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

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

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

Post by skierincolorado »

Hfearless wrote: Sun Oct 24, 2021 8:53 am With the obvious caveat that market timing methods that do better than chance are few and far between, and that the ones that backtest well don’t come with any guarantees either—
skierincolorado wrote: Fri Oct 22, 2021 8:38 pm By diversifying we can maintain the same expected value with lower variance. Market timing is inherently less diversified and therefore has higher variance.
Conditional on the existence of a market timing method that actually works, doesn’t switching from one diversified portfolio to a different diversified portfolio reduce risk?

Note that the entire concept of rebalancing has a market timing element to it—you trade what performed better for what performed worse and expect the latter to revert to the mean, bringing in some profit.
skierincolorado wrote: Fri Oct 22, 2021 8:38 pm Also, and this comes to your second point, performing *different* than the market is a risk. If most retail investors are rolling in it because they bought stocks, and you are missing out because you bought much more bonds than most, you may be priced out of buying a home for example. It's also a psychological burden and increases the odds you will abandon ship - further increasing your risk.
So primarily a psychological concern.

To what extent does your suggestion of 125:200 include this concern? What would the allocation be under the sole goal of maximizing the long-term gain, assuming the mental fortitude not to deviate from the IPS?
I don't think switching from one diversified portfolio to another reduces risk. They can't both be fully diversified. In the example of switching from ITT to STT, you're switching from missing one huge segment of the global bond and financial market, to missing a different segment. The variance in returns will be greater than if you owned both.

I also don't think rebalancing is market timing, depending on how it is done. In leveraged strategies it is often necessary to rebalance back to market weights just due to the volatility decay or low-volatility boost. One can very easily stray from market weights. The bond market is also always growing as new debt is issued.

What backtested best since 1955 in terms of sharpe ratio is something like 125/250. I actually don't do 125/200 personally, more like 130/180.

The above reason is one of the reasons why. But I and other posters also found that there is less variance in the final total return from stocks than from bonds. There were several other reasons as well for shifting a little more to stocks. It still backtests extremely well. Gotta skim the thread and hte thread that millenialmillions started.
L2F
Posts: 26
Joined: Fri Sep 24, 2021 9:33 am

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

Post by L2F »

How much USD should I keep for /ZF collateral to be safe (enough)?
I'm using cash account on IBKR (as I'm afraid of using margin) - is it really recommended to switch to Portfolio Margin or Reg T Margin accounts?
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

L2F wrote: Sun Oct 24, 2021 12:25 pm How much USD should I keep for /ZF collateral to be safe (enough)?
I'm using cash account on IBKR (as I'm afraid of using margin) - is it really recommended to switch to Portfolio Margin or Reg T Margin accounts?
I would seriously recommend switching to margin. Reg T or Port Margin if eligible. You don't have to use the margin, but just turning it on would allow you to keep much less cash in the account. You are currently experiencing "cash drag" which will lower returns. You could keep the cash balance at $500 or $1,000. Then if ZF lost value, your cash balance would drop but you would not have to fear the position being liquidated. So you wouldn't actually be borrowing on margin, but it's just there IF ZF drops a lot and you need it.

However if you don't want to turn margin on:

I've calculated max drawdowns in the last 10 years around 5%. So 6k in cash (+CME requirement = ~7k) to endure a max-draw event. The previous max-drawdowns were over an extended period, so you'd have time to create more cash. So you could probably get away with a bit less, like 3-4k, if you were willing to occasionally monitor the account. By turning on margin though you could drop it to $500 or so, and then if ZF drops it would allow the cash to go negative rather than liquidating the ZF(assuming there are other assets in the account that can be used as collateral for a margin loan).
L2F
Posts: 26
Joined: Fri Sep 24, 2021 9:33 am

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

Post by L2F »

skierincolorado wrote: Mon Oct 25, 2021 10:39 am
L2F wrote: Sun Oct 24, 2021 12:25 pm How much USD should I keep for /ZF collateral to be safe (enough)?
I'm using cash account on IBKR (as I'm afraid of using margin) - is it really recommended to switch to Portfolio Margin or Reg T Margin accounts?
I would seriously recommend switching to margin. Reg T or Port Margin if eligible. You don't have to use the margin, but just turning it on would allow you to keep much less cash in the account. You are currently experiencing "cash drag" which will lower returns. You could keep the cash balance at $500 or $1,000. Then if ZF lost value, your cash balance would drop but you would not have to fear the position being liquidated. So you wouldn't actually be borrowing on margin, but it's just there IF ZF drops a lot and you need it.

However if you don't want to turn margin on:

I've calculated max drawdowns in the last 10 years around 5%. So 6k in cash (+CME requirement = ~7k) to endure a max-draw event. The previous max-drawdowns were over an extended period, so you'd have time to create more cash. So you could probably get away with a bit less, like 3-4k, if you were willing to occasionally monitor the account. By turning on margin though you could drop it to $500 or so, and then if ZF drops it would allow the cash to go negative rather than liquidating the ZF(assuming there are other assets in the account that can be used as collateral for a margin loan).
Thanks, that makes a lot of sense, I definitely will consider switching to margin account.
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

L2F wrote: Mon Oct 25, 2021 11:14 am
skierincolorado wrote: Mon Oct 25, 2021 10:39 am
L2F wrote: Sun Oct 24, 2021 12:25 pm How much USD should I keep for /ZF collateral to be safe (enough)?
I'm using cash account on IBKR (as I'm afraid of using margin) - is it really recommended to switch to Portfolio Margin or Reg T Margin accounts?
I would seriously recommend switching to margin. Reg T or Port Margin if eligible. You don't have to use the margin, but just turning it on would allow you to keep much less cash in the account. You are currently experiencing "cash drag" which will lower returns. You could keep the cash balance at $500 or $1,000. Then if ZF lost value, your cash balance would drop but you would not have to fear the position being liquidated. So you wouldn't actually be borrowing on margin, but it's just there IF ZF drops a lot and you need it.

However if you don't want to turn margin on:

I've calculated max drawdowns in the last 10 years around 5%. So 6k in cash (+CME requirement = ~7k) to endure a max-draw event. The previous max-drawdowns were over an extended period, so you'd have time to create more cash. So you could probably get away with a bit less, like 3-4k, if you were willing to occasionally monitor the account. By turning on margin though you could drop it to $500 or so, and then if ZF drops it would allow the cash to go negative rather than liquidating the ZF(assuming there are other assets in the account that can be used as collateral for a margin loan).
Thanks, that makes a lot of sense, I definitely will consider switching to margin account.
The cash drag is measured relative to a portfolio with the exact same AA. Options 1 and 2 below maintain the exact same AA. Option 3 changes the AA. All three options eliminate the cash drag.

1. Turn on margin and hold the cash in a STT / money market type fund. The AA is unchanged, but the cash drag is eliminated.

2. Turn on margin, and use the cash to buy stocks or bonds directly. Sell MES, other futures to maintain the same AA. This eliminates the borrowing cost implict in the futures contracts. Not only does this eliminate the cash drag, but it also eliminates the financing cost of futures that is ~.2-.4% above rates on cash.

3. Turn on margin and just buy more VGIT or VTI. This changes your AA, but also eliminates cash drag.

The one that usually makes the most sense is #3 unless we're dealing with very large amounts of money and looking to pinpoint a very specific AA.
Last edited by skierincolorado on Mon Oct 25, 2021 12:14 pm, edited 6 times in total.
secondopinion
Posts: 6011
Joined: Wed Dec 02, 2020 12:18 pm

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

Post by secondopinion »

skierincolorado wrote: Mon Oct 25, 2021 12:02 pm
L2F wrote: Mon Oct 25, 2021 11:14 am
skierincolorado wrote: Mon Oct 25, 2021 10:39 am
L2F wrote: Sun Oct 24, 2021 12:25 pm How much USD should I keep for /ZF collateral to be safe (enough)?
I'm using cash account on IBKR (as I'm afraid of using margin) - is it really recommended to switch to Portfolio Margin or Reg T Margin accounts?
I would seriously recommend switching to margin. Reg T or Port Margin if eligible. You don't have to use the margin, but just turning it on would allow you to keep much less cash in the account. You are currently experiencing "cash drag" which will lower returns. You could keep the cash balance at $500 or $1,000. Then if ZF lost value, your cash balance would drop but you would not have to fear the position being liquidated. So you wouldn't actually be borrowing on margin, but it's just there IF ZF drops a lot and you need it.

However if you don't want to turn margin on:

I've calculated max drawdowns in the last 10 years around 5%. So 6k in cash (+CME requirement = ~7k) to endure a max-draw event. The previous max-drawdowns were over an extended period, so you'd have time to create more cash. So you could probably get away with a bit less, like 3-4k, if you were willing to occasionally monitor the account. By turning on margin though you could drop it to $500 or so, and then if ZF drops it would allow the cash to go negative rather than liquidating the ZF(assuming there are other assets in the account that can be used as collateral for a margin loan).
Thanks, that makes a lot of sense, I definitely will consider switching to margin account.
with margin on you can hold the cash in a risk free money market or short term bond fund.. that's where the drag is relative to. Or you could buy more stock and sell an MES contract, which means you are borrowing less because borrowing cost is implicit in all futures contracts... better to invest all cash before using futures to borrow.
Definitely. You will never get called into margin calls off of that or even need a margin account. If you want to take major risks, take with your own money first.
Passive investing: not about making big bucks but making profits. Active investing: not about beating the market but meeting goals. Speculation: not about timing the market but taking profitable risks.
comeinvest
Posts: 2709
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 »

skierincolorado wrote: Mon Oct 25, 2021 12:02 pm
L2F wrote: Mon Oct 25, 2021 11:14 am
skierincolorado wrote: Mon Oct 25, 2021 10:39 am
L2F wrote: Sun Oct 24, 2021 12:25 pm How much USD should I keep for /ZF collateral to be safe (enough)?
I'm using cash account on IBKR (as I'm afraid of using margin) - is it really recommended to switch to Portfolio Margin or Reg T Margin accounts?
I would seriously recommend switching to margin. Reg T or Port Margin if eligible. You don't have to use the margin, but just turning it on would allow you to keep much less cash in the account. You are currently experiencing "cash drag" which will lower returns. You could keep the cash balance at $500 or $1,000. Then if ZF lost value, your cash balance would drop but you would not have to fear the position being liquidated. So you wouldn't actually be borrowing on margin, but it's just there IF ZF drops a lot and you need it.

However if you don't want to turn margin on:

I've calculated max drawdowns in the last 10 years around 5%. So 6k in cash (+CME requirement = ~7k) to endure a max-draw event. The previous max-drawdowns were over an extended period, so you'd have time to create more cash. So you could probably get away with a bit less, like 3-4k, if you were willing to occasionally monitor the account. By turning on margin though you could drop it to $500 or so, and then if ZF drops it would allow the cash to go negative rather than liquidating the ZF(assuming there are other assets in the account that can be used as collateral for a margin loan).
Thanks, that makes a lot of sense, I definitely will consider switching to margin account.
The cash drag is measured relative to a portfolio with the exact same AA. Options 1 and 2 below maintain the exact same AA. Option 3 changes the AA. All three options eliminate the cash drag.

