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: 2728
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 »

stunning_linguist wrote: Tue Feb 13, 2024 11:09 pm A bit of a tangent:

If anyone recalls the repo market liquidity crisis of 2019, SOFR rates spiked surprisingly high during that time, and SOFR rates overall became very volatile. What's the general consensus on the likelihood of such liquidity crises reoccurring in the future, and their impact on SOFR futures vs. Treasury futures returns in mHFEA?

Example: In September 2019, the SOFR rate spiked up above 5%. If I had held a SOFR future whose maturity / reference period coincided with that event, it would surely have eaten at my returns "for no good reason"

And a follow-up question: Are investors adequately compensated for taking on that specific risk? Because I currently imagine that SOFR futures returns have 2 components: 1 component tied to the underlying (and shifting) term structure of risk free interest rates, and 1 component that captures these idiosyncratic black swan risks (like liquidity crunches) that arise purely out of inefficiencies in the market. And knowing how bad the market is at pricing/predicting black swan events, I'm feeling kind of skeptical of SOFR futures
I don't know much about the 2019 liquidity crisis, and I think it has not been discussed in this thread. Can you provide some data what exactly happened? How did SOFR futures fare in comparison to other fixed income like treasuries? Term SOFR rates and/or futures at what expiration spiked?
If it was just the short-term rates or the futures with near expiration, it should have not affected your overall positions from duration exposure a lot. On the flip side, it might have lifted your instruments with implied financing (equity index futures, treasury futures, and SPX box spreads) a bit, if the mark to market valuation of the implied loan decreased (i.e. the value of your short exposure to the loan increased) - although during some abnormal periods, different derivatives reacted in different ways; so that effect might have been obscured, or exaggerated (for example when options implied financing rates rise).
From what I recall, there were some episodes where SOFR (back then Eurodollar) futures were a bit more stable than treasury futures for example, possibly explaining their sometimes slightly lower expected return in very recent history. There are some papers which I still wanted to read and post.
SOFR and Eurodollar futures data is available for free on Yahoo, investing.com, and at some brokers; so you can backtest.
bond93
Posts: 25
Joined: Mon Jan 22, 2024 11:38 pm

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

Post by bond93 »

comeinvest wrote: Mon Feb 12, 2024 10:58 pm
bond93 wrote: Mon Feb 12, 2024 1:30 pm As far as the gold is concerned, isn't that precisely why Dalio and others use it? Its lack of correlation to anything? Note that the 3rd portfolio in that test did the best and had 15% gold.
1. I doubt that something with no correlation to anything and 0% expected return will be of any use if I add to any portfolio.
2. Like I said before. Take say 10 commodities, or anything else like used cars, fine art, or plastic straws - and backtest over whatever limited backtesting timeframe - you will probably find more than one, maybe 3-5, that improve your portfolio by Sharpe ratio or whatever, during that timeframe. High risk of parameter fitting the past!
Thanks for the tip about your commodities posts.

I might push back on the idea that Gold has 0 expected return. It's demand can only rise as world population increases, whether that be for jewelry or use in technology/computers/etc. Over the last 100 years it has almost matched stocks, and since we got off the gold standard it has grown from $35 to $2000. Gold has a market cap of 13 trillion, Microsoft 3 trillion. The difference between gold and plastic straws/used cars etc. is that gold, unlike most other companies that exist today, has been highly valued for 5000 years. I probably wouldn't put it in an unleveraged stock portfolio as yes 100% stocks wins out in the end, but it may help to tame a 3x leveraged one using a small percentage. Also I lol-ed reading your comment about Dalio/Swensen etc. "needing to justify their job positions" by adding seemingly useless assets. I like your sense of humour. However, 7.5% of Bridgewater's assets would be close to 10 billion in gold. That's a lot of dough in a useless asset. I don't think he grew the world's most successful hedge fund with the idea that by tossing in some gold he would "look busier."

A test going back to 1987 with unlevered gold as a hedge:
https://www.portfoliovisualizer.com/bac ... xgfRddxQsD

Even 50/50 unlevered gold and stocks have nearly identical CAGR:
https://www.portfoliovisualizer.com/bac ... vSnaQ5zb76

Oh another fun fact about gold: its virtually indestructible
comeinvest
Posts: 2728
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 »

bond93 wrote: Wed Feb 14, 2024 1:23 pm
comeinvest wrote: Mon Feb 12, 2024 10:58 pm
bond93 wrote: Mon Feb 12, 2024 1:30 pm As far as the gold is concerned, isn't that precisely why Dalio and others use it? Its lack of correlation to anything? Note that the 3rd portfolio in that test did the best and had 15% gold.
1. I doubt that something with no correlation to anything and 0% expected return will be of any use if I add to any portfolio.
2. Like I said before. Take say 10 commodities, or anything else like used cars, fine art, or plastic straws - and backtest over whatever limited backtesting timeframe - you will probably find more than one, maybe 3-5, that improve your portfolio by Sharpe ratio or whatever, during that timeframe. High risk of parameter fitting the past!
Thanks for the tip about your commodities posts.

I might push back on the idea that Gold has 0 expected return. It's demand can only rise as world population increases, whether that be for jewelry or use in technology/computers/etc. Over the last 100 years it has almost matched stocks, and since we got off the gold standard it has grown from $35 to $2000. Gold has a market cap of 13 trillion, Microsoft 3 trillion. The difference between gold and plastic straws/used cars etc. is that gold, unlike most other companies that exist today, has been highly valued for 5000 years. I probably wouldn't put it in an unleveraged stock portfolio as yes 100% stocks wins out in the end, but it may help to tame a 3x leveraged one using a small percentage. Also I lol-ed reading your comment about Dalio/Swensen etc. "needing to justify their job positions" by adding seemingly useless assets. I like your sense of humour. However, 7.5% of Bridgewater's assets would be close to 10 billion in gold. That's a lot of dough in a useless asset. I don't think he grew the world's most successful hedge fund with the idea that by tossing in some gold he would "look busier."

A test going back to 1987 with unlevered gold as a hedge:
https://www.portfoliovisualizer.com/bac ... xgfRddxQsD

Even 50/50 unlevered gold and stocks have nearly identical CAGR:
https://www.portfoliovisualizer.com/bac ... vSnaQ5zb76

Oh another fun fact about gold: its virtually indestructible
Your first PV link has symbols that I don't have access to - do you have a pro version or something? Your second PV simulation looks appealing, but read the articles about gold by Larry Swedroe here and on Seeking Alpha.
I think the theory is that gold has no tail risk as it can be hidden, and it is hard to destroy even in a war; and the drawdown risk is not correlated to most other assets; as it has no risk, therefore it has no risk premium, i.e. its long-term expected real return is zero. Don't be fooled by creating your own little theories if you don't have insider knowledge in the industry; it's hard to predict supply and demand, and the future of any market is often different from what you think. If supply and demand were that clear cut, many portfolio managers would pile into it.
There is also a possibility that your chart starts at a low valuation point.
You could argue that if it has no risk in that sense i.e. doesn't add to the drawdown risk in a diversified portfolio, and also no expected return, you can just as well add it. But it in reality there is some storage cost, and it would reduce the capital efficiency because you have nominal leverage constraints in your brokerage account.
The fact that some famous dude has 10% gold in a portfolio that he markets in many media and that his brand name is attached to, is not a good argument. Think of survivorship bias - possibly one dude had a lucky strike; and think of why have the many other superstar investors no gold in their portfolios, if it was that clear cut. All superstar investors, along with all endowments and all family trusts and hedge funds have access to the gold charts, and most of them probably looked at them and deliberated it at some point.
stunning_linguist
Posts: 6
Joined: Sun Dec 24, 2023 2:20 pm

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

Post by stunning_linguist »

comeinvest wrote: Wed Feb 14, 2024 4:31 am
stunning_linguist wrote: Tue Feb 13, 2024 11:09 pm A bit of a tangent:

If anyone recalls the repo market liquidity crisis of 2019, SOFR rates spiked surprisingly high during that time, and SOFR rates overall became very volatile. What's the general consensus on the likelihood of such liquidity crises reoccurring in the future, and their impact on SOFR futures vs. Treasury futures returns in mHFEA?

Example: In September 2019, the SOFR rate spiked up above 5%. If I had held a SOFR future whose maturity / reference period coincided with that event, it would surely have eaten at my returns "for no good reason"

And a follow-up question: Are investors adequately compensated for taking on that specific risk? Because I currently imagine that SOFR futures returns have 2 components: 1 component tied to the underlying (and shifting) term structure of risk free interest rates, and 1 component that captures these idiosyncratic black swan risks (like liquidity crunches) that arise purely out of inefficiencies in the market. And knowing how bad the market is at pricing/predicting black swan events, I'm feeling kind of skeptical of SOFR futures
I don't know much about the 2019 liquidity crisis, and I think it has not been discussed in this thread. Can you provide some data what exactly happened? How did SOFR futures fare in comparison to other fixed income like treasuries? Term SOFR rates and/or futures at what expiration spiked?
If it was just the short-term rates or the futures with near expiration, it should have not affected your overall positions from duration exposure a lot. On the flip side, it might have lifted your instruments with implied financing (equity index futures, treasury futures, and SPX box spreads) a bit, if the mark to market valuation of the implied loan decreased (i.e. the value of your short exposure to the loan increased) - although during some abnormal periods, different derivatives reacted in different ways; so that effect might have been obscured, or exaggerated (for example when options implied financing rates rise).
From what I recall, there were some episodes where SOFR (back then Eurodollar) futures were a bit more stable than treasury futures for example, possibly explaining their sometimes slightly lower expected return in very recent history. There are some papers which I still wanted to read and post.
SOFR and Eurodollar futures data is available for free on Yahoo, investing.com, and at some brokers; so you can backtest.
Here's the chart of the SOFR rate during 2019, pulled from https://www.newyorkfed.org/markets/reference-rates/sofr :

Image

There's a spike in September

However, strangely, here's how the futures price behaved in 2019 -- there's no noticeable dip at all!

Image

In comparison, the ZT (2 year treasury) futures price was far more volatile:

Image

I hope I'm not looking at the wrong data here... but I guess the effect of transient spikes in the SOFR rate gets washed out when you take the full 3 month average for the SOFR3 reference calculation; and there's only cause for concern if the rate spike persists across a long period of time (?)

Now I wonder what kinds of scenarios might cause a spike like that to persist ... a majority of the US's financial institutions becoming undercapitalized? a financial meltdown like 2008?

Looking at the historical divergence between the FFR and LIBOR (as a proxy, since SOFR wasn't around in 2008), it does seem that the LIBOR did maintain a sizable, sustained spread above the FFR during some part of the 2008 crisis:

Image

However, this data won't tell the full story, because of the Libor Scandal (which I just learned about 3 minutes ago during my research for this post...) : https://en.wikipedia.org/wiki/Libor_scandal -- where banks under-reported their actualized interest rates for LIBOR.

I know I'm rambling at this point, but I'm just thinking out loud here-- So I guess the missing piece of the thesis is whether we think our financial institutions have enough guardrails to prevent another financial crisis at the scale of 2008. There's the Dodd-Frank Act of course. And it seems the SEC is even brewing up guardrails against another 2019/2020-like repo liquidity crunch: https://finance.yahoo.com/news/billiona ... 16905.html

So is this enough data for SOFR holders to rest easy? I don't know ... maybe?
Last edited by stunning_linguist on Thu Feb 15, 2024 12:10 am, edited 1 time in total.
comeinvest
Posts: 2728
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 »

stunning_linguist wrote: Thu Feb 15, 2024 12:05 am
You have to include the image URL, not the Imgur page, if you want the chart to show in the post.
stunning_linguist
Posts: 6
Joined: Sun Dec 24, 2023 2:20 pm

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

Post by stunning_linguist »

comeinvest wrote: Thu Feb 15, 2024 12:08 am
stunning_linguist wrote: Thu Feb 15, 2024 12:05 am
You have to include the image URL, not the Imgur page, if you want the chart to show in the post.
ah ... oops 😅
comeinvest
Posts: 2728
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 »

stunning_linguist wrote: Thu Feb 15, 2024 12:05 am Here's the chart of the SOFR rate during 2019, pulled from https://www.newyorkfed.org/markets/reference-rates/sofr :

...
There's a spike in September

However, strangely, here's how the futures price behaved in 2019 -- there's no noticeable dip at all!

...
In comparison, the ZT (2 year treasury) futures price was far more volatile:

...
I hope I'm not looking at the wrong data here... but I guess the effect of transient spikes in the SOFR rate gets washed out when you take the full 3 month average for the SOFR3 reference calculation; and there's only cause for concern if the rate spike persists across a long period of time (?)

