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

[comeinvest]

Has anyone here explored dynamic leverage ratio determination? Not sure if there's a better term for it. Basically the idea is adjust leverage based on expected returns for stocks/treasuries. I recently read this post: https://elmwealth.com/earnings-yield-dy ... llocation/, which focuses on asset allocation and only very briefly touches on leverage, but moreso as a constraint rather than a "what if."

Right now I just have target ratios of 1.65 for equities and 1.5 for ITTs, and I generally rebalance quarterly whenever I rollover futures. However, there seems to be a strong case for dynamic ratios... I'm just not sure where to get started (what bands, what indicators, and how to track/apply them).

This was a great article. Critical point that is not commonly discussed in finance analysis: Equity earnings yield ratios need to be compared with real interest rates, not nominal bond yields.

And to your question, I've explored dynamic leverage ratios based solely on cost of borrowing. This could help with the expected return part of the calculus. However, I think static leverage ratios are great bang for the work involved in dynamic allocations.

Very difficult to analyze the reason for low cost of borrowing. You would have to analyze whether it's due to low growth expectations, whether those low growth expectations are justified, or whether due to excessive QE, and when is it "excessive". The list goes on. You don't want to increase leverage when growth prospects for equities decrease. I feel I would quickly become a full-time macroeconomic researcher if I were to attempt such market timing. Basic assumption is that ERP is a (static or dynamic) spread on top of the risk-free rate.

[comeinvest]

Has anyone here explored dynamic leverage ratio determination? Not sure if there's a better term for it. Basically the idea is adjust leverage based on expected returns for stocks/treasuries. I recently read this post: https://elmwealth.com/earnings-yield-dy ... llocation/, which focuses on asset allocation and only very briefly touches on leverage, but moreso as a constraint rather than a "what if."

Right now I just have target ratios of 1.65 for equities and 1.5 for ITTs, and I generally rebalance quarterly whenever I rollover futures. However, there seems to be a strong case for dynamic ratios... I'm just not sure where to get started (what bands, what indicators, and how to track/apply them).

I think I cited the same article before. With leverage you can implement a dynamic allocation for both equities and ITT independently, i.e. without the 100% constraint. So you probably want to come up with two somewhat independent valuation or risk based dynamic allocations, which is a more complex optimization problem than that in the article. Not easy. Keep in mind the backtests in the article are backfitted, not out of sample. For ITT it would basically come down to estimating term premia, as I pointed out a few posts up.
I think in the end it comes down to how you model the stochastic processes underlying equities performance and interest rates. With static allocations and rebalancing to static allocations, we implicitly made already some assumptions, but nobody says static allocations are optimal, neither empirically nor theoretically. If we introduce mean-reverting components in the model, like for example Vasicek for interest rates, another modeling approach for optimal asset allocation, different from purely risk and return considerations, might be with Ornstein-Uhlenbeck processes which reflect trend, mean reversion, and random noise, and for which so-called optimal stopping problems can be solved to determine optimal rebalancing thresholds.
In other words, use mHFEA to make the white noise work for you.

[gamthe]

This was a great article. Critical point that is not commonly discussed in finance analysis: Equity earnings yield ratios need to be compared with real interest rates, not nominal bond yields.

It is difficult to time the market based on valuation alone (real or nominal).



This chart looks busy. This is trying to recreate what elms was talking about. We have real yield data from US since 2003. Elm's used UK inflation data from 1992 as proxy. US and UK inflation tracked similar till 2013, so reasonable assumption.

CAPE earning yield was below 10 year real yield in June 1996. If we timed the market and out of the market or reduced stock allocation because CAPE yield is less than real yield, then you would have missed specatacular run till 2000. Even 10 year forward return from 1996 till 2006 is around 6.5% CAGR.

[ipparkos]

...

45 bps would be roughly the same as the average of box spreads over the last few years.
However box spreads are currently around 75 bps (compounded interest similar to treasuries) or 85 bps (daily geometric growth method [[ending value]/[beginning value]]^365/daycount - 1) above Term SOFR. Sorry I still have to wrap my head around day count and compounding methods.
The expected funding cost of equity index futures should be easy to calculate from the current futures price as long as you have reliable expected dividend data. Treasury futures are a bit more complicated, as the funding cost has some volatility model dependencies due to various optionalities. User physicist tried to calculate the funding cost of treasury futures by hand as per the conversations between him and myself above in this thread.
But in most scenarios I don't see how any of this would be actionable. In IRAs the only reasonable way of leveraging equity markets is with equity index futures. Call options for equity leverage have a number of cost items, issues, and uncertainties, their pricing model is hard to reverse-engineer, and it is unclear if their funding cost is any cheaper. You also don't want to be in and out of the market based on funding cost.
European and Japanese markets seem to be more stable in terms of money market rates and derivatives funding costs.

comeinvest, can I ask if you ever looked at the implied financing cost of the MSCI World Index futures, FMWP (price) and FMWO (net return)? Your analysis already indicates that MSCI EAFE and Emerging Market index futures have lower financing cost compared to MES, I wonder the case for the World index futures. I am not really looking for a lower cost, buying the global stock market may soon be the easiest option for me, so I am happy as long as their liquidity and cost is comparable to those of MES. Thanks!

[comeinvest]

Happy new year!
I sold significant amounts (several $M) March 2025 expiration SPX box spreads at ca. 5.3% (daily compounded rate) on the quarterly SPX options expiration day 12/20/2024. Yesterday 12/31/2024 I sold a $50k March 2025 expiration SPX box spread at 494.75 (4.871%), and that was even with price improvement. 3-month SOFR decreased insignificantly from ca. 4.33% to ca. 4.3%.
Obviously I am not happy about the unfortunate timing of my trades.
Might this be a pattern at year end or at quarterly expirations?