1. Turn on margin and hold the cash in a STT / money market type fund. The AA is unchanged, but the cash drag is eliminated.

2. Turn on margin, and use the cash to buy stocks or bonds directly. Sell MES, other futures to maintain the same AA. This eliminates the borrowing cost implict in the futures contracts. Not only does this eliminate the cash drag, but it also eliminates the financing cost of futures that is ~.2-.4% above rates on cash.

3. Turn on margin and just buy more VGIT or VTI. This changes your AA, but also eliminates cash drag.

The one that usually makes the most sense is #3 unless we're dealing with very large amounts of money and looking to pinpoint a very specific AA.
4. Buy short-term SPX box spreads with the extra cash. Same as (1.) except you cut out the middle man. Get a slightly higher rate than with money market funds, and a slightly higher rate than the implied financing rate of treasury futures (if the conclusions on the last few pages of this thread are correct), and a rate about equal to the implied financing rate of equity index futures. Is (1.) or (4.) more flexible? Depends on the amount of cash and other parameters. Both incur trading cost. I would tend to prefer (4.) over (1.). I found I need 2-4 weeks to recoup the bid/ask spread of MINT, but had some good experiences with box spreads of only 1-2 weeks duration, and of course they could be laddered to produce less work.
Hfearless
Posts: 135
Joined: Sat Aug 14, 2021 9:53 am

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

Post by Hfearless »

I wonder, by the way, how does the tax treatment of box spreads affect all this. It depends on jurisdiction, of course, but I’d guess most countries will treat it as a position that you hold and eventually close at a loss. Wouldn’t that loss reduce the taxable base a little?
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

Hfearless wrote: Mon Oct 25, 2021 4:33 pm I wonder, by the way, how does the tax treatment of box spreads affect all this. It depends on jurisdiction, of course, but I’d guess most countries will treat it as a position that you hold and eventually close at a loss. Wouldn’t that loss reduce the taxable base a little?
In the original lifecycle investing thread, it was suggested boxes are good because the interest/loss is deductible. That's true, but the interest/loss that other forms of leverage experience, like futures, is just subtracted from the gain so also results in a tax saving. Nobody disagreed when I pointed this out. Still, in a taxable account a box advantage over futures due to the long-term captial gains treatment of equities vs 60/40 tax treatment of futures. Here's my calculation on the tax advantages of each in this post and in the one above it:

viewtopic.php?f=10&t=357281&p=6225902&h ... x#p6225902
comeinvest
Posts: 2709
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 »

skierincolorado wrote: Fri Oct 22, 2021 1:01 pm
comeinvest wrote: Fri Oct 22, 2021 3:36 am I made two adjustments to the treasury futures portion of my portfolio tonight. They are mostly not meant to time the market (although I also played around a little bit), but to transition to my final strategic allocation:
- Eliminated all remaining short ZT positions that I established as a partial hedge of my long treasury futures to rising rates, when 2-year rates were on the floor i.e. close to nil. I did this today because I found out that the 2-year carry (ca. 0.8%, with yield to maturity of ca. 0.45%) just surpassed the carry of the 5-year (ca. 1.85% if my math was right) on a duration-adjusted basis, which creeped me out, and I couldn't expect the alpha from the pair trade any more that I expected based on papers advocating a dynamic positioning based on current carry per duration. I pocketed a nice profit, and the hedge did its job. I'm still playing a similar game in the European market.
- Transitioned from some of the ZB to TN+ZN. I should have done this on Monday directly from UB instead of moving from UB to ZB, but someone from Colorado ;) convinced me just this week to go with even shorter durations, so I ended up doing it in two steps. In the meantime, I made a tiny profit ("profit" as in "less loss") by holding ZB instead of TN+ZN for those 4 days. TN+ZN has almost the same DV01 as ZB, which made the transition easy on a DV01-neutral basis. I went with TN and ZN for diversification, because I have already a larger position in ZF.

I used limit orders at the bid (for buy) and ask (for sell) in the night. The volume in the night is lower, but I think the bid and ask sizes are lower too, such that passive limit orders might profit from liquidity based fills ("patient trading") just like by day. The spread is the same as by day. I'm not sure if it's statistically better to trade by day or night for minimizing transaction cost, but it seems to work well either way.

I was kind of sad to let my UB go this week, but it dawned me that it became a curse word in this message thread ;) My allocation comes now closer to an equidistribution between ZF/ZN/TN/ZB on duration-adjusted basis (it's currently still somewhere between a duration-based and notional-value-based equidistribution). I still have to define strict ratios and a suitable rebalancing strategy, so that I don't have to think too much every time in the future. Comments are welcome.

Assuming the duraiton on ZB is 11.5 years, I think this is good, but I think ZB is more like 17 years. Rules for the dynamic and static parts would be good too of course.

For ZB though did we ever figure out what the duration is? CME says it's currently 11.5... but the historical performance was more volatile that the S&P 10-20 year bond index, which has duration of 14.99 years. That would suggest it has duration of around 17 years which is quite similar to VUSTX which is the one I use in my backtesting since 1991 to show how much better VFITX is. While I am fully on-board with the diversification of not just being ZF, I would feel strongly that 17 years is too long if taking equal nominal amounts of each. If equal weighted by risk I would be on board I think. It might look something like this. I'm blue, you're orange, red is what you were before (or worse). I can't confidently say blue is better than orange - it's close enough. And this assumes ZB is 17 years duration. If it's shorter like 11.5 years, it would probably be the same as blue.

If you do do it, I like scaling on DV01 because it would mean less ZB than other measures of risk, like historical stdev or max-draw. As long as you're staying at or above that orange line, I think you're good. There are even a couple short periods where the orange line outperforms blue (2021, 1992-1994).

https://www.portfoliovisualizer.com/bac ... on4_3=-110
My proposed solution, an ITT strategy diversified across ZF/ZN/TN/ZB (weights tbd), would not have anything to do with VUSTX which would represent UB. I think we still have not found an ETF equivalent to ZB. A duration-neutral alternative to ZF-only equidistributed based on current durations would be (omitting ZB for the moment, but that would make little difference if we weigh by constant duration risk): https://www.portfoliovisualizer.com/bac ... on4_3=-163

Which shows that
1. It makes little difference; ZF-only wins, but the time series seems random so that I would not feel confident extrapolating the small cumulative difference into the future based on a 1-country 20-year-only time series; there was practically no difference in the first third of the 20-year period
2. There is little diversification, but maybe one could avoid the risk of potential future funding cost anomalies by diversifying across maturities (diversification of implementation risk); only a small change in funding cost / slippage e.g. in ZF could have a relatively large effect on results due to leverage
3. Yearly differences between the ITT varieties were up to about 3%. If you want do avoid being on the losing end for a year, then invest in several varieties of ITT; however, most of the difference is typically made up soon after a losing year.
4. It's a challenge to do meaningful backtests without having past durations of futures, ETFs, or both.

Have we figured out yet what is the current duration of ZB? The stddev aficionados in this thread said that it must be much higher than what the treasury analytics page says. (Can we verify this for the stddev of the current contract?) I personally feel as long as I can't figure out what should be such a simple thing, I don't really understand treasury futures and I don't feel confident deciding on a treasure futures allocation and distribution ratios, even if it doesn't include ZB.
comeinvest
Posts: 2709
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 »

skierincolorado wrote: Mon Oct 25, 2021 6:04 pm
Hfearless wrote: Mon Oct 25, 2021 4:33 pm I wonder, by the way, how does the tax treatment of box spreads affect all this. It depends on jurisdiction, of course, but I’d guess most countries will treat it as a position that you hold and eventually close at a loss. Wouldn’t that loss reduce the taxable base a little?
In the original lifecycle investing thread, it was suggested boxes are good because the interest/loss is deductible. That's true, but the interest/loss that other forms of leverage experience, like futures, is just subtracted from the gain so also results in a tax saving. Nobody disagreed when I pointed this out. Still, in a taxable account a box advantage over futures due to the long-term captial gains treatment of equities vs 60/40 tax treatment of futures. Here's my calculation on the tax advantages of each in this post and in the one above it:

viewtopic.php?f=10&t=357281&p=6225902&h ... x#p6225902
I read your posting that you linked. Comments:
1. Correct me if I'm wrong, but I think you could only deduct $3,000 per year of the box financing cost, unless you have other gains that you can net the options losses from the box against. Many who run a modified HFEA with their entire NAV would have options losses many times higher if they do boxes. EDIT: If you run HFEA with treasury futures, net the 1256 gains from the treasury futures against the box losses?
2. I think equity index options are treated as 1256 contracts just like futures, not LTCG. However, consider (1.)
3. With boxes, you can basically defer capital gains forever, if you do it right, and if boxes are not disallowed by law or some rule at some point in the future. Buy a house? Write a box secured by your portfolio! (keep your total leverage in check!!) On the other hand, if they are outlawed, you would be "stuck" with brokers' margin rates with your appreciated positions - not fun ;)
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

comeinvest wrote: Mon Oct 25, 2021 7:55 pm
skierincolorado wrote: Mon Oct 25, 2021 6:04 pm
Hfearless wrote: Mon Oct 25, 2021 4:33 pm I wonder, by the way, how does the tax treatment of box spreads affect all this. It depends on jurisdiction, of course, but I’d guess most countries will treat it as a position that you hold and eventually close at a loss. Wouldn’t that loss reduce the taxable base a little?
In the original lifecycle investing thread, it was suggested boxes are good because the interest/loss is deductible. That's true, but the interest/loss that other forms of leverage experience, like futures, is just subtracted from the gain so also results in a tax saving. Nobody disagreed when I pointed this out. Still, in a taxable account a box advantage over futures due to the long-term captial gains treatment of equities vs 60/40 tax treatment of futures. Here's my calculation on the tax advantages of each in this post and in the one above it:

viewtopic.php?f=10&t=357281&p=6225902&h ... x#p6225902
I read your posting that you linked. Comments:
1. Correct me if I'm wrong, but I think you could only deduct $3,000 per year of the box financing cost, unless you have other gains that you can net the options losses from the box against. Many who run a modified HFEA with their entire NAV would have options losses many times higher if they do boxes. EDIT: If you run HFEA with treasury futures, net the 1256 gains from the treasury futures against the box losses?
2. I think equity index options are treated as 1256 contracts just like futures, not LTCG. However, consider (1.)
3. With boxes, you can basically defer capital gains forever, if you do it right, and if boxes are not disallowed by law or some rule at some point in the future. Buy a house? Write a box secured by your portfolio! (keep your total leverage in check!!) On the other hand, if they are outlawed, you would be "stuck" with brokers' margin rates with your appreciated positions - not fun ;)
1. Good point. They could probably sell some gains so the losses aren't totally wasted.
3. Yeah that would not be fun. Probably best to use contributions to pay off loans at that point.
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

comeinvest wrote: Mon Oct 25, 2021 6:10 pm
skierincolorado wrote: Fri Oct 22, 2021 1:01 pm
comeinvest wrote: Fri Oct 22, 2021 3:36 am I made two adjustments to the treasury futures portion of my portfolio tonight. They are mostly not meant to time the market (although I also played around a little bit), but to transition to my final strategic allocation:
- Eliminated all remaining short ZT positions that I established as a partial hedge of my long treasury futures to rising rates, when 2-year rates were on the floor i.e. close to nil. I did this today because I found out that the 2-year carry (ca. 0.8%, with yield to maturity of ca. 0.45%) just surpassed the carry of the 5-year (ca. 1.85% if my math was right) on a duration-adjusted basis, which creeped me out, and I couldn't expect the alpha from the pair trade any more that I expected based on papers advocating a dynamic positioning based on current carry per duration. I pocketed a nice profit, and the hedge did its job. I'm still playing a similar game in the European market.
- Transitioned from some of the ZB to TN+ZN. I should have done this on Monday directly from UB instead of moving from UB to ZB, but someone from Colorado ;) convinced me just this week to go with even shorter durations, so I ended up doing it in two steps. In the meantime, I made a tiny profit ("profit" as in "less loss") by holding ZB instead of TN+ZN for those 4 days. TN+ZN has almost the same DV01 as ZB, which made the transition easy on a DV01-neutral basis. I went with TN and ZN for diversification, because I have already a larger position in ZF.