Now I wonder what kinds of scenarios might cause a spike like that to persist ... a majority of the US's financial institutions becoming undercapitalized? a financial meltdown like 2008?

Looking at the historical divergence between the FFR and LIBOR (as a proxy, since SOFR wasn't around in 2008), it does seem that the LIBOR did maintain a sizable, sustained spread above the FFR during some part of the 2008 crisis:

...
However, this data won't tell the full story, because of the Libor Scandal (which I just learned about 3 minutes ago during my research for this post...) : https://en.wikipedia.org/wiki/Libor_scandal -- where banks under-reported their actualized interest rates for LIBOR.

I know I'm rambling at this point, but I'm just thinking out loud here-- So I guess the missing piece of the thesis is whether we think our financial institutions have enough guardrails to prevent another financial crisis at the scale of 2008. There's the Dodd-Frank Act of course. And it seems the SEC is even brewing up guardrails against another 2019/2020-like repo liquidity crunch: https://finance.yahoo.com/news/billiona ... 16905.html

So is this enough data for SOFR holders to rest easy? I don't know ... maybe?
I think the SOFR would not have exhibited the behavior of the LIBOR in 2008. It was an artifact of the LIBOR-FFR spread, which would probably not translate into a SOFR-FFR spread, but more of a LIBOR-SOFR spread. SOFR has no credit risk component.
stunning_linguist
Posts: 6
Joined: Sun Dec 24, 2023 2:20 pm

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

Post by stunning_linguist »

comeinvest wrote: Thu Feb 15, 2024 12:16 am
stunning_linguist wrote: Thu Feb 15, 2024 12:05 am Here's the chart of the SOFR rate during 2019, pulled from https://www.newyorkfed.org/markets/reference-rates/sofr :

...
There's a spike in September

However, strangely, here's how the futures price behaved in 2019 -- there's no noticeable dip at all!

...
In comparison, the ZT (2 year treasury) futures price was far more volatile:

...
I hope I'm not looking at the wrong data here... but I guess the effect of transient spikes in the SOFR rate gets washed out when you take the full 3 month average for the SOFR3 reference calculation; and there's only cause for concern if the rate spike persists across a long period of time (?)

Now I wonder what kinds of scenarios might cause a spike like that to persist ... a majority of the US's financial institutions becoming undercapitalized? a financial meltdown like 2008?

Looking at the historical divergence between the FFR and LIBOR (as a proxy, since SOFR wasn't around in 2008), it does seem that the LIBOR did maintain a sizable, sustained spread above the FFR during some part of the 2008 crisis:

...
However, this data won't tell the full story, because of the Libor Scandal (which I just learned about 3 minutes ago during my research for this post...) : https://en.wikipedia.org/wiki/Libor_scandal -- where banks under-reported their actualized interest rates for LIBOR.

I know I'm rambling at this point, but I'm just thinking out loud here-- So I guess the missing piece of the thesis is whether we think our financial institutions have enough guardrails to prevent another financial crisis at the scale of 2008. There's the Dodd-Frank Act of course. And it seems the SEC is even brewing up guardrails against another 2019/2020-like repo liquidity crunch: https://finance.yahoo.com/news/billiona ... 16905.html

So is this enough data for SOFR holders to rest easy? I don't know ... maybe?
I think the SOFR would not have exhibited the behavior of the LIBOR in 2008. It was an artifact of the LIBOR-FFR spread, which would probably not translate into a SOFR-FFR spread, but more of a LIBOR-SOFR spread. SOFR has no credit risk component.
That last part worries me-- and I guess similar thoughts have already been raised in this thread-- If SOFR has no credit risk component (since it's collateralized by Treasuries), then it should *in theory* be "risk free". And yet a spread exists. What gives? I'm not brave or knowledgeable enough in the debt markets to assert this is an arbitrage opportunity ... however tempting it is. hopefully someone someday can shed the light this thread needs.

*Edit: actually, I'm dumb. I shouldn't be confusing "credit risk" with the apparent "liquidity premium" (might be mis-using terms here) that SOFR exhibits during liquidity crunches.

**Edit 2: wait ... maybe the spread is just the "liquidity premium" plus a correction factor due to differences in volatility of SOFR vs. treasuries: in the front of the curve, a positive spread exists because there are insufficient regulatory guardrails to prevent liquidity crunches in the short term, which near-term SOFR contracts would have less time to avoid; whereas in the back of the curve, a negative spread exists because of the assumption that financial regulations will improve by then, and so the volatility factor becomes a dominating term (which Treasuries have more of).
Last edited by stunning_linguist on Thu Feb 15, 2024 1:09 am, edited 1 time in total.
comeinvest
Posts: 2728
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 »

stunning_linguist wrote: Thu Feb 15, 2024 12:38 am
comeinvest wrote: Thu Feb 15, 2024 12:16 am I think the SOFR would not have exhibited the behavior of the LIBOR in 2008. It was an artifact of the LIBOR-FFR spread, which would probably not translate into a SOFR-FFR spread, but more of a LIBOR-SOFR spread. SOFR has no credit risk component.
That last part worries me-- and I guess similar thoughts have already been raised in this thread-- If SOFR has no credit risk component (since it's collateralized by Treasuries), then it should *in theory* be "risk free". And yet a spread exists. What gives? I'm not brave or knowledgeable enough in the debt markets to assert this is an arbitrage opportunity ... however tempting it is. hopefully someone someday can shed the light this thread needs
I came to the same conclusion. Read my posts a number of pages up on the SOFR futures vs treasury futures comparison, and the charts of the swap spread.
unemployed_pysicist
Posts: 220
Joined: Sat Oct 09, 2021 2:32 pm
Location: Amsterdam

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

Post by unemployed_pysicist »

comeinvest wrote: Sat Jan 27, 2024 10:19 pm
unemployed_pysicist wrote: Sat Jan 27, 2024 6:44 am
SongOfTheFates wrote: Sun Jan 21, 2024 4:57 am Hey folks! Thanks for all the research and discussion you've all documented here.

I've been making my way through the thread, but has there been much discussion about using treasury spreads to hedge instead of going long treasuries? I've been eyeing using the 10y-2y and 10y-3mo SOFR futures and on a surface level it seems to be pretty cohesive as long as you're treating it as a tool you apply sparingly instead of a comprehensive 'portfolio' you can hold and backtest for 30+ years.

I'll elaborate more when it's not 3am where I'm at :)
I would be interested to hear more details about your yield spread (overlay?) and how you plan to systematically apply your exposure to this spread in the context of a mHFEA portfolio.

As I recall, the 10y-2y spread has about a 0.5 Pearson correlation coefficient to the unemployment rate from about 1953. So it is a proxy for the business cycle. However, going long this spread would not have helped in the most recent, 2022-2023 drawdown in equities and bonds, for example. In fact it would have been an additional drag on the portfolio. So I am curious how this additional tool could be used to improve mHFEA.
I don't think it makes sense, because if you read the papers posted earlier in this thread, the Sharpe ratio is higher the lower the maturity of the treasuries down to about 12 months. So you would be short the most profitable part of the yield curve, and also the part that is the most un-correlated or anti-correlated to equities during equity market crashes.
Earlier in this thread I instead proposed a curve steepener as a perpetual or strategic position, i.e. an exaggerated version of mHFEA that is not constrained by >0% allocations on the yield curve. The 0% constraint as you move from HFEA to mHFEA seems arbitrary.
For example, +200% ITT and -50% ITT-equivalent LTT as a partial hedge of the ITT for a 150% ITT-equivalent total target duration exposure, but with better risk-adjusted performance as LTT has lousy risk-adjusted performance. LTT, for example the UB future or perhaps even better the 30-year swap futures contract, is the black sheep, if the premise of mHFEA is right.

The slope between 20y and 30y has historically been rather negative - almost the entire history, except ca. 5 years during ZIRP: viewtopic.php?p=7470297#p7470297
A position that exploits the 10y-2y spread is a steepener, is it not? I assumed that was what user SongOfTheFates meant. To make money from the widening spread of the 10 year and 2 year, which is what generally happens around recessions, you would need to go long the 2 year note while shorting the 10 year note.

A steepener did not help during the 2022-2023 drawdown in stocks and bonds. I made this comment just in case this overlay to mHFEA had been suggested as a way to help in a bonds falling/stocks falling scenario. I could imagine that overlaying a steepener onto mHFEA should enhance returns around recessions, and perhaps diminish returns around the peak of the business cycle. But I am curious what is meant by:
SongOfTheFates wrote: Sun Jan 21, 2024 4:57 am as long as you're treating it as a tool you apply sparingly instead of a comprehensive 'portfolio' you can hold and backtest for 30+ years.
To me, this implies some kind of systematic approach.

Thinking out loud, I wonder if there is a more reliable negative correlation between equity returns and the 10s 2s (duration neutral) steepener than there is between short term treasury returns and equity returns. Maybe worth having a look.
couldn't afford the h | | BUY BONDS | WEAR DIAMONDS
unemployed_pysicist
Posts: 220
Joined: Sat Oct 09, 2021 2:32 pm
Location: Amsterdam

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

Post by unemployed_pysicist »

stunning_linguist wrote: Thu Feb 15, 2024 12:05 am
Here's the chart of the SOFR rate during 2019, pulled from https://www.newyorkfed.org/markets/reference-rates/sofr :
...
There's a spike in September
...
However, strangely, here's how the futures price behaved in 2019 -- there's no noticeable dip at all!
...
I hope I'm not looking at the wrong data here... but I guess the effect of transient spikes in the SOFR rate gets washed out when you take the full 3 month average for the SOFR3 reference calculation; and there's only cause for concern if the rate spike persists across a long period of time (?)
Thanks for posting the SOFR rate and resulting returns (I assume it was from the front contract?) The results are what I expected to see (the brief blowup in rates gets washed out in the full 3 month average). But I was too lazy to check it myself :happy
couldn't afford the h | | BUY BONDS | WEAR DIAMONDS
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 »

Quarterly check-in, since I'm about to roll over futures/rebalance. Taxable and IRA are in Interactive Brokers:

Taxable:
Net liquidation: $506k. Equities financed via $300k box spread. Net cash is -$14k.
Equities: $808k, split between ITOT (total stock market), VBR (small cap value), VOO (S&P500), VXUS (total international ex-US). Ratio to net equity is 1.6:1.
Treasuries: $745k, all ZF. Ratio to net equity is 1.47:1

IRA:
Net liquidation: $312k. Net cash is +$163k.
Equities: $452k, split between ES, MES, and VT (total international plus US). Ratio to net equity is 1.45:1.
Treasuries: $477k, split between ZF and VGIT. Ratio to net equity is 1.53:1.

401k (Schwab, allows for self-directed brokerage account):
$110k UPRO (61%)
$69k TYA (39%)

Thoughts/next steps:
  • First time treasury futures have closed positive before a rollover, I think? Feels good.
  • Going to rebalance taxable and IRA to 1.65:1 for both equities and treasuries.
  • I have too much cash in the IRA. Original plan was 15% of equities exposure 5% of treasury exposure in cash, but I think that's too much, and would rather stay closer to a $0 balance and just sell if needed. I looked at buying t-bills, but it's not quite clear how that would work - is there a way to set it to automatically liquidate t-bills if more cash is needed due to futures losses?
Any thoughts/suggestions to what I have? I have read this thread... or tried to, at least. Most of the meat is in the first 5-10 pages. After that, the content is very very dense.
diving6403
Posts: 12
Joined: Mon Jun 12, 2023 11:05 am

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

Post by diving6403 »

DMoogle wrote: Thu Feb 22, 2024 10:29 am Taxable:
Net liquidation: $506k. Equities financed via $300k box spread. Net cash is -$14k.
Equities: $808k, split between ITOT (total stock market), VBR (small cap value), VOO (S&P500), VXUS (total international ex-US). Ratio to net equity is 1.6:1.
Treasuries: $745k, all ZF. Ratio to net equity is 1.47:1
I can't answer any of your questions but I'm looking forward to seeing other's answers. I have a question about this though. I'm still finishing wrapping my head around this thread and futures trading, but don't you need to have a positive cash balance as collateral for your futures? Are you being charged margin interest on that -$14k?

