HEDGEFUNDIE's excellent adventure Part II: The next journey

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
stockmaster
Posts: 33
Joined: Thu Oct 08, 2020 7:46 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by stockmaster »

Uncorrelated wrote: Sat Oct 31, 2020 5:52 pm A correlation of zero means that there is a 50% chance they will move in unison. If you want to evaluate the correlation of two different funds, then you'll need to review many different events, not just one.

There never was solid evidence that TMF was negatively correlated with the stock market, there definitely isn't any evidence that the situation changed recently. Evaluating correlations is extremely difficult, even over a time period of 20 years or more you'll often end up with estimates of the correlation that are complete garbage.
Here's some evidence... Note that TMF has literally negative correlation as well, which is better than gold's positive correlation when looking for a hedge.

I realize the level of correlation changes with time, but the big question here is "what's the best hedge available?"... Some posters are saying gold is better than TMF, most are saying TMF is bad and has too high correlation now, but nobody explains why. Also, nobody outside of this forum seems to agree with this idea that TMF is going to correlate with stocks now.

This was the closest thing I could find discussing the topic recently, but again, it says bonds continue to inversely correlate with stocks.
000 wrote: Sat Oct 31, 2020 5:57 pm
As always, it depends on exactly when one enters and exits one's positions. TMF also "crashed" in March. If you bought TMF in Februrary, you'd have been underwater on its worst day in March:
The article I linked above mentions that crash.
User avatar
Uncorrelated
Posts: 1105
Joined: Sun Oct 13, 2019 3:16 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Uncorrelated »

stockmaster wrote: Sun Nov 01, 2020 2:02 am
Uncorrelated wrote: Sat Oct 31, 2020 5:52 pm A correlation of zero means that there is a 50% chance they will move in unison. If you want to evaluate the correlation of two different funds, then you'll need to review many different events, not just one.

There never was solid evidence that TMF was negatively correlated with the stock market, there definitely isn't any evidence that the situation changed recently. Evaluating correlations is extremely difficult, even over a time period of 20 years or more you'll often end up with estimates of the correlation that are complete garbage.
Here's some evidence... Note that TMF has literally negative correlation as well, which is better than gold's positive correlation when looking for a hedge.
That is an anecdote, not evidence. If the actual correlation in that time period was zero, how likely do you think it is that the measured correlation is -0.4? Quite likely.

When evaluated over sufficiently long time horizons, the measured correlation is sometimes positive and sometimes negative. A very probably conclusion is that these measurements pick up random market gyrations rather than actual negative correlation in the underlying statistical process. That's what I meant with a lack of evidence.

Even if there is a correlation between stocks and bonds, it isn't at all obvious what you should do as a result of that. The efficient market is very good at pricing in market expectations. If you believe bond returns in the future are positively correlated with stock returns, you better believe that the market also knows that and has priced treasuries accordingly.
stockmaster
Posts: 33
Joined: Thu Oct 08, 2020 7:46 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by stockmaster »

Uncorrelated wrote: Sun Nov 01, 2020 3:11 am
stockmaster wrote: Sun Nov 01, 2020 2:02 am
Uncorrelated wrote: Sat Oct 31, 2020 5:52 pm A correlation of zero means that there is a 50% chance they will move in unison. If you want to evaluate the correlation of two different funds, then you'll need to review many different events, not just one.

There never was solid evidence that TMF was negatively correlated with the stock market, there definitely isn't any evidence that the situation changed recently. Evaluating correlations is extremely difficult, even over a time period of 20 years or more you'll often end up with estimates of the correlation that are complete garbage.
Here's some evidence... Note that TMF has literally negative correlation as well, which is better than gold's positive correlation when looking for a hedge.
That is an anecdote, not evidence. If the actual correlation in that time period was zero, how likely do you think it is that the measured correlation is -0.4? Quite likely.

When evaluated over sufficiently long time horizons, the measured correlation is sometimes positive and sometimes negative. A very probably conclusion is that these measurements pick up random market gyrations rather than actual negative correlation in the underlying statistical process. That's what I meant with a lack of evidence.