I used limit orders at the bid (for buy) and ask (for sell) in the night. The volume in the night is lower, but I think the bid and ask sizes are lower too, such that passive limit orders might profit from liquidity based fills ("patient trading") just like by day. The spread is the same as by day. I'm not sure if it's statistically better to trade by day or night for minimizing transaction cost, but it seems to work well either way.

I was kind of sad to let my UB go this week, but it dawned me that it became a curse word in this message thread ;) My allocation comes now closer to an equidistribution between ZF/ZN/TN/ZB on duration-adjusted basis (it's currently still somewhere between a duration-based and notional-value-based equidistribution). I still have to define strict ratios and a suitable rebalancing strategy, so that I don't have to think too much every time in the future. Comments are welcome.

Assuming the duraiton on ZB is 11.5 years, I think this is good, but I think ZB is more like 17 years. Rules for the dynamic and static parts would be good too of course.

For ZB though did we ever figure out what the duration is? CME says it's currently 11.5... but the historical performance was more volatile that the S&P 10-20 year bond index, which has duration of 14.99 years. That would suggest it has duration of around 17 years which is quite similar to VUSTX which is the one I use in my backtesting since 1991 to show how much better VFITX is. While I am fully on-board with the diversification of not just being ZF, I would feel strongly that 17 years is too long if taking equal nominal amounts of each. If equal weighted by risk I would be on board I think. It might look something like this. I'm blue, you're orange, red is what you were before (or worse). I can't confidently say blue is better than orange - it's close enough. And this assumes ZB is 17 years duration. If it's shorter like 11.5 years, it would probably be the same as blue.

If you do do it, I like scaling on DV01 because it would mean less ZB than other measures of risk, like historical stdev or max-draw. As long as you're staying at or above that orange line, I think you're good. There are even a couple short periods where the orange line outperforms blue (2021, 1992-1994).

https://www.portfoliovisualizer.com/bac ... on4_3=-110
My proposed solution, an ITT strategy diversified across ZF/ZN/TN/ZB (weights tbd), would not have anything to do with VUSTX which would represent UB. I think we still have not found an ETF equivalent to ZB. A duration-neutral alternative to ZF-only equidistributed based on current durations would be (omitting ZB for the moment, but that would make little difference if we weigh by constant duration risk): https://www.portfoliovisualizer.com/bac ... on4_3=-163

Which shows that
1. It makes little difference; ZF-only wins, but the time series seems random so that I would not feel confident extrapolating the small cumulative difference into the future based on a 1-country 20-year-only time series; there was practically no difference in the first third of the 20-year period
2. There is little diversification, but maybe one could avoid the risk of potential future funding cost anomalies by diversifying across maturities (diversification of implementation risk); only a small change in funding cost / slippage e.g. in ZF could have a relatively large effect on results due to leverage
3. Yearly differences between the ITT varieties were up to about 3%. If you want do avoid being on the losing end for a year, then invest in several varieties of ITT; however, most of the difference is typically made up soon after a losing year.
4. It's a challenge to do meaningful backtests without having past durations of futures, ETFs, or both.

Have we figured out yet what is the current duration of ZB? The stddev aficionados in this thread said that it must be much higher than what the treasury analytics page says. (Can we verify this for the stddev of the current contract?) I personally feel as long as I can't figure out what should be such a simple thing, I don't really understand treasury futures and I don't feel confident deciding on a treasure futures allocation and distribution ratios, even if it doesn't include ZB.
I think I figured it out partially. If you remember, I had theorized that the duration of ZB has changed. I spent a lot of time trying to find historical duration or DV01 data and couldn't find anything. If that didn't work my plan was to calculate relative stdev in different periods other than the previous 10 years. But this morning I remembered zkn had also done 3 and 5 year comparisons. Well guess what? They seem to confirm my suspicion.

Here are the stdev of ZB vs the 10-20 year cash bond index:

10 years: 8.40% / 7.75%
5 years: 8.80% / 8.60%
3 years: 9.50% / 9.94%

You can see over the full 10 years it is .65% higher but the last three it is .44% lower. This would suggest that over the first 5 years it was over 1% higher. Which is a pretty big swing from being over 1% higher to .44% lower. That's a pretty big change in duration. None of the other contracts show anything like that. This would tend to confirm that the current duration of ZB is the 11.5 on the CME page, but that it used to be higher. Maybe we could get exact stdev for the last 3-12 months vs a 5 year period 5-10 years ago. And then do the proper math to convert to relative durations.

If ZB is only 11.5 currently, then it's probably not too bad. But a couple things to consider about your IEF backtest:

1 You chose a 2:3 ratio to replace IEF with VFITX. The conclusions are highly sensitive on the ratio you pick. 2:3 seems correct. The durations are 5.2:8 which would be a bit higher, but I think a lower ratio is correct for reasons we've discussed before.

3. There was a flattening of close to 1% during this period, giving a slight advantage to longer durations.

4. I we bump the ratio of VFITX up slightly and start in 2005 (per 1 and 3), VFITX wins solidly. Add in some ZN and it might start to look significant.

5. Even the current ZB duration is quite a bit longer than IEF (11.5 vs 8).


All that said I have absolutely no problem with holding equal risk in ZB as in ZF, ZN, and TN assuming that the duration is 11.5 years. I think you'd see some small dropoff in the backtesting, but not a convincing amount. If it was equal nominal amounts, I think you'd see convincing dropoff.

You'd have to have a lot invested because you'd need a lot of contracts to get that balance. For 1 ZB you'd need 1.5 TN, 2 ZN and 3 ZF. If you round the TN down to 1, that's still 7 contracts with pretty long duration - more than I need personally.
Hfearless
Posts: 135
Joined: Sat Aug 14, 2021 9:53 am

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

Post by Hfearless »

So those looking to diversify among durations but without the capital to buy that many futures contracts should instead use ETFs and box spreads?
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

Hfearless wrote: Tue Oct 26, 2021 10:27 am So those looking to diversify among durations but without the capital to buy that many futures contracts should instead use ETFs and box spreads?
Given that diversifying to such long durations (11.5 years) shows no benefit in backtesting and is in fact slightly detrimental I would not worry about it. The financing cost of a box is too high and would make it even worse. Don’t buy bonds on a box. Decent replacements for zb (11.5y) and tn (8.8) are tmf (17 y) and tyd (8 y), but the fees are pretty equal to the finance premium on a box at 0.3% per $ of bond exposure. And as we know the risk adjusted returns at the tmf duration of 17 years are pretty bad. Risk adjusted returns start dropping historically at anything past like 6 years duration. It’s all a moot point to me because zf and maybe some zn and or zt is plenty diversification. 1 zf contract and 20k of tyd isn’t bad either. Personally I just own zf and zn.

What would one even buy that had the same duration as zb other than direct treasuries? Is there a fund with that duration?

If someone doesn't have the capital to diversify even at the shorter durations, it's fine to just own 1 ZF. This still backtests much better than owning TMF for example. Just maybe a hair worse than diversifying across ZT, ZF, ZN. One could buy some VFISX, IEF, or TYD to supplement to get to desired AA. But just ZF is fine too IMO given how well it backtests.

For example with 100k in capital:

60k VTI
3 MES (68k)
1 ZF ( 121k) = ~100k @ 5 yr duration
20k TYD (60k) = ~100k @ 5 yr duration

would be a 140/200 stock/bond allocation with the bond duration standardized at 5 years, but taking ~50/50 risk at 4.3 (ZF) and 8 years (TYD).

It would arguably be more efficient to implement the above as

6 MES (135k)
1 ZF ( 121k) = ~100k @ 5 yr duration
60k IEF (60k) = ~100k @ 5 yr duration

depending on what you assume the financing costs internal to TYD are. If you assume there is .2% finance premium above LIBOR internal to TYD, then TYD costs .5% above LIBOR including fees per $. Since that is likely higher than MES, it would be better to swap the TYD + VTI for MES and + IEF. Switching the TYD to IEF nets .35% (.5-.15) in cost above LIBOR, while switching from VTI to MES loses you maybe .3% (-.05% + .35%).

Either implementation would be fine really. The main point is both are plenty diversified without going longer than 8 years on the curve (TYD/IEF).

I'd actually probably me more inclined to diverisfy in the lower direction. For example buying ZF (4.3) or ZN (6.3) and diversifying with BSV/VFISX (1-4). Or buying ZT (1.8) and diversifying with VGIT/VFITX (4-7).
User avatar
Bentonkb
Posts: 83
Joined: Sun Mar 07, 2021 5:47 am

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

Post by Bentonkb »

skierincolorado wrote: Tue Oct 26, 2021 10:10 am
I think I figured it out partially. If you remember, I had theorized that the duration of ZB has changed. I spent a lot of time trying to find historical duration or DV01 data and couldn't find anything. If that didn't work my plan was to calculate relative stdev in different periods other than the previous 10 years. But this morning I remembered zkn had also done 3 and 5 year comparisons. Well guess what? They seem to confirm my suspicion.

Here are the stdev of ZB vs the 10-20 year cash bond index:

10 years: 8.40% / 7.75%
5 years: 8.80% / 8.60%
3 years: 9.50% / 9.94%

You can see over the full 10 years it is .65% higher but the last three it is .44% lower. This would suggest that over the first 5 years it was over 1% higher. Which is a pretty big swing from being over 1% higher to .44% lower. That's a pretty big change in duration. None of the other contracts show anything like that. This would tend to confirm that the current duration of ZB is the 11.5 on the CME page, but that it used to be higher. Maybe we could get exact stdev for the last 3-12 months vs a 5 year period 5-10 years ago. And then do the proper math to convert to relative durations.

If ZB is only 11.5 currently, then it's probably not too bad. But a couple things to consider about your IEF backtest:
I think what you are seeing is the CTD bond for the contract is shifting to closer maturities. The range of acceptable maturities for the ZB basket is very wide. As the rate drops well below the notional 6% rate, the CTD tends to favor closer maturities with higher coupons. There are several deliverable bonds that are nearly CTD that have even lower duration (DV01) than the current CTD.

