Let's Talk SPX Box Spreads

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
LoveTheBogle
Posts: 188
Joined: Tue Aug 13, 2019 5:53 pm

Re: Let's Talk SPX Box Spreads

Post by LoveTheBogle »

I see the issue. Fidelity's option details page PAIRED two of my short box spread's WRONG. I had another short box spread in there with the same expiration as the one showing a GAIN and the strike prices are different than my order history as they were not filled on the same day. If I was to close all my short box trades today that I opened a few months ago then I would be down more than if I just let them creep toward the payoff amount of expiration date.
adamhg
Posts: 218
Joined: Sat Apr 10, 2021 8:40 am

Re: Let's Talk SPX Box Spreads

Post by adamhg »

LoveTheBogle wrote: Mon Aug 08, 2022 2:10 pm Is it possible to close a SHORT box spread for a GAIN before expiration? In my Fidelity account the options summary page correctly paired the four positions of a short box trade where I got a credit and it is showing a + and in green for unrealized gain/loss. It isn't much, but it isn't zero and all of my other short box trades that are open right now are red and a negative number for the unrealized gain/loss. I'm thinking it is because of bad quotes even though the market is open because maybe the actual strike and expiration has little to no volume.

If it is possible to close a short box trade for a gain then can someone explain what took place for this to happen?
Once I left an opposite GTC order (I was long, so it was a STC order) for at or near the max gain for the box spread just to see what would happen. I forgot all about it until I went to re-open the box at exp only to realize my trade executed and closed it out weeks ago.
comeinvest
Posts: 2709
Joined: Mon Mar 12, 2012 6:57 pm

Re: Let's Talk SPX Box Spreads

Post by comeinvest »

adamhg wrote: Mon Aug 08, 2022 5:07 pm
LoveTheBogle wrote: Mon Aug 08, 2022 2:10 pm Is it possible to close a SHORT box spread for a GAIN before expiration? In my Fidelity account the options summary page correctly paired the four positions of a short box trade where I got a credit and it is showing a + and in green for unrealized gain/loss. It isn't much, but it isn't zero and all of my other short box trades that are open right now are red and a negative number for the unrealized gain/loss. I'm thinking it is because of bad quotes even though the market is open because maybe the actual strike and expiration has little to no volume.

If it is possible to close a short box trade for a gain then can someone explain what took place for this to happen?
Once I left an opposite GTC order (I was long, so it was a STC order) for at or near the max gain for the box spread just to see what would happen. I forgot all about it until I went to re-open the box at exp only to realize my trade executed and closed it out weeks ago.
The closer you come to expiration, orders just below the size of the spread will probably be filled sooner or later. You have additional trading cost, so I wouldn't do it unless you unexpectedly no longer need the money that you "borrowed".
comeinvest
Posts: 2709
Joined: Mon Mar 12, 2012 6:57 pm

Re: Let's Talk SPX Box Spreads

Post by comeinvest »

adamhg wrote: Fri Aug 05, 2022 11:11 am Here's a new project to calculate live box spread rate spreads. Quote times are UTC and were live as of the screenshot.

Image

With the large interest rate movement today, I thought it was a good opportunity post this before the daily boxtrades.com dump tomorrow and keep me honest. Looked pretty consistent except for the first one and the last few starting around Dec'24. The short duration makes it harder to find a good mid-point and the last few appears to be when liquidity starts to dry up on the ask side.
Interesting to see the tabular presentation with the synthetic box quotes, but what is the takeaway? I see that the bid and ask quotes get closer to each other with time to expiration, then get wider again. Is that because of volumes of the individual options for the different expirations?
All of the rates have significantly wider i.e. worse indicative bid/ask spreads than what you would get with a combination order. I'm not sure if we can learn anything; even the midpoint is not necessarily a good representation of the mark price, based on some articles that I read the other day.
If we want to compare the mid price implied rates to the treasuries rates, I don't see the drop in implied rates between June 2024 and Dec 2024 reflected in the treasury curve. But again, the options midprice data may be skewed by quotes from market participants, and other intricacies of options pricing.
comeinvest
Posts: 2709
Joined: Mon Mar 12, 2012 6:57 pm

Re: Let's Talk SPX Box Spreads

Post by comeinvest »

Looks like during the period of about one month before expiration of the Aug 2022 box spreads, the Aug 2022 box spreads traded at higher implied rates than the Sep 2022 box spreads, even though the treasury yield curve had a positive slope in that segment. Thoughts? Shall we conclude to not go below 1 month to expiration when borrowing?

Image

Image

Image
comeinvest
Posts: 2709
Joined: Mon Mar 12, 2012 6:57 pm

Re: Let's Talk SPX Box Spreads

Post by comeinvest »

Today I tried to buy 3000-4000 SPX 08/31/2022 expiration box @998.9 (2.54% yield), no fill. Comparable treasuries per WSJ are at ca. 1.45% ASK yield.
Today I bought 3000-4000 SPX 09/30/2022 expiration box @996.5 (2.82% yield) with price improvement. WSJ quotes show 09/30 treasuries at ca. 1.93% ASK yield https://www.wsj.com/market-data/bonds/treasuries .
My spread to treasuries would be ca. 0.9%. Can that be right? Can someone please verify my math?

The yield curve http://www.worldgovernmentbonds.com/cou ... ed-states/ seems to show consistently about 0.2% higher yields than WSJ.
User avatar
Kevin M
Posts: 15787
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: Let's Talk SPX Box Spreads

Post by Kevin M »

comeinvest wrote: Mon Aug 15, 2022 5:39 pm Today I tried to buy 3000-4000 SPX 08/31/2022 expiration box @998.9 (2.54% yield), no fill. Comparable treasuries per WSJ are at ca. 1.45% ASK yield.
Today I bought 3000-4000 SPX 09/30/2022 expiration box @996.5 (2.82% yield) with price improvement. WSJ quotes show 09/30 treasuries at ca. 1.93% ASK yield https://www.wsj.com/market-data/bonds/treasuries .
My spread to treasuries would be ca. 0.9%. Can that be right? Can someone please verify my math?

The yield curve http://www.worldgovernmentbonds.com/cou ... ed-states/ seems to show consistently about 0.2% higher yields than WSJ.
I see high ask yield of 1.998% for 9/30/22 from Fidelity yesterday, but all of those have coupons. Hi yield for close maturities was 2.185% for a T bill maturing 9/27/22. I would use the T bill yield for comparison, since that's what I'd buy if I were buying a Treasury of that maturity.

For your box trade I get a yield of 2.85%, so 66 bps higher than Treasury yield. Not quite as high as your calculations show, but still excellent. I was debating transferring more cash into Fidelity to buy another box spread this week, or wait until the Friday expiration of my 8/19/22 box spread, and buy more next week. I'm inspired buy your purchase to do the former.

Kevin
If I make a calculation error, #Cruncher probably will let me know.
comeinvest
Posts: 2709
Joined: Mon Mar 12, 2012 6:57 pm

Re: Let's Talk SPX Box Spreads

Post by comeinvest »

Kevin M wrote: Tue Aug 16, 2022 11:27 am
comeinvest wrote: Mon Aug 15, 2022 5:39 pm Today I tried to buy 3000-4000 SPX 08/31/2022 expiration box @998.9 (2.54% yield), no fill. Comparable treasuries per WSJ are at ca. 1.45% ASK yield.
Today I bought 3000-4000 SPX 09/30/2022 expiration box @996.5 (2.82% yield) with price improvement. WSJ quotes show 09/30 treasuries at ca. 1.93% ASK yield https://www.wsj.com/market-data/bonds/treasuries .
My spread to treasuries would be ca. 0.9%. Can that be right? Can someone please verify my math?

The yield curve http://www.worldgovernmentbonds.com/cou ... ed-states/ seems to show consistently about 0.2% higher yields than WSJ.
I see high ask yield of 1.998% for 9/30/22 from Fidelity yesterday, but all of those have coupons. Hi yield for close maturities was 2.185% for a T bill maturing 9/27/22. I would use the T bill yield for comparison, since that's what I'd buy if I were buying a Treasury of that maturity.

For your box trade I get a yield of 2.85%, so 66 bps higher than Treasury yield. Not quite as high as your calculations show, but still excellent. I was debating transferring more cash into Fidelity to buy another box spread this week, or wait until the Friday expiration of my 8/19/22 box spread, and buy more next week. I'm inspired buy your purchase to do the former.

Kevin
I just now realized that WSJ has another tab for T-bills.
But what is the reason for the different yields to maturity of the various bills, notes, and bonds? The yields are even different when the coupons are almost identical, and with such a short time to maturity the coupons should not measurably alter duration in the first place.
User avatar
Kevin M
Posts: 15787
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: Let's Talk SPX Box Spreads

Post by Kevin M »

comeinvest wrote: Tue Aug 16, 2022 6:31 pm I just now realized that WSJ has another tab for T-bills.
But what is the reason for the different yields to maturity of the various bills, notes, and bonds? The yields are even different when the coupons are almost identical, and with such a short time to maturity the coupons should not measurably alter duration in the first place.
Yes, the Treasury yield curve is very "noisy". Explanations I can think are duration, liquidity, cash flow preference, and maturity preference. Here are ask yields from Fidelity for maturities up to about 6 months:

Image

Typically it seems that bills have the highest yield for a given approximate maturity. Higher coupon bonds and STRIPS seem to typically have lower yields.

