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.
comeinvest
Posts: 1458
Joined: Mon Mar 12, 2012 6:57 pm

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

Post by comeinvest »

I calculated the average of the ACM term premium estimates per unit of duration from 1961 to now, per maturity. Since the duration of treasuries with fixed maturity varied greatly over time, I just assumed durations of 1 for the 1y, 1.8 for the 2y, 4 for the 5y, and 8 for the 10y. The results are very similar for the 1y, 2y, and 5y, and slightly lower for the 10y.

1y: 0.27
2y: 0.28
5y: 0.25
10y: 0.20

Image
User avatar
LTCM
Posts: 380
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 these eurodollar futures...

If I buy a 4 part strip for Dec22/Mar23/Jun23/Sept23 is that closer to owning a 6 month treasury than a 1 year treasury? I need to buy 4 x Dec23 if I want a 1 year treasury right?

The reason to buy strips is to save on trading costs & commisions? You only trade into the position and then get paid out when it expires? Otherwise you constantly have to sell your 4x and move it forward every 3 months by buying back in?

If a single ZT contract is worth 200k and a single ZF is worth 100k what is the equivalent amount of Eurodollar contracts to take on the same amount of risk? How many Dec24 eurodollars contracts would I need to match a single ZT contract? (or how many ZT contracts to match a eurodollar contract). I'm reading they're worth 1,000,000 each but I'm not sure its apples to apples.
60% VUG - 20% VEA - 10% EDV - 10% I-Bonds
comeinvest
Posts: 1458
Joined: Mon Mar 12, 2012 6:57 pm

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

Post by comeinvest »

LTCM wrote: Fri Sep 23, 2022 3:50 am So these eurodollar futures...

If I buy a 4 part strip for Dec22/Mar23/Jun23/Sept23 is that closer to owning a 6 month treasury than a 1 year treasury? I need to buy 4 x Dec23 if I want a 1 year treasury right?

The reason to buy strips is to save on trading costs & commisions? You only trade into the position and then get paid out when it expires? Otherwise you constantly have to sell your 4x and move it forward every 3 months by buying back in?

If a single ZT contract is worth 200k and a single ZF is worth 100k what is the equivalent amount of Eurodollar contracts to take on the same amount of risk? How many Dec24 eurodollars contracts would I need to match a single ZT contract? (or how many ZT contracts to match a eurodollar contract). I'm reading they're worth 1,000,000 each but I'm not sure its apples to apples.
I'm not an expert on the subject, but I tried to wrap my head around this a while ago.
First off, as was mentioned before, there is no direct investable "underlying" for this contract, as it is cash-settled to a rate, which in and by itself is not directly investable.
Therefore, the term "notional value" also becomes fuzzy in the case of rate futures. Some places I see it referenced as representing a $1,000,000 "notional" value worth of 90-day deposits; other places I see it referenced as representing a "notional" value of $250,000 worth of 1-year duration exposure, i.e. the "notional value" is annualized.
BUT like Beliavsky mentioned above, it looks like it does NOT mimic the investment performance of 90-day eurodollar deposits. Maybe we can say it mimics the interest rate sensitivity of 90-day eurodollar deposits during a future period?
The LIBOR rate that the settlement is based on is an annualized rate.
Since the eurodollar futures contract represents a forward interest rate during a future 90-day period, you would need an 8 quarter strip to have the equivalent of a 2-year treasury futures contract. Bond interest rates are the mathematical average over the short-term rates between now and maturity (expectations hypothesis).
Last edited by comeinvest on Fri Sep 23, 2022 5:35 am, edited 2 times in total.
User avatar
LTCM
Posts: 380
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 »

Small follow up...

Paper trading:
I bought 5 x ZT around 4am EST and they're down approx $2000
I bought 5 x GEZ24 around 4am EST and they're down approx $2500
I sold 2 x ZF around 4am EST and I'm up approx $600
I sold 2 x GEZ27 around 4am EST and I'm up approx $250

So its seems GE are "bigger" but not in a uniform way so far (or I didn't match the durations close enough).
60% VUG - 20% VEA - 10% EDV - 10% I-Bonds
User avatar
LTCM
Posts: 380
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 »

comeinvest wrote: Fri Sep 23, 2022 5:24 am you would need an 8 quarter strip to have the equivalent of a 2-year treasury futures contract. Bond interest rates are the mathematical average over the short-term rates between now and maturity (expectations hypothesis).
But which 8 quarters? From now until 2 years out? Or from 1 year out to 3 years out (avg 2 years).
60% VUG - 20% VEA - 10% EDV - 10% I-Bonds
comeinvest
Posts: 1458
Joined: Mon Mar 12, 2012 6:57 pm

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

Post by comeinvest »

LTCM wrote: Fri Sep 23, 2022 5:39 am
comeinvest wrote: Fri Sep 23, 2022 5:24 am you would need an 8 quarter strip to have the equivalent of a 2-year treasury futures contract. Bond interest rates are the mathematical average over the short-term rates between now and maturity (expectations hypothesis).
But which 8 quarters? From now until 2 years out? Or from 1 year out to 3 years out (avg 2 years).
Now to 2y. But if you need to ask these questions, I strongly recommend reading up the basics before you trade these. You don't want to invest in something and have no clue what you are doing.
User avatar
LTCM
Posts: 380
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 ask questions to understand better. I'm not fully convinced on that 8 strip not being significantly shorter duration than a ZT yet.

If a 1 part strip is approx 3 month. A 2 part strip is 6 months then what is a 1 part strip 6 months out? 9 months?

A 1 part strip 5 years out is significantly longer than 5 year duration? That doesn't seem right to me. It would have to be approx 10 year duration for a 20 part strip to average out at 5 years.
60% VUG - 20% VEA - 10% EDV - 10% I-Bonds
User avatar
LTCM
Posts: 380
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 think I wrapped my head around it now. Since each eurobond is only a 3 month deposit you need string them together to get the compounding. You can't just have 4x 3 month deposits 12 months out and get the same result.
60% VUG - 20% VEA - 10% EDV - 10% I-Bonds
SCraw
Posts: 28
Joined: Thu Dec 19, 2019 3:22 pm

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

Post by SCraw »

LTCM wrote: Fri Sep 23, 2022 3:50 am So these eurodollar futures...