If we ever had a situation where rates started moving up near 6% I think you would find that the duration for ZF and TN make small shifts because they have narrow acceptable ranges of maturity. ZN and ZB would have big swings in duration.

I found a blog post from several years ago that suggested that ZB could be traded against TLT just like ES can be traded against SPY. That doesn't seem to be the case anymore. UB would be a better match for TLT now.
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

Bentonkb wrote: Tue Oct 26, 2021 2:09 pm
skierincolorado wrote: Tue Oct 26, 2021 10:10 am
I think I figured it out partially. If you remember, I had theorized that the duration of ZB has changed. I spent a lot of time trying to find historical duration or DV01 data and couldn't find anything. If that didn't work my plan was to calculate relative stdev in different periods other than the previous 10 years. But this morning I remembered zkn had also done 3 and 5 year comparisons. Well guess what? They seem to confirm my suspicion.

Here are the stdev of ZB vs the 10-20 year cash bond index:

10 years: 8.40% / 7.75%
5 years: 8.80% / 8.60%
3 years: 9.50% / 9.94%

You can see over the full 10 years it is .65% higher but the last three it is .44% lower. This would suggest that over the first 5 years it was over 1% higher. Which is a pretty big swing from being over 1% higher to .44% lower. That's a pretty big change in duration. None of the other contracts show anything like that. This would tend to confirm that the current duration of ZB is the 11.5 on the CME page, but that it used to be higher. Maybe we could get exact stdev for the last 3-12 months vs a 5 year period 5-10 years ago. And then do the proper math to convert to relative durations.

If ZB is only 11.5 currently, then it's probably not too bad. But a couple things to consider about your IEF backtest:
I think what you are seeing is the CTD bond for the contract is shifting to closer maturities. The range of acceptable maturities for the ZB basket is very wide. As the rate drops well below the notional 6% rate, the CTD tends to favor closer maturities with higher coupons. There are several deliverable bonds that are nearly CTD that have even lower duration (DV01) than the current CTD.

If we ever had a situation where rates started moving up near 6% I think you would find that the duration for ZF and TN make small shifts because they have narrow acceptable ranges of maturity. ZN and ZB would have big swings in duration.

I found a blog post from several years ago that suggested that ZB could be traded against TLT just like ES can be traded against SPY. That doesn't seem to be the case anymore. UB would be a better match for TLT now.
Yeah I knew it does that but I don't really understand why it would shift more in the last 10 years since the yield has been under 6% that whole time. But it does appear to have shifted, while other durations have not, possibly related to the wider bands as you mention.
User avatar
Bentonkb
Posts: 83
Joined: Sun Mar 07, 2021 5:47 am

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

Post by Bentonkb »

skierincolorado wrote: Tue Oct 26, 2021 2:38 pm
Bentonkb wrote: Tue Oct 26, 2021 2:09 pm
skierincolorado wrote: Tue Oct 26, 2021 10:10 am
I think I figured it out partially. If you remember, I had theorized that the duration of ZB has changed. I spent a lot of time trying to find historical duration or DV01 data and couldn't find anything. If that didn't work my plan was to calculate relative stdev in different periods other than the previous 10 years. But this morning I remembered zkn had also done 3 and 5 year comparisons. Well guess what? They seem to confirm my suspicion.

Here are the stdev of ZB vs the 10-20 year cash bond index:

10 years: 8.40% / 7.75%
5 years: 8.80% / 8.60%
3 years: 9.50% / 9.94%

You can see over the full 10 years it is .65% higher but the last three it is .44% lower. This would suggest that over the first 5 years it was over 1% higher. Which is a pretty big swing from being over 1% higher to .44% lower. That's a pretty big change in duration. None of the other contracts show anything like that. This would tend to confirm that the current duration of ZB is the 11.5 on the CME page, but that it used to be higher. Maybe we could get exact stdev for the last 3-12 months vs a 5 year period 5-10 years ago. And then do the proper math to convert to relative durations.

If ZB is only 11.5 currently, then it's probably not too bad. But a couple things to consider about your IEF backtest:
I think what you are seeing is the CTD bond for the contract is shifting to closer maturities. The range of acceptable maturities for the ZB basket is very wide. As the rate drops well below the notional 6% rate, the CTD tends to favor closer maturities with higher coupons. There are several deliverable bonds that are nearly CTD that have even lower duration (DV01) than the current CTD.

If we ever had a situation where rates started moving up near 6% I think you would find that the duration for ZF and TN make small shifts because they have narrow acceptable ranges of maturity. ZN and ZB would have big swings in duration.

I found a blog post from several years ago that suggested that ZB could be traded against TLT just like ES can be traded against SPY. That doesn't seem to be the case anymore. UB would be a better match for TLT now.
Yeah I knew it does that but I don't really understand why it would shift more in the last 10 years since the yield has been under 6% that whole time. But it does appear to have shifted, while other durations have not, possibly related to the wider bands as you mention.
The yield has been below 6% for a long time, true, but the basket used to include bonds near 6%, so it didn't have a huge effect. As the supply of bonds with coupons near 6% was depleted, or just dropped out of the deliverable basket, the ZB duration had to shift to the extreme short end of the acceptable range. It takes a while for all of the high coupon bonds to drop out of the deliverable basket.
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

Bentonkb wrote: Tue Oct 26, 2021 3:06 pm
skierincolorado wrote: Tue Oct 26, 2021 2:38 pm
Bentonkb wrote: Tue Oct 26, 2021 2:09 pm
skierincolorado wrote: Tue Oct 26, 2021 10:10 am
I think I figured it out partially. If you remember, I had theorized that the duration of ZB has changed. I spent a lot of time trying to find historical duration or DV01 data and couldn't find anything. If that didn't work my plan was to calculate relative stdev in different periods other than the previous 10 years. But this morning I remembered zkn had also done 3 and 5 year comparisons. Well guess what? They seem to confirm my suspicion.

Here are the stdev of ZB vs the 10-20 year cash bond index:

10 years: 8.40% / 7.75%
5 years: 8.80% / 8.60%
3 years: 9.50% / 9.94%

You can see over the full 10 years it is .65% higher but the last three it is .44% lower. This would suggest that over the first 5 years it was over 1% higher. Which is a pretty big swing from being over 1% higher to .44% lower. That's a pretty big change in duration. None of the other contracts show anything like that. This would tend to confirm that the current duration of ZB is the 11.5 on the CME page, but that it used to be higher. Maybe we could get exact stdev for the last 3-12 months vs a 5 year period 5-10 years ago. And then do the proper math to convert to relative durations.

If ZB is only 11.5 currently, then it's probably not too bad. But a couple things to consider about your IEF backtest:
I think what you are seeing is the CTD bond for the contract is shifting to closer maturities. The range of acceptable maturities for the ZB basket is very wide. As the rate drops well below the notional 6% rate, the CTD tends to favor closer maturities with higher coupons. There are several deliverable bonds that are nearly CTD that have even lower duration (DV01) than the current CTD.

If we ever had a situation where rates started moving up near 6% I think you would find that the duration for ZF and TN make small shifts because they have narrow acceptable ranges of maturity. ZN and ZB would have big swings in duration.

I found a blog post from several years ago that suggested that ZB could be traded against TLT just like ES can be traded against SPY. That doesn't seem to be the case anymore. UB would be a better match for TLT now.
Yeah I knew it does that but I don't really understand why it would shift more in the last 10 years since the yield has been under 6% that whole time. But it does appear to have shifted, while other durations have not, possibly related to the wider bands as you mention.
The yield has been below 6% for a long time, true, but the basket used to include bonds near 6%, so it didn't have a huge effect. As the supply of bonds with coupons near 6% was depleted, or just dropped out of the deliverable basket, the ZB duration had to shift to the extreme short end of the acceptable range. It takes a while for all of the high coupon bonds to drop out of the deliverable basket.
But if aything the shorter durations would have higher coupon bonds in them... like ZF could have a 26 year old 30 year bond in its basket. So it should have shifted too. But there's really no shift for the other contracts and a pretty big shift for ZB. Maybe it is just the wider range.
comeinvest
Posts: 2709
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 »

skierincolorado wrote: Tue Oct 26, 2021 10:10 am
comeinvest wrote: Mon Oct 25, 2021 6:10 pm
skierincolorado wrote: Fri Oct 22, 2021 1:01 pm
comeinvest wrote: Fri Oct 22, 2021 3:36 am I made two adjustments to the treasury futures portion of my portfolio tonight. They are mostly not meant to time the market (although I also played around a little bit), but to transition to my final strategic allocation:
- Eliminated all remaining short ZT positions that I established as a partial hedge of my long treasury futures to rising rates, when 2-year rates were on the floor i.e. close to nil. I did this today because I found out that the 2-year carry (ca. 0.8%, with yield to maturity of ca. 0.45%) just surpassed the carry of the 5-year (ca. 1.85% if my math was right) on a duration-adjusted basis, which creeped me out, and I couldn't expect the alpha from the pair trade any more that I expected based on papers advocating a dynamic positioning based on current carry per duration. I pocketed a nice profit, and the hedge did its job. I'm still playing a similar game in the European market.
- Transitioned from some of the ZB to TN+ZN. I should have done this on Monday directly from UB instead of moving from UB to ZB, but someone from Colorado ;) convinced me just this week to go with even shorter durations, so I ended up doing it in two steps. In the meantime, I made a tiny profit ("profit" as in "less loss") by holding ZB instead of TN+ZN for those 4 days. TN+ZN has almost the same DV01 as ZB, which made the transition easy on a DV01-neutral basis. I went with TN and ZN for diversification, because I have already a larger position in ZF.

I used limit orders at the bid (for buy) and ask (for sell) in the night. The volume in the night is lower, but I think the bid and ask sizes are lower too, such that passive limit orders might profit from liquidity based fills ("patient trading") just like by day. The spread is the same as by day. I'm not sure if it's statistically better to trade by day or night for minimizing transaction cost, but it seems to work well either way.

I was kind of sad to let my UB go this week, but it dawned me that it became a curse word in this message thread ;) My allocation comes now closer to an equidistribution between ZF/ZN/TN/ZB on duration-adjusted basis (it's currently still somewhere between a duration-based and notional-value-based equidistribution). I still have to define strict ratios and a suitable rebalancing strategy, so that I don't have to think too much every time in the future. Comments are welcome.

Assuming the duraiton on ZB is 11.5 years, I think this is good, but I think ZB is more like 17 years. Rules for the dynamic and static parts would be good too of course.

For ZB though did we ever figure out what the duration is? CME says it's currently 11.5... but the historical performance was more volatile that the S&P 10-20 year bond index, which has duration of 14.99 years. That would suggest it has duration of around 17 years which is quite similar to VUSTX which is the one I use in my backtesting since 1991 to show how much better VFITX is. While I am fully on-board with the diversification of not just being ZF, I would feel strongly that 17 years is too long if taking equal nominal amounts of each. If equal weighted by risk I would be on board I think. It might look something like this. I'm blue, you're orange, red is what you were before (or worse). I can't confidently say blue is better than orange - it's close enough. And this assumes ZB is 17 years duration. If it's shorter like 11.5 years, it would probably be the same as blue.