Even if there is a correlation between stocks and bonds, it isn't at all obvious what you should do as a result of that. The efficient market is very good at pricing in market expectations. If you believe bond returns in the future are positively correlated with stock returns, you better believe that the market also knows that and has priced treasuries accordingly.
The scale is from -1 to 1. So assuming TMF has 0% correlation with stocks, the likelihood of it displaying -40% correlation for some large time period is very low.

However, TMF consistently shows negative correlation across multiple time horizons. Therefore the correlation is not 0%, but negative. Try different time values in that link I provided.
perfectuncertainty
Posts: 143
Joined: Sun Feb 04, 2018 7:44 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by perfectuncertainty »

drock wrote: Sat Oct 31, 2020 12:21 am
perfectuncertainty wrote: Fri Oct 30, 2020 8:46 am
stormcrow wrote: Fri Oct 30, 2020 8:25 am
Perfect Uncertainty wrote: Fri Oct 30, 2020 1:23 am Fills are super easy. Lots of volume.
So you are just letting them expire then - profiting if ITM at expiry? (I.e., not cashing any out ahead of time.)
Yes since it’s a hedge I'm not trying to manage them for a profit. Most of the time they expire worthless, but if they happen to be ITM there is an automatic account credit adjustment since there is no assignment like options on equities. Feels horrible but when you get your brain around it as good insurance versus an options play it feels a bit better.
So if there is a vol spike you don't plan to take advantage and sell at some point while vol is high to lock in the gains before vol returns closer to the mean?
I backtested the VIX Black Swan Hedge (vBSH) against the 2008 Crisis.

Using the setup criteria of the hedge I laid out previously for an account worth 100k, the vBSH would have paid just under $732k during the 2008 financial crisis.

During periods like late 2018, the hedge did not pay out because volatility did not rise sufficiently. We can also argue that 2018 never rose to Black Swan proportions too.

It is somewhat tedious to test the vBSH since options data carries quite a bit more complexity due to different strike prices etc and is not readily available. To perform the tests I used a 3rd party tool called eDeltapro.
drock
Posts: 26
Joined: Sat Mar 23, 2019 4:47 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by drock »

perfectuncertainty wrote: Sun Nov 01, 2020 8:00 pm
drock wrote: Sat Oct 31, 2020 12:21 am
perfectuncertainty wrote: Fri Oct 30, 2020 8:46 am
stormcrow wrote: Fri Oct 30, 2020 8:25 am
Perfect Uncertainty wrote: Fri Oct 30, 2020 1:23 am Fills are super easy. Lots of volume.
So you are just letting them expire then - profiting if ITM at expiry? (I.e., not cashing any out ahead of time.)
Yes since it’s a hedge I'm not trying to manage them for a profit. Most of the time they expire worthless, but if they happen to be ITM there is an automatic account credit adjustment since there is no assignment like options on equities. Feels horrible but when you get your brain around it as good insurance versus an options play it feels a bit better.
So if there is a vol spike you don't plan to take advantage and sell at some point while vol is high to lock in the gains before vol returns closer to the mean?
I backtested the VIX Black Swan Hedge (vBSH) against the 2008 Crisis.

Using the setup criteria of the hedge I laid out previously for an account worth 100k, the vBSH would have paid just under $732k during the 2008 financial crisis.

During periods like late 2018, the hedge did not pay out because volatility did not rise sufficiently. We can also argue that 2018 never rose to Black Swan proportions too.

It is somewhat tedious to test the vBSH since options data carries quite a bit more complexity due to different strike prices etc and is not readily available. To perform the tests I used a 3rd party tool called eDeltapro.
I appreciate your efforts to help quantify where the strategy would have paid off. I think for folks nearing or in retirement it sounds like a decent way to hedge your portfolio and reduce sequence of returns risk in a time where bonds may not provide as much of a portfolio ballast as they have in the past.

For someone still in the accumulation phase I think it is more of a dilemma. Will there be enough times where the hedge will pay off huge to account for the sizable amount it will reduce your compounding? I'm not sure how to answer that because my crystal ball is in the shop right now.
Mother_McCrankel
Posts: 4
Joined: Fri Oct 30, 2020 9:36 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Mother_McCrankel »

Some good movements in TMF today with the indicies rising. May be a good time to average down on long term treasuries prior to election outcomes.
nehawk87
Posts: 16
Joined: Sun Jun 30, 2019 8:49 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by nehawk87 »

TMF is back from the dead! Up 7%+. UPRO 8%+.