Logically, the outcome of what you're doing makes sense to me; it doesn't make sense to carry a positive cash balance while also carrying a box spread, but I thought it was required (hence the conversation about using t bills as collateral instead). But I'm not sure logistically how it works.
comeinvest
Posts: 2728
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 »

DMoogle wrote: Thu Feb 22, 2024 10:29 am Quarterly check-in, since I'm about to roll over futures/rebalance. Taxable and IRA are in Interactive Brokers:

Taxable:
Net liquidation: $506k. Equities financed via $300k box spread. Net cash is -$14k.
Equities: $808k, split between ITOT (total stock market), VBR (small cap value), VOO (S&P500), VXUS (total international ex-US). Ratio to net equity is 1.6:1.
Treasuries: $745k, all ZF. Ratio to net equity is 1.47:1

IRA:
Net liquidation: $312k. Net cash is +$163k.
Equities: $452k, split between ES, MES, and VT (total international plus US). Ratio to net equity is 1.45:1.
Treasuries: $477k, split between ZF and VGIT. Ratio to net equity is 1.53:1.

401k (Schwab, allows for self-directed brokerage account):
$110k UPRO (61%)
$69k TYA (39%)

Thoughts/next steps:
  • First time treasury futures have closed positive before a rollover, I think? Feels good.
  • Going to rebalance taxable and IRA to 1.65:1 for both equities and treasuries.
  • I have too much cash in the IRA. Original plan was 15% of equities exposure 5% of treasury exposure in cash, but I think that's too much, and would rather stay closer to a $0 balance and just sell if needed. I looked at buying t-bills, but it's not quite clear how that would work - is there a way to set it to automatically liquidate t-bills if more cash is needed due to futures losses?
Any thoughts/suggestions to what I have? I have read this thread... or tried to, at least. Most of the meat is in the first 5-10 pages. After that, the content is very very dense.
I think you did a good job with your asset allocation. Only some minor thoughts:

I would think about lowering the equities percentage in the taxable account, because with 1:1.65 you will have to deleverage occasionally, and as a result incur tax drag.

You also have to reflect the higher market beta of VBR in your asset allocation. I think market beta instead of equities percentage is the better risk and allocation measure for the definition of a consistent asset allocation target of mHFEA. Whether the leverage (beta) with respect to the market index is internal, external (explicit), or intrinsic, makes no difference.

Similarly, duration exposure rather than the nominal bond percentage is the more consistent risk measure for fixed income.

With that in mind, I would think about getting rid of UPRO with its associated fees in the 401k, and instead move your VBR and other equities with >1 market beta to the 401k, while maintaining your overall targets across accounts. This way you could possibly create a similar risk exposure within and across accounts, while reducing or eliminating the fees. I'm allergic to fees, because they constitute return-free risk. It is questionable whether the benefit of leveraged fixed income with its small term premium will be bigger than the fee drag from LETFs, if you have both in your portfolio. (Money is fungible, so which asset class has the fee attached is not relevant, as long as any fee product and any low expected return [above the risk-free rate] asset class is in your portfolio, and the elimination of one could eliminate the other.)
cometqq
Posts: 13
Joined: Mon Jul 20, 2020 7:36 am

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

Post by cometqq »

comeinvest wrote: Thu Oct 19, 2023 4:44 pm
sharukh wrote: Thu Oct 19, 2023 2:45 pm in an IRA with 200k:

100k VXUS
50k VTI
50k cash
4 SOFR (1M)
7 MES (150k)

Do we still say the above portfolio is around 140% equities and 160% bonds ?

Thank you.
Without doing the exact math, if your question is whether the SOFR futures constitute a fixed income percentage allocation of the portfolio, I would say yes they are "equivalent", based on DV01 which is what primarily matters, and for the most part based on risk exposures and return premia.
You read the discussion on the SOFR swap spread, so you are aware of the differences. Pending further research and insight on the swap spreads and treasury futures implied financing cost which I plan on doing once I have time, I personally would use SOFR futures strips only up to 2 years out at the moment.
Hi, comeinvest. Why would you only use SOFR futures strips within a 2-year timeframe? Is this based on the notion that mHFEA does better with STT, or are there specific risks or problems associated with SOFR futures strips exceeding the 2-year threshold? I don't remember seeing any problems with SOFR futures strips extending beyond the 2-year mark in the discussion here.
comeinvest
Posts: 2728
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 »

cometqq wrote: Fri Feb 23, 2024 10:24 pm
comeinvest wrote: Thu Oct 19, 2023 4:44 pm
sharukh wrote: Thu Oct 19, 2023 2:45 pm in an IRA with 200k:

100k VXUS
50k VTI
50k cash
4 SOFR (1M)
7 MES (150k)

Do we still say the above portfolio is around 140% equities and 160% bonds ?

Thank you.
Without doing the exact math, if your question is whether the SOFR futures constitute a fixed income percentage allocation of the portfolio, I would say yes they are "equivalent", based on DV01 which is what primarily matters, and for the most part based on risk exposures and return premia.
You read the discussion on the SOFR swap spread, so you are aware of the differences. Pending further research and insight on the swap spreads and treasury futures implied financing cost which I plan on doing once I have time, I personally would use SOFR futures strips only up to 2 years out at the moment.
Hi, comeinvest. Why would you only use SOFR futures strips within a 2-year timeframe? Is this based on the notion that mHFEA does better with STT, or are there specific risks or problems associated with SOFR futures strips exceeding the 2-year threshold? I don't remember seeing any problems with SOFR futures strips extending beyond the 2-year mark in the discussion here.
Go back and read the pages and charts about the SOFR swap spread up in this thread, or check it on chathamfinancial.com. The swap spread gets bigger with increasing maturity. Last time I checked, the swap spread was positive up to ca. 1 year and slightly negative at 2 years. We didn't finish the discussion; I still have some pending research to do. There might be risk based explanations. There are related papers on the internet, if you want to do some reading.
CostcoBoxWine
Posts: 9
Joined: Wed Jul 12, 2023 8:50 pm

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

Post by CostcoBoxWine »

DMoogle wrote: Thu Feb 22, 2024 10:29 am Quarterly check-in, since I'm about to roll over futures/rebalance. Taxable and IRA are in Interactive Brokers:

Taxable:
Net liquidation: $506k. Equities financed via $300k box spread. Net cash is -$14k.
Equities: $808k, split between ITOT (total stock market), VBR (small cap value), VOO (S&P500), VXUS (total international ex-US). Ratio to net equity is 1.6:1.
Treasuries: $745k, all ZF. Ratio to net equity is 1.47:1

IRA:
Net liquidation: $312k. Net cash is +$163k.
Equities: $452k, split between ES, MES, and VT (total international plus US). Ratio to net equity is 1.45:1.
Treasuries: $477k, split between ZF and VGIT. Ratio to net equity is 1.53:1.

401k (Schwab, allows for self-directed brokerage account):
$110k UPRO (61%)
$69k TYA (39%)

Thoughts/next steps:
  • First time treasury futures have closed positive before a rollover, I think? Feels good.
  • Going to rebalance taxable and IRA to 1.65:1 for both equities and treasuries.
  • I have too much cash in the IRA. Original plan was 15% of equities exposure 5% of treasury exposure in cash, but I think that's too much, and would rather stay closer to a $0 balance and just sell if needed. I looked at buying t-bills, but it's not quite clear how that would work - is there a way to set it to automatically liquidate t-bills if more cash is needed due to futures losses?
Any thoughts/suggestions to what I have? I have read this thread... or tried to, at least. Most of the meat is in the first 5-10 pages. After that, the content is very very dense.
I'm curious if you can provide any estimates of what your ZF futures performance was for the last quarter in comparison to if you had a direct position in a treasury ETF, like VGIT for example.

I know these aren't directly comparable due to differences between the single duration of the CTD treasury vs the basket of durations in VGIT, but from an implementation comparison, i would think these would be close competitors. (similar to MillenialMillions last yearly performance benchmark)

Currently sitting on the Go button for starting a ZF roll myself, but i keep worrying myself with the inverted yield curve thinking this is a bad time to start. I can convince myself not to time the market in equities, but when I notice NTSX is posting negative yields on their futures ladder, it gives me cold feet...
comeinvest
Posts: 2728
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 »

CostcoBoxWine wrote: Sun Feb 25, 2024 2:32 pm I'm curious if you can provide any estimates of what your ZF futures performance was for the last quarter in comparison to if you had a direct position in a treasury ETF, like VGIT for example.

I know these aren't directly comparable due to differences between the single duration of the CTD treasury vs the basket of durations in VGIT, but from an implementation comparison, i would think these would be close competitors. (similar to MillenialMillions last yearly performance benchmark)

Currently sitting on the Go button for starting a ZF roll myself, but i keep worrying myself with the inverted yield curve thinking this is a bad time to start. I can convince myself not to time the market in equities, but when I notice NTSX is posting negative yields on their futures ladder, it gives me cold feet...
The negative current cash flow is the wrong way to think of it, especially on the short end of the yield curve. You need to consider the forward rates in time and maturity space. Trying to reverse-engineer the yield curve and term structure dynamics yourself, quickly gets you deep into term premium estimation theories.
For mid- and long-term treasuries, the slope is an indication, although not foolproof, of the expected term premium.

Regarding the ZF vs VGIT comparison: This would only be meaningful over long periods, and even then a comparison is problematic on a risk-adjusted basis. Over short periods, the performance differences will be dominated by the mean-reverting effect of even small yield curve slope changes.
You would also have to adjust for the implied cash rate, which results in an additional estimation error that contributes to noise in any short-term performance comparison.
You don't know the exact composition of VGIT, let alone the exact historical composition. So it all comes down to calculating the implied financing rate of ZF based on the underlying CTD, with entails estimating all the unknowns like the various optionalities embedded in the futures contracts; or long-term comparisons based on observed risk-adjusted returns (basically implying the historical durations, or partially filtering out the effect of the composition of historical maturities of the ETF).
diving6403
Posts: 12
Joined: Mon Jun 12, 2023 11:05 am

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

Post by diving6403 »

I've read through this entire thread at least one, and various sections a dozen times while trying to understand it. I've also read the Lifecycle Investing book. My overall target is 140/160 ITT, with equities divided 62/38 based on world market cap. I may consider adding factor tilts in later, for now I want to make sure I have the basics down.

Having said that, is there any sort of TL;DR or consolidated resources for deciding on a target bond duration? I don't have any sort of finance background and I'm a buy-and-hold investor, so I'm not very familiar with many of the terms thrown around in this thread. High level I understand the efficient frontier and that shorter term bonds have better risk-adjusted returns than LTTs. But the more I read this thread, the more confused I get about the detailed specifics of the various financial instruments. I've played around a lot with PV and Simba's sheet and have read about SOFRs and term premiums, but it just hasn't "clicked" yet for me how term premia, forward rates, DV01s, SOFR, etc all combine. Maybe a general bond duration exposure of ~3-7 years is good enough for most and comeinvest, skier and others are just microptimizing, but it's difficult for an idiot like me to tell. And to be clear, I know investing is risky and I'm not expecting there to be a "solution" to the perfect portfolio that will definitely not lose money. It's more that I'd like to truly understand more why I might not choose a ~50:50 ZF:ZN ratio. For example,how would using a 75:25 ZF:ZN ratio to target a 4.5 year duration change things? There's an immense amount of extremely dense analysis in this thread and I promise I'm not trying to be lazy and skip reading, it's just that so far I'm having troubles wrapping my head around the intricacies.

This is my allocation plan per account:

Taxable (IBKR):
  • ZF/ZN up to 160% exposure, in ratios equaling 5 years per CME's analytics tool (right now it's basically 50/50).
  • VTI/VXUS weighted by world market cap (same as VT but gives me foreign tax credit)
  • Box spreads to finance equities.
  • 5-10% cash for futures collateral.
Roth IRA (IBKR):
  • DJ600, ES/MES, and EMD for equities, padded with VB (Vanguard small cap ETF) and VWO (Vanguard emerging markets ETF) as necessary to bring total exposure to 140%.
  • ZF/ZN, padded with VGIT.
  • ~20-25% cash as collateral for futures.
401k (Fidelity brokeragelink):
  • TYA: 14% (~6 yr duration)
  • UPRO: 6%
  • VXUS: 10%
  • RSSB: 70% (50% whole world, 50% ITT (calculated as ~5.5 yr duration))
Overall Ratio: 98/112 ITT, with a weighted ER of 0.37% (or 0.18% when divided by leverage)

I have access to mega backdoor Roth and no solo 401k so this is the best I can do. I'm able to transfer my Roth funds to my IRA regularly so this account grows the slowest of them all. I might want to use some TUA to bring the bond duration exposure down a bit.

