on forward rates. But as I said in my last post we definitely can and should include roll. Which gives 1.5% and 2.2% for ITT and LTT respectively. After that, it's just a matter of how much do you want to revise down based on forward rates.

The biggest thing to understand is that using current YTM is absolutely an incorrect metric for two main reasons:

1) It doesn't include roll. We are not holding to maturity, we are rolling. Factoring in roll brings ITT to 1.5% and LTT to 2.2%.

2) The current yield curve has priced in future changes in rates that will affect returns. Even if we can't calculate what these changes are exactly, we know that they are there and built into market expectations and current interest rates. Given this, we probably should not assume that LTT will have better risk adjusted returns than ITT for the first time in history simply because the interest rate is higher. There's a high probability the interest rate is higher because the market expect long-term rates to increase, which will hold down returns for owning LTT in the short-term. We should either try to estimate this from forward rates (which seem to confirm this suspicion even if we don't have the math perfect yet) or we should stick with the historical fact that ITT have consistently had better risk-adjusted returns.
billb
Posts: 275
Joined: Fri Jun 12, 2009 10:46 am
Location: Kennesaw, GA

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by billb »

skierincolorado wrote: Fri Sep 10, 2021 9:14 am
Yes if you had 100k, with a 50/50 AA leveraged 3x to 150/150, you would want 1.5 ZFs. Which isn't possible of course, but there are other ways to own 150k of bonds.

1) You could buy 50k of VGIT (ITT ETF) and 1 ZF. This gets you to 150k of bonds, but you still need 150k of stock. You have 50k of cash remaining. You could buy 6 MES contracts (each expose 22.5k of S&P500). That gets you to 135k of stock. Then buy 15k of VTI or VBR. You'd have 35k of cash remaining as collateral for 1 ZF and 6 MES. Each ZF has a collateral requirement of 800 in brokerage, 1600 in IRA. Each MES is 1500 in taxable, 3000 in IRA. Assuming this is in an IRA the collateral requirement would be 19,800. You would hold 35k in cash to meet this requirement, with $15.2k to spare. Or you could sell all the VTI/VBR (15k) and buy a 7th MES. This would give you 50k of cash and the requirement would go up by 3k to 22.8k (19.8k +3k). Now you have 50k of cash with a 22.8k requirement, leaving 27.2k to spare in case the market goes down.

2) You could buy a ZN instead of a ZF. The average duration of ZN is 7 years instead of 4.5. This makes it approximately 1.5x riskier, and also approximately 1.5x higher return/interest. So it's functionally very similar to holding 1 ZF, and is still "ITT". So you have 100k of ITT w/ 7 year duration now, but functionally 150k of ITT w/ 4.5 year duration. Now you need 150k of stock. I'd buy 60k of VTI and 4 MES (90k). Leaving you with 40k of cash collateral for 4 MES and 1 ZN.

I would prefer method #2.

The goal is simply to own 150k of stock and 150k of ITT (converted to 5-yr duration). There's a lot of ways to do this. The higher your NW is the easier it gets (although doing across multiple accounts as I do gets tricky), although it can really be done with any account size.

Also I would say a 1:1 ratio of stock to ITT is far too low. The ITT will have some effect on the portfolio, but not much. Historically a 3:7 ratio was best since 1978. I do 1:2 to keep it simple and because I don't want to take a ton of risk in bonds while everyone else owns stocks. It's bad to lose money when everyone else is making money.
Thank you, thank you. This was tripping me up with regard to implementation. This clears up a lot as far as setup and future management. 1:1 was just to simplify, I'd rather not have such a large bond holding. Method 2 makes a lot of sense!
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by skierincolorado »

billb wrote: Fri Sep 10, 2021 9:41 am
skierincolorado wrote: Fri Sep 10, 2021 9:14 am
Yes if you had 100k, with a 50/50 AA leveraged 3x to 150/150, you would want 1.5 ZFs. Which isn't possible of course, but there are other ways to own 150k of bonds.