If I buy a 4 part strip for Dec22/Mar23/Jun23/Sept23 is that closer to owning a 6 month treasury than a 1 year treasury? I need to buy 4 x Dec23 if I want a 1 year treasury right?

The reason to buy strips is to save on trading costs & commisions? You only trade into the position and then get paid out when it expires? Otherwise you constantly have to sell your 4x and move it forward every 3 months by buying back in?

If a single ZT contract is worth 200k and a single ZF is worth 100k what is the equivalent amount of Eurodollar contracts to take on the same amount of risk? How many Dec24 eurodollars contracts would I need to match a single ZT contract? (or how many ZT contracts to match a eurodollar contract). I'm reading they're worth 1,000,000 each but I'm not sure its apples to apples.
Bear in mind that ED futures are tied to 3m libor rather than treasuries themselves
Beliavsky
Posts: 1155
Joined: Sun Jun 29, 2014 10:21 am

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

Post by Beliavsky »

Term premium estimates, as studied in this article, may help you decide how much to leverage 2-year Treasuries or Eurodollar/SOFR strips.

Short-Dated Term Premia and the Level of Inflation
by Richard Crump, Charles Smith, and Peter Van Tassel
Federal Reserve Bank of New York
September 28, 2022

Some statement in the post are
Term Premiums Were High when Interest Rates and Inflation Were High

Short-Dated, Survey-Based Term Premia in June 2022 Were Small

Term Premia were higher in the 1980s and early 1990s

Varlous Measures of Short-Dated Term Premia Have Remained Low
LazyOverthinker
Posts: 114
Joined: Wed Jun 08, 2022 5:03 am

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

Post by LazyOverthinker »

Unrelated to the current conversation, but I stumbled upon Ben Felix referencing a paper asserting that from 1927-2020, the total market return was 4.91% during high-inflation while the small-cap-value return was 11.95%. Very relevant for mHFEA imo and it looks like I'll be sticking to my 100% SCV.

Timestamped link to the quote: https://youtu.be/65ug05JtgcY?t=3441
User avatar
LTCM
Posts: 380
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 need some help with backtesting a position in Simba.

100% VTSAX
1250% STT
-250% ITT

Initially I simply entered that into Simba and then added -1000% T-Bills TR but I'm no longer confident that's accurate. Do I really get to fully deduct 250% T-Bills TR when I short 250% ITT? Because it's really improving the Sharpe ratio (and allowing it to out compete shorting LTT) by reducing the borrowing requirement.

I'm paper trading at IBKR and if I put 100,000 into an IRA I end up with these positions (approx):
+430 VTI ($80,000)
+5 /ZT ($1,000,000 nominal) ($7500 collateral)
-1 /ZF ($100,000 nominal) ($2000 collateral)
-1 /ZN ($100,000 nominal) ($2500 collateral)
$8000 cash balance to avoid regular margin calls

So...could that be better described as follows?
80% VTSAX
1000% STT
-200% ITT
20% Hard Cash (collateral & cash)
-800% T-Bills TR

Do I have enough T-Bills TR minused out? Do I really get that +200 borrowing benefit from shorting ITT?
Should i be using VUSXX instead of T-Bills TR?

Anything else I've missed?
60% VUG - 20% VEA - 10% EDV - 10% I-Bonds
comeinvest
Posts: 1458
Joined: Mon Mar 12, 2012 6:57 pm

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

Post by comeinvest »

LazyOverthinker wrote: Wed Sep 28, 2022 2:14 pm Unrelated to the current conversation, but I stumbled upon Ben Felix referencing a paper asserting that from 1927-2020, the total market return was 4.91% during high-inflation while the small-cap-value return was 11.95%. Very relevant for mHFEA imo and it looks like I'll be sticking to my 100% SCV.

Timestamped link to the quote: https://youtu.be/65ug05JtgcY?t=3441
The problem is you don't know how long high inflation will persist. If or when inflation expectations (which probably matter more than actual inflation) will subside, you might get hosed (again!) with your SCV tilt like the last 2-3 decades. I personally will probably settle for a mix - some equity index futures, and some SCV or other factor tilted ETFs, so to have my eggs in many baskets.
comeinvest
Posts: 1458
Joined: Mon Mar 12, 2012 6:57 pm

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

Post by comeinvest »

Beliavsky wrote: Wed Sep 28, 2022 7:34 am Term premium estimates, as studied in this article, may help you decide how much to leverage 2-year Treasuries or Eurodollar/SOFR strips.

Short-Dated Term Premia and the Level of Inflation
by Richard Crump, Charles Smith, and Peter Van Tassel
Federal Reserve Bank of New York
September 28, 2022

Some statement in the post are
Term Premiums Were High when Interest Rates and Inflation Were High

Short-Dated, Survey-Based Term Premia in June 2022 Were Small

Term Premia were higher in the 1980s and early 1990s

Varlous Measures of Short-Dated Term Premia Have Remained Low
It doesn't say what the realized term premia were historically, does it? It is well known that term premia, not only for STT, are correlated with inflation uncertainty.
comeinvest
Posts: 1458
Joined: Mon Mar 12, 2012 6:57 pm

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

Post by comeinvest »

LTCM wrote: Wed Sep 28, 2022 5:18 pm I need some help with backtesting a position in Simba.

100% VTSAX
1250% STT
-250% ITT

Initially I simply entered that into Simba and then added -1000% T-Bills TR but I'm no longer confident that's accurate. Do I really get to fully deduct 250% T-Bills TR when I short 250% ITT? Because it's really improving the Sharpe ratio (and allowing it to out compete shorting LTT) by reducing the borrowing requirement.