Duration does not seem to explain much if anything for Treasuries with less than six months to maturity, since there are no coupon payments before maturity. Personally, I don't care much about liquidity, nor do I have any cash flow preferences, so I'll use the highest yield with the same approximate maturity as the expiration date for the box spread for comparison purposes.

Kevin
If I make a calculation error, #Cruncher probably will let me know.
User avatar
Kevin M
Posts: 15787
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: Let's Talk SPX Box Spreads

Post by Kevin M »

For box spreads that expire on Friday (settle on Monday), does anyone know for sure that one could buy a replacement box spread on Friday at Fidelity with no cash, using the settled funds on Monday to cover the new box spread? I would assume it works this way, but had a bad experience with Fidelity charging me some margin interest when they shouldn't have, and they could not explain why they did so.

Thanks,

Kevin
If I make a calculation error, #Cruncher probably will let me know.
User avatar
indexfundfan
Posts: 3962
Joined: Tue Feb 20, 2007 10:21 am
Contact:

Re: Let's Talk SPX Box Spreads

Post by indexfundfan »

Kevin M wrote: Wed Aug 17, 2022 12:07 pm For box spreads that expire on Friday (settle on Monday), does anyone know for sure that one could buy a replacement box spread on Friday at Fidelity with no cash, using the settled funds on Monday to cover the new box spread? I would assume it works this way, but had a bad experience with Fidelity charging me some margin interest when they shouldn't have, and they could not explain why they did so.

Thanks,

Kevin
I buy and sell ETFs often on the same day at Fidelity, so that their credit and debit cancel each other. I have not been charged margin interest. I'm not sure why it would be different for option trades. But I have not had a specific example with matching up option trade settlements.

If/When they charge you margin interest, can you point them to the settlement dates on the confirmation statement to argue that there should not be margin interest charged?
My signature has been deleted.
User avatar
Kevin M
Posts: 15787
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: Let's Talk SPX Box Spreads

Post by Kevin M »

indexfundfan wrote: Wed Aug 17, 2022 12:23 pm
Kevin M wrote: Wed Aug 17, 2022 12:07 pm For box spreads that expire on Friday (settle on Monday), does anyone know for sure that one could buy a replacement box spread on Friday at Fidelity with no cash, using the settled funds on Monday to cover the new box spread? I would assume it works this way, but had a bad experience with Fidelity charging me some margin interest when they shouldn't have, and they could not explain why they did so.

Thanks,

Kevin
I buy and sell ETFs often on the same day at Fidelity, so that their credit and debit cancel each other. I have not been charged margin interest. I'm not sure why it would be different for option trades. But I have not had a specific example with matching up option trade settlements.

If/When they charge you margin interest, can you point them to the settlement dates on the confirmation statement to argue that there should not be margin interest charged?
Thanks, that's what I'd think too.

The problem at Fidelity was that I bought multiple box spreads, which of course show up as four option trades, so they just kept saying something like, "You had all of these large option trades and there must have been a point at which the margin interest was due", but they could never point to the date on which it was incurred. I had my transaction history in front of me, and they wouldn't even go through it with me--it was just too overwhelming for them.

For one trade, I used a cash deposit I made, which had not settled but was included in my available to trade balance. I also transferred some to FZDXX, which also serves as a settlement fund, and the timing of this might have thrown things off. I finally asked to speak to a supervisor, and he credited me the interest, which is what I had asked for at the beginning of this hour + charade.

Kevin
If I make a calculation error, #Cruncher probably will let me know.
User avatar
indexfundfan
Posts: 3962
Joined: Tue Feb 20, 2007 10:21 am
Contact:

Re: Let's Talk SPX Box Spreads

Post by indexfundfan »

Kevin M wrote: Wed Aug 17, 2022 12:33 pm I also transferred some to FZDXX, which also serves as a settlement fund, ...
I think this is where the trouble happens.

If you submit a purchase of FZDXX, in a margin account, and there was a debit balance, Fidelity wouldn't sell FZDXX to settle the debit. This results in a debit balance that will accrue margin interest.

I knew this because I was charged margin interest among several trades I made. I couldn't figured out why I was charged margin interest (because I always had a positive balance in FZDXX), until I created a spreadsheet and track exactly every trade and their settlement during that period. The margin interest was around $4 so I didn't even bother to call them.
My signature has been deleted.
comeinvest
Posts: 2709
Joined: Mon Mar 12, 2012 6:57 pm

Re: Let's Talk SPX Box Spreads

Post by comeinvest »

Kevin M wrote: Wed Aug 17, 2022 12:33 pm
indexfundfan wrote: Wed Aug 17, 2022 12:23 pm
Kevin M wrote: Wed Aug 17, 2022 12:07 pm For box spreads that expire on Friday (settle on Monday), does anyone know for sure that one could buy a replacement box spread on Friday at Fidelity with no cash, using the settled funds on Monday to cover the new box spread? I would assume it works this way, but had a bad experience with Fidelity charging me some margin interest when they shouldn't have, and they could not explain why they did so.

Thanks,

Kevin
I buy and sell ETFs often on the same day at Fidelity, so that their credit and debit cancel each other. I have not been charged margin interest. I'm not sure why it would be different for option trades. But I have not had a specific example with matching up option trade settlements.

If/When they charge you margin interest, can you point them to the settlement dates on the confirmation statement to argue that there should not be margin interest charged?
Thanks, that's what I'd think too.

The problem at Fidelity was that I bought multiple box spreads, which of course show up as four option trades, so they just kept saying something like, "You had all of these large option trades and there must have been a point at which the margin interest was due", but they could never point to the date on which it was incurred. I had my transaction history in front of me, and they wouldn't even go through it with me--it was just too overwhelming for them.

For one trade, I used a cash deposit I made, which had not settled but was included in my available to trade balance. I also transferred some to FZDXX, which also serves as a settlement fund, and the timing of this might have thrown things off. I finally asked to speak to a supervisor, and he credited me the interest, which is what I had asked for at the beginning of this hour + charade.

Kevin
I don't mean to take sides, but I think it's unlikely that their interest calc was incorrect. If you bought a fund, like FZDXX, this transaction has a 2 day settlement period as far as I know. Your options have 1 day settlement. Then you said you made deposits, which have their own settlement times. Etc. Re-do your math. They probably just credited your account to get rid of you before they incurred too much customer service cost.

EDIT: "Money market funds close and settles on the same day as the trade date." Nevertheless, do the accurate math with the accurate settlement dates.
Last edited by comeinvest on Wed Aug 17, 2022 6:00 pm, edited 2 times in total.
adamhg
Posts: 218
Joined: Sat Apr 10, 2021 8:40 am

Re: Let's Talk SPX Box Spreads

Post by adamhg »

Any interest in seeing this updated regularly?

Image

Box spread vs Treasury yields (8/16/22)
comeinvest
Posts: 2709
Joined: Mon Mar 12, 2012 6:57 pm

Re: Let's Talk SPX Box Spreads

Post by comeinvest »

adamhg wrote: Wed Aug 17, 2022 5:53 pm Any interest in seeing this updated regularly?

Box spread vs Treasury yields (8/16/22)
Yes! But please use the zero coupon yield curve, not the CMT or other curves, as only zero coupon treasuries are comparable to zero coupon box spreads. Which one did you use?

Zero coupon yields: https://data.nasdaq.com/data/FED/SVENY- ... ield-curve click on "Table" to see the zero coupon data per maturity

Your treasury yields in the 3y maturity area seem too high.

Interesting charts related to your effort are in this post: viewtopic.php?p=6503540#p6503540 which references this document: https://www.stern.nyu.edu/sites/default ... a_19_1.pdf
Last edited by comeinvest on Wed Aug 17, 2022 6:36 pm, edited 2 times in total.
comeinvest
Posts: 2709
Joined: Mon Mar 12, 2012 6:57 pm

Re: Let's Talk SPX Box Spreads

Post by comeinvest »

Kevin M wrote: Tue Aug 16, 2022 7:03 pm
comeinvest wrote: Tue Aug 16, 2022 6:31 pm I just now realized that WSJ has another tab for T-bills.
But what is the reason for the different yields to maturity of the various bills, notes, and bonds? The yields are even different when the coupons are almost identical, and with such a short time to maturity the coupons should not measurably alter duration in the first place.
Yes, the Treasury yield curve is very "noisy". Explanations I can think are duration, liquidity, cash flow preference, and maturity preference. Here are ask yields from Fidelity for maturities up to about 6 months:

Image

Typically it seems that bills have the highest yield for a given approximate maturity. Higher coupon bonds and STRIPS seem to typically have lower yields.

Duration does not seem to explain much if anything for Treasuries with less than six months to maturity, since there are no coupon payments before maturity. Personally, I don't care much about liquidity, nor do I have any cash flow preferences, so I'll use the highest yield with the same approximate maturity as the expiration date for the box spread for comparison purposes.