Nice to see a day like this but would really sting if the inverse ever occurred.
parval
Posts: 30
Joined: Tue Oct 22, 2019 9:23 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by parval »

nehawk87 wrote: Wed Nov 04, 2020 10:30 am TMF is back from the dead! Up 7%+. UPRO 8%+.

Nice to see a day like this but would really sting if the inverse ever occurred.
This is super worrying actually, if TMF continues to behave as it has for the past week or so, we're essentially 100% in UPRO. Desperately need correlation to go the other way
User avatar
Ramjet
Posts: 509
Joined: Thu Feb 06, 2020 11:45 am

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Ramjet »

parval wrote: Wed Nov 04, 2020 11:55 am
nehawk87 wrote: Wed Nov 04, 2020 10:30 am TMF is back from the dead! Up 7%+. UPRO 8%+.

Nice to see a day like this but would really sting if the inverse ever occurred.
This is super worrying actually, if TMF continues to behave as it has for the past week or so, we're essentially 100% in UPRO. Desperately need correlation to go the other way
TMF negative correlation is primarily for equity crashes not so much daily activity
perfectuncertainty
Posts: 143
Joined: Sun Feb 04, 2018 7:44 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by perfectuncertainty »

drock wrote: Tue Nov 03, 2020 3:22 pm I appreciate your efforts to help quantify where the strategy would have paid off. I think for folks nearing or in retirement it sounds like a decent way to hedge your portfolio and reduce sequence of returns risk in a time where bonds may not provide as much of a portfolio ballast as they have in the past.

For someone still in the accumulation phase I think it is more of a dilemma. Will there be enough times where the hedge will pay off huge to account for the sizable amount it will reduce your compounding? I'm not sure how to answer that because my crystal ball is in the shop right now.
As you say, it's a personal decision. I'm as greedy as the next guy, and I want like an HFEA risk-parity approach to compound growth faster, but the risk/reward axiom never seemed more necessary than in this case. There is no guarantee that we will always have a V-shaped recovery from a black swan event and since we are benefitting from the juiced-up compounding, for me I think the hedge makes sense personally.
RovenSkyfall
Posts: 104
Joined: Wed Apr 01, 2020 11:40 am

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by RovenSkyfall »

Today is a reminder day to not try and time the market. Almost held back from investing in HFEA with some IRA money due to election and COVID cases increasing, but decided yesterday that was not consistent with my investment plan. Boy was I lucky and today served as a nice cautionary tale as to why that would have been a bad idea....
User avatar
nisiprius
Advisory Board
Posts: 42201
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by nisiprius »

stockmaster wrote: Sun Nov 01, 2020 5:33 am...The scale is from -1 to 1. So assuming TMF has 0% correlation with stocks, the likelihood of it displaying -40% correlation for some large time period is very low...
It might not be. Without a more precise statement it's hard to say anything sensible.

But, for example, suppose you have a twenty-year period and you believe that the annual returns are independent random samples from a distribution with ρ = 0.0, zero correlation. You calculate the actual correlation over the annual returns; N =20. Then there is a 5% chance that the calculated correlation coefficient between the two twenty-year samples will be between ρ = -0.48 to ρ = 0.48. That is, an observed correlation of ρ = -0.48 is not statistically different from 0. There's a 5% chance of seeing that far a departure from 0.

Some people seem to assume assume that any change in measured correlations over limited time periods means "the correlation really changed" when it could just have been luck. It takes a lot of data to put a tight confidence limits around a correlation coefficient.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.
jeremyl
Posts: 231
Joined: Sat Dec 20, 2014 8:38 am
Location: Indiana

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by jeremyl »

I've been combing through this thread and was considering this with using about 3% of my portfolio but through all of the posts I had a couple of questions.

What's the best type of account to do this in (Roth or Taxable)? My thought was to start in January with my 2021 Roth IRA contribution and adding $4,000 to be all in for $10K. But I'd like to use taxable for flexibility but I'm guessing this is not ideal.

If done in Roth, I'd consider using my yearly contribution as part of my strategy going forward. Anyone else doing this?

If done in Roth IRA, do you reinvest dividends or not?