1) You could buy 50k of VGIT (ITT ETF) and 1 ZF. This gets you to 150k of bonds, but you still need 150k of stock. You have 50k of cash remaining. You could buy 6 MES contracts (each expose 22.5k of S&P500). That gets you to 135k of stock. Then buy 15k of VTI or VBR. You'd have 35k of cash remaining as collateral for 1 ZF and 6 MES. Each ZF has a collateral requirement of 800 in brokerage, 1600 in IRA. Each MES is 1500 in taxable, 3000 in IRA. Assuming this is in an IRA the collateral requirement would be 19,800. You would hold 35k in cash to meet this requirement, with $15.2k to spare. Or you could sell all the VTI/VBR (15k) and buy a 7th MES. This would give you 50k of cash and the requirement would go up by 3k to 22.8k (19.8k +3k). Now you have 50k of cash with a 22.8k requirement, leaving 27.2k to spare in case the market goes down.

2) You could buy a ZN instead of a ZF. The average duration of ZN is 7 years instead of 4.5. This makes it approximately 1.5x riskier, and also approximately 1.5x higher return/interest. So it's functionally very similar to holding 1 ZF, and is still "ITT". So you have 100k of ITT w/ 7 year duration now, but functionally 150k of ITT w/ 4.5 year duration. Now you need 150k of stock. I'd buy 60k of VTI and 4 MES (90k). Leaving you with 40k of cash collateral for 4 MES and 1 ZN.

I would prefer method #2.

The goal is simply to own 150k of stock and 150k of ITT (converted to 5-yr duration). There's a lot of ways to do this. The higher your NW is the easier it gets (although doing across multiple accounts as I do gets tricky), although it can really be done with any account size.

Also I would say a 1:1 ratio of stock to ITT is far too low. The ITT will have some effect on the portfolio, but not much. Historically a 3:7 ratio was best since 1978. I do 1:2 to keep it simple and because I don't want to take a ton of risk in bonds while everyone else owns stocks. It's bad to lose money when everyone else is making money.
Thank you, thank you. This was tripping me up with regard to implementation. This clears up a lot as far as setup and future management. 1:1 was just to simplify, I'd rather not have such a large bond holding. Method 2 makes a lot of sense!
Just wanted to make sure you saw the discussion of the stock to bond ratio. Historically it's 3:7 stock to bond. The bond holding being larger. 50/50 stock bond is very low. Holding 50% in bonds would help diversify and improve risk-adjusted returns, leave a lot of room for improvement. The sharpe ratio of 30:70 is .7, the sharpe ratio of 50:50 is .66, and the sharpe ratio of 100:0 is .54. (since 1978). So I like portfolios of 125/250, 150/300, or 125/200 if wanting to trim the bonds a little.

I suppose the .66 sharpe ratio of 50:50 doesn't seem bad, but 1978-present was an incredible period for stocks. It was good for bonds too, but incredible for stocks. The problem with 50:50 to me is that it's still way more risk in stocks than bonds. Stocks are ~5x more risky than 5-yr bonds. The benefit of a 33:67 stock:bond is that if stocks do really poorly (like 1990s Japan) the next 30 years, we'd still be OK if bonds did OK, since we are taking nearly as much risk in bonds as in stocks.

Also I left out international stocks in my example, but I think roughly 1/3 of the stock holding should be intl.
User avatar
Bentonkb
Posts: 83
Joined: Sun Mar 07, 2021 5:47 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by Bentonkb »

skierincolorado wrote: Fri Sep 10, 2021 9:19 am
I've looked into micros too. The volume seems a bit thin and I don't want to be the guinea pig either, but are you sure that you short them to profit from interest rates? I know they are quoted in terms of yield, but just because they are quoted like that doesn't really mean that's what the actual value is. I assume they are just quoting them that way to make them easier to understand, but that you make money by buying them not shorting them (unless betting on an interest rate increase or doing some kind of hedging).
The small exchange has the current quote for their /S10Y at 13.32 with a notional size of $1332 per contract. They say this is the YTM of the on-the-run 10 year note.

It doesn't really look like it is equivalent to /TN at all because /S10Y has no expected return if you buy and hold it during a period when the yield curve is constant. It a pure play on the changes in the yield curve.
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by skierincolorado »

Bentonkb wrote: Fri Sep 10, 2021 10:24 am
skierincolorado wrote: Fri Sep 10, 2021 9:19 am
I've looked into micros too. The volume seems a bit thin and I don't want to be the guinea pig either, but are you sure that you short them to profit from interest rates? I know they are quoted in terms of yield, but just because they are quoted like that doesn't really mean that's what the actual value is. I assume they are just quoting them that way to make them easier to understand, but that you make money by buying them not shorting them (unless betting on an interest rate increase or doing some kind of hedging).
The small exchange has the current quote for their /S10Y at 13.32 with a notional size of $1332 per contract. They say this is the YTM of the on-the-run 10 year note.