Kevin
"Typically it seems that bills have the highest yield for a given approximate maturity" - I am referencing a post in the sister thread to this thread, that seems to show the opposite. Look at the charts.

viewtopic.php?p=6503540#p6503540 which references this study: https://www.stern.nyu.edu/sites/default ... a_19_1.pdf
klaus14 wrote: Tue Feb 08, 2022 12:03 am
comeinvest wrote: Mon Feb 07, 2022 9:43 pm
klaus14 wrote: Mon Feb 07, 2022 6:03 pm
adamhg wrote: Mon Feb 07, 2022 2:39 pm
boxspreads wrote: Sat Feb 05, 2022 8:22 pm Also, do I understand the commissions field on boxtrades.com correctly? Adding a commission seems to reduce the rate with commissions? Should it not increase the effective rate?
indexfundfan wrote: Mon Feb 07, 2022 2:05 pm For a long box, the commission decreases the rate (the earnings interest rate) but for a short box, it increases the rate (the cost of borrowing).
I see what you're both saying, and you are right. I am actually long boxes right now, so I wasn't even looking/thinking through the short box calculations much. Anyways, this has been fixed. Please let me know if its calculating to your expectations
Can you explain why you are long boxes?

I am also considering being long in boxes instead of holding treasuries or CDs because
- "interest" is capital gains.
- higher interest than treasuries. and currently higher than CDs

I am wondering if there is any downside to this compared to treasuries. I guess at times of crisis the spread between treasuries and boxes would widen so the boxes you hold would be worth less. To mitigate this i am considering having a "box ladder" to fund retirement expenses like: 1y box + 2y box + 3y box + 4y box + 5y box. in this way you wouldn't have to sell them prematurely.
I posted a similar idea here: viewtopic.php?p=6493042&sid=dda3dcb49e1 ... 2#p6493042

"I guess at times of crisis the spread between treasuries and boxes would widen so the boxes you hold would be worth less." - I'm not sure about that. On one hand, the TED spread (LIBOR minus T-Bills) sometimes increased during times of crisis in the past. But then again, while LIBOR is sometimes mentioned in connection with futures and options financing rates, futures and options boxes are not LIBOR. LIBOR is an unsecured interbank rate, while boxes are secured by margin accounts and guaranteed by the OCC. During a macroeconomic crash and stock market drawdown, first off all the Fed typically reduces short-term rates soon, which benefits long bond or long box positions. Secondly, the demand for (leveraged) equities typically decreases during crashes, which means implied financing cost with equities as collateral might be less expensive. (But the margin requirement typically increases.) Don't quote me on any of that, as I am not an expert on the subject; but you can google for general collateral, special collateral, repo market, and you will find papers that examine the history of those with charts.

I think the options financing rate is somewhat related to the futures financing rate, and both are related to the "repo" market. To understand the dynamics, it might be worthwhile looking at futures financing rates in March 2020. The only roll data that I found is that of the MME (MSCI Emerging Markets) and MFS (MSCI EAFE) futures:
https://www.theice.com/pace-of-roll/MSC ... 0-20200619
https://www.theice.com/pace-of-roll/MSC ... 0-20200619
Surprisingly, the MEE behaved normally with spreads above LIBOR of around 0.5% on average, while MFS had a spike in the spread above LIBOR topping out at more than 6% above LIBOR.
Hard to understand the different behavior.

I have no idea how long maturity box rates (e.g. 5 years) might be affected by a macroeconomic crash or capital market problem, vs. shorter maturities. I don't know of a data source for historical data.
See:
Fxmove88 wrote: Sat Dec 11, 2021 12:28 am One has to be patient in getting a box filled. Just do a GTC limit order for selling the box. If you insist on getting filled immediately then you will be paying high interest.

The following study shows that Box interest rate has been about 0.38% above corresponding US treasury yield for various tenor. And during the crisis of 2008, the spread goes up triple.

https://www.stern.nyu.edu/sites/default ... a_19_1.pdf

Image

Image

Dec 2022 Box rate should be around 0.65%, given that the 1 year US treasury yield is at 0.27%
https://ycharts.com/indicators/1_year_treasury_rate

Correspondingly, Dec 2023 Box rate should be around 1.05% with the current 2 year US treasury yield at 0.67%
https://ycharts.com/indicators/2_year_treasury_rate
comeinvest
Posts: 2709
Joined: Mon Mar 12, 2012 6:57 pm

Re: Let's Talk SPX Box Spreads

Post by comeinvest »

I find the charts that show the behavior of box spreads during stock market crashes very interesting actually.

viewtopic.php?p=6503540#p6503540 which references this study: https://www.stern.nyu.edu/sites/default ... a_19_1.pdf

They imply that box spreads come with a perceived tail risk, or not? The risk would be incurred by the parties who bought the boxes, as the options have option value that would be lost if CBOE were to fail due to a systemic crisis and the government were to not bail them out, e.g. there is a stock market flash crash bigger than the risk parameters that margin requirements are based on, and some big players who financed using box spreads can't pay back their debt. Which may help explaining the slightly higher implied yield in comparison to treasuries.

The party who sold the boxes i.e. the party who used boxes for financing, seems to benefit from an upward revaluation of their short box spread positions during times of crisis. (I'm not sure if the margin requirement of the short box increased during market crashes, though.) The price paid by the short box spread position holders for that anti-correlation to the stock market would be the additional financing spread to treasuries of ca. 0.35% on average, which is still cheaper than any other financing method available to retail investors.

Comments?
User avatar
indexfundfan
Posts: 3962
Joined: Tue Feb 20, 2007 10:21 am
Contact:

Re: Let's Talk SPX Box Spreads

Post by indexfundfan »

adamhg wrote: Wed Aug 17, 2022 5:53 pm Any interest in seeing this updated regularly?

Image

Box spread vs Treasury yields (8/16/22)
+1 That would be nice!
My signature has been deleted.
User avatar
indexfundfan
Posts: 3962
Joined: Tue Feb 20, 2007 10:21 am
Contact:

Re: Let's Talk SPX Box Spreads

Post by indexfundfan »

indexfundfan wrote: Wed Aug 17, 2022 12:42 pm
Kevin M wrote: Wed Aug 17, 2022 12:33 pm I also transferred some to FZDXX, which also serves as a settlement fund, ...
I think this is where the trouble happens.

If you submit a purchase of FZDXX, in a margin account, and there was a debit balance, Fidelity wouldn't sell FZDXX to settle the debit. This results in a debit balance that will accrue margin interest.

I knew this because I was charged margin interest among several trades I made. I couldn't figured out why I was charged margin interest (because I always had a positive balance in FZDXX), until I created a spreadsheet and track exactly every trade and their settlement during that period. The margin interest was around $4 so I didn't even bother to call them.
Interestingly, I just saw someone report such a case on Reddit.

https://www.reddit.com/r/fidelityinvest ... ing_money/

So if you have a margin account, be very careful when placing a trade to purchase FZDXX if you have another trade that is settling on that same day. Fidelity will not liquidate FZDXX to settle for that trade; instead it will draw from margin.

For myself, I have since removed FZDXX in the margin account and use it only in a non-margin account.
My signature has been deleted.
petulant
Posts: 3601
Joined: Thu Sep 22, 2016 1:09 pm

Re: Let's Talk SPX Box Spreads

Post by petulant »

adamhg wrote: Wed Aug 17, 2022 5:53 pm Any interest in seeing this updated regularly?

Image

Box spread vs Treasury yields (8/16/22)
Could it be generated on the fly with inputs for marginal tax rate applicable to box spreads and marginal tax rate applicable to treasury bonds?
adamhg
Posts: 218
Joined: Sat Apr 10, 2021 8:40 am

Re: Let's Talk SPX Box Spreads

Post by adamhg »

comeinvest wrote: Wed Aug 17, 2022 6:02 pm Yes! But please use the zero coupon yield curve, not the CMT or other curves, as only zero coupon treasuries are comparable to zero coupon box spreads. Which one did you use?

Zero coupon yields: https://data.nasdaq.com/data/FED/SVENY- ... ield-curve click on "Table" to see the zero coupon data per maturity

Your treasury yields in the 3y maturity area seem too high.
Thanks. Bills only go up to 1 year and SVENY only starts at 1 year and only released weekly. Here's all 4 plotted together for reference. You can see a .0008% drop from bills to SVENY, but I believe that's because SVENY quotes are from 8/12 whereas everything else is as of 8/16.

Image

comeinvest wrote: Wed Aug 17, 2022 6:02 pm Interesting charts related to your effort are in this post: viewtopic.php?p=6503540#p6503540 which references this document: https://www.stern.nyu.edu/sites/default ... a_19_1.pdf
Yep! I'm familiar with this study. It's great and one of the reasons I started this. I followed their similar methodology to build the bid/mid/ask spread in my previous post.
The data set contains the bid price, the ask price, the strike and the maturity date for a large range of strike prices for each minute. We compute risk-free rate estimates at the minute level using the mid prices of puts and calls of all strike prices with a particular maturity. To compute daily estimates, we then take a median over the minute-level estimates in the day
But the difference is that I'm using actual trades and not the mid points which their study is based on. I didn't get a chance to post the outcome of that previous post, but I'll do that here in a sec.
adamhg
Posts: 218
Joined: Sat Apr 10, 2021 8:40 am

Re: Let's Talk SPX Box Spreads

Post by adamhg »

comeinvest wrote: Thu Aug 11, 2022 12:28 am
adamhg wrote: Fri Aug 05, 2022 11:11 am Here's a new project to calculate live box spread rate spreads. Quote times are UTC and were live as of the screenshot.