I don't understand all of the other things people are throwing out there with other funds and VIX, etc. Is it as simple as setting it up at the asset allocation (55/45) and rebalancing every quarter?


I could do the $10K in January and let it ride and see what happens. Although, when back testing and using the yearly Roth IRA contributions does increase the balance significantly.

Is this easy to set up in the Vanguard Roth IRA?

Any other details I am missing?

Thanks!
DetroitAvant
Posts: 14
Joined: Sat Jan 07, 2017 9:39 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by DetroitAvant »

Vanguard doesn't offer these funds, you have to look somewhere else You can have multiple Roth IRAs. I think tax advantaged was suggested due to the possibility of tax events during rebalancing.
RovenSkyfall
Posts: 104
Joined: Wed Apr 01, 2020 11:40 am

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by RovenSkyfall »

jeremyl wrote: Thu Nov 05, 2020 6:12 am I've been combing through this thread and was considering this with using about 3% of my portfolio but through all of the posts I had a couple of questions.

What's the best type of account to do this in (Roth or Taxable)? My thought was to start in January with my 2021 Roth IRA contribution and adding $4,000 to be all in for $10K. But I'd like to use taxable for flexibility but I'm guessing this is not ideal.

If done in Roth, I'd consider using my yearly contribution as part of my strategy going forward. Anyone else doing this?

If done in Roth IRA, do you reinvest dividends or not?

I don't understand all of the other things people are throwing out there with other funds and VIX, etc. Is it as simple as setting it up at the asset allocation (55/45) and rebalancing every quarter?


I could do the $10K in January and let it ride and see what happens. Although, when back testing and using the yearly Roth IRA contributions does increase the balance significantly.

Is this easy to set up in the Vanguard Roth IRA?

Any other details I am missing?

Thanks!
Vanguard doesnt allow leveraged ETFs so you might think about something like M1 (which allows easy rebalancing and a lot of people use that) or Fidelity. Most people suggest a tax advantaged account because rebalancing in a taxable account will decrease your efficiency depending on how you sell you shares. It seems like quarterly (on the actual quarter, not quarterly from when you invest) is the best in back testing (who knows going forward) for a standard allocation. That being said, some people have it in a taxable and then rebalance with contributions. There are a lot of fancy ways to invest in these leveraged ETFs, but the original adventure was last modified to a 55/45 allocation.
stockmaster
Posts: 33
Joined: Thu Oct 08, 2020 7:46 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by stockmaster »

nisiprius wrote: Wed Nov 04, 2020 9:29 pmIt might not be. Without a more precise statement it's hard to say anything sensible.

But, for example, suppose you have a twenty-year period and you believe that the annual returns are independent random samples from a distribution with ρ = 0.0, zero correlation. You calculate the actual correlation over the annual returns; N =20. Then there is a 5% chance that the calculated correlation coefficient between the two twenty-year samples will be between ρ = -0.48 to ρ = 0.48. That is, an observed correlation of ρ = -0.48 is not statistically different from 0. There's a 5% chance of seeing that far a departure from 0.
A 5% chance is nothing to brag about... And as you keep getting the same number across multiple time horizons, the chances of it being a fluke go even lower.

The most likely level of correlation is roughly -0.4. Not zero, or any other arbitrary number. That's all I was saying.
nisiprius wrote: Wed Nov 04, 2020 9:29 pm Some people seem to assume assume that any change in measured correlations over limited time periods means "the correlation really changed" when it could just have been luck. It takes a lot of data to put a tight confidence limits around a correlation coefficient.
I agree.
User avatar
firebirdparts
Posts: 1881
Joined: Thu Jun 13, 2019 4:21 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by firebirdparts »

jeremyl wrote: Thu Nov 05, 2020 6:12 am I've been combing through this thread and was considering this with using about 3% of my portfolio but through all of the posts I had a couple of questions.

What's the best type of account to do this in (Roth or Taxable)? My thought was to start in January with my 2021 Roth IRA contribution and adding $4,000 to be all in for $10K. But I'd like to use taxable for flexibility but I'm guessing this is not ideal.

If done in Roth, I'd consider using my yearly contribution as part of my strategy going forward. Anyone else doing this?