I'm paper trading at IBKR and if I put 100,000 into an IRA I end up with these positions (approx):
+430 VTI ($80,000)
+5 /ZT ($1,000,000 nominal) ($7500 collateral)
-1 /ZF ($100,000 nominal) ($2000 collateral)
-1 /ZN ($100,000 nominal) ($2500 collateral)
$8000 cash balance to avoid regular margin calls

So...could that be better described as follows?
80% VTSAX
1000% STT
-200% ITT
20% Hard Cash (collateral & cash)
-800% T-Bills TR

Do I have enough T-Bills TR minused out? Do I really get that +200 borrowing benefit from shorting ITT?
Should i be using VUSXX instead of T-Bills TR?

Anything else I've missed?
Your numbers are theoretically right as long as the total is 100%. But the performance of highly leverage STT, and/or relative positioning, are highly sensitive to implied financing rates, which means I would not be confident enough in your model that uses T-bills as a proxy of the financing rate. As you probably now, we determined ca. 0.2% - 0.35% implied financing rate of most ITT and LTT futures. (I think it slightly increases with the maturity of the futures product, and was measurably higher for /UB.)

You also have to consider increasing commissions and (perhaps) cost from bid/ask spreads if you use STT or relative positioning. And don't forget that the more free parameters you use, the higher the risk of parameter fitting a.k.a. "predicting the past".

The paper https://coexpartnersaig.files.wordpress ... -carry.pdf shows that STT or Eurodollar strips were valid implementations to harvest the STT term premium, but they don't show the implied financing rates; at least I couldn't find them.
User avatar
LTCM
Posts: 380
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 »

comeinvest wrote: Thu Sep 29, 2022 4:51 am But the performance of highly leverage STT, and/or relative positioning, are highly sensitive to implied financing rates, which means I would not be confident enough in your model that uses T-bills as a proxy of the financing rate. As you probably now, we determined ca. 0.2% - 0.35% implied financing rate of most ITT and LTT futures.
I've seen people mention this +0.2% feature of futures financing but I've never seen any actual paper/evidence and I've asked a few times. Personally I'm still pretty convinced the implied repo rate displayed at CME is the financing cost and that moves all over the place depending on demand/expectations. Right now the IRR for the 2 year is 2.88% so you're actually getting a discount on the financing compared to T-bills.
60% VUG - 20% VEA - 10% EDV - 10% I-Bonds
LazyOverthinker
Posts: 114
Joined: Wed Jun 08, 2022 5:03 am

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

Post by LazyOverthinker »

comeinvest wrote: Thu Sep 29, 2022 4:29 am The problem is you don't know how long high inflation will persist. If or when inflation expectations (which probably matter more than actual inflation) will subside, you might get hosed (again!) with your SCV tilt like the last 2-3 decades. I personally will probably settle for a mix - some equity index futures, and some SCV or other factor tilted ETFs, so to have my eggs in many baskets.
Oh I wouldn't say I'm betting on high inflation, rather that I'm very happy to believe SCV offers protection from the worst-case scenarios of my treasury allocation. Behaviorally, I am prepared to endure tracking error because I expect my leverage to at least make up the difference, and I find it much easier to justify diversifying globally as an effort to hunt the premium in more than one locale. Before, I was finding it very difficult to justify less than 70% US stocks to myself.
comeinvest
Posts: 1458
Joined: Mon Mar 12, 2012 6:57 pm

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

Post by comeinvest »

LazyOverthinker wrote: Thu Sep 29, 2022 1:55 pm
comeinvest wrote: Thu Sep 29, 2022 4:29 am The problem is you don't know how long high inflation will persist. If or when inflation expectations (which probably matter more than actual inflation) will subside, you might get hosed (again!) with your SCV tilt like the last 2-3 decades. I personally will probably settle for a mix - some equity index futures, and some SCV or other factor tilted ETFs, so to have my eggs in many baskets.
Oh I wouldn't say I'm betting on high inflation, rather that I'm very happy to believe SCV offers protection from the worst-case scenarios of my treasury allocation. Behaviorally, I am prepared to endure tracking error because I expect my leverage to at least make up the difference, and I find it much easier to justify diversifying globally as an effort to hunt the premium in more than one locale. Before, I was finding it very difficult to justify less than 70% US stocks to myself.
It's always about risk *and* expected return. Yes you may be safer with you SCV-only. But in case that the global race to zero for interest rates resumes, you may be significantly underperforming the mHFEA people - because you both don't have treasuries, AND your SCV might underperform the large cap growth controlled (or at least more diversified) index.
Of course, the opposite might happen and you win. I'm just saying maybe it's good to split the difference.
comeinvest
Posts: 1458
Joined: Mon Mar 12, 2012 6:57 pm

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

Post by comeinvest »

LTCM wrote: Thu Sep 29, 2022 11:46 am
comeinvest wrote: Thu Sep 29, 2022 4:51 am But the performance of highly leverage STT, and/or relative positioning, are highly sensitive to implied financing rates, which means I would not be confident enough in your model that uses T-bills as a proxy of the financing rate. As you probably now, we determined ca. 0.2% - 0.35% implied financing rate of most ITT and LTT futures.
I've seen people mention this +0.2% feature of futures financing but I've never seen any actual paper/evidence and I've asked a few times. Personally I'm still pretty convinced the implied repo rate displayed at CME is the financing cost and that moves all over the place depending on demand/expectations. Right now the IRR for the 2 year is 2.88% so you're actually getting a discount on the financing compared to T-bills.
I think there might be a problem with the numbers on the CME site. They seem to fluctuate a lot on a daily basis, and often don't make sense. I think on the site itself there is a disclaimer that those data points might not be accurate and should not be relied on. I asked about this high up in this thread. I don't fully understand the numbers on the CME site, maybe it's a timing issue of the data points.
But I think the ca. 0.2% - 0.35% implied financing cost of ITT has been thoroughly verified in this thread, and in many papers including some publications of CME. If you read the mHFEA thread, you will find the pages where we compared the performance of the futures to that of ETFs of similar duration, and adjusted for duration differences. One user also did a backtest of various forms of treasuries leverage and summarized the results in a table.
It may be that STT have a similar financing spread, either nominal or duration-adjusted. But I'm saying that because of the slim yield margin that is highly leveraged, the sensitivity of the total performance to the financing spread is much higher than for ITT, so a simulation with an assumption on the spread might not be very accurate. For STT, I would be more confident in backtests with actual futures.
LazyOverthinker
Posts: 114
Joined: Wed Jun 08, 2022 5:03 am