It doesn't really look like it is equivalent to /TN at all because /S10Y has no expected return if you buy and hold it during a period when the yield curve is constant. It a pure play on the changes in the yield curve.
Ugh that's too bad, useless for buy and hold.

"For example, the Micro Ten-Year Yield futures trading at an index level of 1.613 represents a 1.613% yield on the current 10-Year OTR US Treasury note. If the yield index value increases by 0.1 basis points to 1.614, this represents a gain to a long position of $1 or a loss of $1 to a short position."
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by skierincolorado »

Posting this here too because it's the best summary of treasury futures financing costs I've found yet:

https://www.financialresearch.gov/brief ... Trades.pdf
AllomancerJack
Posts: 18
Joined: Fri May 21, 2021 11:58 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by AllomancerJack »

Is there any "rule of thumb" regarding futures delivery date?
User avatar
Bentonkb
Posts: 83
Joined: Sun Mar 07, 2021 5:47 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by Bentonkb »

AllomancerJack wrote: Fri Sep 10, 2021 12:01 pm Is there any "rule of thumb" regarding futures delivery date?
Are you wondering which contract to buy? The right answer is always the contract that will expire first unless it only has a week to go, then you would want the next contract. The contracts are March, June, September, and December.
User avatar
LTCM
Posts: 412
Joined: Wed Sep 09, 2020 3:58 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by LTCM »

skierincolorado wrote: Fri Sep 10, 2021 9:32 am
This isn't quite right. For one, long-term rates are not merely the average of short term rates over the time horizon. There's also the term premium. This was an approximate quote from the paper you posted yesterday

"Long term rates are the average of expected short term rates plus a term premium"

The current 30 year rate should already reflect that.. not just the 30 year rate 1 year from now. So since 1 year rates are clearly supposed to average over 2% for the next 30 years, the term premium on the 30 year is currently negative. The term premium 1 year from now could also be negative. But at some point it will likely become positive, and rates on the 30 year would exceed the 2.14% you calculate, possibly going to 3 or 4% depending on how large the term premium gets.

It's also not exactly clear to me whether forward rates also include a term premium within themselves.
I think forward rates do include a term premium. I think agreeing to a 1 year loan in 30 years includes a term premium. Which is what the forward is. Agreeing to a 2 year loan in 29 years includes term premium. 3 year loan in 28 years… 30 year loan in 1 year.

I’m increasingly confident in those numbers I worked out last night. I think the instantaneous forwards are exactly the tool needed to temporarily shift the yield curve.
55% VUG - 20% VEA - 20% EDV - 5% BNDX
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by skierincolorado »

LTCM wrote: Fri Sep 10, 2021 1:13 pm
skierincolorado wrote: Fri Sep 10, 2021 9:32 am
This isn't quite right. For one, long-term rates are not merely the average of short term rates over the time horizon. There's also the term premium. This was an approximate quote from the paper you posted yesterday

"Long term rates are the average of expected short term rates plus a term premium"

The current 30 year rate should already reflect that.. not just the 30 year rate 1 year from now. So since 1 year rates are clearly supposed to average over 2% for the next 30 years, the term premium on the 30 year is currently negative. The term premium 1 year from now could also be negative. But at some point it will likely become positive, and rates on the 30 year would exceed the 2.14% you calculate, possibly going to 3 or 4% depending on how large the term premium gets.

It's also not exactly clear to me whether forward rates also include a term premium within themselves.
I think forward rates do include a term premium. I think agreeing to a 1 year loan in 30 years includes a term premium. Which is what the forward is. Agreeing to a 2 year loan in 29 years includes term premium. 3 year loan in 28 years… 30 year loan in 1 year.

I’m increasingly confident in those numbers I worked out last night. I think the instantaneous forwards are exactly the tool needed to temporarily shift the yield curve.
Oh I see, so since the forwards include a term premium, and the 30-yr rate 1 year from now ALSO includes a term premium, the average of the forward IFs 1-30 rates should be the 30-yr rate 1-yr from now. That makes sense. Does the current 3-month rate averaged with the next 29 IFs equal the current 30-yr rate?