My attempt is to keep each of my accounts as close to my overall AA as possible, reducing the need to balance across accounts (and across tax-advantaged space). I haven't done it yet, but I'll likely modify my spreadsheet to increase leverage in my non-401k to make up for the slightly reduced leverage there, but considering it's my smallest account, it probably doesn't matter too much. I plan on rebalancing every quater as I roll futures and box spreads, as well as rebalancing with contributions. I will let my overall leverage float a bit, only deleveraging after hitting 325% overall leverage. I plan on deleveraging overtime as per the lifecycle book, ending up near 150% at retirement.

I'd appreciate any feedback anybody could offer.

A few questions:
  • My understanding is that the IBKR commodities account will pull cash from securities for commodities collateral (I have it set to sweep to securities). For example, 2 ZN + 2 ZF would require ~$11k in collateral, let's round up to $20k for a buffer. In my taxable account, instead of keeping $20k cash, I could buy $20k in T bills. That would net my cash out to zero. My commodities account would then pull $20k from securities margin, showing me a negative cash balance. But because I have $20k in t-bills, I wouldn't get charged interest because the t-bills would be the collateral. Is that correct?
  • I've seen it said here multiple times to "adjust for bond allocation". Does that mean that any bond durations should be normalized to the LTT equivilent? For example, if I decided I wanted the classic 60/40 portfolio with $1mm, instead of using 40% towards LTTs, I would calculate how many of the lower duration would be equivilent to the higher? So VGLT's (Vanguard LTT ETF) maturity is 22.7 yrs, and VGIT (Vanguard ITT ETF) is 5.6. So if I wanted to take advantage of the ITT efficiency over LTT, instead of buying $400k of VGLT, I would buy $1600k of VGIT (22.7/5.6)? And if I decided I wanted to target 2 year treasuries (VGSH [Vanguard STT ETF]), I would buy $4.5mm of VGSH to get 22.7 years of exposure but only using STT? (And 3.5 years would be somewhere in the middle)
comeinvest
Posts: 2728
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 »

diving6403 wrote: Sun Feb 25, 2024 10:13 pm I plan on rebalancing every quater as I roll futures and box spreads, as well as rebalancing with contributions. I will let my overall leverage float a bit, only deleveraging after hitting 325% overall leverage. I plan on deleveraging overtime as per the lifecycle book, ending up near 150% at retirement.
For starters. Mixing equities and bonds nominal exposure to arrive at an "overall leverage" number makes little to no sense to me, let alone using it as a discriminant for rebalancing decisions.
Setting that aside, especially for bonds I would define the asset allocation and do calculations in duration exposure per NAV instead of nominal bond values; or alternatively convert all nominal numbers to 5y or 25y duration or any other standardization of your choice.

Consider however that the same duration exposure in STT is somewhat more risky than ITT, which in turn is a bit more risky than LTT, if the duration risk per NAV is held constant. This was discussed towards the beginning of this thread. So to be consistent, I would settle on a duration risk allocation to a maturity bucket for each maturity range, and target that duration risk allocation per maturity bucket consistently over time.

Regarding most of your questions how to split among the treasuries futures, I think there was no definite conclusion in this thread. There was a discussion where skier favored ZF. I personally later revised my original views based on new insight, as I started thinking more in terms of exposure to forward rates instead of exposure to bond maturities, as bonds of different maturities are very correlated by nature (via common forward rates on the timeline). It's hard to come up with a rational justification for a split between the various STT and ITT maturities, as both of the extremes "STT only" (implemented with ZT or SOFR futures), and "ZN only" (covering exposure to the 0y-7y forward rates), could make a ton of sense, depending on how strongly you believe in STT outperformance.

There were 2 studies cited several times in this thread, both with backtests during about 20-25 years ending ca. 2013, that favored STT, and one explicitly recommends going with STT only, with no diversification benefit of adding ITT. Skier tested STT during the 1970ies rising rates environment in this or the HFEA thread, or was it the side discussion thread. If you want more definite answers or backtests over longer timeframes, you probably have to find more studies, or do studies yourself.
comeinvest
Posts: 2728
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 »

diving6403 wrote: Sun Feb 25, 2024 10:13 pm [*] DJ600, ES/MES, and EMD for equities, padded with VB (Vanguard small cap ETF) and VWO (Vanguard emerging markets ETF) as necessary to bring total exposure to 140%.
Don't neglect the N225M (Nikkei futures). It has consistently near zero implied financing cost, and the opportunity cost of the collateral in Yen is also still near zero, due to near zero interest rates.
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 »

diving6403 wrote: Thu Feb 22, 2024 11:03 pm I can't answer any of your questions but I'm looking forward to seeing other's answers. I have a question about this though. I'm still finishing wrapping my head around this thread and futures trading, but don't you need to have a positive cash balance as collateral for your futures? Are you being charged margin interest on that -$14k?

Logically, the outcome of what you're doing makes sense to me; it doesn't make sense to carry a positive cash balance while also carrying a box spread, but I thought it was required (hence the conversation about using t bills as collateral instead). But I'm not sure logistically how it works.
I don't have to have a positive cash balance in taxable, because I can use margin. Yes, I am charged interest on the -$14k. Ideally, I would have a balance of $0, but cash balance changes every day because of futures cash sweeps. When I do a box spread rebalance (likely 1x/year or more often if there have been significant swings), then I aim for a $0 balance.

I partially agree that carrying a positive cash balance doesn't make sense, but keep in mind:
1. You can't go to negative cash balance in IRA (at least, as far as I'm aware).
2. I am paid interest on the cash balance, although obviously there's a spread on that rate vs. box spreads/margin.
3. It's mostly for convenience, so I don't have to constantly rebalance to account for losses. My cash balance has more significant swings than my taxable, because I'm using both equity futures and treasury futures, not just treasury futures.

But like I said, I do think my cash balance is way too high.
comeinvest wrote: Fri Feb 23, 2024 12:07 am
I think you did a good job with your asset allocation. Only some minor thoughts:

I would think about lowering the equities percentage in the taxable account, because with 1:1.65 you will have to deleverage occasionally, and as a result incur tax drag.

You also have to reflect the higher market beta of VBR in your asset allocation. I think market beta instead of equities percentage is the better risk and allocation measure for the definition of a consistent asset allocation target of mHFEA. Whether the leverage (beta) with respect to the market index is internal, external (explicit), or intrinsic, makes no difference.

Similarly, duration exposure rather than the nominal bond percentage is the more consistent risk measure for fixed income.

With that in mind, I would think about getting rid of UPRO with its associated fees in the 401k, and instead move your VBR and other equities with >1 market beta to the 401k, while maintaining your overall targets across accounts. This way you could possibly create a similar risk exposure within and across accounts, while reducing or eliminating the fees. I'm allergic to fees, because they constitute return-free risk. It is questionable whether the benefit of leveraged fixed income with its small term premium will be bigger than the fee drag from LETFs, if you have both in your portfolio. (Money is fungible, so which asset class has the fee attached is not relevant, as long as any fee product and any low expected return [above the risk-free rate] asset class is in your portfolio, and the elimination of one could eliminate the other.)
Thanks for the thoughts - you're one of the kings of this thread, and I appreciate all the insights. To address your points:

On equities ratio: Understood. I already had to delever a bit over the past 2 years. However, I have a ton of realized losses ($150k or so, I think), plus I'll have the realized losses from box spreads. So I should be good for a while. I've also leaned toward having a maximum leverage ratio of 1:2 for equities - so I wouldn't rebalance unless it actually went above that ratio.

On beta and 401k construction: Good point, hadn't considered that since going down this journey. That said, VBR seems to have a beta of 1.02. 1% is a lot over a long period of time, but (1) at some point I'll leave my job, and at that point I'll roll over the Roth portion of my 401k to my IRA and follow the portfolio construction there, and (2) I believe in the strategy - it's not alpha, it's just leverage, and 1% is part of the cost of it.

On duration exposure: Yeah, I know that's the "right" way to do it. Downside is this thread is several books long and there's little consensus as to appropriate ratios. But a lot of folks fall back on ITTs with ZF being in the sweet spot. The SOFR discussion makes my head spin, and there's no real summary that I've seen on it. I understand it's not the ideal way to think about it, but it's the simplest to track. I'm open to suggestions here. I don't mind having to do more active management of my portfolio - it's worth the time.
diving6403
Posts: 12
Joined: Mon Jun 12, 2023 11:05 am

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

Post by diving6403 »

Thank you for the responses!
comeinvest wrote: Mon Feb 26, 2024 12:20 am
diving6403 wrote: Sun Feb 25, 2024 10:13 pm I plan on rebalancing every quater as I roll futures and box spreads, as well as rebalancing with contributions. I will let my overall leverage float a bit, only deleveraging after hitting 325% overall leverage. I plan on deleveraging overtime as per the lifecycle book, ending up near 150% at retirement.
For starters. Mixing equities and bonds nominal exposure to arrive at an "overall leverage" number makes little to no sense to me, let alone using it as a discriminant for rebalancing decisions.
Setting that aside, especially for bonds I would define the asset allocation and do calculations in duration exposure per NAV instead of nominal bond values; or alternatively convert all nominal numbers to 5y or 25y duration or any other standardization of your choice.

Consider however that the same duration exposure in STT is somewhat more risky than ITT, which in turn is a bit more risky than LTT, if the duration risk per NAV is held constant. This was discussed towards the beginning of this thread. So to be consistent, I would settle on a duration risk allocation to a maturity bucket for each maturity range, and target that duration risk allocation per maturity bucket consistently over time.
In case it wasn't clear by my questions, this is the part that makes the least sense to me as well. My intention is to come up with concrete rules that I can follow as my allocation shifts to avoid second guessing and timing the market. Though I think your point about not targeting leverage is that I should instead determine what exposure I want (to everything, bonds, intl, factors, etc) and find the right combination of investments that can get me that exposure. Because money is fungible, leverage isn't something I should think about with any particular investment, but something to keep in mind.

I want to understand your suggestion and concrete examples help me. I can think of two ways to calculate duration exposure to NAV and I'm not sure which is correct:
  • To find the exposure per NAV, I'm taking the market value of a contract (ZF: $106k, ZN: $110k), divided by the modified duration (ZF: 3.98, ZN: 5.98). This gets me ~$27k for ZF and ~$18k for ZN. I can then choose the right combination of those (and others like ZT and ZN) that gets me to my desired bond exposure
  • I can calculate the equivalent DV01's between different assets as in viewtopic.php?p=7223945#p7223945. I have set up my spreadsheet to do those calculations, though I'm not sure how to use them to target a bond duration. To target 5 years for example, would I weight the DV01's of ZF and ZN (which at a 50:50 ratio are roughly 5 years) which gives me a weighted DV01 of ~$52. I can then choose the right ratio of any of the futures that would also give me a DV01 of $52 (such as ZT/ZN; ZT/ZF/ZN; ZT/UB). I would then add those futures in that ratio until their market values is equivalent to my desired allocation. I know there were some conversations higher up about how to calculate the DV01, but for now I've just been using the CME analytics tool to get my numbers. I also know that ZT/UB is likely not a great target, it's just an example, but my understanding is that theoretically I can calculate the mathematical equivalent of a ZF/ZN combo with ZT/UB.
If I'm way off base or you have a concrete example for how I can go about calculating my duration exposure per NAV, I'd appreciate any help.

comeinvest wrote: Mon Feb 26, 2024 12:54 am Don't neglect the N225M (Nikkei futures). It has consistently near zero implied financing cost, and the opportunity cost of the collateral in Yen is also still near zero, due to near zero interest rates.
I'm only neglecting them while I make sure i have the foundations down. I have N225M, MXEA, and MXEF as other international futures I want to do more research on, as well as figuring out how I want to calculate the right ratio between all of them.
DMoogle wrote: Mon Feb 26, 2024 9:30 am On duration exposure: Yeah, I know that's the "right" way to do it. Downside is this thread is several books long and there's little consensus as to appropriate ratios. But a lot of folks fall back on ITTs with ZF being in the sweet spot. The SOFR discussion makes my head spin, and there's no real summary that I've seen on it. I understand it's not the ideal way to think about it, but it's the simplest to track. I'm open to suggestions here. I don't mind having to do more active management of my portfolio - it's worth the time.
I agree with everything DMoogle said here.
comeinvest
Posts: 2728
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 »

diving6403 wrote: Mon Feb 26, 2024 9:36 am Thank you for the responses!
comeinvest wrote: Mon Feb 26, 2024 12:20 am
diving6403 wrote: Sun Feb 25, 2024 10:13 pm I plan on rebalancing every quater as I roll futures and box spreads, as well as rebalancing with contributions. I will let my overall leverage float a bit, only deleveraging after hitting 325% overall leverage. I plan on deleveraging overtime as per the lifecycle book, ending up near 150% at retirement.
For starters. Mixing equities and bonds nominal exposure to arrive at an "overall leverage" number makes little to no sense to me, let alone using it as a discriminant for rebalancing decisions.
Setting that aside, especially for bonds I would define the asset allocation and do calculations in duration exposure per NAV instead of nominal bond values; or alternatively convert all nominal numbers to 5y or 25y duration or any other standardization of your choice.