If you do do it, I like scaling on DV01 because it would mean less ZB than other measures of risk, like historical stdev or max-draw. As long as you're staying at or above that orange line, I think you're good. There are even a couple short periods where the orange line outperforms blue (2021, 1992-1994).

https://www.portfoliovisualizer.com/bac ... on4_3=-110
My proposed solution, an ITT strategy diversified across ZF/ZN/TN/ZB (weights tbd), would not have anything to do with VUSTX which would represent UB. I think we still have not found an ETF equivalent to ZB. A duration-neutral alternative to ZF-only equidistributed based on current durations would be (omitting ZB for the moment, but that would make little difference if we weigh by constant duration risk): https://www.portfoliovisualizer.com/bac ... on4_3=-163

Which shows that
1. It makes little difference; ZF-only wins, but the time series seems random so that I would not feel confident extrapolating the small cumulative difference into the future based on a 1-country 20-year-only time series; there was practically no difference in the first third of the 20-year period
2. There is little diversification, but maybe one could avoid the risk of potential future funding cost anomalies by diversifying across maturities (diversification of implementation risk); only a small change in funding cost / slippage e.g. in ZF could have a relatively large effect on results due to leverage
3. Yearly differences between the ITT varieties were up to about 3%. If you want do avoid being on the losing end for a year, then invest in several varieties of ITT; however, most of the difference is typically made up soon after a losing year.
4. It's a challenge to do meaningful backtests without having past durations of futures, ETFs, or both.

Have we figured out yet what is the current duration of ZB? The stddev aficionados in this thread said that it must be much higher than what the treasury analytics page says. (Can we verify this for the stddev of the current contract?) I personally feel as long as I can't figure out what should be such a simple thing, I don't really understand treasury futures and I don't feel confident deciding on a treasure futures allocation and distribution ratios, even if it doesn't include ZB.
I think I figured it out partially. If you remember, I had theorized that the duration of ZB has changed. I spent a lot of time trying to find historical duration or DV01 data and couldn't find anything. If that didn't work my plan was to calculate relative stdev in different periods other than the previous 10 years. But this morning I remembered zkn had also done 3 and 5 year comparisons. Well guess what? They seem to confirm my suspicion.

Here are the stdev of ZB vs the 10-20 year cash bond index:

10 years: 8.40% / 7.75%
5 years: 8.80% / 8.60%
3 years: 9.50% / 9.94%

You can see over the full 10 years it is .65% higher but the last three it is .44% lower. This would suggest that over the first 5 years it was over 1% higher. Which is a pretty big swing from being over 1% higher to .44% lower. That's a pretty big change in duration. None of the other contracts show anything like that. This would tend to confirm that the current duration of ZB is the 11.5 on the CME page, but that it used to be higher. Maybe we could get exact stdev for the last 3-12 months vs a 5 year period 5-10 years ago. And then do the proper math to convert to relative durations.

If ZB is only 11.5 currently, then it's probably not too bad. But a couple things to consider about your IEF backtest:

1 You chose a 2:3 ratio to replace IEF with VFITX. The conclusions are highly sensitive on the ratio you pick. 2:3 seems correct. The durations are 5.2:8 which would be a bit higher, but I think a lower ratio is correct for reasons we've discussed before.

3. There was a flattening of close to 1% during this period, giving a slight advantage to longer durations.

4. I we bump the ratio of VFITX up slightly and start in 2005 (per 1 and 3), VFITX wins solidly. Add in some ZN and it might start to look significant.

5. Even the current ZB duration is quite a bit longer than IEF (11.5 vs 8).


All that said I have absolutely no problem with holding equal risk in ZB as in ZF, ZN, and TN assuming that the duration is 11.5 years. I think you'd see some small dropoff in the backtesting, but not a convincing amount. If it was equal nominal amounts, I think you'd see convincing dropoff.

You'd have to have a lot invested because you'd need a lot of contracts to get that balance. For 1 ZB you'd need 1.5 TN, 2 ZN and 3 ZF. If you round the TN down to 1, that's still 7 contracts with pretty long duration - more than I need personally.
Regarding the ZB mystery: Thanks for your investigation. I hope zkn or somebody can generate stddev calcs over some more fine-grained time periods.

If you look at the treasury analytics page, you see that the current CTD has a coupon of 5%, while the OTR has a coupon of 1-3/4%. The high coupon of the current CTD might be the reason for a relatively low duration compared to history. I think the OTR will never become a CTD, because all CTDs seem to have a maturity close to the shortest allowable. But the coupon of the CTD and therefore the duration fluctuates from contract to contract. That's my understanding at least.

I am still not comprehending why the current OTR with its longer maturity AND lower coupon has a much lower DV01 than the CTD. Shouldn't it be the other way around? Both longer maturity and lower coupon should result in higher interest rate sensitivity, or not? I also posted a question that nobody commented on yet: viewtopic.php?p=6287887#p6287887

Regarding the charts: I tried to do a DV01-neutral replacement of VFITX, but I was off by 2% or so. The DV01-neutral "diversified" replacement of VFITS using the 5.2:8 ratio looks almost the same:
https://www.portfoliovisualizer.com/bac ... on4_3=-160

Regarding the accuracy of the balancing between ZF/ZN/TN/ZB in the "diversified" scenario for smaller portfolios: Given the already small differences in performance between the contracts, and the fact that we have 4 (not 2 or 3) contracts in our ensemble, I think the ratios don't have to be exact, as long as the balancing / rebalancing is halfway consistent. Round one contract down, round the neighboring contract up, etc., until the total duration exposure according to the asset allocation plan is achieved. If you are concerned about more accurate ratios, you might look at the new micro treasury futures https://www.cmegroup.com/trading/intere ... tures.html ; I think the folks on Elitetrader speculate they might be quite efficient as they would basically mimic the traditional treasury futures, and inefficiencies would be arbitraged away. But I personally am already confused enough about the traditional futures, so I don't want to add yet another type of interest rate product at the moment.
Last edited by comeinvest on Tue Oct 26, 2021 6:57 pm, edited 3 times in total.
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

comeinvest wrote: Tue Oct 26, 2021 6:17 pm

Regarding the ZB mystery: Thanks for your investigation. I hope zkn or somebody can generate stddev calcs over some more fine-grained time periods.

If you look at the treasury analytics page, you see that the current CTD has a coupon of 5%, while the OTR has a coupon of 1-3/4%. The high coupon of the current CTD might be the reason for a relatively low duration compared to history. I think the OTR will never become a CTD, because all CTDs seem to have a maturity close to the shortest allowable. But the coupon of the CTD and therefore the duration fluctuates from contract to contract. That's my understanding at least.

I am still not comprehending why the current OTR with its longer maturity AND lower coupon has a much lower DV01 than the CTD. Shouldn't it be the other way around? Both longer maturity and lower coupon should result in higher interest rate sensitivity, or not? I also posted a question that nobody commented on yet: viewtopic.php?p=6287887#p6287887

Regarding the charts: I tried to do a DV01-neutral replacement of VFITX, but I was off by 2% or so. The DV01-neutral "diversified" replacement of VFITS using the 5.2:8 ratio looks almost the same:
https://www.portfoliovisualizer.com/bac ... on4_3=-160

Regarding the accuracy of the balancing between ZF/ZN/TN/ZB in the "diversified" scenario: I'm probably a bit older than you and therefore might have saved a little more. However, given the already small differences between the contract, and the fact that we have 4 (not 2 or 3) contracts in our ensemble, I think the ratios don't have to be exact, as long as the balancing / rebalancing is halfway consistent. Round one contract down, round the neighboring up, etc., until the total duration exposure according to the asset allocation plan is achieved. If you are concerned about more accurate ratios, you might look at the new micro treasury futures https://www.cmegroup.com/trading/intere ... tures.html ; I think the folks on Elitetrader speculate they might be quite efficient as they would basically mimic the traditional treasury futures and inefficiencies would be arbitraged away. But I personally am already confused enough so I don't want to add yet another type of interest rate product at the moment.
Yeah like you I'm not too concerned about exact ratios, but I would be concerned about being too heavy into ZB consistently which would happen if one didn't have enough to buy a few ZFs and a couple ZNs to go with it. I personally wouldn't want to be longer than ZN or TN, but if going into ZB it definitely should be no more than proportional risk with 2 or 3 other contracts, but with limited funds that could be tough. For a dynamic market timing one could do more ZB occasionally, but for a long term position the backtest would drop off too much if holding more than ~30% of one's risk in ZB.

The DV01 is higher for the futures contract than the OTR due to the CF. The DV01 for the CTD is lower, but after dividing by the CF it is much higher.

My understanding is the Micros are very different and are basically a zero sum game. There is no net profit to be had. Seems weird to me but I read for a while and that was my conclusion and a few other people had the same conclusion. But yeah don't want to waste anymore time on them personally since they aren't needed.
comeinvest
Posts: 2709
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 »

skierincolorado wrote: Tue Oct 26, 2021 6:38 pm
comeinvest wrote: Tue Oct 26, 2021 6:17 pm

Regarding the ZB mystery: Thanks for your investigation. I hope zkn or somebody can generate stddev calcs over some more fine-grained time periods.

If you look at the treasury analytics page, you see that the current CTD has a coupon of 5%, while the OTR has a coupon of 1-3/4%. The high coupon of the current CTD might be the reason for a relatively low duration compared to history. I think the OTR will never become a CTD, because all CTDs seem to have a maturity close to the shortest allowable. But the coupon of the CTD and therefore the duration fluctuates from contract to contract. That's my understanding at least.

I am still not comprehending why the current OTR with its longer maturity AND lower coupon has a much lower DV01 than the CTD. Shouldn't it be the other way around? Both longer maturity and lower coupon should result in higher interest rate sensitivity, or not? I also posted a question that nobody commented on yet: viewtopic.php?p=6287887#p6287887

Regarding the charts: I tried to do a DV01-neutral replacement of VFITX, but I was off by 2% or so. The DV01-neutral "diversified" replacement of VFITS using the 5.2:8 ratio looks almost the same:
https://www.portfoliovisualizer.com/bac ... on4_3=-160

Regarding the accuracy of the balancing between ZF/ZN/TN/ZB in the "diversified" scenario: I'm probably a bit older than you and therefore might have saved a little more. However, given the already small differences between the contract, and the fact that we have 4 (not 2 or 3) contracts in our ensemble, I think the ratios don't have to be exact, as long as the balancing / rebalancing is halfway consistent. Round one contract down, round the neighboring up, etc., until the total duration exposure according to the asset allocation plan is achieved. If you are concerned about more accurate ratios, you might look at the new micro treasury futures https://www.cmegroup.com/trading/intere ... tures.html ; I think the folks on Elitetrader speculate they might be quite efficient as they would basically mimic the traditional treasury futures and inefficiencies would be arbitraged away. But I personally am already confused enough so I don't want to add yet another type of interest rate product at the moment.
Yeah like you I'm not too concerned about exact ratios, but I would be concerned about being too heavy into ZB consistently which would happen if one didn't have enough to buy a few ZFs and a couple ZNs to go with it. I personally wouldn't want to be longer than ZN or TN, but if going into ZB it definitely should be no more than proportional risk with 2 or 3 other contracts, but with limited funds that could be tough. For a dynamic market timing one could do more ZB occasionally, but for a long term position the backtest would drop off too much if holding more than ~30% of one's risk in ZB.