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

Post by LazyOverthinker »

comeinvest wrote: Thu Sep 29, 2022 6:49 pm It's always about risk *and* expected return. Yes you may be safer with you SCV-only. But in case that the global race to zero for interest rates resumes, you may be significantly underperforming the mHFEA people - because you both don't have treasuries, AND your SCV might underperform the large cap growth controlled (or at least more diversified) index.
Of course, the opposite might happen and you win. I'm just saying maybe it's good to split the difference.
comeinvest wrote: Thu Sep 29, 2022 6:49 pm ...because you both don't have treasuries, AND your SCV might underperform the large cap growth controlled (or at least more diversified) index.
Sorry, to be clear I also still plan to hold a hefty ratio of treasuries @ 5 duration. By "100% SCV" I meant that my equity glidepath looks more like 1.6x->1x all in global SCV, rather than 2x->1x MCW
User avatar
LTCM
Posts: 380
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 IRR changes so much because its a function of the futures price.

https://www.investopedia.com/terms/i/im ... s%20repaid.

https://en.m.wikipedia.org/wiki/Implied_repo_rate
An implied repo rate is the rate of return that can be earned by owning a bond and simultaneously shorting a futures or forward contract against it.
So the cost of longing a future while simultaneously not owning a bond is the IRR.
This net return, or repo rate, tends to be close to the risk-free rate as the buying and selling involved amount to arbitrage.
Speculating but i think the markets are way too efficient for skimming 0.2% on a 3 month rate at 3%.
60% VUG - 20% VEA - 10% EDV - 10% I-Bonds
comeinvest
Posts: 1458
Joined: Mon Mar 12, 2012 6:57 pm

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

Post by comeinvest »

LazyOverthinker wrote: Thu Sep 29, 2022 7:10 pm
comeinvest wrote: Thu Sep 29, 2022 6:49 pm It's always about risk *and* expected return. Yes you may be safer with you SCV-only. But in case that the global race to zero for interest rates resumes, you may be significantly underperforming the mHFEA people - because you both don't have treasuries, AND your SCV might underperform the large cap growth controlled (or at least more diversified) index.
Of course, the opposite might happen and you win. I'm just saying maybe it's good to split the difference.
comeinvest wrote: Thu Sep 29, 2022 6:49 pm ...because you both don't have treasuries, AND your SCV might underperform the large cap growth controlled (or at least more diversified) index.
Sorry, to be clear I also still plan to hold a hefty ratio of treasuries @ 5 duration. By "100% SCV" I meant that my equity glidepath looks more like 1.6x->1x all in global SCV, rather than 2x->1x MCW
I think this is reasonable. It comes down the the old question of tilting vs not tilting. There are gazillions of other threads on this subject, we don't have to reiterate the arguments here, except for whatever is specific to SCV in combination with leverage. I think somewhere above you showed that there is an additional benefit (beyond the pure SCV benefit) of leveraged SCV + treasuries vs leveraged index + treasuries, risk-adjusted to reflect increased SCV risk. I think your strategy is solid, although more comprehensive backtesting is always good, like longer time frames or international, to show systematic outperformance.

As you are well aware, there are some who say the SCV is an "independent" source of return. But there are other posts in this forum who have shown that in the long run, an SCV tilt has an effect almost identical to a slightly higher leverage ratio.

In IRAs you can have only very limited SCV if you equities allocation is > 100%, because futures are the only reasonable way of leverage. In my taxable account I personally do "index sampling" with a limited amount of individual stocks; I refuse to use any fund or ETF, as I don't want to be "locked" (via unrealized capital gains) into any product or fund company or strategy for decades. (How many times have mainstream opinions and popular investment products or strategies changed in the last 40 years?!) And I want to do tax loss harvesting, which is a guaranteed risk-free return, in contrast to SCV or to returns from leverage. As a result, I can implement only a quite small amount of SCV within my mHFEA. Although my taxable stock selection I could slightly tilt towards value.
comeinvest
Posts: 1458
Joined: Mon Mar 12, 2012 6:57 pm

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

Post by comeinvest »

LTCM wrote: Thu Sep 29, 2022 7:24 pm The IRR changes so much because its a function of the futures price.

https://www.investopedia.com/terms/i/im ... s%20repaid.

https://en.m.wikipedia.org/wiki/Implied_repo_rate
An implied repo rate is the rate of return that can be earned by owning a bond and simultaneously shorting a futures or forward contract against it.
So the cost of longing a future while simultaneously not owning a bond is the IRR.
This net return, or repo rate, tends to be close to the risk-free rate as the buying and selling involved amount to arbitrage.
Speculating but i think the markets are way too efficient for skimming 0.2% on a 3 month rate at 3%.
The markets might be efficient, but there are a number of possible explanations for the spread of derivatives' implied financing rate to T-bills. Although no definite theory was offered to explain it, this has been discussed at length in this and in the mHFEA thread. Regardless of explanation, it is a fact. The S&P500 futures contract has an even bigger, and relatively consistent implied financing spread to T-bills of about 0.5%. You cannot argue this away. I encourage you to read the entire thread again.
User avatar
LTCM
Posts: 380
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 »

comeinvest wrote: Thu Sep 29, 2022 10:59 pm
LTCM wrote: Thu Sep 29, 2022 7:24 pm The IRR changes so much because its a function of the futures price.

https://www.investopedia.com/terms/i/im ... s%20repaid.