Image

With the large interest rate movement today, I thought it was a good opportunity post this before the daily boxtrades.com dump tomorrow and keep me honest. Looked pretty consistent except for the first one and the last few starting around Dec'24. The short duration makes it harder to find a good mid-point and the last few appears to be when liquidity starts to dry up on the ask side.
Interesting to see the tabular presentation with the synthetic box quotes, but what is the takeaway? I see that the bid and ask quotes get closer to each other with time to expiration, then get wider again. Is that because of volumes of the individual options for the different expirations?
All of the rates have significantly wider i.e. worse indicative bid/ask spreads than what you would get with a combination order. I'm not sure if we can learn anything; even the midpoint is not necessarily a good representation of the mark price, based on some articles that I read the other day.
If we want to compare the mid price implied rates to the treasuries rates, I don't see the drop in implied rates between June 2024 and Dec 2024 reflected in the treasury curve. But again, the options midprice data may be skewed by quotes from market participants, and other intricacies of options pricing.
I was trying to see how accurate extrapolating the box spread rates could be from the individual bid/ask spreads of each SPX put and call option. The rate would have to be within the spread, but I was hoping that the mid point would allow us to pinpoint the going rate of a box spread and take the guess work out of it.

Image

Here are the rates side by side, the actuals are actual box spread trades, rate (avg), high and low rates of the day. The SPX box spread columns are from the original table I posted, but the live mid point rates between the best bid and best ask for a box spread at each tenor. The last column tell us how close the midpoint related to the average box spread rate that traded on 8/5.

Anyways, I'm not exactly sure how I feel about the results. They seem large in certain ways, but actually average out to about 1% difference which isn't bad in aggregate. It also makes a much nicer looking yield curve (orange) than the actual trades (blue):

Image
adamhg
Posts: 218
Joined: Sat Apr 10, 2021 8:40 am

Re: Let's Talk SPX Box Spreads

Post by adamhg »

...and one last chart to combine the last two posts. Here's the SPX bid/ask mid-point method to estimate box spread rates plotted against the 8/16 actual trade data

Image
comeinvest
Posts: 2709
Joined: Mon Mar 12, 2012 6:57 pm

Re: Let's Talk SPX Box Spreads

Post by comeinvest »

adamhg wrote: Wed Aug 17, 2022 9:51 pm
comeinvest wrote: Wed Aug 17, 2022 6:02 pm Yes! But please use the zero coupon yield curve, not the CMT or other curves, as only zero coupon treasuries are comparable to zero coupon box spreads. Which one did you use?

Zero coupon yields: https://data.nasdaq.com/data/FED/SVENY- ... ield-curve click on "Table" to see the zero coupon data per maturity

Your treasury yields in the 3y maturity area seem too high.
Thanks. Bills only go up to 1 year and SVENY only starts at 1 year and only released weekly. Here's all 4 plotted together for reference. You can see a .0008% drop from bills to SVENY, but I believe that's because SVENY quotes are from 8/12 whereas everything else is as of 8/16.

Image

comeinvest wrote: Wed Aug 17, 2022 6:02 pm Interesting charts related to your effort are in this post: viewtopic.php?p=6503540#p6503540 which references this document: https://www.stern.nyu.edu/sites/default ... a_19_1.pdf
Yep! I'm familiar with this study. It's great and one of the reasons I started this. I followed their similar methodology to build the bid/mid/ask spread in my previous post.
The data set contains the bid price, the ask price, the strike and the maturity date for a large range of strike prices for each minute. We compute risk-free rate estimates at the minute level using the mid prices of puts and calls of all strike prices with a particular maturity. To compute daily estimates, we then take a median over the minute-level estimates in the day
But the difference is that I'm using actual trades and not the mid points which their study is based on. I didn't get a chance to post the outcome of that previous post, but I'll do that here in a sec.
First of all, very interesting chart, thank you! Might come in very useful for box trade positioning, once we improve confidence in the data and in the chart.

Some thoughts:

- The chart shows a sharp dropoff in the T-bill curve in the ca. 0-2 mo. area. I'm not sure what I shall make out of it. It might suggest to not borrow using box spreads at <= 2 mo. maturities, or perhaps the T-bills don't really represent the expected rates per expectation hypothesis in that segment. Aug 31 SOFR futures are currently at ca. 2.3% implied yield, and Sep 30 SOFR futures are at ca. 2.45% implied yield, Oct 31 at 2.9%, Nov 30 at 3.25%. Kind of consistent with the T-bill rates on your chart. (SOFR futures are rearwards looking and expire a month after the date that they are named after.) I still don't see how this can not violate some non-arbitrage principle or expectation hypothesis, for example if I were to invest (buy) now in a < 2 mo. box, and after it expires roll into the then prevailing box rate assuming I then get a similar spread to the T-bill rates, all the while shorting now a 4-month box for example (violation of no-arbitrage principle via yield curve expectation hypothesis). Re-phrasing the conundrum, you could also say that with the anomaly in the 0-2 mo. segment, the expectation hypothesis can not hold true for both the blue and the red chart at the same time.

- It seems like your blue line does not reflect your plots on boxtrades.com : For example both 06/2024 and 12/2024 boxes on boxtrades.com are pretty accurately at ca. 3.5%, but your blue line is at 2.625% and 2.55-2.6% respectively at these same expirations. Probably interpolation error -> use more data points

- I have not read the paper on bootstrapping zero coupon curves, but there might be some intricacies. For example I am having some trouble explaining the upward sloping zero coupon curve in the 20-30y segment, all the while the CMT curve is downward sloping in the same segment.

- The interpolation of the grey line does not instill confidence. More data points would be needed.

- Your charts don't seem to reflect the almost static spread of box spreads to treasuries with varying maturities as shown on the chart in viewtopic.php?p=6829477#p6829477 , although the latter is an average over many years, so they are not in direct contradiction if we assume that there are both downward and upward fluctuations in the spread at different maturities at the same time.

- You said SVENY is only released weekly. Make sure that the dates of the various curves are the same; if not either make adjustments, or collect the data when all are available on the same day.
Last edited by comeinvest on Thu Aug 18, 2022 8:45 pm, edited 1 time in total.
comeinvest
Posts: 2709
Joined: Mon Mar 12, 2012 6:57 pm

Re: Let's Talk SPX Box Spreads

Post by comeinvest »

adamhg wrote: Thu Aug 18, 2022 12:11 am ...and one last chart to combine the last two posts. Here's the SPX bid/ask mid-point method to estimate box spread rates plotted against the 8/16 actual trade data
What is the difference between this and the chart in your previous post?
comeinvest
Posts: 2709
Joined: Mon Mar 12, 2012 6:57 pm

Re: Let's Talk SPX Box Spreads

Post by comeinvest »

adamhg wrote: Wed Aug 17, 2022 10:58 pm
comeinvest wrote: Thu Aug 11, 2022 12:28 am
adamhg wrote: Fri Aug 05, 2022 11:11 am Here's a new project to calculate live box spread rate spreads. Quote times are UTC and were live as of the screenshot.

Image

With the large interest rate movement today, I thought it was a good opportunity post this before the daily boxtrades.com dump tomorrow and keep me honest. Looked pretty consistent except for the first one and the last few starting around Dec'24. The short duration makes it harder to find a good mid-point and the last few appears to be when liquidity starts to dry up on the ask side.
Interesting to see the tabular presentation with the synthetic box quotes, but what is the takeaway? I see that the bid and ask quotes get closer to each other with time to expiration, then get wider again. Is that because of volumes of the individual options for the different expirations?
All of the rates have significantly wider i.e. worse indicative bid/ask spreads than what you would get with a combination order. I'm not sure if we can learn anything; even the midpoint is not necessarily a good representation of the mark price, based on some articles that I read the other day.
If we want to compare the mid price implied rates to the treasuries rates, I don't see the drop in implied rates between June 2024 and Dec 2024 reflected in the treasury curve. But again, the options midprice data may be skewed by quotes from market participants, and other intricacies of options pricing.
I was trying to see how accurate extrapolating the box spread rates could be from the individual bid/ask spreads of each SPX put and call option. The rate would have to be within the spread, but I was hoping that the mid point would allow us to pinpoint the going rate of a box spread and take the guess work out of it.

Image

Here are the rates side by side, the actuals are actual box spread trades, rate (avg), high and low rates of the day. The SPX box spread columns are from the original table I posted, but the live mid point rates between the best bid and best ask for a box spread at each tenor. The last column tell us how close the midpoint related to the average box spread rate that traded on 8/5.

Anyways, I'm not exactly sure how I feel about the results. They seem large in certain ways, but actually average out to about 1% difference which isn't bad in aggregate. It also makes a much nicer looking yield curve (orange) than the actual trades (blue):

Image
I think to be meaningful, please add information on the number of trades that were used for the averages of the trades. Some averages suggest that they might be based on only one trade. That trade could be either a dealer (sell side) or dealer (buy side) trade, which would heavily skew the results. Even when only a few trades for an expiration were used, I would not expect the averages to be representative.