Consider however that the same duration exposure in STT is somewhat more risky than ITT, which in turn is a bit more risky than LTT, if the duration risk per NAV is held constant. This was discussed towards the beginning of this thread. So to be consistent, I would settle on a duration risk allocation to a maturity bucket for each maturity range, and target that duration risk allocation per maturity bucket consistently over time.
In case it wasn't clear by my questions, this is the part that makes the least sense to me as well. My intention is to come up with concrete rules that I can follow as my allocation shifts to avoid second guessing and timing the market. Though I think your point about not targeting leverage is that I should instead determine what exposure I want (to everything, bonds, intl, factors, etc) and find the right combination of investments that can get me that exposure. Because money is fungible, leverage isn't something I should think about with any particular investment, but something to keep in mind.

I want to understand your suggestion and concrete examples help me. I can think of two ways to calculate duration exposure to NAV and I'm not sure which is correct:
  • To find the exposure per NAV, I'm taking the market value of a contract (ZF: $106k, ZN: $110k), divided by the modified duration (ZF: 3.98, ZN: 5.98). This gets me ~$27k for ZF and ~$18k for ZN. I can then choose the right combination of those (and others like ZT and ZN) that gets me to my desired bond exposure
  • I can calculate the equivalent DV01's between different assets as in viewtopic.php?p=7223945#p7223945. I have set up my spreadsheet to do those calculations, though I'm not sure how to use them to target a bond duration. To target 5 years for example, would I weight the DV01's of ZF and ZN (which at a 50:50 ratio are roughly 5 years) which gives me a weighted DV01 of ~$52. I can then choose the right ratio of any of the futures that would also give me a DV01 of $52 (such as ZT/ZN; ZT/ZF/ZN; ZT/UB). I would then add those futures in that ratio until their market values is equivalent to my desired allocation. I know there were some conversations higher up about how to calculate the DV01, but for now I've just been using the CME analytics tool to get my numbers. I also know that ZT/UB is likely not a great target, it's just an example, but my understanding is that theoretically I can calculate the mathematical equivalent of a ZF/ZN combo with ZT/UB.
If I'm way off base or you have a concrete example for how I can go about calculating my duration exposure per NAV, I'd appreciate any help.
I think you are on track, but I think you need to multiply and not divide by the modified duration, if you want to use duration and not DV01 for your calculations.

Regarding your last bullet point. I think you are on track, but your thinking might still be a bit convoluted. I'm not sure what "target 5 years" means, and what you are trying to achieve with your ratios.

Like I said, I personally would do these steps:
1. Determine target duration risk exposure, expressed in 5y equivalents, 25y equivalents, 1-year equivalents, or "portfolio duration". The portfolio duration (duration risk exposure) is what mostly determines the performance and risk of the fixed income part of your portfolio.
2. Determine how you want to allocate your duration exposure from (1.) across different treasury maturity buckets: STT, ITT, ZT, ZF, ZN, etc. While doing this, adjust for the different risk levels, if you want. (Shorter maturities somewhat higher risk, but also higher Sharpe ratio if duration exposure held constant.) This is what, secondary to the total duration risk exposure per (1.), will also determine but to a lesser extent, the performance and risk of the fixed income part of your portfolio. (1.) is like saying "I want to eat fruits". (2.) is like saying "I prefer apples over oranges".
(1.) and (2.) is your asset allocation plan, and static (except for lifetime glide path of applicable).
3. At time of rebalancing, convert your target for each bucket into a dollar duration, or DV01, based on the current NAV.
4. Determine the number of contracts (SOFR, ZT, ZF, etc.) and/or ETF shares for each bucket, based on (3.) and the published DV01 for each contract. (Alternatively you can use modified durations for your math.)
5. Round up or down in each bucket, if needed, without distorting too much your overall duration risk target.
I'm not a licensed mHFEA instructor, so please critique.
(1.) and (2.) of course are the decisions for which I think there is no conclusive consensus in this thread, so you have to make up your own mind, or do more research.
I personally also use tax-exempt bonds in my taxable account, and I have other stuff like preferreds, equity options, and index options in some of my accounts. But the principle of keeping the main risk exposure characteristics of the overall portfolio (stock market beta, options delta, portfolio duration, etc.) on target is the same.
diving6403
Posts: 12
Joined: Mon Jun 12, 2023 11:05 am

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

Post by diving6403 »

comeinvest wrote: Mon Feb 26, 2024 3:44 pm
diving6403 wrote: Mon Feb 26, 2024 9:36 am ...
I think you are on track, but I think you need to multiply and not divide by the modified duration, if you want to use duration and not DV01 for your calculations.

Regarding your last bullet point. I think you are on track, but your thinking might still be a bit convoluted. I'm not sure what "target 5 years" means, and what you are trying to achieve with your ratios.

Like I said, I personally would do these steps:
1. Determine target duration risk exposure, expressed in 5y equivalents, 25y equivalents, 1-year equivalents, or "portfolio duration". The portfolio duration (duration risk exposure) is what mostly determines the performance and risk of the fixed income part of your portfolio.
2. Determine how you want to allocate your duration exposure from (1.) across different treasury maturity buckets: STT, ITT, ZT, ZF, ZN, etc. While doing this, adjust for the different risk levels, if you want. (Shorter maturities somewhat higher risk, but also higher Sharpe ratio if duration exposure held constant.) This is what, secondary to the total duration risk exposure per (1.), will also determine but to a lesser extent, the performance and risk of the fixed income part of your portfolio. (1.) is like saying "I want to eat fruits". (2.) is like saying "I prefer apples over oranges".
(1.) and (2.) is your asset allocation plan, and static (except for lifetime glide path of applicable).
3. At time of rebalancing, convert your target for each bucket into a dollar duration, or DV01, based on the current NAV.
4. Determine the number of contracts (SOFR, ZT, ZF, etc.) and/or ETF shares for each bucket, based on (3.) and the published DV01 for each contract. (Alternatively you can use modified durations for your math.)
5. Round up or down in each bucket, if needed, without distorting too much your overall duration risk target.
I'm not a licensed mHFEA instructor, so please critique.
(1.) and (2.) of course are the decisions for which I think there is no conclusive consensus in this thread, so you have to make up your own mind, or do more research.
I personally also use tax-exempt bonds in my taxable account, and I have other stuff like preferreds, equity options, and index options in some of my accounts. But the principle of keeping the main risk exposure characteristics of the overall portfolio (stock market beta, options delta, portfolio duration, etc.) on target is the same.
Thank you, that's helpful and I think what you've been saying about targeting durations not leverage % has finally clicked (though I'm sure you'll tell me something I've missed). My "target 5 years" doesn't make sense, I was thinking about it the wrong way. What I meant was a 50:50 weighted ratio of the durations of ZF and ZN is 5 years, then I'd leverage that up to 160%. Conceptually I understand why that was the wrong way to think about it, even if the final numbers ended up close to what I'd want.

Image

This is the spreadsheet I came up with, trying to follow along with your steps on the left.

1. I determine I want 20 years of bond exposure
2.a. Using a weird mix of futures and VGIT. (It probably doesn't make sense as an actual target, but being able to play with the numbers helped me understand)
2.b. I decide what percentage of my duration exposure I want to come from each bucket.
3. Given a $1mm portfolio, 40% of which should be 20 years of bond exposure, I calculate the bond exposure per bucket using (EE)
4. I calculate the shares required to reach that exposure
5. I round to the nearest whole share.

FF-II are simply there to help me double check that the math makes sense. This doesn't include any advanced rounding, but that's easy to add. This was all done using durations. From this post I mentioned above I know how to relate DV01's to each other, but only if I have an "anchor", such as UB. I think for your (1.), I could instead of using year equivalents, I could use "UB-equivilents", and then do all my math based off the UB DV01. I'm not sure if it would make a meaningful difference, though I'd appreciate any advice on that.

I do appreciate your help, especially considering you aren't an mHFEA instructor. Or maybe I'm still not getting it and I should wait to provide my thanks until after I do :happy
director84
Posts: 50
Joined: Thu Dec 23, 2010 9:59 am

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

Post by director84 »

diving6403 wrote: Mon Feb 26, 2024 11:13 pm
comeinvest wrote: Mon Feb 26, 2024 3:44 pm
diving6403 wrote: Mon Feb 26, 2024 9:36 am ...
I think you are on track, but I think you need to multiply and not divide by the modified duration, if you want to use duration and not DV01 for your calculations.

Regarding your last bullet point. I think you are on track, but your thinking might still be a bit convoluted. I'm not sure what "target 5 years" means, and what you are trying to achieve with your ratios.

Like I said, I personally would do these steps:
1. Determine target duration risk exposure, expressed in 5y equivalents, 25y equivalents, 1-year equivalents, or "portfolio duration". The portfolio duration (duration risk exposure) is what mostly determines the performance and risk of the fixed income part of your portfolio.
2. Determine how you want to allocate your duration exposure from (1.) across different treasury maturity buckets: STT, ITT, ZT, ZF, ZN, etc. While doing this, adjust for the different risk levels, if you want. (Shorter maturities somewhat higher risk, but also higher Sharpe ratio if duration exposure held constant.) This is what, secondary to the total duration risk exposure per (1.), will also determine but to a lesser extent, the performance and risk of the fixed income part of your portfolio. (1.) is like saying "I want to eat fruits". (2.) is like saying "I prefer apples over oranges".
(1.) and (2.) is your asset allocation plan, and static (except for lifetime glide path of applicable).
3. At time of rebalancing, convert your target for each bucket into a dollar duration, or DV01, based on the current NAV.
4. Determine the number of contracts (SOFR, ZT, ZF, etc.) and/or ETF shares for each bucket, based on (3.) and the published DV01 for each contract. (Alternatively you can use modified durations for your math.)
5. Round up or down in each bucket, if needed, without distorting too much your overall duration risk target.
I'm not a licensed mHFEA instructor, so please critique.
(1.) and (2.) of course are the decisions for which I think there is no conclusive consensus in this thread, so you have to make up your own mind, or do more research.
I personally also use tax-exempt bonds in my taxable account, and I have other stuff like preferreds, equity options, and index options in some of my accounts. But the principle of keeping the main risk exposure characteristics of the overall portfolio (stock market beta, options delta, portfolio duration, etc.) on target is the same.
Thank you, that's helpful and I think what you've been saying about targeting durations not leverage % has finally clicked (though I'm sure you'll tell me something I've missed). My "target 5 years" doesn't make sense, I was thinking about it the wrong way. What I meant was a 50:50 weighted ratio of the durations of ZF and ZN is 5 years, then I'd leverage that up to 160%. Conceptually I understand why that was the wrong way to think about it, even if the final numbers ended up close to what I'd want.

Image

This is the spreadsheet I came up with, trying to follow along with your steps on the left.

1. I determine I want 20 years of bond exposure
2.a. Using a weird mix of futures and VGIT. (It probably doesn't make sense as an actual target, but being able to play with the numbers helped me understand)
2.b. I decide what percentage of my duration exposure I want to come from each bucket.
3. Given a $1mm portfolio, 40% of which should be 20 years of bond exposure, I calculate the bond exposure per bucket using (EE)
4. I calculate the shares required to reach that exposure
5. I round to the nearest whole share.

FF-II are simply there to help me double check that the math makes sense. This doesn't include any advanced rounding, but that's easy to add. This was all done using durations. From this post I mentioned above I know how to relate DV01's to each other, but only if I have an "anchor", such as UB. I think for your (1.), I could instead of using year equivalents, I could use "UB-equivilents", and then do all my math based off the UB DV01. I'm not sure if it would make a meaningful difference, though I'd appreciate any advice on that.

I do appreciate your help, especially considering you aren't an mHFEA instructor. Or maybe I'm still not getting it and I should wait to provide my thanks until after I do :happy
I started on this journey last year and had a lot of the same questions you do (and still have some of them). I'll try to provide some commentary about how I think about this.