If done in Roth IRA, do you reinvest dividends or not?

I don't understand all of the other things people are throwing out there with other funds and VIX, etc. Is it as simple as setting it up at the asset allocation (55/45) and rebalancing every quarter?


I could do the $10K in January and let it ride and see what happens. Although, when back testing and using the yearly Roth IRA contributions does increase the balance significantly.

Is this easy to set up in the Vanguard Roth IRA?

Any other details I am missing?

Thanks!
I would by all means do it in a Roth, and in fact I am. I don't see any reason not to reinvest dividends in a Roth. That part of it doesn't matter a whole lot.
Last edited by firebirdparts on Tue Nov 10, 2020 7:03 am, edited 1 time in total.
A fool and your money are soon partners
nolegs
Posts: 24
Joined: Mon Mar 23, 2020 12:06 am

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by nolegs »

In another thread, someone posted that you can buy and roll Nasdaq futures as an alternative to holding TQQQ, and not see the decreased returns from volatility.

What are the pros and cons of doing so?
User avatar
Uncorrelated
Posts: 1105
Joined: Sun Oct 13, 2019 3:16 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Uncorrelated »

nolegs wrote: Thu Nov 05, 2020 5:29 pm In another thread, someone posted that you can buy and roll Nasdaq futures as an alternative to holding TQQQ, and not see the decreased returns from volatility.

What are the pros and cons of doing so?
Daily leveraged ETF's do not result in lower expected return before expense ratio. The lower expected utility comes from a difference between your target leverage and actual leverage. If your target leverage is 3x, then using a 3x daily leveraged ETF is optimal because the target leverage is always equal to the actual leverage. If you purchase futures, then the leverage will be equal to 3x only on the day you brought it. The following days the leverage will drift until you rebalance. For obvious reasons: if your target leverage does not match your actual leverage, that is bad.

I did the math in this post. The assumption is that there is a 1% expense ratio difference between LETF and options/futures. If you rebalance futures daily, then futures always win. If you rebalance monthly, futures are a better choice if your target leverage is below 2.6x. If you rebalance semi-annually, the breakpoint is at a target leverage of 2.1x. This analysis assumes 100% equity allocation since that's the allocation generally used by lifecycle investing.

This analysis ignores taxes, assumes you're following a lifecycle investing strategy (you should) and assumes complete overview of your portfolio (i.e. no bucketing). Since HFEA only targets ~1.65x equity leverage, I suspect futures are always better than LETF's, but I don't have enough data to do any meaningful math on portfolio's with leveraged treasury products.
jeremyl
Posts: 231
Joined: Sat Dec 20, 2014 8:38 am
Location: Indiana

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by jeremyl »

RovenSkyfall wrote: Thu Nov 05, 2020 8:38 am
jeremyl wrote: Thu Nov 05, 2020 6:12 am I've been combing through this thread and was considering this with using about 3% of my portfolio but through all of the posts I had a couple of questions.

What's the best type of account to do this in (Roth or Taxable)? My thought was to start in January with my 2021 Roth IRA contribution and adding $4,000 to be all in for $10K. But I'd like to use taxable for flexibility but I'm guessing this is not ideal.

If done in Roth, I'd consider using my yearly contribution as part of my strategy going forward. Anyone else doing this?

If done in Roth IRA, do you reinvest dividends or not?

I don't understand all of the other things people are throwing out there with other funds and VIX, etc. Is it as simple as setting it up at the asset allocation (55/45) and rebalancing every quarter?


I could do the $10K in January and let it ride and see what happens. Although, when back testing and using the yearly Roth IRA contributions does increase the balance significantly.

Is this easy to set up in the Vanguard Roth IRA?

Any other details I am missing?

Thanks!
Vanguard doesnt allow leveraged ETFs so you might think about something like M1 (which allows easy rebalancing and a lot of people use that) or Fidelity. Most people suggest a tax advantaged account because rebalancing in a taxable account will decrease your efficiency depending on how you sell you shares. It seems like quarterly (on the actual quarter, not quarterly from when you invest) is the best in back testing (who knows going forward) for a standard allocation. That being said, some people have it in a taxable and then rebalance with contributions. There are a lot of fancy ways to invest in these leveraged ETFs, but the original adventure was last modified to a 55/45 allocation.
I have been thinking of transferring my roth from Vanguard to Fidelity to consolidate accounts anyway. I just need to convert my mutual funds to ETF's before doing so (all VTSAX). I am planning on doing this soon so everything is moved over before or by January when I contribute the full amount to my Roth and I can start this strategy.