https://en.m.wikipedia.org/wiki/Implied_repo_rate
An implied repo rate is the rate of return that can be earned by owning a bond and simultaneously shorting a futures or forward contract against it.
So the cost of longing a future while simultaneously not owning a bond is the IRR.
This net return, or repo rate, tends to be close to the risk-free rate as the buying and selling involved amount to arbitrage.
Speculating but i think the markets are way too efficient for skimming 0.2% on a 3 month rate at 3%.
The markets might be efficient, but there are a number of possible explanations for the spread of derivatives' implied financing rate to T-bills. Although no definite theory was offered to explain it, this has been discussed at length in this and in the mHFEA thread. Regardless of explanation, it is a fact. The S&P500 futures contract has an even bigger, and relatively consistent implied financing spread to T-bills of about 0.5%. You cannot argue this away. I encourage you to read the entire thread again.
I can only repeat my request to show the source of this 0.2% claim or alternatively a refutation of the IRR as the financing rate.
60% VUG - 20% VEA - 10% EDV - 10% I-Bonds
Topic Author
skierincolorado
Posts: 1737
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 29, 2022 11:54 pm
comeinvest wrote: Thu Sep 29, 2022 10:59 pm
LTCM wrote: Thu Sep 29, 2022 7:24 pm The IRR changes so much because its a function of the futures price.

https://www.investopedia.com/terms/i/im ... s%20repaid.

https://en.m.wikipedia.org/wiki/Implied_repo_rate
An implied repo rate is the rate of return that can be earned by owning a bond and simultaneously shorting a futures or forward contract against it.
So the cost of longing a future while simultaneously not owning a bond is the IRR.
This net return, or repo rate, tends to be close to the risk-free rate as the buying and selling involved amount to arbitrage.
Speculating but i think the markets are way too efficient for skimming 0.2% on a 3 month rate at 3%.
The markets might be efficient, but there are a number of possible explanations for the spread of derivatives' implied financing rate to T-bills. Although no definite theory was offered to explain it, this has been discussed at length in this and in the mHFEA thread. Regardless of explanation, it is a fact. The S&P500 futures contract has an even bigger, and relatively consistent implied financing spread to T-bills of about 0.5%. You cannot argue this away. I encourage you to read the entire thread again.
I can only repeat my request to show the source of this 0.2% claim or alternatively a refutation of the IRR as the financing rate.
I think it's in one of the papers in the op. If not I'll find it tomorrow
User avatar
LTCM
Posts: 380
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 »

Thanks. Here's another description of IRR.
It is called the implied repo rate (IRR). It is a theoretical yield produced by buying the cash security, selling the futures contract, lending the cash security in the repo market and finally, delivering the security into the futures contract on last delivery day.
We can rearrange this formula like any other equation:

The IRR is the rate PAID by SELLING the cash security, BUYING the futures contract, BORROWING the cash security in the repo market and finally RECEIVING the cash security from the futures contract on delivery day.

This is the same as rolling a futures contract. You sell the bond, borrow it for 3 months and then receive it again. The IRR is the financing cost/yield of the operation.
60% VUG - 20% VEA - 10% EDV - 10% I-Bonds
comeinvest
Posts: 1458
Joined: Mon Mar 12, 2012 6:57 pm

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

Post by comeinvest »

LTCM wrote: Fri Sep 30, 2022 12:31 am Thanks. Here's another description of IRR.
It is called the implied repo rate (IRR). It is a theoretical yield produced by buying the cash security, selling the futures contract, lending the cash security in the repo market and finally, delivering the security into the futures contract on last delivery day.
We can rearrange this formula like any other equation:

The IRR is the rate PAID by SELLING the cash security, BUYING the futures contract, BORROWING the cash security in the repo market and finally RECEIVING the cash security from the futures contract on delivery day.

This is the same as rolling a futures contract. You sell the bond, borrow it for 3 months and then receive it again. The IRR is the financing cost/yield of the operation.
You can slice and dice and re-arrange the terms of this simple equation, and you can use whatever terminology you want. But we are talking about and we are interested in the spread of the implied financing rate to T-bills when using futures, as if I were to borrow money and buy the underlying instead. There are several pages somewhere on pages 5-20 roughly of this thread, where we examined the spread in detail, which is roughly in the 0.2% - 0.3% area for ITT. Like I said, STT may have the same or a slightly different spread, but due to the high sensitivity of the performance of highly leveraged STT on the spread, I personally would not be confident using an assumption on the spread as input parameter to a backtest simulation. I would have more confidence when I see backtests with actual futures which removes any doubt about possible systematic implementation biases and intricacies, rather than a simulated futures strategy. The paper https://coexpartnersaig.files.wordpress ... -carry.pdf shows results with actual historical futures data. The paper is quite good and covers about 19 years, but unfortunately I have not found any other studies on STT term premium strategy implementations with futures that cover pre-1995 or post-2013 periods.
User avatar
LTCM
Posts: 380
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 »

When I short ITT do I get an extra 0.2% over the T-bill rate?
60% VUG - 20% VEA - 10% EDV - 10% I-Bonds
User avatar
LTCM
Posts: 380
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 »

Another reason STT futures financing should be less than ITT (and Equities).

If I leant you a million dollar asset for 3 months that I was certain would be worth exactly $1m in 3 months what would I reasonably charge for that?

If I leant you a million dollar asset for 3 months that would be worth on average $1m in 3 months time but there was a 50% chance it was worth $2m and a 50% chance it was worth zero.

Which one would I charge you more to borrow? And why?

Which one would be STT in this scenario?
60% VUG - 20% VEA - 10% EDV - 10% I-Bonds
ivgrivchuck
Posts: 1382
Joined: Sun Sep 27, 2020 6:20 pm

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

Post by ivgrivchuck »

LazyOverthinker wrote: Wed Sep 28, 2022 2:14 pm Unrelated to the current conversation, but I stumbled upon Ben Felix referencing a paper asserting that from 1927-2020, the total market return was 4.91% during high-inflation while the small-cap-value return was 11.95%. Very relevant for mHFEA imo and it looks like I'll be sticking to my 100% SCV.
Are you not worried that you are carrying too much idiosyncratic risk (SCV funds often concentrate only on a few sectors)? You are not getting compensated for this risk (as a well-diversified investor can diversify the risk away), but it hurts your expected returns (as unrewarded risk always does).

Of course one can argue that negative correlations with treasuries, extreme factor exposure more than make up for this extra risk, but this view requires a lot of faith on the ultimate superiority of SCV (more than I have)...
25% VTI | 25% VXUS | 12.5% AVUV | 10% AVDV | 2.5% VWO | 25% I/EE-bonds
User avatar
LTCM
Posts: 380
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 »

Another thought exercise for this 0.2% rate

If I can borrow at the fed rate and buy 1,000,000 in treasuries I can earn $2000 per year lending them to the futures market? So If I buy $100,000,000 I can make $200,000 in a year? Totally risk free. Doesn't that seem high?