I don't think this issue will persist. Even I was spending time to think how to efficiently arbitrage this difference. By the time I figured it out, the opportunity is gone.

If ever this happens again, it is better to capture that over any carry strategies

How would you arbitrage this? I don't think it was clear if the excessive spread would go away or persist until the expiration of the options. Even if you eventually "win", by arbitraging different spreads for different expirations you would also incur drawdown risk if the "trade" moves against you; so you would have to calculate the risk-adjusted return for your balance sheet exposure, if it makes sense at all.

[comeinvest]

comeinvest, can I ask if you ever looked at the implied financing cost of the MSCI World Index futures, FMWP (price) and FMWO (net return)? Your analysis already indicates that MSCI EAFE and Emerging Market index futures have lower financing cost compared to MES, I wonder the case for the World index futures. I am not really looking for a lower cost, buying the global stock market may soon be the easiest option for me, so I am happy as long as their liquidity and cost is comparable to those of MES. Thanks!

2-3 years ago I looked at all equity index worldwide. First off, U.S. residents are not allowed to trade all foreign futures for example on EUREX, but only a select list that is maintained by some U.S. government agency that has nothing else to do but to monitor worldwide derivative products, and determines which ones are "safe" for U.S. citizens (crazy!), to "protect" its citizens from themselves, while collecting a fee for their "services" from the foreign exchanges. It's due to some ancient law that I guess retail investors don't have a strong enough lobby to get rid of. If I remember right, the MSCI World index futures are not on this list. (Never mind you are allowed to put 100% of your money into GameStop.) Yes, this is what is supposed to be the "country of the free"
If you are not in the U.S.: From what I remember, the MSCI World futures are less liquid than the major U.S., European, and Japanese index futures, and the calendar rolls might cost you more due to larger minimum tick sizes of calendar rolls. Furthermore, the relative tax advantage of foreign futures vs. international ETF with respect to foreign withholding tax (for U.S. investors) would partially go away if you use a "commingled" futures product that has U.S. and international combined. I don't have data for the actual implied financing spread of MSCI World futures.

[sharukh]

I don't think it was clear if the excessive spread would go away or persist until the expiration of the options.

I got a bit confused, I was still hanging on the other issue of ES having higher financing rate. You are right about spx option spread persist until expiration.

How would you arbitrage this?

I was thinking about arbitrage of ES higher financing rate, by going synthetic long SPX and short 2 ES in futures. If we use portfolio margin account do this, it needs about $100k for each ES future. If everything goes well and after all the transaction costs gone, if a similar spread like last time exist like 0.5% from SPX box and ES roll, then it will net about $300 relatively risk free.

Even if you eventually "win", by arbitraging different spreads for different expirations you would also incur drawdown risk if the "trade" moves against you; so you would have to calculate the risk-adjusted return for your balance sheet exposure, if it makes sense at all.

correct, provided the spread doesn't get even wider or intraday movement is not too big and push me into a margin call.

[ipparkos]

[...]

45 bps would be roughly the same as the average of box spreads over the last few years.
However box spreads are currently around 75 bps (compounded interest similar to treasuries) or 85 bps (daily geometric growth method [[ending value]/[beginning value]]^365/daycount - 1) above Term SOFR. Sorry I still have to wrap my head around day count and compounding methods.
The expected funding cost of equity index futures should be easy to calculate from the current futures price as long as you have reliable expected dividend data. Treasury futures are a bit more complicated, as the funding cost has some volatility model dependencies due to various optionalities. User physicist tried to calculate the funding cost of treasury futures by hand as per the conversations between him and myself above in this thread.
But in most scenarios I don't see how any of this would be actionable. In IRAs the only reasonable way of leveraging equity markets is with equity index futures. Call options for equity leverage have a number of cost items, issues, and uncertainties, their pricing model is hard to reverse-engineer, and it is unclear if their funding cost is any cheaper. You also don't want to be in and out of the market based on funding cost.
European and Japanese markets seem to be more stable in terms of money market rates and derivatives funding costs.

In case you or anyone else is interested, Return Stacked people have a short writeup on the implied financing costs of S&P500 and US 10-year treasury futures since 2000. Their findings agree with what comeinvest has found so far. They curiously do not report any numbers directly, but their graphs clearly show that the spread is ~50bp for S&P500 and much less (much closer to 0) for the US 10-year. I just learnt about this writeup at the RR forum.
https://www.returnstacked.com/return-st ... -leverage/

[gamthe]

In case you or anyone else is interested, Return Stacked people have a short writeup on the implied financing costs of S&P500 and US 10-year treasury futures since 2000. Their findings agree with what comeinvest has found so far. They curiously do not report any numbers directly, but their graphs clearly show that the spread is ~50bp for S&P500 and much less (much closer to 0) for the US 10-year. I just learnt about this writeup at the RR forum.
https://www.returnstacked.com/return-st ... -leverage/

I do not understand what Y axis represent in his charts. Is daily rebalancing correct way to measure roll funding?. Most of the traders are rolling during the expiration week. Last hicup happened during the expiration week in December for the equity futures

Where exactly actual repo cost posted? (edit) Is it same as Implied repo with some haircut or o/n rate? It is certainly different than what he is showing.

[comeinvest]

Nice curve now

[comeinvest]

In case you or anyone else is interested, Return Stacked people have a short writeup on the implied financing costs of S&P500 and US 10-year treasury futures since 2000. Their findings agree with what comeinvest has found so far. They curiously do not report any numbers directly, but their graphs clearly show that the spread is ~50bp for S&P500 and much less (much closer to 0) for the US 10-year. I just learnt about this writeup at the RR forum.
https://www.returnstacked.com/return-st ... -leverage/

I do not understand what Y axis represent in his charts. Is daily rebalancing correct way to measure roll funding?. Most of the traders are rolling during the expiration week. Last hicup happened during the expiration week in December for the equity futures

Where exactly actual repo cost posted? (edit) Is it same as Implied repo with some haircut or o/n rate? It is certainly different than what he is showing.