The bid/ask quotes could be skewed by customer bids or asks within the dealer bid/ask spread. Also, I read somewhere that the bid/ask quotes of options are typically not arithmetically centered around the fair value estimated market price, but geometrically centered (forgot the details), which means the midpoint would not necessarily be the fair value estimated market price (where trades would likely happen), especially when the bid/ask spreads are as big as in your data. Perhaps if you use synthetic spreads at strike prices closer to the current index value, you would get more accurate results due to smaller bid/ask spreads.

Obviously your data and charts seem to have a lot of noise, and don't instill confidence. Both the actual trade data as well as the bid/ask data would probably need interpolation over many strike prices and expirations, and perhaps over observation times adjusted for market movements, to be more accurate, as single options don't have enough volume to be representative. I think that is what brokers and other quantitative options people do to calculate "mark prices" of options. The options market is a naturally very fragmented market due to the many strike price and expiration combinations. Interactive Brokers and probably most other brokers show the mark price of box spreads, if you enter them in IB's TWS. They don't disclose their algo that they use to arrive at their mark prices (I asked them). But you can cross-check your results with theirs. I think you are trying to embark on a relatively complex numerical problem here, which is basically trying to calculate reliable mark prices from a large amount of disparate data points across strike prices, expirations, and observation times.
adamhg
Posts: 218
Joined: Sat Apr 10, 2021 8:40 am

Re: Let's Talk SPX Box Spreads

Post by adamhg »

comeinvest wrote: Thu Aug 18, 2022 5:21 am ...snip
- The chart shows a sharp dropoff in the T-bill curve in the ca. 0-2 mo. area. I'm not sure what I shall make out of it. It might suggest to not borrow using box spreads at <= 2 mo. maturities, or perhaps the T-bills don't really represent the expected rates per expectation hypothesis in that segment. Aug 31 SOFR futures are currently at ca. 2.3% implied yield, and Sep 30 SOFR futures are at ca. 2.45% implied yield, Oct 31 at 2.9%, Nov 30 at 3.25%. Kind of consistent with the T-bill rates on your chart. (SOFR futures are rearwards looking and expire a month after the date that they are named after.) I still don't see how this can not violate some non-arbitrage principle or expectation hypothesis, for example if I were to invest (buy) now in a < 2 mo. box, and after it expires roll into the then prevailing box rate assuming I then get a similar spread to the T-bill rates, all the while shorting now a 4-month box for example (violation of no-arbitrage principle via yield curve expectation hypothesis). Re-phrasing the conundrum, you could also say that with the anomaly in the 0-2 mo. segment, the expectation hypothesis can not hold true for both the blue and the red chart at the same time.
The box spread rates are pretty wonky at <2 mo, so that could just be noisy data. I implement a filter for the charts, so that might explain some of it. Once the options get a few weeks from expiry, you need really wide boxes for the yield calculations to make sense, otherwise the denominator (i.e. 365 / dte) makes the rates fluctuate too wildly. So there might be something here, but it could just be noise from being close to expiry.
comeinvest wrote: Thu Aug 18, 2022 5:21 am - It seems like your blue line does not reflect your plots on boxtrades.com : For example both 06/2024 and 12/2024 boxes on boxtrades.com are pretty accurately at ca. 3.5%, but your blue line is at 2.625% and 2.55-2.6% respectively at these same expirations. Probably interpolation error -> use more data points
I'm not seeing what you're seeing. Blue line for Dec '24 was 3.59 (with a single trade on that day). June '24 is actually missing from the data because there weren't any trades there since early Aug, so what you're seeing is just connecting the dots.
comeinvest wrote: Thu Aug 18, 2022 5:21 am - Your charts don't seem to reflect the almost static spread of box spreads to treasuries with varying maturities as shown on the chart in viewtopic.php?p=6829477#p6829477 , although the latter is an average over many years, so they are not in direct contradiction if we assume that there are both downward and upward fluctuations in the spread at different maturities at the same time.
I noticed that too. I think the narrowing of the gap seems to happen more farther out and CBOE only recently opened up 4-5 yr SPX options so they wouldn't have been available at the time of this study.
comeinvest wrote: Thu Aug 18, 2022 5:21 am - You said SVENY is only released weekly. Make sure that the dates of the various curves are the same; if not either make adjustments, or collect the data when all are available on the same day.
Agreed, but WSJ (where I got the bond/bill yields) doesn't post historical data so I'll have to re-run it next time SVENY is released. I'll try to remember, but if not, if somebody could grab them for me, I can run the box spreads against it.
comeinvest wrote: Thu Aug 18, 2022 5:41 am
adamhg wrote: Thu Aug 18, 2022 12:11 am ...and one last chart to combine the last two posts. Here's the SPX bid/ask mid-point method to estimate box spread rates plotted against the 8/16 actual trade data
What is the difference between this and the chart in your previous post?
Forgot to explain. The chart in quotes is based on daily reported (near EOD) bid/ask. I think its like 10m before close, where as the previous chart is based off my live brokerage data. So the difference is mid-day vs EOD quotes and the EOD quotes I can run historical whereas I won't be able to with real-time data
comeinvest wrote: Thu Aug 18, 2022 5:53 am I think to be meaningful, please add information on the number of trades that were used for the averages of the trades. Some averages suggest that they might be based on only one trade. That trade could be either a dealer (sell side) or dealer (buy side) trade, which would heavily skew the results. Even when only a few trades for an expiration were used, I would not expect the averages to be representative.
Unfortunately there simply aren't many data points for most days so this will be fairly limited and you're right, most of the longer duration / less liquid options really do only have a single trade or none at all. To get some more context, there are about 20 active expirations right now for SPX and on any given day only about 7-8 are traded where most of them only have a single trade. There only maybe 2-3 that have multiple trades a day, so the data is going to inherently be noisy which is why I've been trying to use the bid/asks to build better data points.
comeinvest wrote: Thu Aug 18, 2022 5:53 amThe bid/ask quotes could be skewed by customer bids or asks within the dealer bid/ask spread. Also, I read somewhere that the bid/aks quotes of options are typically not centered around the fair value, but logarithmically skewed (forgot the details), which means the midpoint would not necessarily be the fair value or where trades would likely happen, especially when the bid/ask spreads are as big as in your data. Perhaps if you use synthetic spreads at strike prices closer to the current index value, you would get more accurate results due to smaller bid/ask spreads.

Obviously your data and charts seem to have a lot of noise, and don't instill confidence. Both the actual trade data as well as the bid/ask data would probably need interpolation over many strike prices and expirations, to be more accurate, as single options don't have enough volume to be representative. I think that is what brokers and other quantitative options people do to calculate "mark prices" of options. The options market is a naturally very fragmented market due to the many strike price and expiration combinations. Interactive Brokers and probably most other brokers show the mark price of box spreads, if you enter them in IB's TWS. They don't disclose their algo that they use to arrive at their mark prices (I asked them). But you can cross-check your results with theirs. I think you are trying to embark on a relatively complex numerical task here, which is basically trying to calculate reliable mark prices from a large amount of disparate data points across strike prices, expirations, and times of observation.
With the synthetic spreads, I'm using the mid points between the best bid and best ask. So I'm calculating every combination of box spread possible and using the best bid and ask so this should be the natural outcome. For instance, Aug '22 there are 414 strikes, so I'm testing 414 * 413 = 170,982 combinations of spreads to find the best bid and best ask. It's also almost certainly the case that the best bid and ask aren't from the same strikes. So assuming there's no systemic reason (big if) the "fair market value" to be closer to the bid or ask, then this should based on averages hopefully put the midpoint closer to the actual value. If you have a paper to show how to better mark bid/ask spreads, I'd love to read it.
comeinvest
Posts: 2709
Joined: Mon Mar 12, 2012 6:57 pm

Re: Let's Talk SPX Box Spreads

Post by comeinvest »

adamhg wrote: Thu Aug 18, 2022 11:53 pm With the synthetic spreads, I'm using the mid points between the best bid and best ask. So I'm calculating every combination of box spread possible and using the best bid and ask so this should be the natural outcome. For instance, Aug '22 there are 414 strikes, so I'm testing 414 * 413 = 170,982 combinations of spreads to find the best bid and best ask. It's also almost certainly the case that the best bid and ask aren't from the same strikes. So assuming there's no systemic reason (big if) the "fair market value" to be closer to the bid or ask, then this should based on averages hopefully put the midpoint closer to the actual value. If you have a paper to show how to better mark bid/ask spreads, I'd love to read it.
I think your approach is flawed. If you take the "best" quotes, you are likely to capture "stray" quotes, e.g. retail customer quotes. If anything I would take averages over larger samples. But with the wide spreads I think the midpoint is unlikely the mark price and not even necessarily very close to it. I think to be of any value, you would need to apply sophisticated interpolation techniques across strike prices, expirations, and observation times, perhaps incorporating general market movements. (Those parameters come to mind.) I have a lot more confidence in your actual box trade charts. The biggest improvement that I would like to see for your boxtrades site is to differentiate between dealer buy and dealer sell trades of the native combination orders and fills. Does the raw options exchange data have any information on this? I know for bond trades for example market places are required to disclose which side was the dealer.
gougou
Posts: 1317
Joined: Thu Sep 28, 2017 7:42 pm

Re: Let's Talk SPX Box Spreads

Post by gougou »

Today I'm rolling my SPX Box Spreads forward for 3 months. I just sold Nov 17, 3000 - 5000 Box at 1985.1. I'm not sure how to calculate the exact APR, but it should be about 3%.