It doesn't quite work out for me.. I would predict the current 30-yr rate would be 2.06% using the current short-term rate and the next 29 instantaneous forward rates.

There could be some reason it's slightly off? But the gap should be accurate for the current 30-yr vs the 30-yr 1 year from now. The gap is .09%. So that means 30-yr rates are expected to rise .09% over the next year... very close to the .1% I estimated. Although the forwards don't predict the current rate perfectly, the gap between the each year should be accurate. This is great, thank you. So would we agree the 30-yr rate is supposed to rise .09% in the next year?

Basically you got 2.15 for the 30 year 1 year from now. Using the same method to calculate the current 30-yr rate you would get 2.06. Which isn't accurate. But the delta between the predicted current 30-yr rate, and the predicted 30-yr rate 1-yr from now, should accurately reflect market expectations. The gap is .09%!

We can do the same for the 5 year. The predicted 5-yr 1-yr from now is .96%. The predicted 5-yr currently is .64% (.03+.15+.6+1.04+1.39)/5. That's a .32% rate increase.

Although I think this is more like the 4-yr rate because it's the current 3-mo rate averaged with the next 4 IFs. To center it over 5 years, we would need to use the current 3-mo, and the next 5 IFs, and coung IFs 1-4 twice each. Doing that I get .802% for the current 5 year (.03+.15*2+.6*2+1.04*2+1.39*2+1.63)/10) . .... And I get 1.127 for the 5-yr rate 1-yr from now ( .15+.6*2+1.04*2+1.39*2+1.63*2+1.8)/10). Which is still a .325% rate increase.


So 30-yr rates are expected to go up .09%. 5 year rates are expected to go up .325%. Plugging this into a bond calculator, that's a 1.6% capital loss for the 5-yr, and a 2.0% capital loss for the 30 yr. After you factor in return from roll, that brings both of their returns very close to 0%.
Last edited by skierincolorado on Fri Sep 10, 2021 1:53 pm, edited 2 times in total.
User avatar
LTCM
Posts: 412
Joined: Wed Sep 09, 2020 3:58 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by LTCM »

skierincolorado wrote: Fri Sep 10, 2021 1:23 pm
LTCM wrote: Fri Sep 10, 2021 1:13 pm
skierincolorado wrote: Fri Sep 10, 2021 9:32 am
This isn't quite right. For one, long-term rates are not merely the average of short term rates over the time horizon. There's also the term premium. This was an approximate quote from the paper you posted yesterday

"Long term rates are the average of expected short term rates plus a term premium"

The current 30 year rate should already reflect that.. not just the 30 year rate 1 year from now. So since 1 year rates are clearly supposed to average over 2% for the next 30 years, the term premium on the 30 year is currently negative. The term premium 1 year from now could also be negative. But at some point it will likely become positive, and rates on the 30 year would exceed the 2.14% you calculate, possibly going to 3 or 4% depending on how large the term premium gets.

It's also not exactly clear to me whether forward rates also include a term premium within themselves.
I think forward rates do include a term premium. I think agreeing to a 1 year loan in 30 years includes a term premium. Which is what the forward is. Agreeing to a 2 year loan in 29 years includes term premium. 3 year loan in 28 years… 30 year loan in 1 year.

I’m increasingly confident in those numbers I worked out last night. I think the instantaneous forwards are exactly the tool needed to temporarily shift the yield curve.
Oh I see, so since the forwards include a term premium, and the 30-yr rate 1 year from now all includes a term premium, the average of the forward 1-yr rates should be the 30-yr rate 1-yr from now. That makes sense. Does the current 1-yr rate averaged with the next 29 1-yr forwards equal the current 30-yr rate?

It doesn't quite work out for me.. I would predict the current 30-yr rate would be 2.06% using the current short-term rate and the next 29 instantaneous forward rates.