I think the y axis in the first two charts is the capital growth of the different versions of risk-free rates, one of them being the "cash and carry" hedged portfolio consisting of the cash S&P 500 components and a short position in the futures, i.e. the implied financing rate.

I don't know what roll assumption they made, but usually similar charts assume a calendar roll on the busiest roll day, or a volume weighted average over the roll period.

I'm not qualified to answer your repo question, but my understanding is that "actual" repo is the rate from actual repo transactions (google), and "implied" repo is basically the implied financing rate of the futures, which should theoretically be equal or very close to the actual repo rate, but which we learned is not always the case.

I'm still not sure what I shall make of the low 3.5% implied repo rate in your chart, and the one in your previous posts. It would imply a hugely negative spread between implied financing rate and 3-month Term SOFR or 3-month treasuries; but the spread has been positive for a long time as per many of my and others' posts, along with other financing rates and negative swap spreads. 3-month SOFR is currently at ca. 4.29%, and 3-month T-bill are currently at ca. 4.33%. Something cannot be right.

[gamthe]

I think the y axis in the first two charts is the capital growth of the different versions of risk-free rates, one of them being the "cash and carry" hedged portfolio consisting of the cash S&P 500 components and a short position in the futures, i.e. the implied financing rate.

I don't know what roll assumption they made, but usually similar charts assume a calendar roll on the busiest roll day, or a volume weighted average over the roll period.

I'm not qualified to answer your repo question, but my understanding is that "actual" repo is the rate from actual repo transactions (google), and "implied" repo is basically the implied financing rate of the futures, which should theoretically be equal or very close to the actual repo rate, but which we learned is not always the case.

I'm still not sure what I shall make of the low 3.5% implied repo rate in your chart, and the one in your previous posts. It would imply a hugely negative spread between implied financing rate and 3-month Term SOFR or 3-month treasuries; but the spread has been positive for a long time as per many of my and others' posts, along with other financing rates and negative swap spreads. 3-month SOFR is currently at ca. 4.29%, and 3-month T-bill are currently at ca. 4.33%. Something cannot be right.

I used the exact same series and same dates for this chart



From this chart, I can say the equity funding rate is around 3mo bill + 20bp. ES1 Index rolls on EOM before expiry month. No traders rolls around the same point. Roll funding cost is slightly higher during the expiration week. So in reality 3mo + 30 bp sounds about right.

I wish he had given the series for actual repo he is using. O/N rate is basically EFPR. It will have jagged edge like SOFR rate he is using. His "Actual repo rate" is pretty smooth

[sharukh]

Hi,

Its time for me to take a mortgage. Does mHEFA followers prefer ARM over 30-year mortgage ?
Please note: this is not my mortgage specific question. I am trying to understand how mHEFA followers think or need to think about taking on debt. So posted here instead of a new thread.

Does people who believe in Carry from long duration bonds. i.e. going long Treasure futures, borrow short and lend long, will take only ARM mortgages ?

Why borrow from 30 years(in mortgage) and buy a long Treasury future in mHFEA. It essentially becomes a ARM. I might as well just take ARM mortgage, right ?

Buy taking ARM and not creating a carry from long treasury futures, I also save on the taxes that needs to be paid on the generated carry.

ARM comes with lower interest rate, the difference is equivalent to carry generated by bonds in mHFEA, I guess.
10/6M ARM - 6.55% / 6.87% APR
30 Year Fixed - 6.80% / 6.83% APR

Is this thinking/philosophy right ?

Thank you.

[comeinvest]

Hi,

Its time for me to take a mortgage. Does mHEFA followers prefer ARM over 30-year mortgage ?
Please note: this is not my mortgage specific question. I am trying to understand how mHEFA followers think or need to think about taking on debt. So posted here instead of a new thread.

Does people who believe in Carry from long duration bonds. i.e. going long Treasure futures, borrow short and lend long, will take only ARM mortgages ?

Why borrow from 30 years(in mortgage) and buy a long Treasury future in mHFEA. It essentially becomes a ARM. I might as well just take ARM mortgage, right ?

Buy taking ARM and not creating a carry from long treasury futures, I also save on the taxes that needs to be paid on the generated carry.

ARM comes with lower interest rate, the difference is equivalent to carry generated by bonds in mHFEA, I guess.
10/6M ARM - 6.55% / 6.87% APR
30 Year Fixed - 6.80% / 6.83% APR

Is this thinking/philosophy right ?

Thank you.

Rates are hard to compare, because 30 years mortgage loans come with a refinance option.
I think 30 year mortgage loans are priced based on the average time before homeowners move. If you plan on owning the house longer, then the option might be worth more to you than the price based on the length of stay that the market assumes.
Other than that, your thinking that the total duration exposure is what matters, is correct. But I personally would generally get a 30 year mortgage loan if I plan to keep the house forever, and especially if I can deduct the mortgage e.g. from rents received or using the personal mortgage deduction. Note that you can deduct the "coupon", but not the yield curve carry return of the mortgage loan.

[comeinvest]

I think the y axis in the first two charts is the capital growth of the different versions of risk-free rates, one of them being the "cash and carry" hedged portfolio consisting of the cash S&P 500 components and a short position in the futures, i.e. the implied financing rate.

I don't know what roll assumption they made, but usually similar charts assume a calendar roll on the busiest roll day, or a volume weighted average over the roll period.