The DV01 is higher for the futures contract than the OTR due to the CF. The DV01 for the CTD is lower, but after dividing by the CF it is much higher.

My understanding is the Micros are very different and are basically a zero sum game. There is no net profit to be had. Seems weird to me but I read for a while and that was my conclusion and a few other people had the same conclusion. But yeah don't want to waste anymore time on them personally since they aren't needed.
I hope to let things run more on their own in the long run, but I watch my portfolio almost daily these days just out of curiosity, and to learn about the futures products. I'm noticing that on many days the ZF moves in the opposite direction to the other 3 contracts that we are considering. However this perceived diversification unfortunately doesn't result in higher sharpe ratios for the diversified portfolio as per the PV charts. My main argument for having longer durations in the mix is diversification. The instant carry of the ZF is currently only about equal, but not bigger than that of the longer durations. I would like to have some dry powder to shift to the shorter end, once the belly has higher carry (as in coupon yield to maturity minus funding cost plus rolldown yield), and/or the curve becomes really flat, both of which I think were often the case historically. The other argument, as we said before, is diversification in case of permanent flattening (similar to most other developed countries - the U.S. is kind of still the exception), and diversification of unforeseen implementation cost. Consider this: There was a measurable increase in funding cost of equity index futures around 2015; before 2015 it was on average about zero and often negative. The same might happen to treasury futures at any time, what do I know. Which would tilt our results from this thread and change the case for treasury futures. Another analogy for the diversification argument: Australia's stock market had the best risk-adjusted return worldwide 1900-2016. Do you invest all your equity allocation in Australia?
comeinvest
Posts: 2709
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 »

skierincolorado wrote: Tue Oct 26, 2021 6:38 pm My understanding is the Micros are very different and are basically a zero sum game. There is no net profit to be had. Seems weird to me but I read for a while and that was my conclusion and a few other people had the same conclusion. But yeah don't want to waste anymore time on them personally since they aren't needed.
My conclusion was the opposite. On the surface it looks like the Micros are just a bet on interest rates. But if you think of it, the Micros must mimic the return of traditional treasuries, including coupon yield and rolldown return, because of a simple no-arbitrage argument: If they were not mimicking the return of traditional treasuries, I could invest in a nice portfolio of treasury futures, and hedge all my interest rate risk with Micros of the corresponding duration, and would pocket a nice return with no risk. Similar to LIBOR or SOFR futures: In one of the papers that you read, they use interest rate futures to get almost the exact same return as with treasury futures. Interest rate futures are not (not just) a bet on interest rates, but reflect the term structure of the treasury yield curve, including the term premium. One of the posters in the HFEA thread used strips of STIR futures to implement HFEA.

But I'm happy to stand corrected. Do you have a link to the rationale or discussion that says there is no net profit to be had from Micros?
Last edited by comeinvest on Tue Oct 26, 2021 9:50 pm, edited 1 time in total.
Kbg
Posts: 355
Joined: Thu Mar 23, 2017 11:33 am

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

Post by Kbg »

+1 on the micros information.

"Normalizing" them to the larger contracts would be super nice to know (e.g. X micros = 1 large)
comeinvest
Posts: 2709
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 »

Kbg wrote: Tue Oct 26, 2021 9:30 pm +1 on the micros information.

"Normalizing" them to the larger contracts would be super nice to know (e.g. X micros = 1 large)
Keep in mind that as you go with shorter durations like ZF as SIC suggests, your granularity is already pretty fine. The ZF currently trades at about 121k. Normalized to LTT (UB or Vanguard's VGLT), the duration of one ZF corresponds to only about $30k worth of LTT. (SIC suggests a lightly different ratio of relative risk.) And you can always do "padding" with ETFs. Question is how fine a granularity to we need in relation to NAV for effective rebalancing?
Kbg
Posts: 355
Joined: Thu Mar 23, 2017 11:33 am

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

Post by Kbg »

Sorry if I misread your reply…by normalize I meant (notionally)

Short 10 5yr micro treasury futures = 1 ZF long

Maybe there’s no relationship other than contract length, IDK?
comeinvest
Posts: 2709
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 »

skierincolorado wrote: Tue Oct 26, 2021 10:43 am What would one even buy that had the same duration as zb other than direct treasuries? Is there a fund with that duration?

If someone doesn't have the capital to diversify even at the shorter durations, it's fine to just own 1 ZF. This still backtests much better than owning TMF for example. Just maybe a hair worse than diversifying across ZT, ZF, ZN. One could buy some VFISX, IEF, or TYD to supplement to get to desired AA. But just ZF is fine too IMO given how well it backtests.

For example with 100k in capital:

60k VTI
3 MES (68k)
1 ZF ( 121k) = ~100k @ 5 yr duration
20k TYD (60k) = ~100k @ 5 yr duration

would be a 140/200 stock/bond allocation with the bond duration standardized at 5 years, but taking ~50/50 risk at 4.3 (ZF) and 8 years (TYD).

It would arguably be more efficient to implement the above as

6 MES (135k)
1 ZF ( 121k) = ~100k @ 5 yr duration
60k IEF (60k) = ~100k @ 5 yr duration

depending on what you assume the financing costs internal to TYD are. If you assume there is .2% finance premium above LIBOR internal to TYD, then TYD costs .5% above LIBOR including fees per $. Since that is likely higher than MES, it would be better to swap the TYD + VTI for MES and + IEF. Switching the TYD to IEF nets .35% (.5-.15) in cost above LIBOR, while switching from VTI to MES loses you maybe .3% (-.05% + .35%).

Either implementation would be fine really. The main point is both are plenty diversified without going longer than 8 years on the curve (TYD/IEF).

I'd actually probably me more inclined to diverisfy in the lower direction. For example buying ZF (4.3) or ZN (6.3) and diversifying with BSV/VFISX (1-4). Or buying ZT (1.8) and diversifying with VGIT/VFITX (4-7).
I personally would not diversify using any treasury ETFs. It takes away from space for other potential sources of return and arguably more meaningful portfolio diversifiers like factor tilted equities. Otherwise I agree with your very accurate analysis.
comeinvest
Posts: 2709
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 »

skierincolorado wrote: Tue Oct 26, 2021 3:17 pm But if aything the shorter durations would have higher coupon bonds in them... like ZF could have a 26 year old 30 year bond in its basket. So it should have shifted too. But there's really no shift for the other contracts and a pretty big shift for ZB. Maybe it is just the wider range.
Why did CME actually specify the eligible maturity range for each treasury bond? Wouldn't it make more sense to specify the eligible duration range? For more stable, predictable, and meaningful characterization of each futures contract.
User avatar
Bentonkb
Posts: 83
Joined: Sun Mar 07, 2021 5:47 am

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

Post by Bentonkb »

Kbg wrote: Tue Oct 26, 2021 11:29 pm Sorry if I misread your reply…by normalize I meant (notionally)

Short 10 5yr micro treasury futures = 1 ZF long

Maybe there’s no relationship other than contract length, IDK?
Here is the response (8/17/2021) I got from TastyWorks regarding the congruency between /S30Y and /ZB or /UB:

"With regard to duration, the /S30Y index is calculated using the on-the-run issue, and therefore the index is not duration adjusted. We need to take that into consideration since /ZB's deliverable includes off-the-run issues / cheapest to deliver with 15-25 years remaining until maturity. And, yes, you're correct with the DV01 for /ZB is currently $195.62 as of today.
https://www.cmegroup.com/tools-informat ... ytics.html

One can argue that a more true comparison would be the /UB as it is what's considered the true 30-year (25 year + deliverable). The /S30Y is approximately 1/100th of the /UB for what it's worth. I believe the /ZB was chosen as a benchmark simply because it's more universal throughout the bond futures world given product tenure and the fact options trade on it.

As far a DV01 for the /S30Y, I would have to defer to our folks at The Small Exchange as I'm not sure what they are using for their figures but here is my best guess as to the benchmark comparison:

[ (.01x 21.9074) x 101,660 ] x .01 = 244.98 in lieu of the 195.44 for /ZB

* I used the modified duration of the /UB given it's congruency
** $101,660 is the current price of the on-the-run 30 year to my knowledge, sans a Bloomberg terminal."
User avatar
Bentonkb
Posts: 83
Joined: Sun Mar 07, 2021 5:47 am

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

Post by Bentonkb »

comeinvest wrote: Tue Oct 26, 2021 6:17 pm I am still not comprehending why the current OTR with its longer maturity AND lower coupon has a much lower DV01 than the CTD. Shouldn't it be the other way around? Both longer maturity and lower coupon should result in higher interest rate sensitivity, or not?
The cash DV01 is lower, but the futures DV01 is higher. That is the CF at work again.
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

comeinvest wrote: Wed Oct 27, 2021 2:58 am
skierincolorado wrote: Tue Oct 26, 2021 10:43 am What would one even buy that had the same duration as zb other than direct treasuries? Is there a fund with that duration?

If someone doesn't have the capital to diversify even at the shorter durations, it's fine to just own 1 ZF. This still backtests much better than owning TMF for example. Just maybe a hair worse than diversifying across ZT, ZF, ZN. One could buy some VFISX, IEF, or TYD to supplement to get to desired AA. But just ZF is fine too IMO given how well it backtests.

For example with 100k in capital:

60k VTI
3 MES (68k)
1 ZF ( 121k) = ~100k @ 5 yr duration
20k TYD (60k) = ~100k @ 5 yr duration

would be a 140/200 stock/bond allocation with the bond duration standardized at 5 years, but taking ~50/50 risk at 4.3 (ZF) and 8 years (TYD).

It would arguably be more efficient to implement the above as

6 MES (135k)
1 ZF ( 121k) = ~100k @ 5 yr duration
60k IEF (60k) = ~100k @ 5 yr duration

depending on what you assume the financing costs internal to TYD are. If you assume there is .2% finance premium above LIBOR internal to TYD, then TYD costs .5% above LIBOR including fees per $. Since that is likely higher than MES, it would be better to swap the TYD + VTI for MES and + IEF. Switching the TYD to IEF nets .35% (.5-.15) in cost above LIBOR, while switching from VTI to MES loses you maybe .3% (-.05% + .35%).

Either implementation would be fine really. The main point is both are plenty diversified without going longer than 8 years on the curve (TYD/IEF).