There could be some reason it's slightly off? But the gap should be accurate for the current 30-yr vs the 30-yr 1 year from now. The gap is .08%. So that means 30-yr rates are expected to rise .08% over the next year... reasonably close to the .1% I estimated!
If you set the current rate to zero (overnight rate) and average with the 1 year instantaneous forward you get 0.075 which is the current 1 year rate. I was also finding it slightly overestimated the current 30 year rate though too when using the same method.
55% VUG - 20% VEA - 20% EDV - 5% BNDX
AllomancerJack
Posts: 18
Joined: Fri May 21, 2021 11:58 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by AllomancerJack »

Bentonkb wrote: Fri Sep 10, 2021 1:12 pm Are you wondering which contract to buy? The right answer is always the contract that will expire first unless it only has a week to go, then you would want the next contract. The contracts are March, June, September, and December.
Doesn't it requires more frequent switching from one contract to another? Could you please explain why it is undesirable to buy the contract with latest delivery date (minimizing broker fees and contract switching burden)?
User avatar
LTCM
Posts: 412
Joined: Wed Sep 09, 2020 3:58 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by LTCM »

The other small issue is that the instantaneous forwards only go forward 30 years. The 1Y01F is avg of IF1+IF2. The 2Y01F is IF1+IF2+IF3. So the 30Y01F would need to include IF31 which we don’t have. So you can either make it up (it’s very likely the same as IF30) or accept you’re only able to move the 29 year bond forward 1 year.
55% VUG - 20% VEA - 20% EDV - 5% BNDX
User avatar
LTCM
Posts: 412
Joined: Wed Sep 09, 2020 3:58 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by LTCM »

AllomancerJack wrote: Fri Sep 10, 2021 1:47 pm
Bentonkb wrote: Fri Sep 10, 2021 1:12 pm Are you wondering which contract to buy? The right answer is always the contract that will expire first unless it only has a week to go, then you would want the next contract. The contracts are March, June, September, and December.
Doesn't it requires more frequent switching from one contract to another? Could you please explain why it is undesirable to buy the contract with latest delivery date (minimizing broker fees and contract switching burden)?
The spreads are horrible. Switching between contracts is very cheap and the spreads are incredibly tight. Cheaper and tighter than buying into your initial contract.
55% VUG - 20% VEA - 20% EDV - 5% BNDX
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by skierincolorado »

LTCM wrote: Fri Sep 10, 2021 1:36 pm
skierincolorado wrote: Fri Sep 10, 2021 1:23 pm
LTCM wrote: Fri Sep 10, 2021 1:13 pm
skierincolorado wrote: Fri Sep 10, 2021 9:32 am
This isn't quite right. For one, long-term rates are not merely the average of short term rates over the time horizon. There's also the term premium. This was an approximate quote from the paper you posted yesterday

"Long term rates are the average of expected short term rates plus a term premium"

The current 30 year rate should already reflect that.. not just the 30 year rate 1 year from now. So since 1 year rates are clearly supposed to average over 2% for the next 30 years, the term premium on the 30 year is currently negative. The term premium 1 year from now could also be negative. But at some point it will likely become positive, and rates on the 30 year would exceed the 2.14% you calculate, possibly going to 3 or 4% depending on how large the term premium gets.

It's also not exactly clear to me whether forward rates also include a term premium within themselves.
I think forward rates do include a term premium. I think agreeing to a 1 year loan in 30 years includes a term premium. Which is what the forward is. Agreeing to a 2 year loan in 29 years includes term premium. 3 year loan in 28 years… 30 year loan in 1 year.

I’m increasingly confident in those numbers I worked out last night. I think the instantaneous forwards are exactly the tool needed to temporarily shift the yield curve.
Oh I see, so since the forwards include a term premium, and the 30-yr rate 1 year from now all includes a term premium, the average of the forward 1-yr rates should be the 30-yr rate 1-yr from now. That makes sense. Does the current 1-yr rate averaged with the next 29 1-yr forwards equal the current 30-yr rate?

It doesn't quite work out for me.. I would predict the current 30-yr rate would be 2.06% using the current short-term rate and the next 29 instantaneous forward rates.

There could be some reason it's slightly off? But the gap should be accurate for the current 30-yr vs the 30-yr 1 year from now. The gap is .08%. So that means 30-yr rates are expected to rise .08% over the next year... reasonably close to the .1% I estimated!
If you set the current rate to zero (overnight rate) and average with the 1 year instantaneous forward you get 0.075 which is the current 1 year rate. I was also finding it slightly overestimated the current 30 year rate though too when using the same method.
Could be some reason it's slightly off but, the delta between years should be accurate.

We can do the same for the 5 year. The predicted 5-yr 1-yr from now is .96%. The predicted 5-yr currently is .64% (.03+.15+.6+1.04+1.39)/5. That's a .32% rate increase.