I do have a question about whether to use the Nov 17 (Thursday) or the Nov 18 (Friday) boxes. I'm always doing the trades on Fridays, which box is better for me? Do I get charged margin interests from Nov 17 to Nov 18?
The sillier the market’s behavior, the greater the opportunity for the business like investor.
Lowlim
Posts: 214
Joined: Wed Jun 12, 2019 11:11 pm

Re: Let's Talk SPX Box Spreads

Post by Lowlim »

I'm trading a box for the first time on IBKR. Am I doing this correctly?

Image
gougou
Posts: 1317
Joined: Thu Sep 28, 2017 7:42 pm

Re: Let's Talk SPX Box Spreads

Post by gougou »

Lowlim wrote: Fri Aug 19, 2022 1:40 pm I'm trading a box for the first time on IBKR. Am I doing this correctly?

Image
Looks good. I like to make the numbers positive (instead of 3900 - 3800 box I use 3800 - 3900 box), and I then sell a positive box. Makes things less confusing for me.
The sillier the market’s behavior, the greater the opportunity for the business like investor.
adamhg
Posts: 218
Joined: Sat Apr 10, 2021 8:40 am

Re: Let's Talk SPX Box Spreads

Post by adamhg »

comeinvest wrote: Fri Aug 19, 2022 4:54 am I think your approach is flawed. If you take the "best" quotes, you are likely to capture "stray" quotes, e.g. retail customer quotes. If anything I would take averages over larger samples. But with the wide spreads I think the midpoint is unlikely the mark price and not even necessarily very close to it. I think to be of any value, you would need to apply sophisticated interpolation techniques across strike prices, expirations, and observation times, perhaps incorporating general market movements. (Those parameters come to mind.) I have a lot more confidence in your actual box trade charts. The biggest improvement that I would like to see for your boxtrades site is to differentiate between dealer buy and dealer sell trades of the native combination orders and fills. Does the raw options exchange data have any information on this? I know for bond trades for example market places are required to disclose which side was the dealer.
It made me curious, so if we were worried about catching a customer quote or two here's how the top 10 best bids and asks with the average mid of all 10
Image

...and here is the average of all 10 best bid/ask midpoints imposed upon the earlier actuals vs synthetics on 8/16.
Image

with the only notable differences between the average mids and the best mid for Dec '27 (5.8% different) and Oct '23 (1.1% different which isn't even shown on the site because of how few actual trades there are) and afterwards nothing greater than 0.3% and netting to a 0.01% (not 1%) difference.
Lowlim
Posts: 214
Joined: Wed Jun 12, 2019 11:11 pm

Re: Let's Talk SPX Box Spreads

Post by Lowlim »

gougou wrote: Fri Aug 19, 2022 1:56 pm
Lowlim wrote: Fri Aug 19, 2022 1:40 pm I'm trading a box for the first time on IBKR. Am I doing this correctly?

Image
Looks good. I like to make the numbers positive (instead of 3900 - 3800 box I use 3800 - 3900 box), and I then sell a positive box. Makes things less confusing for me.
A few questions:

1. Do boxes only trade when the market is open?
2. Are there any issues with putting a box up for sale after hours?
3. I understand that the spread is important as it determines the size of your box. But does the price actually matter? Example, SPX is currently trading ~$4230. Are the following strikes equivalent; 4300 - 4100, 4000 - 3800, 2000 - 1800?
4. The commission for my trade is noted as "4.20 ... 6.00." Does this mean that my commission will be between $4.20 and $6.00? If not, what does this mean?
5. In my screenshot, IBKR is indicating that I will have a small change in my initial margin and maintenance margin. The change is negligible, but my understanding was this there should be no change. Why is there going to be a change?
(note: I have portfolio margin and am currently holding mutual funds only.)
gougou
Posts: 1317
Joined: Thu Sep 28, 2017 7:42 pm

Re: Let's Talk SPX Box Spreads

Post by gougou »

Lowlim wrote: Fri Aug 19, 2022 3:14 pm
gougou wrote: Fri Aug 19, 2022 1:56 pm
Lowlim wrote: Fri Aug 19, 2022 1:40 pm I'm trading a box for the first time on IBKR. Am I doing this correctly?

Image
Looks good. I like to make the numbers positive (instead of 3900 - 3800 box I use 3800 - 3900 box), and I then sell a positive box. Makes things less confusing for me.
A few questions:

1. Do boxes only trade when the market is open?
2. Are there any issues with putting a box up for sale after hours?
3. I understand that the spread is important as it determines the size of your box. But does the price actually matter? Example, SPX is currently trading ~$4230. Are the following strikes equivalent; 4300 - 4100, 4000 - 3800, 2000 - 1800?
4. The commission for my trade is noted as "4.20 ... 6.00." Does this mean that my commission will be between $4.20 and $6.00? If not, what does this mean?
5. In my screenshot, IBKR is indicating that I will have a small change in my initial margin and maintenance margin. The change is negligible, but my understanding was this there should be no change. Why is there going to be a change?
(note: I have portfolio margin and am currently holding mutual funds only.)
1. Not sure but I can't seem to trade SPX Boxes outside of regular trading hours.
2. I have never tried trading after hours.
3. I like to pick the options that are round numbers and reasonably active. I don't know whether it matters. I have sold 2000 - 5000 boxes before and got reasonable prices.
4. The commission is going to be $6. $1.5 each leg.
5. If you currently have a negative cash balance, you are going to see some margin changes.
The sillier the market’s behavior, the greater the opportunity for the business like investor.
unemployed_pysicist
Posts: 220
Joined: Sat Oct 09, 2021 2:32 pm
Location: Amsterdam

Re: Let's Talk SPX Box Spreads

Post by unemployed_pysicist »

adamhg wrote: Thu Aug 18, 2022 11:53 pm
comeinvest wrote: Thu Aug 18, 2022 5:21 am - You said SVENY is only released weekly. Make sure that the dates of the various curves are the same; if not either make adjustments, or collect the data when all are available on the same day.
Agreed, but WSJ (where I got the bond/bill yields) doesn't post historical data so I'll have to re-run it next time SVENY is released. I'll try to remember, but if not, if somebody could grab them for me, I can run the box spreads against it.
I quite like this idea of using box spreads to get the yield curve.
I assumed the discount rate for option valuation would be close to the OIS curve (with some spread maybe), but the paper posted/discussed earlier in this thread made some good arguments for why this is not the case/more to it than that. But if you can get it from box spreads, then you get the discount rate for SPX options themselves, no need to figure out the OIS curve, apply a spread, etc.

A few comments regarding SVENY:

The SVENY curve is expressed as a continuously compounding rate, so you will have to convert the bills and box spread rates to continuous compounding to put everything on the same footing. The difference is probably very small, but thought I would mention it if you have not already done this in your chart.

What SVENY maturities do you want to put with the T bills/box spread yield curve? The SVENY01-30 maturities are released every Tuesday at around 4pm EST. The release is based off end-of-day quotes and includes all the yields for these maturities up to the previous friday. E.g., the data will be available up through Friday August 19 starting Tuesday, August 23. The parameters are also given to construct whatever maturities you want. The resulting curve may not match very well to short term issues, and also not very well to T bills due to the segmentation of the T bill market.

The WSJ quotes are from 3pm EST, as compared to the NSS fit of SVENY done with closing prices; hopefully they are close enough in time to get a commensurate measurement.

Here is the SVENY and instantaneous forward curve as of August 12, out to 10 years and sampled monthly (approximately; 0.083 years between data points.)

Image
couldn't afford the h | | BUY BONDS | WEAR DIAMONDS
comeinvest
Posts: 2709
Joined: Mon Mar 12, 2012 6:57 pm

Re: Let's Talk SPX Box Spreads

Post by comeinvest »

adamhg wrote: Thu Aug 18, 2022 11:53 pm With the synthetic spreads, I'm using the mid points between the best bid and best ask. So I'm calculating every combination of box spread possible and using the best bid and ask so this should be the natural outcome. For instance, Aug '22 there are 414 strikes, so I'm testing 414 * 413 = 170,982 combinations of spreads to find the best bid and best ask. It's also almost certainly the case that the best bid and ask aren't from the same strikes. So assuming there's no systemic reason (big if) the "fair market value" to be closer to the bid or ask, then this should based on averages hopefully put the midpoint closer to the actual value. If you have a paper to show how to better mark bid/ask spreads, I'd love to read it.
Two questions here:

1. In your explanation of the table that you first posted, you said you started with the "n" strike prices with the highest OI. Did you change your method between then and now in that you now consider all strike prices?
viewtopic.php?p=6812355#p6812355

2. You are saying you test 414*414 combinations in your example. I'm sure your home computer can handle the computational complexity; but why do you test so many? (I hear energy prices are on an upward trend.) A box can be decomposed into a synthetic long stock positions (the 2 options at the upper strike) and a synthetic short stock position (the 2 options at the lower strike). The optimal synthetic long and the optimal synthetic short should be independent of each other, or not? The synthetic long for example puts a certain dollar amount into your pocket when you buy it. I call that dollar amount "x". At the time of expiration you get exactly the index value at that time minus the strike price, and not a penny more or less than that. To determine the optimal synthetic long position, you can just compare the 414 values of "x - strike", and select the highest. Done. The synthetic long using that exact strike price has the lowest implied financing rate for your synthetic long. Rinse and repeat with the synthetic short. Caveat: I'm new to this. Please correct me if I'm wrong.