I'm not qualified to answer your repo question, but my understanding is that "actual" repo is the rate from actual repo transactions (google), and "implied" repo is basically the implied financing rate of the futures, which should theoretically be equal or very close to the actual repo rate, but which we learned is not always the case.

I'm still not sure what I shall make of the low 3.5% implied repo rate in your chart, and the one in your previous posts. It would imply a hugely negative spread between implied financing rate and 3-month Term SOFR or 3-month treasuries; but the spread has been positive for a long time as per many of my and others' posts, along with other financing rates and negative swap spreads. 3-month SOFR is currently at ca. 4.29%, and 3-month T-bill are currently at ca. 4.33%. Something cannot be right.

I used the exact same series and same dates for this chart



From this chart, I can say the equity funding rate is around 3mo bill + 20bp. ES1 Index rolls on EOM before expiry month. No traders rolls around the same point. Roll funding cost is slightly higher during the expiration week. So in reality 3mo + 30 bp sounds about right.

I wish he had given the series for actual repo he is using. O/N rate is basically EFPR. It will have jagged edge like SOFR rate he is using. His "Actual repo rate" is pretty smooth

Equity funding spread to bills per the Bloomberg data in your chart is 0.16% on average, if I read the data in your screenshot right. I'm not sure why they use EOM before expiry month as roll dates, given that liquidity is probably lower than during expiration week, as the customary roll date is Monday before third Friday expiration, per CME information.
Do you have a source for the information that roll funding cost is higher during the expiration week?
I still don't know what I shall make of your 3.5% implied repo rate per your chart in your earlier post.

[ipparkos]

I think the y axis in the first two charts is the capital growth of the different versions of risk-free rates, one of them being the "cash and carry" hedged portfolio consisting of the cash S&P 500 components and a short position in the futures, i.e. the implied financing rate.

I don't know what roll assumption they made, but usually similar charts assume a calendar roll on the busiest roll day, or a volume weighted average over the roll period.

I'm not qualified to answer your repo question, but my understanding is that "actual" repo is the rate from actual repo transactions (google), and "implied" repo is basically the implied financing rate of the futures, which should theoretically be equal or very close to the actual repo rate, but which we learned is not always the case.

I'm still not sure what I shall make of the low 3.5% implied repo rate in your chart, and the one in your previous posts. It would imply a hugely negative spread between implied financing rate and 3-month Term SOFR or 3-month treasuries; but the spread has been positive for a long time as per many of my and others' posts, along with other financing rates and negative swap spreads. 3-month SOFR is currently at ca. 4.29%, and 3-month T-bill are currently at ca. 4.33%. Something cannot be right.

I used the exact same series and same dates for this chart



From this chart, I can say the equity funding rate is around 3mo bill + 20bp. ES1 Index rolls on EOM before expiry month. No traders rolls around the same point. Roll funding cost is slightly higher during the expiration week. So in reality 3mo + 30 bp sounds about right.

I wish he had given the series for actual repo he is using. O/N rate is basically EFPR. It will have jagged edge like SOFR rate he is using. His "Actual repo rate" is pretty smooth

Aside from comeinvest's question about the spread being higher during the roll period, I am also interested in your finding of a spread of 20-30bp for ES. comeinvest and other people I linked to a bit above all found 45-50bp on average (above 100 in bad times). What gives? This is not just splitting hairs, your spread value is below box spread costs, and the ones below are about the same. So they have material consequences for how we obtain leverage. I am quoting the charts from my previous post.

https://x.com/EconomPic/status/1869796284854612350


https://x.com/EconomPic/status/1869802458219983191

[comeinvest]

Aside from comeinvest's question about the spread being higher during the roll period, I am also interested in your finding of a spread of 20-30bp for ES. comeinvest and other people I linked to a bit above all found 45-50bp on average (above 100 in bad times). What gives? This is not just splitting hairs, your spread value is below box spread costs, and the ones below are about the same. So they have material consequences for how we obtain leverage. I am quoting the charts from my previous post.

I think the funding rate spread of S&P 500 futures was +- zero for most of their history, until after the GFC, in part due to more regulation. (Thank you, regulators, for eating my returns!) The average of zero and 50 bps is roughly the average in gamthe's post over the longer period.

[gamthe]

Do you have a source for the information that roll funding cost is higher during the expiration week?

[km91]

I know there is an issue with IBKR and receiving payments in lieu when using box spreads. I contacted support asking what if anything could be done to recall my shares at dividend dates to maintain the qualified status of dividends. I was told that my shares are not being lent out, they are being pledged to the clearinghouse, and that this is a clearinghouse margin requirement. SEC rule 15c3-3(c)(1)/020 seems to confirm as much:

However, where a customer writes a call option, the proceeds of his writing transaction are included in the “customer” account at OCC. The broker or dealer will be required to deposit additional margin to secure the customer’s writing obligation, and to this extent, fully-paid or excess margin securities held by the broker or dealer to margin such customer’s writing obligations, may be used to the extent of 140% of the amount derived by adding to the customer’s net debit balance the amount of margin required by OCC from the clearing member or the amount of margin required by the broker or dealer’s Designated Examining Authority.

With this said, is it true that other brokers payout significantly less dividends as PIL than IBKR when using box spreads? The customer rep was sure this is a regulatory requirement and little can be done. Are other brokers getting around this?