Although I think this is more like the 4-yr rate because it's the current 3-mo rate averaged with the next 4 IFs. To center it over 5 years, we would need to use the current 3-mo, and the next 5 IFs, and coung IFs 1-4 twice each. Doing that I get .802% for the current 5 year (.03+.15*2+.6*2+1.04*2+1.39*2+1.63)/10) . .... And I get 1.127 for the 5-yr rate 1-yr from now ( .15+.6*2+1.04*2+1.39*2+1.63*2+1.8)/10). Which is still a .325% rate increase.


So 30-yr rates are expected to go up .09%. 5 year rates are expected to go up .325%. Plugging this into a bond calculator, that's a 1.6% capital loss for the 5-yr, and a 2.0% capital loss for the 30 yr. After you factor in return from roll (.9% for the 5 year, and .2% for the 30 year), that brings both of their returns very close to 0%.

For the 5 year: -1.6% due to interest rate increase + .9% from roll + .8% interest = +.1%
For the 30 year: -2.0% due to interest rate increase + .2% from roll + 1.9% interest = +.1%


Roll calculations:
For the 30 year the difference in rate between the 20y and the 30y is .07%. Per year this is less than .01% (I rounded up to .01%). When the yield of a 30-yr bond with a 1.9% coupon falls by .01%, the price increases .2%.

For the 5 year, the difference in rate between the 3y and the 5y is .36%, or .18% per year. When the yield of a 5-yr bond with a .8% coupon falls .18%, the price increases .9%.


If we put .1% for both ITT and LTT in the efficient frontier assumptions we get ~10% each to LTT and ITT. When we do .2% expected return, we get 32% to ITT and nothing to LTT.

In summary, we should expect bond returns for the next year or two to be between 0 and 0.5% for both ITT and LTT. Given the uncertainty, we should go with the fact that historically efficient frontier heavily favors STT and ITT.

Given these low expected bond returns, the EF seems favors 50:50 stocks/ITT using EfficientInvestors assumption of 6% domestic and 7% international stock returns.

My final assumptions are this:

.3% ITT
.4% LTT
4.5% domestic stock
5% international stock

The efficient frontier is: 34/22/44 domestic stock / intl stock / ITT

https://www.portfoliovisualizer.com/eff ... ints=false
Last edited by skierincolorado on Fri Sep 10, 2021 2:29 pm, edited 1 time in total.
User avatar
LTCM
Posts: 412
Joined: Wed Sep 09, 2020 3:58 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by LTCM »

skierincolorado wrote: Fri Sep 10, 2021 1:56 pm For the 5 year: -1.6% due to interest rate increase + .9% from roll + .8% interest = +.1%
For the 30 year: -2.0% due to interest rate increase + .2% from roll + 1.9% interest = +.1%
how's the 2 year?
55% VUG - 20% VEA - 20% EDV - 5% BNDX
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by skierincolorado »

LTCM wrote: Fri Sep 10, 2021 2:23 pm
skierincolorado wrote: Fri Sep 10, 2021 1:56 pm For the 5 year: -1.6% due to interest rate increase + .9% from roll + .8% interest = +.1%
For the 30 year: -2.0% due to interest rate increase + .2% from roll + 1.9% interest = +.1%
how's the 2 year?
Current predicted 2 year: (.03 + .15*2 + .6)/4 = .232
1 year from now predicted 2 year: (.15 + .6*2 + 1.04)/4 = .598

Capital loss due to a .366% rate increase = .73%

Slope is ~.15% so the gain from roll is = .3%

Interest rate is .2%

-.73% + .3% + .2% = -.23% expected return

This might be a hair pessimistic since the .366% rate increase over the next year is probably expected to be backloaded by the market.
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by skierincolorado »

In case anyone missed it, this is my "final" estimate.

.3% ITT
.4% LTT
4.5% domestic stock
5% international stock

The efficient frontier is: 34/22/44 domestic stock / intl stock / ITT

https://www.portfoliovisualizer.com/eff ... ints=false

***This is extremely sensitive to asssumptions. I suggest simply using the historical efficient frontier.***
Last edited by skierincolorado on Fri Sep 10, 2021 8:38 pm, edited 2 times in total.
DMoogle
Posts: 549
Joined: Sat Oct 31, 2020 10:24 pm

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by DMoogle »

skierincolorado wrote: Fri Sep 10, 2021 2:39 pm In case anyone missed it, this is my "final" estimate.