Your trades and/or midpoints vs treasuries plot is not consistent with the plot from the paper that was referenced above.
Your midpoint plots seem pretty consistent, but to gain even more confidence and perhaps insight, I would be curious to see a plot for each expiration: the ca. 414 synthetic long implied yields on the y axis versus the strike prices on the x axis; then the ca. 414 synthetic short implied yields versus the strike prices. Then the same for the inverse spread (or buying the original spread). That's 4 * 414 yields for this expiration if I'm not mistaken.
Then we could plot the average midpoints vs. the expiration dates. Possibly eliminate outliers (stray values) before.
Then we have to think about getting more samples of traded box spreads. Perhaps add the trades that you eliminated due to too few trades, and interpolate all data. Then plot again trades vs. bid/ask midpoints and treasury yields. If still not good enough, plot the spreads of trades vs midpoints over time, and/or vs treasury yields over time, for several expirations separately, to try to reverse-engineer a pattern (for example adjusting for general interest rate movements in each maturity segment); then do the trades vs bid/ask midpoints and treasury yields over expirations plot again using interpolation over more data with that pattern.
comeinvest
Posts: 2709
Joined: Mon Mar 12, 2012 6:57 pm

Re: Let's Talk SPX Box Spreads

Post by comeinvest »

unemployed_pysicist wrote: Sat Aug 20, 2022 3:53 pm A few comments regarding SVENY:

The SVENY curve is expressed as a continuously compounding rate, so you will have to convert the bills and box spread rates to continuous compounding to put everything on the same footing. The difference is probably very small, but thought I would mention it if you have not already done this in your chart.

What SVENY maturities do you want to put with the T bills/box spread yield curve? The SVENY01-30 maturities are released every Tuesday at around 4pm EST. The release is based off end-of-day quotes and includes all the yields for these maturities up to the previous friday. E.g., the data will be available up through Friday August 19 starting Tuesday, August 23. The parameters are also given to construct whatever maturities you want. The resulting curve may not match very well to short term issues, and also not very well to T bills due to the segmentation of the T bill market.

The WSJ quotes are from 3pm EST, as compared to the NSS fit of SVENY done with closing prices; hopefully they are close enough in time to get a commensurate measurement.

Here is the SVENY and instantaneous forward curve as of August 12, out to 10 years and sampled monthly (approximately; 0.083 years between data points.)
I have to do a lot more reading to understand everything that you are writing. But quick question: Is the blue line (instantaneous forward curve) calculated based on the orange curve? If so, would the blue line be comparable to the SOFR futures implied rates, and would that be a good cross-check to do?
Last edited by comeinvest on Sun Aug 21, 2022 4:18 am, edited 1 time in total.
comeinvest
Posts: 2709
Joined: Mon Mar 12, 2012 6:57 pm

Re: Let's Talk SPX Box Spreads

Post by comeinvest »

unemployed_pysicist wrote: Sat Aug 20, 2022 3:53 pm I quite like this idea of using box spreads to get the yield curve.
I assumed the discount rate for option valuation would be close to the OIS curve (with some spread maybe), but the paper posted/discussed earlier in this thread made some good arguments for why this is not the case/more to it than that. But if you can get it from box spreads, then you get the discount rate for SPX options themselves, no need to figure out the OIS curve, apply a spread, etc.
... but I think the whole idea is to compare the implied options discount rate to some other, relatively independent data, to verify fair value and quality of trade execution, and to detect discrepancies or aberrations. Just getting the discount rates from the options trades themselves would be purely descriptive and of little practical value, and would provide little insight.
unemployed_pysicist
Posts: 220
Joined: Sat Oct 09, 2021 2:32 pm
Location: Amsterdam

Re: Let's Talk SPX Box Spreads

Post by unemployed_pysicist »

comeinvest wrote: Sun Aug 21, 2022 12:59 am
unemployed_pysicist wrote: Sat Aug 20, 2022 3:53 pm I quite like this idea of using box spreads to get the yield curve.
I assumed the discount rate for option valuation would be close to the OIS curve (with some spread maybe), but the paper posted/discussed earlier in this thread made some good arguments for why this is not the case/more to it than that. But if you can get it from box spreads, then you get the discount rate for SPX options themselves, no need to figure out the OIS curve, apply a spread, etc.
... but I think the whole idea is to compare the implied options discount rate to some other, relatively independent data, to verify fair value and quality of trade execution, and to detect discrepancies or aberrations. Just getting the discount rates from the options trades themselves would be purely descriptive and of little practical value, and would provide little insight.
I understand that is the focus for this thread. But from my perspective, once the discount rates are known, I could use them to assist with calculating forward prices, implied dividends, and implied probability distributions from the options chain. I use Fed Funds/SOFR for that now, but I think this would be a cleaner way to do things - use discount rates implied from the options chain for all operations with the options chain.
couldn't afford the h | | BUY BONDS | WEAR DIAMONDS
unemployed_pysicist
Posts: 220
Joined: Sat Oct 09, 2021 2:32 pm
Location: Amsterdam

Re: Let's Talk SPX Box Spreads

Post by unemployed_pysicist »

comeinvest wrote: Sun Aug 21, 2022 12:54 am
I have to do a lot more reading to understand everything that you are writing. But quick question: Is the blue line (instantaneous forward curve) calculated based on the orange curve? If so, would be blue line be comparable to the SOFR futures implied rates, and would that be a good cross-check to do?
Yes, you could say that the blue line is based on the orange line. The SOFR and treasury curves are similar, but not the same:

Image

I think the SOFR data should be ignored in this plot starting around 2030, because I see no open interest where I get the data. Do you see any open interest for SOFR 2030-2032 contracts in your broker?

Apart from this, my SOFR forward plots generally agree with the forward curves I see on Chatham's website:

https://www.chathamfinancial.com/techno ... ard-curves
couldn't afford the h | | BUY BONDS | WEAR DIAMONDS
comeinvest
Posts: 2709
Joined: Mon Mar 12, 2012 6:57 pm

Re: Let's Talk SPX Box Spreads

Post by comeinvest »

unemployed_pysicist wrote: Sun Aug 21, 2022 4:10 am
comeinvest wrote: Sun Aug 21, 2022 12:54 am
I have to do a lot more reading to understand everything that you are writing. But quick question: Is the blue line (instantaneous forward curve) calculated based on the orange curve? If so, would be blue line be comparable to the SOFR futures implied rates, and would that be a good cross-check to do?
Yes, you could say that the blue line is based on the orange line. The SOFR and treasury curves are similar, but not the same:

I think the SOFR data should be ignored in this plot starting around 2030, because I see no open interest where I get the data. Do you see any open interest for SOFR 2030-2032 contracts in your broker?

Apart from this, my SOFR forward plots generally agree with the forward curves I see on Chatham's website:

https://www.chathamfinancial.com/techno ... ard-curves
I was asking because I thought the "instantaneous forward rates" might be by definition the same as the market-implied future yields of Treasury notes of short duration. If so, shouldn't they be very close to the SOFR forward rates? The curves are not close together.
unemployed_pysicist
Posts: 220
Joined: Sat Oct 09, 2021 2:32 pm
Location: Amsterdam

Re: Let's Talk SPX Box Spreads

Post by unemployed_pysicist »

comeinvest wrote: Sun Aug 21, 2022 4:25 am
I was asking because I thought the "instantaneous forward rates" might be by definition the same as the market-implied future yields of Treasury notes of short duration. If so, shouldn't they be very close to the SOFR forward rates? The curves are not close together.
I would not necessarily expect the instantaneous forward rates for SOFR and treasuries to be the same. I think a treasury (even an on the run) has a higher yield than a corresponding OIS tenor for longer maturities (>1 year).
I assume it will be the same story with SOFR rates.

I read that the US Treasury was going to issue SOFR indexed floating rate notes, but I don't know whatever happened with that. Presumably those would be in line with other SOFR-based market instruments.
couldn't afford the h | | BUY BONDS | WEAR DIAMONDS
User avatar
Kevin M
Posts: 15787
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: Let's Talk SPX Box Spreads

Post by Kevin M »

Bought a 11/18/22 box spread at 992.70 for gross yield of 3.09% compared to Treasury at 2.71%, so 38 basis point spread over Treasury. Not great, but I worked up from 992.00 in 0.05 increments, so it took awhile. I was just about to replace the order with a 11/30/22 box, but by the time I had worked out the numbers, the 11/18/22 had filled. It took about 4 minutes to fill.

Strikes were 3,600 and 4,600.

My net TEY is 3.26%, which is only 18 bps over Treasury TEY of 3.08%.

I've now filled a missing slot in my "ladder" with about 3 months to expiration, which I really wanted to do.