[comeinvest]

I know there is an issue with IBKR and receiving payments in lieu when using box spreads. I contacted support asking what if anything could be done to recall my shares at dividend dates to maintain the qualified status of dividends. I was told that my shares are not being lent out, they are being pledged to the clearinghouse, and that this is a clearinghouse margin requirement. SEC rule 15c3-3(c)(1)/020 seems to confirm as much:

However, where a customer writes a call option, the proceeds of his writing transaction are included in the “customer” account at OCC. The broker or dealer will be required to deposit additional margin to secure the customer’s writing obligation, and to this extent, fully-paid or excess margin securities held by the broker or dealer to margin such customer’s writing obligations, may be used to the extent of 140% of the amount derived by adding to the customer’s net debit balance the amount of margin required by OCC from the clearing member or the amount of margin required by the broker or dealer’s Designated Examining Authority.

With this said, is it true that other brokers payout significantly less dividends as PIL than IBKR when using box spreads? The customer rep was sure this is a regulatory requirement and little can be done. Are other brokers getting around this?

They were trying to shut you up by confusing you with irrelevant nuances of terminology, and trying to find some lame excuse for their greedy behavior.
It is true that the broker is allowed to de-segregate securities worth up to 140% of a customer's debit balance from the segregated customers' assets and put them into the broker's own account, and lend them out for the broker's own benefit.
It is also true that "borrowing" via options has the same effect as borrowing via margin loan, due to regulatory provisions to that effect. In other words, the broker industry understood how to cover their bases and protect their income streams, and steered the regulator accordingly.
To me the SEC rule that you cited doesn't seem to be directly related to how many shares the broker can lend out, although it references the same 140%. I'm not sure if the shares are actually "physically" moved from the broker to the OCC, and if OCC would lend out shares that it holds as security for options writing obligations, but in the end it doesn't matter whether the broker or the OCC lends them out.
What the customer support that you talked to conveniently held back is that nobody forces either the broker or the OCC to actually lend out shares, even if they hold them as security for a margin loan or an options writing obligation. All major U.S. brokers (Schwab, Fidelity, etc.) except IB either recall shares before the ex dividend date, or compensate their customers for the additional tax burden in taxable accounts. It would appear that only IB is so greedy that they don't recall shares and collect the lending fee also on the ex dividend date, with no compensation to their customers.

"The customer rep was sure this is a regulatory requirement and little can be done." - Lol. I never received even one cent of PIL in years, when leveraging my sizable Schwab account with SPX options spreads to multiple times the account value. This suggests that even in the case of leverage with SPX options instead of broker margin, the broker appears to be under control of when to lend out the shares and when to recall them. Never mind, IB doesn't recall shares around ex dividend dates, even when broker margin is used for leverage. Next time, you can suggest that IB management ask Schwab for help, if they want to figure out "what can be done"

In the meantime, draw your own conclusions. You can move your account to Schwab or another broker.

[km91]

I know there is an issue with IBKR and receiving payments in lieu when using box spreads. I contacted support asking what if anything could be done to recall my shares at dividend dates to maintain the qualified status of dividends. I was told that my shares are not being lent out, they are being pledged to the clearinghouse, and that this is a clearinghouse margin requirement. SEC rule 15c3-3(c)(1)/020 seems to confirm as much:



With this said, is it true that other brokers payout significantly less dividends as PIL than IBKR when using box spreads? The customer rep was sure this is a regulatory requirement and little can be done. Are other brokers getting around this?

They were trying to confuse you with irrelevant nuances of terminology, and trying to find some lame excuse for their greedy behavior.
It is true that the broker is allowed to de-segregate securities worth up to 140% of a customer's debit balance from the segregated customers' assets and put them into the broker's own account, and lend them out for the broker's own benefit.
It is also true that "borrowing" via options has the same effect as borrowing via margin loan, due to regulatory provisions to that effect. In other words, the broker industry understood how to cover their bases and protect their income streams, and steered the regulator accordingly.
To me the SEC rule that you cited doesn't seem to be directly related to how many shares the broker can lend out, although it references the same 140%. I'm not sure if the shares are actually "physically" moved from the broker to the OCC, and if OCC would lend out shares that it holds as security for options writing obligations, but in the end it doesn't matter whether the broker or the OCC lends them out.
What the customer support that you talked to conveniently held back is that nobody forces either the broker or the OCC to actually lend out shares, even if they hold them as security for a margin loan or an options writing obligation. All major U.S. brokers (Schwab, Fidelity, etc.) except IB either recall shares before the ex dividend date, or compensate their customers for the additional tax burden in taxable accounts. It would appear that only IB is so greedy that they don't recall shares and collect the lending fee also on the ex dividend date, with no compensation to their customers.

"The customer rep was sure this is a regulatory requirement and little can be done." - Lol. I never received even one cent of PIL in years, when leveraging my sizable Schwab account with SPX options spreads to multiple times the account value. This suggests that even in the case of leverage with SPX options instead of broker margin, the broker appears to be under control of when to lend out the shares and when to recall them. Never mind, IB doesn't recall shares around ex dividend dates, even when broker margin is used for leverage. Next time, you can suggest that IB ask Schwab for help, if they want to figure out "what can be done"

Draw your own conclusions. You can move your account to Schwab or another broker.

Yes, your experience confirms what I suspected, they have control over when and how many share they pull from your account, and when they give them back, and would rather earn the extra day of fees. I will probably move my account, I hold 5 Avantis ETFs and received 98% of 2024 dividends as PIL. It's truly egregious. Any point in pushing IBKR harder, have you ever heard of them setting up an account to receive no PIL?