.3% ITT
.4% LTT
4.5% domestic stock
5% international stock

The efficient frontier is: 34/22/44 domestic stock / intl stock / ITT

https://www.portfoliovisualizer.com/eff ... ints=false
It's worth noting that this efficient frontier is EXTREMELY sensitive to those expected return values. If you change domestic return to 5.5%, then international allocation becomes 7%. If you change it to 6%, then it nearly removes international altogether.
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by skierincolorado »

DMoogle wrote: Fri Sep 10, 2021 3:02 pm
skierincolorado wrote: Fri Sep 10, 2021 2:39 pm In case anyone missed it, this is my "final" estimate.

.3% ITT
.4% LTT
4.5% domestic stock
5% international stock

The efficient frontier is: 34/22/44 domestic stock / intl stock / ITT

https://www.portfoliovisualizer.com/eff ... ints=false
It's worth noting that this efficient frontier is EXTREMELY sensitive to those expected return values. If you change domestic return to 5.5%, then international allocation becomes 7%. If you change it to 6%, then it nearly removes international altogether.
Yes I should have put that caveat in. If you lower the bond returns to what LTCM and I calculated (.1%) it mostly eliminates all bonds.

Which is yet another argument for going with the historical efficient frontier rather than trying to project forward expectations which are sensitive to assumptions. Since 1978 looks like this:

https://www.portfoliovisualizer.com/eff ... ngTreasury


International isn't available over that time period, and did poorly over the time period it is available (1986-present). Common sense says we should include at least some intl.
adamhg
Posts: 218
Joined: Sat Apr 10, 2021 8:40 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by adamhg »

Been researching this ever since skiers posts in hfea a few weeks back. I'm primarily looking to do this in my taxable, so I've been looking into the box spread method. But what advantage does futures present over box spreads, even in tax advantaged accounts? Seems like the borrow rate would be equivalent with much fewer rolls (2-3 years) and you get around imprecise ratios from the large nominal values
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by skierincolorado »

adamhg wrote: Fri Sep 10, 2021 7:52 pm Been researching this ever since skiers posts in hfea a few weeks back. I'm primarily looking to do this in my taxable, so I've been looking into the box spread method. But what advantage does futures present over box spreads, even in tax advantaged accounts? Seems like the borrow rate would be equivalent with much fewer rolls (2-3 years) and you get around imprecise ratios from the large nominal values
For one, I'm pretty sure you can't do box-spreads in an IRA because that level of options writing is restricted.

Besides, in an IRA, futures are preferable because the borrow rate on box-spreads is higher. I get like .7% on boxes, vs .1% on Treasury Futures and .4% on MES futures.

Also, it's hard to buy 200-300% of portfolio value in ITT using margin. I don't know about other brokers, but the margin requirement for bond ETFs are the same as for stock ETFs at IBKR. You can buy actual bonds with low margin requirements, but I found that to be a terrible pain.

The borrow rate on box-spreads is so high, I'm thinking of doing the math on whether it's worth using MES futures in taxable as well. Inferior tax treatment, but lower borrow rate. Maybe if I wrote larger boxes I'd get down to .6% or something.

I already use treasury futures in taxable because of the large amounts being borrowed and the very low implied finance rates ~.1% they offer and because bond interest is taxable anyways. Treasury futures might actually be preferable to actual bonds from a tax perspective - not sure on that one. So I don't use box-spreads except for equity leverage in taxable.

One has to rebalance from time to time anyways, so I do this when rolling. I also find rolling futures to be easier than writing box-spreads. They're much more liquid and I don't have to worry about the bid-ask spread.

Finally, generally speaking, it's best to take the same amount of leverage with the same AA across all accounts, or at least look at the big picture when deciding on an investment plan. For example, it doesn't make sense to be leveraging stock in taxable, paying .4-.7% financing rates, while having an unleveraged 80/20 or 60/40 in your IRA or 401k. You could sell bonds in the IRA and buy stock, and sell stock in the IRA and then buy treasury futures to account for the bonds sold. It would be the exact same AA, but with less borrowing costs. Decide on your target AA for your entire net worth, and then decide the most efficient way to implement it. Finally, since 'modified hfea' should do better in the long-run, we probably want to protect those gains. Conversely, if somehow all the risk blows up in our faces and we bail out after 3 years, we might not be happy having our remaining money all in our IRA/401k and none in taxable (although it would be a good deduction for a while). Instead, we could take a lower amount of risk across all accounts, rather than a lot of risk in some accounts and much less in others. For example, my SO has been doing this in taxable because the 401k is very restrictive, and the taxable account is already getting quite large relative to the 401k. We'd be happier taking less leverage in taxable, and some leverage in the 401k.