Kevin
If I make a calculation error, #Cruncher probably will let me know.
gougou
Posts: 1317
Joined: Thu Sep 28, 2017 7:42 pm

Re: Let's Talk SPX Box Spreads

Post by gougou »

Kevin M wrote: Mon Aug 22, 2022 1:51 pm Bought a 11/18/22 box spread at 992.70 for gross yield of 3.09% compared to Treasury at 2.71%, so 38 basis point spread over Treasury. Not great, but I worked up from 992.00 in 0.05 increments, so it took awhile. I was just about to replace the order with a 11/30/22 box, but by the time I had worked out the numbers, the 11/18/22 had filled. It took about 4 minutes to fill.

Strikes were 3,600 and 4,600.

My net TEY is 3.26%, which is only 18 bps over Treasury TEY of 3.08%.

I've now filled a missing slot in my "ladder" with about 3 months to expiration, which I really wanted to do.

Kevin
I think we should trade against each other to save some money. I sold Nov 17, 3000 - 5000 Box at 1985.1 last Friday.
The sillier the market’s behavior, the greater the opportunity for the business like investor.
comeinvest
Posts: 2709
Joined: Mon Mar 12, 2012 6:57 pm

Re: Let's Talk SPX Box Spreads

Post by comeinvest »

gougou wrote: Mon Aug 22, 2022 2:40 pm
Kevin M wrote: Mon Aug 22, 2022 1:51 pm Bought a 11/18/22 box spread at 992.70 for gross yield of 3.09% compared to Treasury at 2.71%, so 38 basis point spread over Treasury. Not great, but I worked up from 992.00 in 0.05 increments, so it took awhile. I was just about to replace the order with a 11/30/22 box, but by the time I had worked out the numbers, the 11/18/22 had filled. It took about 4 minutes to fill.

Strikes were 3,600 and 4,600.

My net TEY is 3.26%, which is only 18 bps over Treasury TEY of 3.08%.

I've now filled a missing slot in my "ladder" with about 3 months to expiration, which I really wanted to do.

Kevin
I think we should trade against each other to save some money. I sold Nov 17, 3000 - 5000 Box at 1985.1 last Friday.
Kevin bought a 1000 size box at 992.70, which would correspond to 1985.4 for a 2000 size box. You sold a 2000 size box at 1985.1 one settlement day prior, which would correspond to ca. 1985.25 had you traded today. So the effective buy/sell spread was 0.15 per 2000, or ca. 0.0075%, or (with ca. 3 months to maturity) ca. 0.03% p.a. This is consistent with the results from my box spread yield calculation spreadsheet. I did not account for market rates movement between Friday and today, and one data point is not enough for a meaningful estimation of the average buy/sell spread. It's just one data point.
I am avoiding the term "bid/ask spread", because the option quotes are not the effective filled prices.
Last edited by comeinvest on Mon Aug 22, 2022 7:22 pm, edited 2 times in total.
gougou
Posts: 1317
Joined: Thu Sep 28, 2017 7:42 pm

Re: Let's Talk SPX Box Spreads

Post by gougou »

comeinvest wrote: Mon Aug 22, 2022 7:07 pm
gougou wrote: Mon Aug 22, 2022 2:40 pm
Kevin M wrote: Mon Aug 22, 2022 1:51 pm Bought a 11/18/22 box spread at 992.70 for gross yield of 3.09% compared to Treasury at 2.71%, so 38 basis point spread over Treasury. Not great, but I worked up from 992.00 in 0.05 increments, so it took awhile. I was just about to replace the order with a 11/30/22 box, but by the time I had worked out the numbers, the 11/18/22 had filled. It took about 4 minutes to fill.

Strikes were 3,600 and 4,600.

My net TEY is 3.26%, which is only 18 bps over Treasury TEY of 3.08%.

I've now filled a missing slot in my "ladder" with about 3 months to expiration, which I really wanted to do.

Kevin
I think we should trade against each other to save some money. I sold Nov 17, 3000 - 5000 Box at 1985.1 last Friday.
Kevin bought a 1000 size box at 992.70, which would correspond to 1985.4 for a 2000 size box. You sold a 2000 size box at 1985.1 one settlement day prior, which would correspond to ca. 1985.25 had you traded today. So the effective buy/sell spread was 0.15 per 2000, or ca. 0.0075%, or (with ca. 3 months to maturity) ca. 0.03% p.a. This is consistent with the results from my box spread yield calculation spreadsheet. I did not account for market rates movement between Friday and today, and one data point is not enough for a meaningful estimation of the average buy/sell spread. It's just one data point.
I am avoiding the term "bid/ask spread", because the option quotes are not the effective filled prices.
OK 0.15 per 2000 is just $15 bucks. I guess it's too little it doesn't matter. It seems that Box spreads are quite liquid.
The sillier the market’s behavior, the greater the opportunity for the business like investor.
comeinvest
Posts: 2709
Joined: Mon Mar 12, 2012 6:57 pm

Re: Let's Talk SPX Box Spreads

Post by comeinvest »

gougou wrote: Mon Aug 22, 2022 7:20 pm
comeinvest wrote: Mon Aug 22, 2022 7:07 pm
gougou wrote: Mon Aug 22, 2022 2:40 pm
Kevin M wrote: Mon Aug 22, 2022 1:51 pm Bought a 11/18/22 box spread at 992.70 for gross yield of 3.09% compared to Treasury at 2.71%, so 38 basis point spread over Treasury. Not great, but I worked up from 992.00 in 0.05 increments, so it took awhile. I was just about to replace the order with a 11/30/22 box, but by the time I had worked out the numbers, the 11/18/22 had filled. It took about 4 minutes to fill.

Strikes were 3,600 and 4,600.

My net TEY is 3.26%, which is only 18 bps over Treasury TEY of 3.08%.

I've now filled a missing slot in my "ladder" with about 3 months to expiration, which I really wanted to do.

Kevin
I think we should trade against each other to save some money. I sold Nov 17, 3000 - 5000 Box at 1985.1 last Friday.
Kevin bought a 1000 size box at 992.70, which would correspond to 1985.4 for a 2000 size box. You sold a 2000 size box at 1985.1 one settlement day prior, which would correspond to ca. 1985.25 had you traded today. So the effective buy/sell spread was 0.15 per 2000, or ca. 0.0075%, or (with ca. 3 months to maturity) ca. 0.03% p.a. This is consistent with the results from my box spread yield calculation spreadsheet. I did not account for market rates movement between Friday and today, and one data point is not enough for a meaningful estimation of the average buy/sell spread. It's just one data point.
I am avoiding the term "bid/ask spread", because the option quotes are not the effective filled prices.
OK 0.15 per 2000 is just $15 bucks. I guess it's too little it doesn't matter. It seems that Box spreads are quite liquid.
EDIT: Oct 2022 1-month SOFR implied rate increased ca. 0.03%, and Nov 2022 1-month SOFR by ca. 0.07% during the course of Monday. Dec 2022 3-month SOFR implied rate increased by ca. 0.05%. I would take an average of 0.03% for the time period between now and Nov 18. Adjusted for this rate movement, the spread between Kevin's and gougou's trades was about zero. I don't have real time treasury yield data; we can check tomorrow. SOFR rates, Eurodollar rates, and T-bill rates or treasury note or bond rates are not the only determinants of box spread rates. Please someone verify my math.

To know the effective buy/sell spread for sure, somebody would have to bite the bullet and sell and buy the same spread at the same time. The spread is part of the trading cost, together with the commissions. Trading cost is good to know for the folks who are trying to minimize their cash drag with option box spreads.
User avatar
Kevin M
Posts: 15787
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: Let's Talk SPX Box Spreads

Post by Kevin M »

gougou wrote: Mon Aug 22, 2022 2:40 pm
Kevin M wrote: Mon Aug 22, 2022 1:51 pm Bought a 11/18/22 box spread at 992.70 for gross yield of 3.09% compared to Treasury at 2.71%, so 38 basis point spread over Treasury. Not great, but I worked up from 992.00 in 0.05 increments, so it took awhile. I was just about to replace the order with a 11/30/22 box, but by the time I had worked out the numbers, the 11/18/22 had filled. It took about 4 minutes to fill.

Strikes were 3,600 and 4,600.

My net TEY is 3.26%, which is only 18 bps over Treasury TEY of 3.08%.

I've now filled a missing slot in my "ladder" with about 3 months to expiration, which I really wanted to do.

Kevin
I think we should trade against each other to save some money. I sold Nov 17, 3000 - 5000 Box at 1985.1 last Friday.
How would we do that?
If I make a calculation error, #Cruncher probably will let me know.
User avatar
Kevin M
Posts: 15787
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: Let's Talk SPX Box Spreads

Post by Kevin M »

Didn't have time to post it yesterday, but I also bought a 1/20/23 box spread at 986.20 for gross yield of 3.40% compared to Treasury at 2.97%, so 44 basis point spread over Treasury. I worked up from 985.40 in 0.10 increments. I was in the process of replacing the order at 986.30, but when I attempted the replacement I could not, because the order had filled at 986.20.

Strikes were 3,700 and 4,700.

My net TEY is 3.61%, which is 24 bps over Treasury TEY of 3.37%.

This added a $100K, 5-month rung to my "ladder", which now has rungs at approximately 1, 2, 3, 4, 5 and 7 months, with the rung principal amounts ranging from $50K (4 and 7) to $200K (2), with the rest at $100K.

The last two purchases were done mostly with proceeds from expired boxes (8/19/22), with some additional cash added.

Kevin
If I make a calculation error, #Cruncher probably will let me know.
Post Reply