With a treasury futures market totaling $1.5 trillion in notional value outstanding that means those lending the treasuries are making $3bn per year. RISK FREE.
60% VUG - 20% VEA - 10% EDV - 10% I-Bonds
comeinvest
Posts: 1458
Joined: Mon Mar 12, 2012 6:57 pm

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

Post by comeinvest »

LTCM wrote: Fri Sep 30, 2022 2:42 am Another thought exercise for this 0.2% rate

If I can borrow at the fed rate and buy 1,000,000 in treasuries I can earn $2000 per year lending them to the futures market? So If I buy $100,000,000 I can make $200,000 in a year? Totally risk free. Doesn't that seem high?

With a treasury futures market totaling $1.5 trillion in notional value outstanding that means those lending the treasuries are making $3bn per year. RISK FREE.
Again, this was thoroughly investigated and debated for ITT and for derivatives in general, on several pages worth of discussion in this thread. I'm not sure what you are trying to get at. The spread of STT is likely somewhere between zero and 0.2%. The spread is somewhere above zero for most derivatives, which have various margin requirements depending on the risk. And yes, someone makes some small margin from it who is in some way balance sheet or regulatory constrained and cannot hedge infinite amounts for free. What are you trying to prove or to debate? Are you trying to develop your own micro-economic theory for derivative financing spreads? If not, do some reading and let us know if you find some study that as historical numbers.
Topic Author
skierincolorado
Posts: 1737
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 30, 2022 2:42 am Another thought exercise for this 0.2% rate

If I can borrow at the fed rate and buy 1,000,000 in treasuries I can earn $2000 per year lending them to the futures market? So If I buy $100,000,000 I can make $200,000 in a year? Totally risk free. Doesn't that seem high?

With a treasury futures market totaling $1.5 trillion in notional value outstanding that means those lending the treasuries are making $3bn per year. RISK FREE.
Figure 5 here

https://www.financialresearch.gov/brief ... Trades.pdf

The average is something like .15 for the 2 year but as high as .4 some years. The 5 year is a bit higher like .2 or .25.

It makes sense that one could earn .15% risk free. The market is not perfectly efficient and these transactions are not entirely devoid of counterparty risk. Not just anybody can borrow exactly at the risk free rate.

Early in the thread we empirically verified the numbers from this paper by comparing futures and bond fund returns. See the posts by zkn. Search the thread for zkn.
LazyOverthinker
Posts: 114
Joined: Wed Jun 08, 2022 5:03 am

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

Post by LazyOverthinker »

ivgrivchuck wrote: Fri Sep 30, 2022 1:47 am
LazyOverthinker wrote: Wed Sep 28, 2022 2:14 pm Unrelated to the current conversation, but I stumbled upon Ben Felix referencing a paper asserting that from 1927-2020, the total market return was 4.91% during high-inflation while the small-cap-value return was 11.95%. Very relevant for mHFEA imo and it looks like I'll be sticking to my 100% SCV.
Are you not worried that you are carrying too much idiosyncratic risk (SCV funds often concentrate only on a few sectors)? You are not getting compensated for this risk (as a well-diversified investor can diversify the risk away), but it hurts your expected returns (as unrewarded risk always does).

Of course one can argue that negative correlations with treasuries, extreme factor exposure more than make up for this extra risk, but this view requires a lot of faith on the ultimate superiority of SCV (more than I have)...
I actually had this exact concern and posted in the RationalReminder forum about it to much discussion. There were decent arguments in both directions, but overall the crowd over there trusts the SCV process tremendously (and are biased, of course). DFA actually controls for sectors, not allowing their allocations to get more than +/- 10% of MCW sector-weighting. Avantis doesn't, resulting in something like a 30% finance allocation recently.

What calms my nerves about SCV is that it, by definition, can't be in a bubble. It can underperform and get hit extra-hard by macro-economic factors, but it can be the superstar winner of a "bubble environment," like in the dotcom crash (where the market was 40 P/E thanks to tech but SCV was still systematically in <15P/E). So even though AVUV has an eye-popping amount of Finance sector in it right now (which evokes 2008 fears), I'm really not worried that there's a collective "finance [sector] bubble" between the hundreds of small, discounted banks and financial institutions they've systematically invested in. In my eyes, the risk of SCV is much better defined as macro-economic.

Another way to say it is that growth-investing is defined by bad surprises, value-investing is defined by good surprises. Thoughts?
User avatar
LTCM
Posts: 380
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 30, 2022 10:08 am
LTCM wrote: Fri Sep 30, 2022 2:42 am Another thought exercise for this 0.2% rate

If I can borrow at the fed rate and buy 1,000,000 in treasuries I can earn $2000 per year lending them to the futures market? So If I buy $100,000,000 I can make $200,000 in a year? Totally risk free. Doesn't that seem high?

With a treasury futures market totaling $1.5 trillion in notional value outstanding that means those lending the treasuries are making $3bn per year. RISK FREE.
Figure 5 here

https://www.financialresearch.gov/brief ... Trades.pdf

The average is something like .15 for the 2 year but as high as .4 some years. The 5 year is a bit higher like .2 or .25.

It makes sense that one could earn .15% risk free. The market is not perfectly efficient and these transactions are not entirely devoid of counterparty risk. Not just anybody can borrow exactly at the risk free rate.

Early in the thread we empirically verified the numbers from this paper by comparing futures and bond fund returns. See the posts by zkn. Search the thread for zkn.
Thanks. It's hard on a phone but from eyeballing it the 2 year looks like it averages more like 0.02 than 0.2.

You agree the IRR on the CME site is the financing rate?

Do you think we receive the IRR when shorting futures?

I'll look up zkn when I get back to a computer.