[comeinvest]

They were trying to confuse you with irrelevant nuances of terminology, and trying to find some lame excuse for their greedy behavior.
It is true that the broker is allowed to de-segregate securities worth up to 140% of a customer's debit balance from the segregated customers' assets and put them into the broker's own account, and lend them out for the broker's own benefit.
It is also true that "borrowing" via options has the same effect as borrowing via margin loan, due to regulatory provisions to that effect. In other words, the broker industry understood how to cover their bases and protect their income streams, and steered the regulator accordingly.
To me the SEC rule that you cited doesn't seem to be directly related to how many shares the broker can lend out, although it references the same 140%. I'm not sure if the shares are actually "physically" moved from the broker to the OCC, and if OCC would lend out shares that it holds as security for options writing obligations, but in the end it doesn't matter whether the broker or the OCC lends them out.
What the customer support that you talked to conveniently held back is that nobody forces either the broker or the OCC to actually lend out shares, even if they hold them as security for a margin loan or an options writing obligation. All major U.S. brokers (Schwab, Fidelity, etc.) except IB either recall shares before the ex dividend date, or compensate their customers for the additional tax burden in taxable accounts. It would appear that only IB is so greedy that they don't recall shares and collect the lending fee also on the ex dividend date, with no compensation to their customers.

"The customer rep was sure this is a regulatory requirement and little can be done." - Lol. I never received even one cent of PIL in years, when leveraging my sizable Schwab account with SPX options spreads to multiple times the account value. This suggests that even in the case of leverage with SPX options instead of broker margin, the broker appears to be under control of when to lend out the shares and when to recall them. Never mind, IB doesn't recall shares around ex dividend dates, even when broker margin is used for leverage. Next time, you can suggest that IB ask Schwab for help, if they want to figure out "what can be done"

Draw your own conclusions. You can move your account to Schwab or another broker.

Yes, your experience confirms what I suspected, they have control over when and how many share they pull from your account, and when they give them back, and would rather earn the extra day of fees. I will probably move my account, I hold 5 Avantis ETFs and received 98% of 2024 dividends as PIL. It's truly egregious. Any point in pushing IBKR harder, have you ever heard of them setting up an account to receive no PIL?

I asked them if they can change their policy to comply with industry standards, or make an exception or reimburse me, or else I have to sadly move my account. They responded they don't care.
I don't mean to bash IB. They are unique with their range of global investment products, and I like their platform. But for a leveraged strategy it simply doesn't work because of the PIL. I still have my international equities with them. They also produce PIL, but relatively little.
If you have a low leverage ratio, you could set up a second account where you use no leverage to avoid some PIL, but generally it will be too complicated. When the market drops you would have to shuffle money around.

[comeinvest]

Yes, your experience confirms what I suspected, they have control over when and how many share they pull from your account, and when they give them back, and would rather earn the extra day of fees. I will probably move my account, I hold 5 Avantis ETFs and received 98% of 2024 dividends as PIL. It's truly egregious. Any point in pushing IBKR harder, have you ever heard of them setting up an account to receive no PIL?

Which Avantis ETFs did you have that produced 98% PIL?

[km91]

Yes, your experience confirms what I suspected, they have control over when and how many share they pull from your account, and when they give them back, and would rather earn the extra day of fees. I will probably move my account, I hold 5 Avantis ETFs and received 98% of 2024 dividends as PIL. It's truly egregious. Any point in pushing IBKR harder, have you ever heard of them setting up an account to receive no PIL?

Which Avantis ETFs did you have that produced 98% PIL?

AVLV, AVUV, AVIV, AVDV, AVES

Not all of that would've been qualified dividends, but 98% of the dividends I received on these in 2024 were received as PIL. I'm a bit baffled by this, almost all my shares were lent out every ex div date?

[comeinvest]

Do you have a source for the information that roll funding cost is higher during the expiration week?

Very informative, thanks! But I don't see any supporting data that shows that roll funding cost is higher during the expiration week.
Too bad they don't show the Stoxx Europe 600 futures.
MSCI EAFE vs. Euro Stoxx 50 + Nikkei 225 would be almost a wash in terms of trading cost, if I read the data right.
Is this publication free for Merrill Edge brokerage customers?

[sharukh]

Hi gamthe/comeinvest,

Are you aware of any chart that shows the difference in SPX box based funding cost VS ES futures roll cost over time?
Thanks.

Is this publication free for Merrill Edge brokerage customers?

Yes, I just looked up.
They have report for each day of the roll. you might like this.




Thanks gamthe . Neve looked into Bofa reports earlier, as you showed their research is good.
They seem to release more than 50 reports a day. No way I would have found what you shared.
Can you please share any other report names/titles you like it there?

[gamthe]

Very informative, thanks! But I don't see any supporting data that shows that roll funding cost is higher during the expiration week.

BofA is calculating VWAP of traded roll volume. Almost all of these happens during last 10 days of the expiration. I was comparing this to Corey Hoffstein's method of daily rebalancing of futures vs SPXT.

Hi gamthe/comeinvest,

Are you aware of any chart that shows the difference in SPX box based funding cost VS ES futures roll cost over time?
Thanks.

Sorry no. I would imagine it should be similar.

Can you please share any other report names/titles you like it there?

Depends on your area of interest. I like anything Savitha Subramanian writes and their derivative research is also good

[comeinvest]

Thanks gamthe . Neve looked into Bofa reports earlier, as you showed their research is good.
They seem to release more than 50 reports a day. No way I would have found what you shared.
Can you please share any other report names/titles you like it there?

I had a Merrill Edge account many years ago to collect some promotional bonus, but I profoundly hated the web site and the trading capabilities (or lack of them) and closed my account. Should I sign up again just for the futures roll and perhaps options research?! The futures roll analysis is interesting, but I'm not sure if anything is actionable. I don't see how the roll analysis is actionable.

[comeinvest]

Do you have a source for the information that roll funding cost is higher during the expiration week?

Any possible reason or logic why they measure the funding spread of U.S. futures relative to "Fed Funds", but USD based foreign futures relative to "USD SOFR"?
Also, assuming they use the spot rates, shouldn't 3-month Term SOFR be the reference rate?