I am going to do the rebalancing every quarter on the actual quarter and do the 55/45 allocation.

Thanks!
User avatar
coingaroo
Posts: 93
Joined: Fri Apr 26, 2019 11:31 am

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by coingaroo »

Direxion says the annual report will be published around the end of the month.

Very keen to see what the financing spread they have been paying.
cooljack4u
Posts: 1
Joined: Mon Nov 09, 2020 9:51 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by cooljack4u »

perfectuncertainty wrote: Thu Oct 29, 2020 12:32 am Here is a detailed description of how HFEA (55/45 UPRO/TMF) portfolio has performed with the protection of a VIX options hedge. The numbers are refined from my prior post which overstated the performance. The reason for the overstatement was due to the fact that I did not derive the current monthly value of HFEA to be used as a basis for purchasing the monthly VIX option tranches that would be in effect in mid-March 2020. Nevertheless, the performance of HFEA with the protection of the VIX options hedge was substantially better than the performance of HFEA without the benefit of the insurance that was in effect since the beginning of the model.

Background Setup.
Start Date of 9/1/2011 based on stockmaster's question on page 132 of this thread.
Starting Capital: 100,000
Allocation: 55% UPRO / 45% TMF
UPRO Peak is the focus and occurred on 2/19/2020

VIX Hedge Setup.
We start with the 100,000 in starting capital
Allocate 0.3% of Net Liq each month and buy 10 delta VIX call options ~ 120 days to expiration. Purchase this on the 3rd Friday of the month.
This created 4 tranches of VIX call options that would be in effect during mid-March 2020.

The tranches were as follows:
11/15/2019 Net Liq: 648,793 to purchase VIX call options expiring on 3/18/2020. Strike price 45.00 Cost per option 0.300. Number purchased 6,400.
12/15/2019 Net Liq: 653,806 to purchase VIX call options expiring on 4/15/2020. Strike price 45.00 Cost per option 0.275. Number purchased 7,100.
01/15/2020 Net Liq: 718,380 to purchase VIX call options expiring on 5/20/2020. Strike price 40.00 Cost per option 0.200. Number purchased 10,700.
02/15/2020 Net Liq: 778,954 to purchase VIX call options expiring on 6/17/2020. Strike price 42.50. Cost per option 0.200. Number purchased 8,400.
Total Cost of the VIX Options Hedge in effect in mid-March 2020: 8,399.80

HFEA Value as of2/19/2020.
As of 2/19/2020 HFEA with the VIX Hedge applied each month since 9/1/2011 was worth 778,954.40.

This is derived using PV and daily prices from TradingView charts as follows:
PV Model with 0.3% monthly withdrawal LINK
As of 1/31/2020 HFEA with VIX Hedge applied had a Final Balance of 718,380
Price UPRO on 1/31/2020 = 69.37
Price of TMF on 1/31/2020 = 31.99
UPRO 55% allocation = 5,696 shares
TMF 45% allocation = 10,105 shares
Price of UPRO on 2/19/2020 = 80.36
Price of TMF on 2/19/2020 = 31.79
Using the # of shares for each calculated above the value for HFEA with the monthly 0.3% allocation for the VIX hedge is 778,954.40.

Performance During Mid-March 2020 for HFEA and the VIX Option Hedge
By 3/9/2020, HFEA declined by 3.57% from the 2/19 highpoint to a value of 751,182.60.
By 3/13/2020, HFEA declined by 25.64% from the 2/19 highpoint to a value of 579,204.14.
By 3/16/2020, HFEA declined by 26.86% from the 2/19 highpoint to a value of 569,763.08.

By 3/9/2020, the VIX hedge increased by 1242.53% to a value of 104,370.00.
By 3/13/2020, the VIX hedge increased by 2549.63% to a value of 214,164.00.
By 3/16/2020, the VIX hedge increased by 6912.03% to a value of 580,597.00.