For starters, I'm not quite following your calculations. The way I think about it, is if I have a $1M portfolio and want to target 1.75x allocation to bonds at 5-year duration with, for example, equal duration adjusted allocation among ZT/ZF/ZN (33% each) then I want $583k (1.75 * $1M / 3) duration equivalent. In terms of contracts, that would be:
ZT: $203,632 * 1.8983 / 5 = $77,095 per contract at 5-years; $583k / $77k = 7.5 ZT contracts
Using the same math, would get you 6.9 ZF and 4.5 ZN contracts (and obviously round to whole number of contracts).

Some other general comments from my experience:
- On a $1M portfolio, you don't need VGIT. You can get close enough to your target allocation with just futures and save the cash for your equity portion.
- I only use ZT/ZF/ZN, no UB. I think it was shown earlier in this thread, and comeinvest alluded to, that longer duration treasuries are not optimal.
- In your spreadsheet, I don't think rows BB and CC are necessary. You just need a target allocation for bonds (175% in this example).
- The numbers you have are close enough, but if you want the latest duration and DV01 values, you can get them from the CME Treasury Analytics.

That's all I have off the top of my head. I'll try to respond to some of your earlier comments upthread when I have more time.
diving6403
Posts: 12
Joined: Mon Jun 12, 2023 11:05 am

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

Post by diving6403 »

director84 wrote: Wed Feb 28, 2024 9:52 am
diving6403 wrote: Mon Feb 26, 2024 11:13 pm ...
I started on this journey last year and had a lot of the same questions you do (and still have some of them). I'll try to provide some commentary about how I think about this.

For starters, I'm not quite following your calculations. The way I think about it, is if I have a $1M portfolio and want to target 1.75x allocation to bonds at 5-year duration with, for example, equal duration adjusted allocation among ZT/ZF/ZN (33% each) then I want $583k (1.75 * $1M / 3) duration equivalent. In terms of contracts, that would be:
ZT: $203,632 * 1.8983 / 5 = $77,095 per contract at 5-years; $583k / $77k = 7.5 ZT contracts
Using the same math, would get you 6.9 ZF and 4.5 ZN contracts (and obviously round to whole number of contracts).

Some other general comments from my experience:
- On a $1M portfolio, you don't need VGIT. You can get close enough to your target allocation with just futures and save the cash for your equity portion.
- I only use ZT/ZF/ZN, no UB. I think it was shown earlier in this thread, and comeinvest alluded to, that longer duration treasuries are not optimal.
- In your spreadsheet, I don't think rows BB and CC are necessary. You just need a target allocation for bonds (175% in this example).
- The numbers you have are close enough, but if you want the latest duration and DV01 values, you can get them from the CME Treasury Analytics.

That's all I have off the top of my head. I'll try to respond to some of your earlier comments upthread when I have more time.
Thank you, that's helpful! I think I basically end up with the same rough math as you, I just go through a lot more steps. How you described it here is how I was thinking about it before, but I wasn't doing the math correctly. This recent spreadsheet gets the same basic math correct, but it's very convoluted compared to what you described. I'm not sure if I overanalyzed comeinvest's comments.

Responses to your comments:
- VGIT & UB: Agreed, on this portfolio I wouldn't use them. They're only in the spreadsheet to verify that the math works out and makes sense to me, not a proposal for what I'd actually do.
- BB and CC: Correct, these are unnecessary when doing math the way you're suggesting.
- CME numbers: I use sellenium in Python to pull the numbers directly from the CME Analytics site, so they were accurate at the time I took the screenshot :) I'm using the "Modified Duration" and "Futures DV01" from the Deliverables; I'm not sure if those need to be modified at all or can be used as is.
comeinvest
Posts: 2728
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 »

diving6403 wrote: Wed Feb 28, 2024 10:27 am - CME numbers: I use sellenium in Python to pull the numbers directly from the CME Analytics site, so they were accurate at the time I took the screenshot :) I'm using the "Modified Duration" and "Futures DV01" from the Deliverables; I'm not sure if those need to be modified at all or can be used as is.
I think the "futures DV01" are generally the relevant numbers for the duration exposure when holding a futures contract, although how exactly they are calculated is not 100% clear. There was a discussion a bit earlier in this thread between physicist and myself, where we tried to replicate the modified duration and DV01 numbers on the CME analytics site.
I can't remember if the "modified duration" is the duration of the CTD or that of the futures contract; I would have to go back up and re-read the discussion. The difference is not that big and you may not bother if it's for your mHFEA asset allocation, if you are not genuinely interested in your mathematically correct exposure.
diving6403
Posts: 12
Joined: Mon Jun 12, 2023 11:05 am

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

Post by diving6403 »

diving6403 wrote: Wed Feb 28, 2024 10:27 am
director84 wrote: Wed Feb 28, 2024 9:52 am
diving6403 wrote: Mon Feb 26, 2024 11:13 pm ...
For starters, I'm not quite following your calculations.
...
...
I think I basically end up with the same rough math as you, I just go through a lot more steps.
...
I updated my spreadsheet and ended up with the exact same math as you, not just roughly the same. director, your method just involves way fewer steps. So thanks for helping me simplify!
comeinvest wrote: Wed Feb 28, 2024 12:02 pm
diving6403 wrote: Wed Feb 28, 2024 10:27 am - CME numbers: I use sellenium in Python to pull the numbers directly from the CME Analytics site, so they were accurate at the time I took the screenshot :) I'm using the "Modified Duration" and "Futures DV01" from the Deliverables; I'm not sure if those need to be modified at all or can be used as is.
I think the "futures DV01" are generally the relevant numbers for the duration exposure when holding a futures contract, although how exactly they are calculated is not 100% clear. There was a discussion a bit earlier in this thread between physicist and myself, where we tried to replicate the modified duration and DV01 numbers on the CME analytics site.
I can't remember if the "modified duration" is the duration of the CTD of that of the futures contract; I would have to go back up and re-read the discussion. The difference is not that big and you may not bother if it's for your mHFEA asset allocation, if you are not genuinely interested in your mathematically correct exposure.
I am interested from an academic standpoint now that I've done so much reading about futures and bonds and durations, but I wouldn't expect it to make a difference in the end consider the other variability in the portfolio, such as rounding futures to whole numbers.
comeinvest
Posts: 2728
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 »

director84 wrote: Wed Feb 28, 2024 9:52 am For starters, I'm not quite following your calculations. The way I think about it, is if I have a $1M portfolio and want to target 1.75x allocation to bonds at 5-year duration with, for example, equal duration adjusted allocation among ZT/ZF/ZN (33% each) then I want $583k (1.75 * $1M / 3) duration equivalent. In terms of contracts, that would be:
ZT: $203,632 * 1.8983 / 5 = $77,095 per contract at 5-years; $583k / $77k = 7.5 ZT contracts
Using the same math, would get you 6.9 ZF and 4.5 ZN contracts (and obviously round to whole number of contracts).
That is one correct way of doing the math. Or you could just calculate the desired/required DV01 in each bucket per your asset allocation (in your case by dividing your total desired portfolio DV01 exposure by 3), and then divide by the published DV01 of the respective contract, to arrive at the number of contracts to buy.

Like I said a few times before, be aware that using the current duration numbers of 1.9 (ZT), 4.0 (ZF), 5.9 (ZN), ca. 1.9/4.0 = 0.475 (ca. 50%) of the forward rate exposure of ZF contracts is identical to the forward rate exposure of ZT contracts, and ca. 1.9/5.9 = 0.32 (ca. 30%) of the forward rate exposure of ZN is identical to the forward rate exposure of ZT on the forward rate timeline. By "identical" I am not referring to the magnitude of duration risk, but to the "qualitative" exposure on the rates timeline between the present and some time in the future, i.e. the exposure to the segments on the timeline that the various contracts provide exposure to, and their overlap.
That means your "effective" exposure to the forward rate time segment covered by ZT is ca. 0.33 + 0.33 * 0.475 + 0.33 * 0.32 -> 0.59 (59%). There is nothing wrong with that; I'm just pointing out that the futures contracts that you are using are not independent variables, i.e. your total exposure to the market's 0-2 years biased forward rate estimates and their term premia is 59% of your total duration risk exposure, or ca. 30% per year on the timeline, if you decide on an equal duration adjusted split among the three contracts. By implication, the totality of your exposure to the market's 2-7 year forward rate estimates and their term premia would be only ca. 41% of your total duration risk exposure, which is ca. 8% per year on the timeline. So you are giving ca. 4 times the weight (per year) to the short forward rate term premia that you give (per year) to the intermediate forward rate term premia.
By holding fixed income futures contracts, you are making bets on the forward rates, hoping they are over-estimated by the market, at various points on the forward curve, trying to harvest roll-down returns.
If you lump years 1-2 and years 3-7 together respectively in those two buckets, perhaps representing "short-term" and "intermediate-term" forward rates depending on the interpretation of those terms, then the weights of the buckets would be 59% and 41%, but the second bucket represents 5 years vs 2 years for the first. I'm not sure which distribution is the most sensical, nor if there is any diversification benefit; I guess only a numerical risk and performance analysis will give a rational answer.
comeinvest
Posts: 2728
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 »

This time I looked at the treasury futures roll again. I think theoretically the curves of the near expiration 3-month SOFR futures and the treasury futures rolls should have an inverse appearance, because higher 3-month financing cost should raise the futures roll price and vice versa, all other things being equal.
I can see that for the most part when I look at the Feb 22 vs Feb 26 level.
But the 3-week movements don't make sense to me at all, especially the movement between Feb 12 and Feb 13 that was rather consistent across all treasury futures roll prices. I would expect the roll prices to jump upward not downward, when the SOFR futures price jumped down (i.e. the expected overnight SOFR for the next 3 months, and by implication the implied financing rate of all futures products based on an underlying index or security, should have gone up).
I rolled all my treasury futures on Feb 26, using limit orders at the prevailing bid (buy) and ask (sell). My limit orders often get gradually consumed over time, due to the CBOT order matching priority logic.
The UB had higher fluctuations than then other futures after Feb 22.
The charts show only the synthetic and not the native calendar spread rolls with the tighter bid/ask spreads, because IB charts can't re-generate historical native futures spread charts.

Image
Image
Image
Image
Image
comeinvest
Posts: 2728
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 »

For whatever reason, the ZB and UB "March first day" and the UB "June last day" CTDs switched to the far end of the maturity range. Big differences between first day vs. last day DV01 numbers for March, even for the ZN (10y) contract. I'm wondering what that means for the "actual" DV01 exposure.

Image

Image

Image

Image

Image
comeinvest
Posts: 2728
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 »

Update: Inspecting stock/bond correlation on various scales

Image

Image

Image

Image
parval
Posts: 153
Joined: Tue Oct 22, 2019 9:23 pm

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

Post by parval »

Folks PLEASE sanity check me on this:

I think we can get cheaper financing than box spreads - Sell deep ITM leaps (puts)

This assumes portfolio margin, so the leveraged portion just dips into buying power instead of paying margin fees.

For example DEC2024 600P is basically all intrinsic:

Image

Cash from short put goes into STT.

This seems like near free leverage? Full exposure to the index, slight risk if we close above 600 but we can just roll up. We're selling on SPX so minimal fees and 60/40 tax treatment. Thoughts?
comeinvest
Posts: 2728
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 »

parval wrote: Sun Mar 03, 2024 8:35 am Folks PLEASE sanity check me on this:

I think we can get cheaper financing than box spreads - Sell deep ITM leaps (puts)

This assumes portfolio margin, so the leveraged portion just dips into buying power instead of paying margin fees.

For example DEC2024 600P is basically all intrinsic:

Cash from short put goes into STT.

This seems like near free leverage? Full exposure to the index, slight risk if we close above 600 but we can just roll up. We're selling on SPX so minimal fees and 60/40 tax treatment. Thoughts?
You are showing SPY options which are American exercise style options, and not SPX options. It may not matter for your argument.
Writing cash-secured puts is a well known strategy, with better risk-adjusted returns than the index in the past. There are many papers about this, including papers from CBOE, and there is even a "put write" index that you can use to examine historical performance. (The put write index uses more ATM options though; but the principle is the same.)
I'm not sure if you can call it "near free leverage", and how you determine "near free". The call options corresponding to your put options at strike price 6000 have a positive price (not zero); so you have to consider volatility, even if the likelihood of not being assigned (i.e. the index at expiration above 6000) is a relatively rare event; making "rare" synonymous with "zero" often works in the short run, but won't work in the long run; or else you could also for example sell the 6000 call options for "free cash" (except that it is not free).
There is a theory that selling volatility provides another independent source of return, i.e. that the market perpetually overestimates volatility, i.e. that implied volatility is, on average, higher than realized volatility. There is another theory that the risk-adjusted excess return of selling options during the last ca. 2 decades may have been the result of an anomaly that efficient markets may have "fixed" for the future; or perhaps that rare catastrophic events don't show in limited backtesting. Do some more reading. I still wanted to investigate if volatility can indeed be considered an independent source of return. If true, it might be a good addition, or variation, of mHFEA; for example look at the CBOE SPX put write index, leverage it to the same risk as plain SPX, and replace your SPX exposure in mHFEA with the put write exposure.
parval
Posts: 153
Joined: Tue Oct 22, 2019 9:23 pm

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

Post by parval »

comeinvest wrote: Sun Mar 03, 2024 2:58 pm
parval wrote: Sun Mar 03, 2024 8:35 am Folks PLEASE sanity check me on this:

I think we can get cheaper financing than box spreads - Sell deep ITM leaps (puts)

This assumes portfolio margin, so the leveraged portion just dips into buying power instead of paying margin fees.