I'd actually probably me more inclined to diverisfy in the lower direction. For example buying ZF (4.3) or ZN (6.3) and diversifying with BSV/VFISX (1-4). Or buying ZT (1.8) and diversifying with VGIT/VFITX (4-7).
I personally would not diversify using any treasury ETFs. It takes away from space for other potential sources of return and arguably more meaningful portfolio diversifiers like factor tilted equities. Otherwise I agree with your very accurate analysis.
Yeah this is just the most efficient implementation of that specific AA. For factor tilts take the first example and swap the VTI for some factor tilts. Would need to weigh against the benefit of factors vs the costs, although as my math showed the costs between the two implementations are nearly identical and at most differ by .05%. Despite the costs I would probably prefer the first example and do some factor tilts. Although in reality I don't prefer either example, because I don't think diversifying all the way out to TYD durations is necessary esp if it is pushing one into higher cost investment vehicles like TYD... swap the TYD for a ZT ZF or ZN. Was more for the poster who wanted to go all the way out to 8 years duration but with limited funds, not necessary in my opinion but that's how one could do it.
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

comeinvest wrote: Tue Oct 26, 2021 7:25 pm
skierincolorado wrote: Tue Oct 26, 2021 6:38 pm My understanding is the Micros are very different and are basically a zero sum game. There is no net profit to be had. Seems weird to me but I read for a while and that was my conclusion and a few other people had the same conclusion. But yeah don't want to waste anymore time on them personally since they aren't needed.
My conclusion was the opposite. On the surface it looks like the Micros are just a bet on interest rates. But if you think of it, the Micros must mimic the return of traditional treasuries, including coupon yield and rolldown return, because of a simple no-arbitrage argument: If they were not mimicking the return of traditional treasuries, I could invest in a nice portfolio of treasury futures, and hedge all my interest rate risk with Micros of the corresponding duration, and would pocket a nice return with no risk. Similar to LIBOR or SOFR futures: In one of the papers that you read, they use interest rate futures to get almost the exact same return as with treasury futures. Interest rate futures are not (not just) a bet on interest rates, but reflect the term structure of the treasury yield curve, including the term premium. One of the posters in the HFEA thread used strips of STIR futures to implement HFEA.

But I'm happy to stand corrected. Do you have a link to the rationale or discussion that says there is no net profit to be had from Micros?
I see the arbitrage argument. That would mean that the micros trade below the benchmark until settlement. As the contract nears settlement, the yield on the contract would converge to the benchmark. We'd have to confirm that it works like this in practice and calculate implied financing costs. Not something I'm interested in doing but others can... 1 ZF is the same risk as a mere 10k of TMF. Anybody investing more than 25-30k could reasonably buy a ZF or ZT. Or buy TYD until they reach 30k.
comeinvest
Posts: 2709
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 »

skierincolorado wrote: Wed Oct 27, 2021 11:36 am
comeinvest wrote: Tue Oct 26, 2021 7:25 pm
skierincolorado wrote: Tue Oct 26, 2021 6:38 pm My understanding is the Micros are very different and are basically a zero sum game. There is no net profit to be had. Seems weird to me but I read for a while and that was my conclusion and a few other people had the same conclusion. But yeah don't want to waste anymore time on them personally since they aren't needed.
My conclusion was the opposite. On the surface it looks like the Micros are just a bet on interest rates. But if you think of it, the Micros must mimic the return of traditional treasuries, including coupon yield and rolldown return, because of a simple no-arbitrage argument: If they were not mimicking the return of traditional treasuries, I could invest in a nice portfolio of treasury futures, and hedge all my interest rate risk with Micros of the corresponding duration, and would pocket a nice return with no risk. Similar to LIBOR or SOFR futures: In one of the papers that you read, they use interest rate futures to get almost the exact same return as with treasury futures. Interest rate futures are not (not just) a bet on interest rates, but reflect the term structure of the treasury yield curve, including the term premium. One of the posters in the HFEA thread used strips of STIR futures to implement HFEA.

But I'm happy to stand corrected. Do you have a link to the rationale or discussion that says there is no net profit to be had from Micros?
I see the arbitrage argument. That would mean that the micros trade below the benchmark until settlement. As the contract nears settlement, the yield on the contract would converge to the benchmark. We'd have to confirm that it works like this in practice and calculate implied financing costs. Not something I'm interested in doing but others can... 1 ZF is the same risk as a mere 10k of TMF. Anybody investing more than 25-30k could reasonably buy a ZF or ZT. Or buy TYD until they reach 30k.
Regarding Micros: What do you mean by benchmark? (EDIT: I think you mean the actual current 10-year / 30-year etc. rates). Yes I would think so. Just like the LIBOR futures forward rate is NOT the actual expected rate, but reflects the term premium.

Regarding granularity of the contracts: The granularity must be small enough not only in relation to the portfolio size, but also to facilitate effective rebalancing e.g. on being +-5% out of sync with respect to the target percentage vs. equities, as rebalancing is part of the performance drivers of this strategy, or not?
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

comeinvest wrote: Wed Oct 27, 2021 12:29 pm
skierincolorado wrote: Wed Oct 27, 2021 11:36 am
comeinvest wrote: Tue Oct 26, 2021 7:25 pm
skierincolorado wrote: Tue Oct 26, 2021 6:38 pm My understanding is the Micros are very different and are basically a zero sum game. There is no net profit to be had. Seems weird to me but I read for a while and that was my conclusion and a few other people had the same conclusion. But yeah don't want to waste anymore time on them personally since they aren't needed.
My conclusion was the opposite. On the surface it looks like the Micros are just a bet on interest rates. But if you think of it, the Micros must mimic the return of traditional treasuries, including coupon yield and rolldown return, because of a simple no-arbitrage argument: If they were not mimicking the return of traditional treasuries, I could invest in a nice portfolio of treasury futures, and hedge all my interest rate risk with Micros of the corresponding duration, and would pocket a nice return with no risk. Similar to LIBOR or SOFR futures: In one of the papers that you read, they use interest rate futures to get almost the exact same return as with treasury futures. Interest rate futures are not (not just) a bet on interest rates, but reflect the term structure of the treasury yield curve, including the term premium. One of the posters in the HFEA thread used strips of STIR futures to implement HFEA.

But I'm happy to stand corrected. Do you have a link to the rationale or discussion that says there is no net profit to be had from Micros?
I see the arbitrage argument. That would mean that the micros trade below the benchmark until settlement. As the contract nears settlement, the yield on the contract would converge to the benchmark. We'd have to confirm that it works like this in practice and calculate implied financing costs. Not something I'm interested in doing but others can... 1 ZF is the same risk as a mere 10k of TMF. Anybody investing more than 25-30k could reasonably buy a ZF or ZT. Or buy TYD until they reach 30k.
Regarding Micros: What do you mean by benchmark? (EDIT: I think you mean the actual current 10-year / 30-year etc. rates). Yes I would think so. Just like the LIBOR futures forward rate is NOT the actual expected rate, but reflects the term premium.

Regarding granularity of the contracts: The granularity must be small enough not only in relation to the portfolio size, but also to facilitate effective rebalancing e.g. on being +-5% out of sync with respect to the target percentage vs. equities, as rebalancing is part of the performance drivers of this strategy, or not?
I think people tend to overestimate the benefit of rebalancing. The primary benefit of diversification is that your risk is reduced because it is unlikely both will have a max-draw at the same time, allowing for more leverage.

For rebalancing to have any benefit, the times when stock prices go down must also be times when future stock returns are higher (relative to bonds). Of course, that's market timing and should not occur in theory, the whole reason stock prices went down was a market belief that future returns are lower.
comeinvest
Posts: 2709
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 »

skierincolorado wrote: Wed Oct 27, 2021 12:56 pm
comeinvest wrote: Wed Oct 27, 2021 12:29 pm Regarding granularity of the contracts: The granularity must be small enough not only in relation to the portfolio size, but also to facilitate effective rebalancing e.g. on being +-5% out of sync with respect to the target percentage vs. equities, as rebalancing is part of the performance drivers of this strategy, or not?
I think people tend to overestimate the benefit of rebalancing. The primary benefit of diversification is that your risk is reduced because it is unlikely both will have a max-draw at the same time, allowing for more leverage.

For rebalancing to have any benefit, the times when stock prices go down must also be times when future stock returns are higher (relative to bonds). Of course, that's market timing and should not occur in theory, the whole reason stock prices went down was a market belief that future returns are lower.
But the EMH aspect is not the whole story. I personally am surely looking forward (in relative terms) to the next macroeconomic crash when equities tank 35% and my treasury futures will be up 20%, which is when I will unload some treasury futures and rebalance for a, say, 20% * 20% = 4% of NAV of excess profit from that episode once everything is back to normal. (I am cognizant that the "once back to normal" may never happen. That's why we need to limit overall leverage. But it always happened in history.) And if rates go again to near zero during that episode, I would unload most if not all of my treasuries. I guess rebalancing should have no effect based on EMH, nor should ITT outperform LTT, but markets are also governed by liquidity, and are known to be mean-reverting. As you say, backtesting is what counts (although the future may be different from the past). I think there was a pretty good study of the rebalancing bonus in the HFEA thread where the additional return was thought to be in the order of 0.5%-1% p.a. if I remember right. I have to read it again. Of course it may be lower if treasuries and equities are more correlated in the future; but I think in case of macroeconomic crashes we can almost surely count on the Fed reducing rates.
Last edited by comeinvest on Wed Oct 27, 2021 4:18 pm, edited 2 times in total.
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

comeinvest wrote: Wed Oct 27, 2021 4:13 pm
skierincolorado wrote: Wed Oct 27, 2021 12:56 pm
comeinvest wrote: Wed Oct 27, 2021 12:29 pm Regarding granularity of the contracts: The granularity must be small enough not only in relation to the portfolio size, but also to facilitate effective rebalancing e.g. on being +-5% out of sync with respect to the target percentage vs. equities, as rebalancing is part of the performance drivers of this strategy, or not?
I think people tend to overestimate the benefit of rebalancing. The primary benefit of diversification is that your risk is reduced because it is unlikely both will have a max-draw at the same time, allowing for more leverage.

For rebalancing to have any benefit, the times when stock prices go down must also be times when future stock returns are higher (relative to bonds). Of course, that's market timing and should not occur in theory, the whole reason stock prices went down was a market belief that future returns are lower.
But the EMH aspect is not the whole story. I personally am surely looking forward (in relative terms) to the next macroeconomic crash when equities tank 35% and my treasury futures will be up 20%, which is when I will unload some treasury futures and rebalance for a, say, 20% * 20% = 4% of NAV of excess profit from that episode once everything is back to normal. And if rates go again to near zero during that episode, I would unload most if not all of my treasuries. I guess rebalancing should have no effect based on EMH, nor should ITT outperform LTT, but markets are also governed by liquidity, and known to be mean-reverting. As you say, backtesting is what counts (although the future may be different from the past). I think there was a pretty good study of the rebalancing bonus in the HFEA thread that was thought to be in the order of 0.5%-1% p.a. if I remember right. I have to read it again. Of course it may be lower if treasuries and equities are more correlated in the future; but I think in case of macroeconomic crashes we can almost surely count on the Fed reducing rates.
Yeah there does seem to be some benefit that does not exist under emh. But still the primary benefit is to reduce risk and allow more leverage. Because of this I would not worry if the rebalancing was crude and less granular. Adding a single zf to a 40k portfolio is an extra 5% of CAGR with almost no extra risk.. the max draw will be due to the equities and could even be reduced by the zf.
comeinvest
Posts: 2709
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 »

skierincolorado wrote: Wed Oct 27, 2021 4:16 pm Yeah there does seem to be some benefit that does not exist under emh. But still the primary benefit is to reduce risk and allow more leverage. Because of this I would not worry if the rebalancing was crude and less granular. Adding a single zf to a 40k portfolio is an extra 5% of CAGR with almost no extra risk.. the max draw will be due to the equities and could even be reduced by the zf.
Over your backtesting period and not adjusted for interest rate levels maybe, but according to expected term premia papers that you can read, the additional CAGR may be close to nothing in the future. IMHO we should not count on significant extra returns. The curve keeps flattening these days. Once it flattens enough and doesn't recover, like most other developed countries' curves (Europe and Japan) flattened and didn't recover, and with a 0.2% funding cost slippage, it would be kind of "game over" for this strategy. At best we could reap the one-time benefit from the rates approaching zero from the current ca. 1-1.5% level of ITTs, and hope that the carry minus cost is not negative while doing so. That's one possible scenario. I don't know if it happens. Just saying we shouldn't count on significant wealth creation from this strategy. Often times things evaporate once you are done with all backtesting and theorizing and start implementing something. I just started this strategy with a reduced percentage that I can increase when interest rates and expected term premia are generally higher. I hope the strategy will survive in the future.
Last edited by comeinvest on Wed Oct 27, 2021 8:18 pm, edited 4 times in total.
comeinvest
Posts: 2709
Joined: Mon Mar 12, 2012 6:57 pm