By 3/9/2020, HFEA combined with the VIX hedge increased by 9.83% to a value of 855,552.60.
By 3/13/2020, HFEA combined with the VIX hedge increased by 1.85% to a value of 793,368.14.
By 3/16/2020, HFEA combined with the VIX hedge increased by 47.68% to a value of 1,150,360.08.

Point of Reference
As a point of reference if one had not utilized any hedging HFEA would have been worth 1,055,118.98 on 2/19/2020 but worth 771,762.57 on 3/16/2020.
This model and analysis does not include any application of hedging comparison for declines in HFEA in August 2015, February of 2018, and December 2018 even though the costs of the protection are modeled as being in effect at that time. The VIX hedge would have provided substantial benefit in 2008 and contained any drawdown. And again in 2010 and 2011 - I have not modeled those periods to ascertain how effective the insurance would have been.

Here is my spreadsheet with the analysis for 3/2020 LINK.
I wonder if this hedging will work with 100 percent UPRO and increasing the hedge amount to 0.3 to 0.6 % every month. The idea is to backtest with different hedge amount to buy the vix options. It would be interesting to see the portfolio performance
perfectuncertainty
Posts: 143
Joined: Sun Feb 04, 2018 7:44 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by perfectuncertainty »

cooljack4u wrote: Mon Nov 09, 2020 10:00 pm
perfectuncertainty wrote: Thu Oct 29, 2020 12:32 am Here is a detailed description of how HFEA (55/45 UPRO/TMF) portfolio has performed with the protection of a VIX options hedge. The numbers are refined from my prior post which overstated the performance. The reason for the overstatement was due to the fact that I did not derive the current monthly value of HFEA to be used as a basis for purchasing the monthly VIX option tranches that would be in effect in mid-March 2020. Nevertheless, the performance of HFEA with the protection of the VIX options hedge was substantially better than the performance of HFEA without the benefit of the insurance that was in effect since the beginning of the model.

Background Setup.
Start Date of 9/1/2011 based on stockmaster's question on page 132 of this thread.
Starting Capital: 100,000
Allocation: 55% UPRO / 45% TMF
UPRO Peak is the focus and occurred on 2/19/2020

VIX Hedge Setup.
We start with the 100,000 in starting capital
Allocate 0.3% of Net Liq each month and buy 10 delta VIX call options ~ 120 days to expiration. Purchase this on the 3rd Friday of the month.
This created 4 tranches of VIX call options that would be in effect during mid-March 2020.

The tranches were as follows:
11/15/2019 Net Liq: 648,793 to purchase VIX call options expiring on 3/18/2020. Strike price 45.00 Cost per option 0.300. Number purchased 6,400.
12/15/2019 Net Liq: 653,806 to purchase VIX call options expiring on 4/15/2020. Strike price 45.00 Cost per option 0.275. Number purchased 7,100.
01/15/2020 Net Liq: 718,380 to purchase VIX call options expiring on 5/20/2020. Strike price 40.00 Cost per option 0.200. Number purchased 10,700.
02/15/2020 Net Liq: 778,954 to purchase VIX call options expiring on 6/17/2020. Strike price 42.50. Cost per option 0.200. Number purchased 8,400.
Total Cost of the VIX Options Hedge in effect in mid-March 2020: 8,399.80

HFEA Value as of2/19/2020.
As of 2/19/2020 HFEA with the VIX Hedge applied each month since 9/1/2011 was worth 778,954.40.

This is derived using PV and daily prices from TradingView charts as follows:
PV Model with 0.3% monthly withdrawal LINK
As of 1/31/2020 HFEA with VIX Hedge applied had a Final Balance of 718,380
Price UPRO on 1/31/2020 = 69.37
Price of TMF on 1/31/2020 = 31.99
UPRO 55% allocation = 5,696 shares
TMF 45% allocation = 10,105 shares
Price of UPRO on 2/19/2020 = 80.36
Price of TMF on 2/19/2020 = 31.79
Using the # of shares for each calculated above the value for HFEA with the monthly 0.3% allocation for the VIX hedge is 778,954.40.