Hope that helps.
adamhg
Posts: 218
Joined: Sat Apr 10, 2021 8:40 am

Re: Modified versions of HFEA with ITT and Futures / Lifecycle Investing with Modern Portfolio Theory

Post by adamhg »

skierincolorado wrote: Fri Sep 10, 2021 8:17 pm
adamhg wrote: Fri Sep 10, 2021 7:52 pm Been researching this ever since skiers posts in hfea a few weeks back. I'm primarily looking to do this in my taxable, so I've been looking into the box spread method. But what advantage does futures present over box spreads, even in tax advantaged accounts? Seems like the borrow rate would be equivalent with much fewer rolls (2-3 years) and you get around imprecise ratios from the large nominal values
For one, I'm pretty sure you can't do box-spreads in an IRA because that level of options writing is restricted.

Besides, in an IRA, futures are preferable because the borrow rate on box-spreads is higher. I get like .7% on boxes, vs .1% on Treasury Futures and .4% on MES futures.

Also, it's hard to buy 200-300% of portfolio value in ITT using margin. I don't know about other brokers, but the margin requirement for bond ETFs are the same as for stock ETFs at IBKR. You can buy actual bonds with low margin requirements, but I found that to be a terrible pain.

The borrow rate on box-spreads is so high, I'm thinking of doing the math on whether it's worth using MES futures in taxable as well. Inferior tax treatment, but lower borrow rate. Maybe if I wrote larger boxes I'd get down to .6% or something.

I already use treasury futures in taxable because of the large amounts being borrowed and the very low implied finance rates ~.1% they offer and because bond interest is taxable anyways. Treasury futures might actually be preferable to actual bonds from a tax perspective - not sure on that one. So I don't use box-spreads except for equity leverage in taxable.

One has to rebalance from time to time anyways, so I do this when rolling. I also find rolling futures to be easier than writing box-spreads. They're much more liquid and I don't have to worry about the bid-ask spread.

Finally, generally speaking, it's best to take the same amount of leverage with the same AA across all accounts, or at least look at the big picture when deciding on an investment plan. For example, it doesn't make sense to be leveraging stock in taxable, paying .4-.7% financing rates, while having an unleveraged 80/20 or 60/40 in your IRA or 401k. You could sell bonds in the IRA and buy stock, and sell stock in the IRA and then buy treasury futures to account for the bonds sold. It would be the exact same AA, but with less borrowing costs. Decide on your target AA for your entire net worth, and then decide the most efficient way to implement it. Finally, since 'modified hfea' should do better in the long-run, we probably want to protect those gains. Conversely, if somehow all the risk blows up in our faces and we bail out after 3 years, we might not be happy having our remaining money all in our IRA/401k and none in taxable (although it would be a good deduction for a while). Instead, we could take a lower amount of risk across all accounts, rather than a lot of risk in some accounts and much less in others. For example, my SO has been doing this in taxable because the 401k is very restrictive, and the taxable account is already getting quite large relative to the 401k. We'd be happier taking less leverage in taxable, and some leverage in the 401k.

Hope that helps.
Yes, thanks as usual for the detailed reply. The margin requirement makes sense so that box spreads probably aren't even feasible in iras.

Personally, I'm optimizing for convenience and plan to automate this whole thing. If I get there, I'll pen source the code. The challenge I'm finding with futures is that there aren't a lot of brokers that support apis to trade futures and equity. IBKR is the obvious choice but I refuse to work with their tech.

Your point about the rate may simply be due to the duration of the loan though. A box spread would be close to the 2-3 yr institutional rate whereas the future at the 1-2mo rate. So while your rate is lower, the box locks it in for longer and you should theoretically always be able to sell it back at the same or similar rate as the future if you wanted to roll monthly even.

But I agree, pin pointing the best rate with a 4 legged option bid/ask isn't going to be as easy as just buying a future. In all honestly, my biggest fear with futures though is that I won't be able to resist going 12x STT
Post Reply