For example DEC2024 600P is basically all intrinsic:

Cash from short put goes into STT.

This seems like near free leverage? Full exposure to the index, slight risk if we close above 600 but we can just roll up. We're selling on SPX so minimal fees and 60/40 tax treatment. Thoughts?
You are showing SPY options which are American exercise style options, and not SPX options. It may not matter for your argument.
Writing cash-secured puts is a well known strategy, with better risk-adjusted returns than the index in the past. There are many papers about this, including papers from CBOE, and there is even a "put write" index that you can use to examine historical performance. (The put write index uses more ATM options though; but the principle is the same.)
I'm not sure if you can call it "near free leverage", and how you determine "near free". The call options corresponding to your put options at strike price 6000 have a positive price (not zero); so you have to consider volatility, even if the likelihood of not being assigned (i.e. the index at expiration above 6000) is a relatively rare event; making "rare" synonymous with "zero" often works in the short run, but won't work in the long run; or else you could also for example sell the 6000 call options for "free cash" (except that it is not free).
There is a theory that selling volatility provides another independent source of return, i.e. that the market perpetually overestimates volatility, i.e. that implied volatility is, on average, higher than realized volatility. There is another theory that the risk-adjusted excess return of selling options during the last ca. 2 decades may have been the result of an anomaly that efficient markets may have "fixed" for the future; or perhaps that rare catastrophic events don't show in limited backtesting. Do some more reading. I still wanted to investigate if volatility can indeed be considered an independent source of return. If true, it might be a good addition, or variation, of mHFEA; for example look at the CBOE SPX put write index, leverage it to the same risk as plain SPX, and replace your SPX exposure in mHFEA with the put write exposure.
Oh I'm using SPY price instead of SPX cause market's not open and SPX is too wide. I'm assuming it should be about the same but could be wrong. Will check Monday morning.

Just to re-iterate:
- Sell SPX put for exposure
- No assignment risk
- Not looking for uncorrelated return, actually want 1 delta
- Not harvesting VRP, way outside of this thread
- Goal is to lower cost of leverage

It's just a replacement for box spread to lever up the equity portion of this strategy, so instead of borrowing 600k @ 5% for VTI. The short put gives same exposure up to 600k, if delta gets <.9 can roll up.
comeinvest
Posts: 2728
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 »

parval wrote: Sun Mar 03, 2024 5:45 pm
comeinvest wrote: Sun Mar 03, 2024 2:58 pm
parval wrote: Sun Mar 03, 2024 8:35 am Folks PLEASE sanity check me on this:

I think we can get cheaper financing than box spreads - Sell deep ITM leaps (puts)

This assumes portfolio margin, so the leveraged portion just dips into buying power instead of paying margin fees.

For example DEC2024 600P is basically all intrinsic:

Cash from short put goes into STT.

This seems like near free leverage? Full exposure to the index, slight risk if we close above 600 but we can just roll up. We're selling on SPX so minimal fees and 60/40 tax treatment. Thoughts?
You are showing SPY options which are American exercise style options, and not SPX options. It may not matter for your argument.
Writing cash-secured puts is a well known strategy, with better risk-adjusted returns than the index in the past. There are many papers about this, including papers from CBOE, and there is even a "put write" index that you can use to examine historical performance. (The put write index uses more ATM options though; but the principle is the same.)
I'm not sure if you can call it "near free leverage", and how you determine "near free". The call options corresponding to your put options at strike price 6000 have a positive price (not zero); so you have to consider volatility, even if the likelihood of not being assigned (i.e. the index at expiration above 6000) is a relatively rare event; making "rare" synonymous with "zero" often works in the short run, but won't work in the long run; or else you could also for example sell the 6000 call options for "free cash" (except that it is not free).
There is a theory that selling volatility provides another independent source of return, i.e. that the market perpetually overestimates volatility, i.e. that implied volatility is, on average, higher than realized volatility. There is another theory that the risk-adjusted excess return of selling options during the last ca. 2 decades may have been the result of an anomaly that efficient markets may have "fixed" for the future; or perhaps that rare catastrophic events don't show in limited backtesting. Do some more reading. I still wanted to investigate if volatility can indeed be considered an independent source of return. If true, it might be a good addition, or variation, of mHFEA; for example look at the CBOE SPX put write index, leverage it to the same risk as plain SPX, and replace your SPX exposure in mHFEA with the put write exposure.
Oh I'm using SPY price instead of SPX cause market's not open and SPX is too wide. I'm assuming it should be about the same but could be wrong. Will check Monday morning.

Just to re-iterate:
- Sell SPX put for exposure
- No assignment risk
- Not looking for uncorrelated return, actually want 1 delta
- Not harvesting VRP, way outside of this thread
- Goal is to lower cost of leverage

It's just a replacement for box spread to lever up the equity portion of this strategy, so instead of borrowing 600k @ 5% for VTI. The short put gives same exposure up to 600k, if delta gets <.9 can roll up.
Like I said, you are still selling some volatility risk if you sell 6000 strike prices. You are basically buying a synthetic long position, but omitting the long call position. Show us the math; you will probably end up with an implied financing rate very similar to box spreads or futures; except the math will be difficult and model-dependent, because you are also selling a bit volatility. You would probably also have higher margin requirements than with a box spread. Lastly, you would not defer capital gains taxes, which you defer with box spreads and VTI or individual equities.
I successfully sold 6000-8000 March 2024 expiration bear put credit spreads in lieu of box spreads, and I got the same or a tiny bit lower rate than with box spreads, and I saved a few dollars in commissions. (Hard to explain why I got a slightly better rate; it should have been slightly worse because I bought net volatility risk premium, i.e. I would have hit the jackpot if the index had moved up to between 6000 and 8000.) But since the index moved higher, I guess this is no longer possible, at least not for the June expiration. Highest June expiration strike price is 6600; September is 8200; December is 9200; March 2025 is 8000; June 2025 is 7600. I don't know who or what determines the range of strike prices, or if CBOE can add additional strike prices progressively. I prefer selling 2-legged credit spreads instead of 4-legged box spreads.
parval
Posts: 153
Joined: Tue Oct 22, 2019 9:23 pm

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

Post by parval »

comeinvest wrote: Sun Mar 03, 2024 8:24 pm
parval wrote: Sun Mar 03, 2024 5:45 pm
comeinvest wrote: Sun Mar 03, 2024 2:58 pm
parval wrote: Sun Mar 03, 2024 8:35 am Folks PLEASE sanity check me on this:

I think we can get cheaper financing than box spreads - Sell deep ITM leaps (puts)

This assumes portfolio margin, so the leveraged portion just dips into buying power instead of paying margin fees.

For example DEC2024 600P is basically all intrinsic:

Cash from short put goes into STT.

This seems like near free leverage? Full exposure to the index, slight risk if we close above 600 but we can just roll up. We're selling on SPX so minimal fees and 60/40 tax treatment. Thoughts?
You are showing SPY options which are American exercise style options, and not SPX options. It may not matter for your argument.
Writing cash-secured puts is a well known strategy, with better risk-adjusted returns than the index in the past. There are many papers about this, including papers from CBOE, and there is even a "put write" index that you can use to examine historical performance. (The put write index uses more ATM options though; but the principle is the same.)
I'm not sure if you can call it "near free leverage", and how you determine "near free". The call options corresponding to your put options at strike price 6000 have a positive price (not zero); so you have to consider volatility, even if the likelihood of not being assigned (i.e. the index at expiration above 6000) is a relatively rare event; making "rare" synonymous with "zero" often works in the short run, but won't work in the long run; or else you could also for example sell the 6000 call options for "free cash" (except that it is not free).
There is a theory that selling volatility provides another independent source of return, i.e. that the market perpetually overestimates volatility, i.e. that implied volatility is, on average, higher than realized volatility. There is another theory that the risk-adjusted excess return of selling options during the last ca. 2 decades may have been the result of an anomaly that efficient markets may have "fixed" for the future; or perhaps that rare catastrophic events don't show in limited backtesting. Do some more reading. I still wanted to investigate if volatility can indeed be considered an independent source of return. If true, it might be a good addition, or variation, of mHFEA; for example look at the CBOE SPX put write index, leverage it to the same risk as plain SPX, and replace your SPX exposure in mHFEA with the put write exposure.
Oh I'm using SPY price instead of SPX cause market's not open and SPX is too wide. I'm assuming it should be about the same but could be wrong. Will check Monday morning.

Just to re-iterate:
- Sell SPX put for exposure
- No assignment risk
- Not looking for uncorrelated return, actually want 1 delta
- Not harvesting VRP, way outside of this thread
- Goal is to lower cost of leverage

It's just a replacement for box spread to lever up the equity portion of this strategy, so instead of borrowing 600k @ 5% for VTI. The short put gives same exposure up to 600k, if delta gets <.9 can roll up.
Like I said, you are still selling some volatility risk if you sell 6000 strike prices. You are basically buying a synthetic long position, but omitting the long call position. Show us the math; you will probably end up with an implied financing rate very similar to box spreads or futures; except the math will be difficult and model-dependent, because you are also selling a bit volatility. You would probably also have higher margin requirements than with a box spread. Lastly, you would not defer capital gains taxes, which you defer with box spreads and VTI or individual equities.
I successfully sold 6000-8000 March 2024 expiration bear put credit spreads in lieu of box spreads, and I got the same or a tiny bit lower rate than with box spreads, and I saved a few dollars in commissions. (Hard to explain why I got a slightly better rate; it should have been slightly worse because I bought net volatility risk premium, i.e. I would have hit the jackpot if the index had moved up to between 6000 and 8000.) But since the index moved higher, I guess this is no longer possible, at least not for the June expiration. Highest June expiration strike price is 6600; September is 8200; December is 9200; March 2025 is 8000; June 2025 is 7600. I don't know who or what determines the range of strike prices, or if CBOE can add additional strike prices progressively. I prefer selling 2-legged credit spreads instead of 4-legged box spreads.
Okay I did the math!

Image

Spreadsheet link: https://docs.google.com/spreadsheets/d/ ... sp=sharing

Skipped taxes for now, we can model that later if necessary.

Just want to make sure this is kosher so far:
1. Sell the put
2. Invest premium in t-bill
3. Take the cost and scale to get the implied %
4. Scale the box implied %

So it's similar but the put is a little cheaper. I think this is because with PM we're shifting the tail risk to the broker. Margin req is 50k ~1.5 mil exposure on 150k base. In the black swan scenario they might not be able to liquidate fast enough
comeinvest
Posts: 2728
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 »

parval wrote: Tue Mar 05, 2024 8:19 pm Okay I did the math!

Image

Spreadsheet link: https://docs.google.com/spreadsheets/d/ ... sp=sharing

Skipped taxes for now, we can model that later if necessary.