[comeinvest]

NIce curves, and surprisingly the major currencies offer similar yield curve carry returns in the middle of the curves.

[sharukh]

They had this in Dec 6th article.
The general averages are:
Q1, Q2,Q3 s Average = 0.44%
Q4s Average = 0.59%
So, it is well known based on the trend from 2013 that Q4 rolls are expensive compared to other quarters.

2024 Q4 was an outlier, see the below chart.

The fact that they had published on Dec 6th was good, When the roll time comes, it wont be a surprise.
If one likes to prepare to do a arbitrage, these figures will help.

Thanks gamthe . Neve looked into Bofa reports earlier, as you showed their research is good.
They seem to release more than 50 reports a day. No way I would have found what you shared.
Can you please share any other report names/titles you like it there?

I had a Merrill Edge account many years ago to collect some promotional bonus, but I profoundly hated the web site and the trading capabilities (or lack of them) and closed my account. Should I sign up again just for the futures roll and perhaps options research?! The futures roll analysis is interesting, but I'm not sure if anything is actionable. I don't see how the roll analysis is actionable.

W.r.t. actionable(among the ones I saw): If I was given well enough notice about this, I would be happy to lockup some of my capital for this risk free arbitrage especially at a very good outlier roll rate.

I was thinking of closing out the Merrill Edge, guess I am going to keep it open for a little longer as this new area is interesting. It doesn't cost much. There are no fees, may be open one account and leave $10 in it.

[comeinvest]

W.r.t. actionable(among the ones I saw): If I was given well enough notice about this, I would be happy to lockup some of my capital for this risk free arbitrage especially at a very good outlier roll rate.

I was thinking of closing out the Merrill Edge, guess I am going to keep it open for a little longer as this new area is interesting. It doesn't cost much. There are no fees, may be open one account and leave $10 in it.

You mean arbitrage short futures vs. synthetic long with SPX options? Even with the worst outlier, Q4 2024, you had only ca. 0.4% APY difference, which is a 0.1% absolute gain. So your total trading friction cost for the futures and the options would need to be below 0.1% of the notional value, right?

Does Merrill / BofA also have articles on SPX options?

[comeinvest]

The contribution of the yield curve rolldown to the carry return, relative to the yield spread
From the paper https://www.returnstackedetfs.com/wp-co ... tation.pdf

[sharukh]

You mean arbitrage short futures vs. synthetic long with SPX options?

Yes

Even with the worst outlier, Q4 2024, you had only ca. 0.4% APY difference, which is a 0.1% absolute gain. So your total trading friction cost for the futures and the options would need to be below 0.1% of the notional value, right?

0.1% absolute gain on the ES notional value. But only 10 to 20% of it is needed for collateral with both SPX-long and ES-short put together. Say 20%, then the return on the collateral is 0.1% * 5 = 0.5% p.a. risk free
hmm.. This was an outlier event. Not a reliable long term strategy any. I shouldn't have thought this much about this.

Does Merrill / BofA also have articles on SPX options?

Looked a lot, couldn't find anything. They are just calling ES rolls as SPX rolls sometimes and are just using it interchangeably.
They just collectively call it as "SPX futures (ES)"

[comeinvest]

Even with the worst outlier, Q4 2024, you had only ca. 0.4% APY difference, which is a 0.1% absolute gain. So your total trading friction cost for the futures and the options would need to be below 0.1% of the notional value, right?

0.1% absolute gain on the ES notional value. But only 10 to 20% of it is needed for collateral with both SPX-long and ES-short put together. Say 20%, then the return on the collateral is 0.1% * 5 = 0.5% p.a. risk free
hmm.. This was an outlier event. Not a reliable long term strategy any. I shouldn't have thought this much about this.

The margin efficiency, or return per margin requirement, is not necessarily the bottleneck. You would want to put that in relation to the drawdown risk anyway.
What I meant is the trading friction cost of the futures and the options position combined. Commissions and bid/ask spreads or effective spreads. In relation to the notional value, it has to be below the expected return per notional value (0.1%) for the trade to be viable.

[comeinvest]

"We calculate MSCI EAFE & EM roll costs using both realized gross and net MSCI dividend" - so which one is it? I'm not comprehending. Realized gross, or realized net? The table shows only one number for each futures contract.

[gamthe]

Any possible reason or logic why they measure the funding spread of U.S. futures relative to "Fed Funds", but USD based foreign futures relative to "USD SOFR"?
Also, assuming they use the spot rates, shouldn't 3-month Term SOFR be the reference rate?

I don't know the answer to your first question.

They are using 3 month forward rates. This is from their Dec10th note.



Regarding Treasury implied repo vs actual repo, this is from Soc. Gen Nov 19th roll notes. I am assuming actual repo estimated is just forward rate between two implied repos

[gamthe]

"We calculate MSCI EAFE & EM roll costs using both realized gross and net MSCI dividend" - so which one is it? I'm not comprehending. Realized gross, or realized net? The table shows only one number for each futures contract.

If i read their note right, they are using combination of past realized gross dividend and forecast of current quarter dividend to come up with dividend rate. For other futures dividend rate assumption is based on dividend forecasts

[comeinvest]

"We calculate MSCI EAFE & EM roll costs using both realized gross and net MSCI dividend" - so which one is it? I'm not comprehending. Realized gross, or realized net? The table shows only one number for each futures contract.

If i read their note right, they are using combination of past realized gross dividend and forecast of current quarter dividend to come up with dividend rate. For other futures dividend rate assumption is based on dividend forecasts

I think their note that you posted last refers to U.S. futures only. I think the MSCI indices roll cost is based on net dividends as they explicitly say in the first screenshot in this viewtopic.php?p=8208842#p8208842 post from you; but the sentence that I cited above is very confusing.
Net dividends vs. gross dividends is probably a difference of 30-40 bps.