That whole paper looks worth reading.
60% VUG - 20% VEA - 10% EDV - 10% I-Bonds
Topic Author
skierincolorado
Posts: 1737
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 30, 2022 11:43 am
skierincolorado wrote: Fri Sep 30, 2022 10:08 am
LTCM wrote: Fri Sep 30, 2022 2:42 am Another thought exercise for this 0.2% rate

If I can borrow at the fed rate and buy 1,000,000 in treasuries I can earn $2000 per year lending them to the futures market? So If I buy $100,000,000 I can make $200,000 in a year? Totally risk free. Doesn't that seem high?

With a treasury futures market totaling $1.5 trillion in notional value outstanding that means those lending the treasuries are making $3bn per year. RISK FREE.
Figure 5 here

https://www.financialresearch.gov/brief ... Trades.pdf

The average is something like .15 for the 2 year but as high as .4 some years. The 5 year is a bit higher like .2 or .25.

It makes sense that one could earn .15% risk free. The market is not perfectly efficient and these transactions are not entirely devoid of counterparty risk. Not just anybody can borrow exactly at the risk free rate.

Early in the thread we empirically verified the numbers from this paper by comparing futures and bond fund returns. See the posts by zkn. Search the thread for zkn.
Thanks. It's hard on a phone but from eyeballing it the 2 year looks like it averages more like 0.02 than 0.2.

You agree the IRR on the CME site is the financing rate?

Do you think we receive the IRR when shorting futures?

I'll look up zkn when I get back to a computer.

That whole paper looks worth reading.
Make sure you are using the left axis. The right axis is for somethint else and has a different scale. Maybe .1 but definitely more than .02. I think the posts by zkn have a figure as well.
User avatar
LTCM
Posts: 380
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 just found another CME paper claiming the spread averages out to close to zero

https://www.google.com/url?sa=t&source= ... XhPs5tJ4x2


Page 16/18 (depending on your reader) has a section "comparing returns of cash to futures and repo"

Now its possible they've picked a bond/time period that makes futures look great but imo more likely is that it can be up or down by the kind of margin you talk about (0.2) but trends toward zero over time.
60% VUG - 20% VEA - 10% EDV - 10% I-Bonds
User avatar
LTCM
Posts: 380
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 30, 2022 11:55 am
LTCM wrote: Fri Sep 30, 2022 11:43 am
skierincolorado wrote: Fri Sep 30, 2022 10:08 am
LTCM wrote: Fri Sep 30, 2022 2:42 am Another thought exercise for this 0.2% rate

If I can borrow at the fed rate and buy 1,000,000 in treasuries I can earn $2000 per year lending them to the futures market? So If I buy $100,000,000 I can make $200,000 in a year? Totally risk free. Doesn't that seem high?

With a treasury futures market totaling $1.5 trillion in notional value outstanding that means those lending the treasuries are making $3bn per year. RISK FREE.
Figure 5 here

https://www.financialresearch.gov/brief ... Trades.pdf

The average is something like .15 for the 2 year but as high as .4 some years. The 5 year is a bit higher like .2 or .25.

It makes sense that one could earn .15% risk free. The market is not perfectly efficient and these transactions are not entirely devoid of counterparty risk. Not just anybody can borrow exactly at the risk free rate.

Early in the thread we empirically verified the numbers from this paper by comparing futures and bond fund returns. See the posts by zkn. Search the thread for zkn.
Thanks. It's hard on a phone but from eyeballing it the 2 year looks like it averages more like 0.02 than 0.2.

You agree the IRR on the CME site is the financing rate?

Do you think we receive the IRR when shorting futures?

I'll look up zkn when I get back to a computer.

That whole paper looks worth reading.
Make sure you are using the left axis. The right axis is for somethint else and has a different scale. Maybe .1 but definitely more than .02. I think the posts by zkn have a figure as well.
Eyeballing it I get:
2010-12: 0
2012-14: +0.2
2014-15: -0.1
2015-17: +0.25
2017-19: -0.2
2019-20: +0.25

So approx +0.65 over 10 years or 0.065
60% VUG - 20% VEA - 10% EDV - 10% I-Bonds
User avatar
LTCM
Posts: 380
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 »

That figure 5 includes "dtcc gcf repo index" as a source. If you google that it has exactly the data we need. Its entire purpose is to record the basis/spread/financing of futures contracts since 2010.
60% VUG - 20% VEA - 10% EDV - 10% I-Bonds
unemployed_pysicist
Posts: 67
Joined: Sat Oct 09, 2021 2:32 pm

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

Post by unemployed_pysicist »

Regarding the .3 spread to Tbills:

I don't understand the rationale for assuming this in general. Look at table 5 in the Hedge Funds and the Treasury Cash-Futures Disconnect paper. The average spread to Tbills is close to zero for most futures contracts (1st, 2nd, and 3rd to deliver) for a 28 year sample. The only positive spread I see is 0.138 for the first to deliver 2 year note future contract in the 1992-2020 sample. Everything else is negative. If you were to extend the sample to the 1980s for the T bond contract, one would see an average spread that is even more negative than in the 1992-2020 sample.

It just happened that there was a recent 10 year period where the 5 year note futures had around a 0.3 spread to the closest Tbill. So what? It makes no sense to me to assume it will be the same going forward. Better to calculate Implied repo rate yourself and know for sure what you are paying for your leverage.
couldn't afford the h | | BUY BONDS | WEAR DIAMONDS
User avatar
LTCM
Posts: 380
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 »

unemployed_pysicist wrote: Fri Sep 30, 2022 1:04 pm Look at table 5 in the Hedge Funds and the Treasury Cash-Futures Disconnect paper. The average spread to Tbills is close to zero for most futures contracts (1st, 2nd, and 3rd to deliver) for a 28 year sample. The only positive spread I see is 0.138 for the first to deliver 2 year note future contract in the 1992-2020 sample.
Unfortunately the 2 year closest to deliver is the one I need. So that means the DEC22 currently right? I wouldn't feel confident in the spreads/liquidity of the MAR23 or JUN23. 0.138 is not ideal. I wonder if there's a reason or if it's just one of those sample size noises.