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

Post by comeinvest »

I know in this forum we don't do market timing, but I find it sometimes amusing to read the ING rate analysis that they publish almost daily. So many words about a simple curve. I learned however, how long-term rates staying stubbornly low could prevent the Fed from increasing short-term rates as planned.
https://seekingalpha.com/article/446230 ... r-us-rates
constructor
Posts: 17
Joined: Fri Sep 03, 2021 6:18 pm

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

Post by constructor »

30 year yield (1.96%) lower than 20 year yield (1.98%) yesterday (October 28). Yield curve flattened in October. Interesting...
https://www.treasury.gov/resource-cente ... data=yield
Kbg
Posts: 355
Joined: Thu Mar 23, 2017 11:33 am

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

Post by Kbg »

Quick question: For 1:3/Stock:ITTs, why not just go 100/300 and not delever?

At least in PV it doesn’t appear one would need to…please punch holes in the above.
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

Kbg wrote: Fri Oct 29, 2021 11:59 am Quick question: For 1:3/Stock:ITTs, why not just go 100/300 and not delever?

At least in PV it doesn’t appear one would need to…please punch holes in the above.
Take it back to 1955 in Simba.
comeinvest
Posts: 2709
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 »

skierincolorado wrote: Fri Oct 29, 2021 12:26 pm
Kbg wrote: Fri Oct 29, 2021 11:59 am Quick question: For 1:3/Stock:ITTs, why not just go 100/300 and not delever?

At least in PV it doesn’t appear one would need to…please punch holes in the above.
Take it back to 1955 in Simba.
Plus, any backtest without meaningful adjustments to current interest rate levels is meaningless.
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

comeinvest wrote: Fri Oct 29, 2021 3:45 pm
skierincolorado wrote: Fri Oct 29, 2021 12:26 pm
Kbg wrote: Fri Oct 29, 2021 11:59 am Quick question: For 1:3/Stock:ITTs, why not just go 100/300 and not delever?

At least in PV it doesn’t appear one would need to…please punch holes in the above.
Take it back to 1955 in Simba.
Plus, any backtest without meaningful adjustments to current interest rate levels is meaningless.
Agree, assuming you mean term premia and not the absolute level
comeinvest
Posts: 2709
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 »

skierincolorado wrote: Fri Oct 29, 2021 4:39 pm
comeinvest wrote: Fri Oct 29, 2021 3:45 pm
skierincolorado wrote: Fri Oct 29, 2021 12:26 pm
Kbg wrote: Fri Oct 29, 2021 11:59 am Quick question: For 1:3/Stock:ITTs, why not just go 100/300 and not delever?

At least in PV it doesn’t appear one would need to…please punch holes in the above.
Take it back to 1955 in Simba.
Plus, any backtest without meaningful adjustments to current interest rate levels is meaningless.
Agree, assuming you mean term premia and not the absolute level
Yes I know the term premia control our returns from this strategy; although probably both absolute levels and term premia are relevant, as they are probably related if you look at expected future term premia. When there is little vertical space on the Y-axis of the yield curve chart to work with between 0 and 30 years, term premia at all maturities will suffer or possibly go negative. Additionally, the gains from yield compression on macroeconomic crashes will be muted or disappear, see for example what the Euro rates did to compensate for the Spring 2020 stock market crash. They did not much, because they couldn't do much as they were already negative before the crash. In short, as I already said in my recent comment in the HFEA thread viewtopic.php?p=6298505#p6298505 , I think the biggest risk to this strategy in the long run are not rising interest rates, but a permanently flattening yield curve and low absolute real rates. Our strategy is dependent on long-term economic growth and demand for capital. I can't predict the future, but we should be cognizant of what is really the risk to this strategy in the long run. Instead of simple, naive backtesting and extrapolation, we need to adjust in a meaningful way any and all performance backtests for the current interest rate level, expected term premia, changed Fed strategy, expected economic growth, global demand and supply of capital, expected correlations between stocks and bonds, etc.
comeinvest
Posts: 2709
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 »

constructor wrote: Fri Oct 29, 2021 9:39 am 30 year yield (1.96%) lower than 20 year yield (1.98%) yesterday (October 28). Yield curve flattened in October. Interesting...
https://www.treasury.gov/resource-cente ... data=yield
If you think that level of inversion is astonishing, look what the future might bring. Look at the British curves, for example: https://www.bankofengland.co.uk/statistics/yield-curves
Unfortunately, the rates didn't follow the script of the media nor the script of the "experts" that the rates "must rise". At least not in October.
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

comeinvest wrote: Fri Oct 29, 2021 6:59 pm
constructor wrote: Fri Oct 29, 2021 9:39 am 30 year yield (1.96%) lower than 20 year yield (1.98%) yesterday (October 28). Yield curve flattened in October. Interesting...
https://www.treasury.gov/resource-cente ... data=yield
If you think that level of inversion is astonishing, look what the future might bring. Look at the British curves, for example: https://www.bankofengland.co.uk/statistics/yield-curves
Unfortunately, the rates didn't follow the script of the media nor the script of the "experts" that the rates "must rise". At least not in October.
Just in case anybody is confused, those are instantaneous forward yields, not actual yields. The actual yield curve is no where near that inverted in the UK right now.. it features a very slight inversion beyond 25 years of ~.05%.
comeinvest
Posts: 2709
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 »

skierincolorado wrote: Fri Oct 29, 2021 7:09 pm
comeinvest wrote: Fri Oct 29, 2021 6:59 pm
constructor wrote: Fri Oct 29, 2021 9:39 am 30 year yield (1.96%) lower than 20 year yield (1.98%) yesterday (October 28). Yield curve flattened in October. Interesting...
https://www.treasury.gov/resource-cente ... data=yield
If you think that level of inversion is astonishing, look what the future might bring. Look at the British curves, for example: https://www.bankofengland.co.uk/statistics/yield-curves
Unfortunately, the rates didn't follow the script of the media nor the script of the "experts" that the rates "must rise". At least not in October.
Just in case anybody is confused, those are instantaneous forward yields, not actual yields. The actual yield curve is no where near that inverted in the UK right now.. it features a very slight inversion beyond 25 years of ~.05%.
I didn't mean to confuse. But according to this chart http://www.worldgovernmentbonds.com/cou ... d-kingdom/, the U.K. yield curve peaks at 20 years at 1.18%, and goes down to 0.87% at 50 years, which is equal to the rate at 7.5 years. At 30 years it's at 1.11%. Despite the high expected inflation numbers in the U.K. The inversion may be a consequence of regulations of pension plans, or the market anticipating long-term muted economic growth, or it may be liquidity driven (a.k.a. no reason at all). What do I know. Arguably what the world looks like in 30+ is hard to predict, let alone interest rates in 30+ years - today's central banks may or may not exist, capitalist countries may become socialist, socialist countries might become capitalist, wars may or may not happen, countries may be wiped out, North Korea and other rogue states may either threaten the world or may have disappeared, the world as we know it today may or may not exist in 30-50 years. I think an inversion at the long end has no impact on our mHFEA strategy as the mHFEA is not invested in those maturities, but a generally low level of interest rates and low term premia has an impact.
User avatar
millennialmillions
Posts: 187
Joined: Sun Aug 18, 2019 6:03 pm

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

Post by millennialmillions »

Catching up on the last few pages, my takeaways are:
  • The treasury futures exposure question has been definitively answered, and we should simply use the market value provided by the broker. Thanks everyone for helping work through that.
  • It's beneficial to diversify duration across treasury futures. This is why I included ZF and ZN. It looks like including some ZT also helps, but the difference seemed small enough to me to simplify to just ZF and ZN.
  • The cost of borrowing using futures is likely slightly higher than the T-Bill rate we had been using in backtesting. I had previously selected a 140% equity 140% bond portfolio that performs well at a higher cost of borrowing, so I am not concerned. However, if we do get a more precise cost, I would certainly want to implement it in my backtesting.

Based on the above (and that I have an extra $100k NLV in my accounts thanks to the market), I wanted to revisit my portfolio plan and post again for review.
Image

Appreciate any feedback on this. Otherwise my next steps are:
  • Early this coming week, shift my whole portfolio to the plan shown above
  • Set up reminders to roll my futures and alerts for accounts getting close to their margin requirements
  • Figure out how to use box spreads to reduce my cost of borrowing (not a pressing issue since I signed up for IBKR using the promo link and my margin interest rate is only 1.08%)
  • Determine how I want to factor buying a house into this. Right now, I'm leaning toward not considering it part of the portfolio at all and doing the standard 20% down and a normal mortgage, but it might make sense to shift at least some of the mortgage to lower interest on the margin account.
Kbg
Posts: 355
Joined: Thu Mar 23, 2017 11:33 am

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

Post by Kbg »

skierincolorado wrote: Fri Oct 29, 2021 12:26 pm
Kbg wrote: Fri Oct 29, 2021 11:59 am Quick question: For 1:3/Stock:ITTs, why not just go 100/300 and not delever?

At least in PV it doesn’t appear one would need to…please punch holes in the above.
Take it back to 1955 in Simba.
Using Simba from 1955-72 and forward of that I'm taking out T-bills with a (-) sign for financing the leverage and nothing looks that great. It seems the basic bottom line here is leveraging successfully assumes we are not going to have a multi-year interest rate run-up.
Topic Author
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

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

Post by skierincolorado »

Kbg wrote: Sun Oct 31, 2021 6:15 pm
skierincolorado wrote: Fri Oct 29, 2021 12:26 pm
Kbg wrote: Fri Oct 29, 2021 11:59 am Quick question: For 1:3/Stock:ITTs, why not just go 100/300 and not delever?

At least in PV it doesn’t appear one would need to…please punch holes in the above.
Take it back to 1955 in Simba.
Using Simba from 1955-72 and forward of that I'm taking out T-bills with a (-) sign for financing the leverage and nothing looks that great. It seems the basic bottom line here is leveraging successfully assumes we are not going to have a multi-year interest rate run-up.
Of course the same thing can happen with stocks (1929-1944 U.S., 1990-2019 Japan). But 100/300 stock/ITT is certainly a bit riskier than 100/0 or 90/10. You can't count on the correlation being as consistently negative as it has been since 1980.
Post Reply