[comeinvest]

Regarding Treasury implied repo vs actual repo, this is from Soc. Gen Nov 19th roll notes. I am assuming actual repo estimated is just forward rate between two implied repos

My interpretation is different. I would say the "Term Repo Est" are actual repo transactions (not related to futures) for reference, to be compared to the futures implied repo rates (futures implied funding rates).
I don't see the calculated effective roll cost anywhere on your screenshot. By my understanding the roll cost would be the implied repo rate (4.62%) of FVH5, adjusted for the implied repo rate (4.18%) of FVZ4. The 4.18% of the front month contract would weight less in the calculation of the roll cost, because the front month contract has much less time to expiration (delivery) than the deferred month contract. (I think the estimated delivery date within the delivery month that a rational actor would deliver would be relevant, per a discussion between physicist and myself earlier in this thread.)
Calendar roll means you buy the deferred month and sell the front month if you want to roll a long position. So the low implied funding rate of the front month would increase the cost of the roll, if my understanding is correct.
The 3-month Term SOFR at the end of November 2024 was about 4.5%.
The 1-month Term SOFR at the end of November 2024 was about 4.6%.
By my estimate the roll cost spread above SOFR of the FV was ca. 0.1% annualized based on the numbers in your screenshot; but don't take my word for it.

B.t.w. I'm not following the logic why large asset manager net long positioning would drive the roll lower. I would have thought the opposite, as the asset managers need to roll their positions vs. constrained dealer balance sheets.

[gamthe]

B.t.w. I'm not following the logic why large asset manager net long positioning would drive the roll lower. I would have thought the opposite, as the asset managers need to roll their positions vs. constrained dealer balance sheets.

I think this one is stright forward. Asset managers are holding ton of long positions in the front and need to sell these contracts to roll. Since most of them try to do at the same window creates downward pressure. These effects do get dislocated when it coincides with events like election.


This regression is general regression not separating out roll period, but shows downward trend with asset manager long position. X axis is the asset manager institutional long futures position in 5 yr futures vs the 5 year note price. Soc Gen recommendation was correct imo

[comeinvest]

B.t.w. I'm not following the logic why large asset manager net long positioning would drive the roll lower. I would have thought the opposite, as the asset managers need to roll their positions vs. constrained dealer balance sheets.

I think this one is stright forward. Asset managers are holding ton of long positions in the front and need to sell these contracts to roll. Since most of them try to do at the same window creates downward pressure. These effects do get dislocated when it coincides with events like election.

This regression is general regression not separating out roll period, but shows downward trend with asset manager long position. X axis is the asset manager institutional long futures position in 5 yr futures vs the 5 year note price. Soc Gen recommendation was correct imo

Downward pressure on the front month contract would make the roll more expense, not cheaper. Calendar roll means sell the front month and buy the deferred month.
It's also intuitive, if you think of it, as a long position holder you will be competing with all other "natural" long position holders during calendar roll. Competition makes stuff more expensive (and stuff to sell cheaper). The back month will get more expensive, and the front month cheaper, making the calendar roll more expensive.
There must be another reasoning that they were referring to.

[gamthe]

[
Downward pressure on the front month contract would make the roll more expense, not cheaper. Calendar roll means sell the front month and buy the deferred month.
It's also intuitive, if you think of it, as a long position holder you will be competing with all other "natural" long position holders during calendar roll. Competition makes stuff more expensive (and stuff to sell cheaper). The back month will get more expensive, and the front month cheaper, making the calendar roll more expensive.
There must be another reasoning that they were referring to.

Both you and soc gen are saying the same thing but are using different nomeclature. You are saying expensive and they are saying spread will roll lower. If you chart the actual calendar spread probably gives the better picture. Their note came out in mid november and said if you are long roll early otherwise it will roll lower. Chart says they were correct

[comeinvest]

[
Downward pressure on the front month contract would make the roll more expense, not cheaper. Calendar roll means sell the front month and buy the deferred month.
It's also intuitive, if you think of it, as a long position holder you will be competing with all other "natural" long position holders during calendar roll. Competition makes stuff more expensive (and stuff to sell cheaper). The back month will get more expensive, and the front month cheaper, making the calendar roll more expensive.
There must be another reasoning that they were referring to.

Both you and soc gen are saying the same thing but are using different nomeclature. You are saying expensive and they are saying spread will roll lower. If you chart the actual calendar spread probably gives the better picture. Their note came out in mid november and said if you are long roll early otherwise it will roll lower. Chart says they were correct

Their language was "large asset manager net long positioning would drive the roll lower". Granted, there might be some ambiguity regarding the definition of the "roll" and/or the meaning of "lower", as they are not using the terminology "cheap roll" and "expensive / rich roll", which I think is unambiguous in the finance community. But your latest screenshot clarifies that indeed the spread "FVH5 - FVZ4" decreased over time, i.e. the calendar roll that the strategic long position holders need to execute got cheaper during the roll period. (The last screenshot seems to show the absolute spread decreasing and not the implied rate, which could have simply been a result of the passage of time due to less remaining time to expiration / delivery; but I assume that the implied rate also decreased during the roll period.)
I personally don't understand any of the reasoning for cheapening or richening *within* one roll period; but common sense would dictate that higher demand for strategic long positions leads to more expensive (richer) calendar rolls altogether, not cheaper ones.
I could only follow their logic if they meant that the "large asset managers" (for whatever reason) only bought the front month (December 2024) contract shortly before its expiration (i.e. during the roll period), with no intention to roll their positions, which then might have made the calendar roll cheaper for those long position holders who indeed intended to roll their positions, as the latter might have sold their front month long positions to the aforementioned "large asset managers".