Paying 0.138 for my STT and then offering a negative rate for shorting ITT is the opposite of what I'm after!
60% VUG - 20% VEA - 10% EDV - 10% I-Bonds
User avatar
LTCM
Posts: 380
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 »

Current IRR for treasury futures is 0.001% under the Fed Rate:

https://www.wsj.com/market-data/bonds/moneyrates

Secured Overnight Financing Rate (SOFR) [As of 9/28/22]
2.98

DTCC GCF Repo Index
Treasury 2.979
MBS 3.087

So this is linking it to the overnight rate, not the 3 month T-Bill. I'm guessing the 3 month T-Bill ends up being a composite of 90 days of the overnight rate?
60% VUG - 20% VEA - 10% EDV - 10% I-Bonds
User avatar
LTCM
Posts: 380
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 »

They have the raw data here:
https://www.dtcc.com/charts/dtcc-gcf-re ... x#download

But you'd need the overnight/3-month t-bill rate data to compare.
60% VUG - 20% VEA - 10% EDV - 10% I-Bonds
comeinvest
Posts: 1458
Joined: Mon Mar 12, 2012 6:57 pm

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

Post by comeinvest »

unemployed_pysicist wrote: Fri Sep 30, 2022 1:04 pm Regarding the .3 spread to Tbills:

I don't understand the rationale for assuming this in general. Look at table 5 in the Hedge Funds and the Treasury Cash-Futures Disconnect paper. The average spread to Tbills is close to zero for most futures contracts (1st, 2nd, and 3rd to deliver) for a 28 year sample. The only positive spread I see is 0.138 for the first to deliver 2 year note future contract in the 1992-2020 sample. Everything else is negative. If you were to extend the sample to the 1980s for the T bond contract, one would see an average spread that is even more negative than in the 1992-2020 sample.

It just happened that there was a recent 10 year period where the 5 year note futures had around a 0.3 spread to the closest Tbill. So what? It makes no sense to me to assume it will be the same going forward. Better to calculate Implied repo rate yourself and know for sure what you are paying for your leverage.
"Better to calculate Implied repo rate yourself and know for sure what you are paying for your leverage" - nice, but 2 things:

First, it's kind of tedious math work, isn't it? You would have to determine the current CTD ytm, the CTD coupon payment amounts and schedule, look up the discount rate to the end of the contract (T-bill rate with same maturity [go the the options box spread thread and you realize that T-bills with same maturity can have much different yields]), then make sure that the timing of your data points coincide, or do adjustments for after-market or intra-day fluctuations, and/or buy real-time data, then do the math. Finally, you would have to account for the value of CTD switching optionality, and certain other intricacies of futures fair value calculations, like volatility and cash flows which have different effects on the futures contract than on the underlying. (Granted the latter may be quite small effects for near-month futures.)
After you are done with all that and discover that the rate is, say, 0.1% higher than it "should" be, how do you know that a T-bill didn't just have a liquidity induced surge on that day?

Second, what would that exercise serve you? Would you be going in and out of the strategy based on daily financing spread fluctuations to T-Bills?
comeinvest
Posts: 1458
Joined: Mon Mar 12, 2012 6:57 pm

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

Post by comeinvest »

LTCM wrote: Fri Sep 30, 2022 12:41 pm
skierincolorado wrote: Fri Sep 30, 2022 11:55 am
LTCM wrote: Fri Sep 30, 2022 11:43 am
skierincolorado wrote: Fri Sep 30, 2022 10:08 am
LTCM wrote: Fri Sep 30, 2022 2:42 am Another thought exercise for this 0.2% rate

If I can borrow at the fed rate and buy 1,000,000 in treasuries I can earn $2000 per year lending them to the futures market? So If I buy $100,000,000 I can make $200,000 in a year? Totally risk free. Doesn't that seem high?

With a treasury futures market totaling $1.5 trillion in notional value outstanding that means those lending the treasuries are making $3bn per year. RISK FREE.
Figure 5 here

https://www.financialresearch.gov/brief ... Trades.pdf

The average is something like .15 for the 2 year but as high as .4 some years. The 5 year is a bit higher like .2 or .25.

It makes sense that one could earn .15% risk free. The market is not perfectly efficient and these transactions are not entirely devoid of counterparty risk. Not just anybody can borrow exactly at the risk free rate.

Early in the thread we empirically verified the numbers from this paper by comparing futures and bond fund returns. See the posts by zkn. Search the thread for zkn.
Thanks. It's hard on a phone but from eyeballing it the 2 year looks like it averages more like 0.02 than 0.2.

You agree the IRR on the CME site is the financing rate?

Do you think we receive the IRR when shorting futures?

I'll look up zkn when I get back to a computer.

That whole paper looks worth reading.
Make sure you are using the left axis. The right axis is for somethint else and has a different scale. Maybe .1 but definitely more than .02. I think the posts by zkn have a figure as well.
Eyeballing it I get:
2010-12: 0
2012-14: +0.2
2014-15: -0.1
2015-17: +0.25
2017-19: -0.2
2019-20: +0.25

So approx +0.65 over 10 years or 0.065
That would roughly be the same as for ITT on a duration-adjusted basis. But I'm not sure if that is a valid logical explanation, at least not a comprehensive one; because the margin requirements of different futures products are already based on and should reflect the volatility and risk of the underlying. The margin requirement should be the safety buffer for the counterparty. (Counterparty is CME for futures, which is considered 100% safe.)
ivgrivchuck
Posts: 1382
Joined: Sun Sep 27, 2020 6:20 pm

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

Post by ivgrivchuck »

LazyOverthinker wrote: Fri Sep 30, 2022 10:33 am Another way to say it is that growth-investing is defined by bad surprises, value-investing is defined by good surprises. Thoughts?
That is perhaps an overstatement. These are small distressed companies. As you said macroeconomic factors can hit these companies especially hard. So a prolonged recession might be especially devastating for these companies. But that's exactly why there is the extra risk-premium that SCV provides. Then again, putting all the eggs in the same basket... non-diversified risk is especially dangerous.

Personally 50%/50% SCV/Total Market is the highest I could tolerate/recommend. But... to each their own...
25% VTI | 25% VXUS | 12.5% AVUV | 10% AVDV | 2.5% VWO | 25% I/EE-bonds
Post Reply