Just want to make sure this is kosher so far:
1. Sell the put
2. Invest premium in t-bill
3. Take the cost and scale to get the implied %
4. Scale the box implied %

So it's similar but the put is a little cheaper. I think this is because with PM we're shifting the tail risk to the broker. Margin req is 50k ~1.5 mil exposure on 150k base. In the black swan scenario they might not be able to liquidate fast enough
I'm not understanding your logic and calcs; but I don't think options pricing depends on a particular broker risk for particular options strategies; I do think however that your calcs are in vain as long as you don't come up with an options pricing model based on stochastic calculus, and as long as you ignore the possibility of the index rising above 6000. In summary, I'm confused what you are trying to accomplish or to demonstrate.
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 »

Are t-bills as good as cash in an IRA in IBKR for collateral purposes? I've found mixed answers on this. If I were to use all the cash in my IRA to buy t-bills, would I still be able to buy futures?
comeinvest
Posts: 2728
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 »

DMoogle wrote: Thu Mar 07, 2024 2:59 pm Are t-bills as good as cash in an IRA in IBKR for collateral purposes? I've found mixed answers on this. If I were to use all the cash in my IRA to buy t-bills, would I still be able to buy futures?
The mechanism how T-bills can serve as futures collateral in IRAs was discussed up in this thread.
Yes I think you can convert all your cash to T-bills; but IB sent an email recently that they are going to strictly enforce positive cash, which means if your futures lose money and your cash balance becomes negative, they will liquidate positions.
So I would leave some cash buffer for that.
After memorial day this year, securities will have T+1 settlement instead of T+2, which makes cash management a bit easier because securities settlement will align with futures daily mark to market settlement; although I'm not sure if IB would only look at instantaneous (including non-settled) total cash for margin check.
Last edited by comeinvest on Thu Mar 07, 2024 3:37 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 »

comeinvest wrote: Thu Mar 07, 2024 3:33 pm
DMoogle wrote: Thu Mar 07, 2024 2:59 pm Are t-bills as good as cash in an IRA in IBKR for collateral purposes? I've found mixed answers on this. If I were to use all the cash in my IRA to buy t-bills, would I still be able to buy futures?
The mechanism how T-bills can serve as futures collateral in IRAs was discussed up in this thread.
Yes I think you can convert all your cash to T-bills; but IB sent an email recently that they are going to strictly enforce positive cash, which means if your futures lose money and your cash balance becomes negative, they will liquidate positions.
But theoretically could I just set it to liquidate the t-bills first?
comeinvest
Posts: 2728
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 »

DMoogle wrote: Thu Mar 07, 2024 3:35 pm
comeinvest wrote: Thu Mar 07, 2024 3:33 pm
DMoogle wrote: Thu Mar 07, 2024 2:59 pm Are t-bills as good as cash in an IRA in IBKR for collateral purposes? I've found mixed answers on this. If I were to use all the cash in my IRA to buy t-bills, would I still be able to buy futures?
The mechanism how T-bills can serve as futures collateral in IRAs was discussed up in this thread.
Yes I think you can convert all your cash to T-bills; but IB sent an email recently that they are going to strictly enforce positive cash, which means if your futures lose money and your cash balance becomes negative, they will liquidate positions.
But theoretically could I just set it to liquidate the t-bills first?
They have no obligation to honor your liquidate first preference, so there is some risk; although the risk is really not big if you have liquid stocks or ETFs in the account except it may mess up whatever position sizes you had. They basically deleverage for you. There will also be some transaction cost, regardless what is being sold either by you or by IB, except perhaps the money market fund TPPXX or some other mutual fund. (But TPPXX cannot be futures collateral as far as I know.) So you have to balance that vs the opportunity cost of zero interest on cash. I was not able to find another no transaction fee money market fund available at IB with reasonable cost but this one.
I would think a very liquid equities or bond ETF has less transaction cost than T-bills, especially if you are on the zero commission plan.
rama13
Posts: 72
Joined: Sun May 15, 2011 2:09 am

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

Post by rama13 »

I was buying an ES box as collateral for my futures in my IRA account at IB. This worked great to keep cash drag to a minimum, since IB previously allowed negative cash balances in IRAs. Now that they don't, I'm wondering if a switch to Schwab might be better because of their lower futures margin requirements which would mean less cash drag. Some questions if anyone is using Schwab for this strategy or is familiar with Schwab's product...

Does Schwab indeed have a lower futures margin requirement than IB (IB is 200%--is Schwab 100% or 125%)?

Does Schwab limit which futures can trade and can be held overnight (including SOFR3) in IRAs?

Does Schwab auto-liquidate without any grace period like IB does, or do you have some time to fix the situation?

Does anyone have an opinion if Schwab, or any other broker, would be preferred over IB due to this new policy IB has for their IRAs?
comeinvest
Posts: 2728
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 »

rama13 wrote: Sun Mar 17, 2024 1:41 am I was buying an ES box as collateral for my futures in my IRA account at IB. This worked great to keep cash drag to a minimum, since IB previously allowed negative cash balances in IRAs. Now that they don't, I'm wondering if a switch to Schwab might be better because of their lower futures margin requirements which would mean less cash drag. Some questions if anyone is using Schwab for this strategy or is familiar with Schwab's product...

Does Schwab indeed have a lower futures margin requirement than IB (IB is 200%--is Schwab 100% or 125%)?

Does Schwab limit which futures can trade and can be held overnight (including SOFR3) in IRAs?

Does Schwab auto-liquidate without any grace period like IB does, or do you have some time to fix the situation?

Does anyone have an opinion if Schwab, or any other broker, would be preferred over IB due to this new policy IB has for their IRAs?
If you use FOPs for futures collateral, then the futures margin requirement shouldn't affect cash drag, should it? You have to maintain a cash buffer to accommodate the daily variation margin (daily settlement) to avoid liquidation due to the positive cash requirement, which should be independent of the futures margin requirement.
I'm for the birds
Posts: 2
Joined: Sat Dec 23, 2023 3:41 pm

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

Post by I'm for the birds »

rama13 wrote: Sun Mar 17, 2024 1:41 am I was buying an ES box as collateral for my futures in my IRA account at IB. This worked great to keep cash drag to a minimum, since IB previously allowed negative cash balances in IRAs. Now that they don't, I'm wondering if a switch to Schwab might be better because of their lower futures margin requirements which would mean less cash drag. Some questions if anyone is using Schwab for this strategy or is familiar with Schwab's product...

Does Schwab indeed have a lower futures margin requirement than IB (IB is 200%--is Schwab 100% or 125%)?

Does Schwab limit which futures can trade and can be held overnight (including SOFR3) in IRAs?

Does Schwab auto-liquidate without any grace period like IB does, or do you have some time to fix the situation?

Does anyone have an opinion if Schwab, or any other broker, would be preferred over IB due to this new policy IB has for their IRAs?
I currently use TDA as my broker for trading futures in an IRA. I called them recently to ask about how the move to Schwab would affect the rules and requirements on this, and the phone rep said that the rules should stay the same, with the caveat that they are within their rights to change them at any time.

The rules at TDA are currently as follows:

125% futures margin requirement in an IRA
$25,000 minimum account value to trade futures in an IRA
No limit on which futures can trade and be held overnight (including SOFR3) in IRAs
- the caveat on this is that they offer a limited selection of available futures in the first place compared to IB
A grace period of one day to meet a margin call


As an aside, I've been meaning to ask a few questions about the mechanics of using an ES box for futures collateral. Apologies if this has been answered before, but after reading this thread and the thread on box spreads, I have only been able to find a few vague-ish references to this but nothing really clear.

Is this something exclusive to IB (like holding Tbills as futures collateral) or should this be possible at other brokers?
How does one determine if the FOPs are European- or American-style? Is it as simple as always just using the EOM contracts for European-style or is there more to it?
Should the size of the box be the same as the futures margin requirement, with some cash buffer still remaining? Or should it be the entire required margin + cash buffer (leaving no cash remaining)?
What should the duration of the box be? Is it preferable to match the duration of the box with the duration of the futures contracts held (ie quarterly)?


Finally, since this is my first post in this thread, I wanted to thank everyone here for all the valuable information and interesting discussion you've provided. This thread and related others have been very helpful for learning about futures, treasury yield curves, lifecycle investing, etc.
rama13
Posts: 72
Joined: Sun May 15, 2011 2:09 am

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

Post by rama13 »

comeinvest wrote: Thu Mar 21, 2024 5:26 am If you use FOPs for futures collateral, then the futures margin requirement shouldn't affect cash drag, should it? You have to maintain a cash buffer to accommodate the daily variation margin (daily settlement) to avoid liquidation due to the positive cash requirement, which should be independent of the futures margin requirement.
Before this new policy, you could buy a box that included that cash buffer and let the cash balance go negative. Now, you need to keep that buffer as cash, and you're not get any interest on the first $10,000. So there is an unfavorable effect on cash drag.

For example, if your margin requirement were $20,000, and you wanted a buffer of $15,000, you could have bought a $35,000 box. If you started to "eat into that buffer", the cash balance would go negative (without having to pay interest). Now, if you were to do that, they'd liquidate once your cash balance went negative, so you'd probably only want to buy a $20,000 box and keep that $15,000 in cash (of which you're only earning interest on $5,000).
rama13
Posts: 72
Joined: Sun May 15, 2011 2:09 am

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

Post by rama13 »

I'm for the birds wrote: Thu Mar 21, 2024 12:30 pm No limit on which futures can trade and be held overnight (including SOFR3) in IRAs
- the caveat on this is that they offer a limited selection of available futures in the first place compared to IB
Is this list available anywhere, or do you know if the futures commonly mentioned in this thread are included?
I'm for the birds wrote: Thu Mar 21, 2024 12:30 pm As an aside, I've been meaning to ask a few questions about the mechanics of using an ES box for futures collateral. Apologies if this has been answered before, but after reading this thread and the thread on box spreads, I have only been able to find a few vague-ish references to this but nothing really clear.

Is this something exclusive to IB (like holding Tbills as futures collateral) or should this be possible at other brokers?
You should be able to buy a box at other brokers, but I don't know if it's going to offset or act as futures collateral (it does at Interactive Brokers). If you try it, please report back.
I'm for the birds wrote: Thu Mar 21, 2024 12:30 pm How does one determine if the FOPs are European- or American-style? Is it as simple as always just using the EOM contracts for European-style or is there more to it?
Each trading platform is going to be different for showing options' exercise style. For IB's TWS, it's under "Description". CME's website also has some information about each contract.
I'm for the birds wrote: Thu Mar 21, 2024 12:30 pm Should the size of the box be the same as the futures margin requirement, with some cash buffer still remaining? Or should it be the entire required margin + cash buffer (leaving no cash remaining)?
See my previous post--it depends on if cash balance is allowed to go negative.
I'm for the birds wrote: Thu Mar 21, 2024 12:30 pm What should the duration of the box be? Is it preferable to match the duration of the box with the duration of the futures contracts held (ie quarterly)?
I don't think it matters. I think it's more related to the interest rate on your cash balances.
comeinvest
Posts: 2728
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 »

rama13 wrote: Sat Mar 23, 2024 6:17 pm
comeinvest wrote: Thu Mar 21, 2024 5:26 am If you use FOPs for futures collateral, then the futures margin requirement shouldn't affect cash drag, should it? You have to maintain a cash buffer to accommodate the daily variation margin (daily settlement) to avoid liquidation due to the positive cash requirement, which should be independent of the futures margin requirement.
Before this new policy, you could buy a box that included that cash buffer and let the cash balance go negative. Now, you need to keep that buffer as cash, and you're not get any interest on the first $10,000. So there is an unfavorable effect on cash drag.

For example, if your margin requirement were $20,000, and you wanted a buffer of $15,000, you could have bought a $35,000 box. If you started to "eat into that buffer", the cash balance would go negative (without having to pay interest). Now, if you were to do that, they'd liquidate once your cash balance went negative, so you'd probably only want to buy a $20,000 box and keep that $15,000 in cash (of which you're only earning interest on $5,000).
I just said that the futures margin requirement doesn't affect cash drag. I didn't say that the positive cash requirement doesn't affect cash drag.

Besides, daily variation margin will never eat into your box, which is your futures collateral, and which you can set at $20k or $35 before or after the new positive cash requirement. ($35k doesn't help you except when the futures margin requirement rises a lot, which doesn't happen very often.) Variation margin will however eat into your cash buffer, whether you have a $20k or a $35k box, where "buffer" is in relation to your cash balance, not in relation to margin requirements. The goal is to minimize the cash balance, in order to minimize cash drag. The positive cash balance requirement makes this a bit more difficult; but this is unrelated to the futures margin requirement or to the size of your box.
comeinvest
Posts: 2728
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 »

New micro treasury futures:
https://www.cmegroup.com/articles/2024/ ... tures.html
https://www.cmegroup.com/articles/2024/ ... tures.html
https://www.cmegroup.com/markets/intere ... tures.html

I just checked: The bid/asks spreads are almost as narrow as those of the regular treasury futures contracts. In contrast to the micro yield futures which had abysmal bid/ask spreads last time I checked.

I think the use cases are limited, because regular ITT futures worked just fine even for quite small accounts. Perhaps the micro futures allow for more granular rebalancing with very small accounts.

They are cash settled, which is also of no major benefit that I can see if strategic positions are typically calendar spread rolled. Except when rebalancing while reducing a position, one could let some contracts expire.
Post Reply