Performance During Mid-March 2020 for HFEA and the VIX Option Hedge
By 3/9/2020, HFEA declined by 3.57% from the 2/19 highpoint to a value of 751,182.60.
By 3/13/2020, HFEA declined by 25.64% from the 2/19 highpoint to a value of 579,204.14.
By 3/16/2020, HFEA declined by 26.86% from the 2/19 highpoint to a value of 569,763.08.

By 3/9/2020, the VIX hedge increased by 1242.53% to a value of 104,370.00.
By 3/13/2020, the VIX hedge increased by 2549.63% to a value of 214,164.00.
By 3/16/2020, the VIX hedge increased by 6912.03% to a value of 580,597.00.

By 3/9/2020, HFEA combined with the VIX hedge increased by 9.83% to a value of 855,552.60.
By 3/13/2020, HFEA combined with the VIX hedge increased by 1.85% to a value of 793,368.14.
By 3/16/2020, HFEA combined with the VIX hedge increased by 47.68% to a value of 1,150,360.08.

Point of Reference
As a point of reference if one had not utilized any hedging HFEA would have been worth 1,055,118.98 on 2/19/2020 but worth 771,762.57 on 3/16/2020.
This model and analysis does not include any application of hedging comparison for declines in HFEA in August 2015, February of 2018, and December 2018 even though the costs of the protection are modeled as being in effect at that time. The VIX hedge would have provided substantial benefit in 2008 and contained any drawdown. And again in 2010 and 2011 - I have not modeled those periods to ascertain how effective the insurance would have been.

Here is my spreadsheet with the analysis for 3/2020 LINK.
I wonder if this hedging will work with 100 percent UPRO and increasing the hedge amount to 0.3 to 0.6 % every month. The idea is to backtest with different hedge amounts to buy the VIX options. It would be interesting to see the portfolio performance
You should be able to extrapolate the performance for early 2020 from the results I posted.
TwoIdenticalIndexes
Posts: 101
Joined: Tue Jan 14, 2020 5:34 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by TwoIdenticalIndexes »

My 45% UPRO 15% UTSL 40% TMF strategy is up ~10.5% since end of July. I've been rebalancing once a month.

Over that period:
SPY +10.5%
UPRO +28.5%
TMF -20.5%
UTSL +28%

I assume I've gotten some tailwind from the poor correlation between spy and utilities over that last 3.5 months (first 2 months were down/flat for UTSL, last 2 months were a boom).

How have the standard UPRO/TMF allocations done?
User avatar
Tinkerer-in-Chief
Posts: 6
Joined: Mon May 25, 2020 4:20 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Tinkerer-in-Chief »

TwoIdenticalIndexes wrote: Sat Nov 14, 2020 9:43 am My 45% UPRO 15% UTSL 40% TMF strategy is up ~10.5% since end of July. I've been rebalancing once a month.

Over that period:
SPY +10.5%
UPRO +28.5%
TMF -20.5%
UTSL +28%

I assume I've gotten some tailwind from the poor correlation between spy and utilities over that last 3.5 months (first 2 months were down/flat for UTSL, last 2 months were a boom).

How have the standard UPRO/TMF allocations done?
I can't precisely answer for the "standard" UPRO/TMF portfolio, but I can tell you how my own twist on the Excellent Adventure is doing so far. I have a small portfolio that is 50%/25%/25% UPRO/TMF/BLV (Vanguard Long-Term Bond Index Fund ETF), and I have been tracking its cashflows. I started investing May 5th and rebalance quarterly from this date.

As of November 5th my portfolio's money-weighted return is 33.0%. As of November 5th, a VTI (Vanguard Total Stock Market Index Fund ETF) portfolio with the same cashflows has a money-weighted return of 24.4%. Hopefully this outperformance continues even it becomes less marked over time! :sharebeer
socoolme
Posts: 2
Joined: Sun Feb 16, 2020 9:47 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by socoolme »

perfectuncertainty wrote: Wed Oct 28, 2020 6:09 pm

Yes, I did mean XVZ. Not UVZ.

On 3/16 the $736k was the value of the hedge only. The returns for HFEA were in addition.

I built the model based on your inception date of 10/2011. Reread my post. But in brief:

The options in the hedge are bought manually. Once a month. Net Liq x 0.003 = X. Find the 10 delta VIX calls 120 days to expiration and get the price. Divide X by that number to give the number of options for the month. Once this is running you will always have 3 months in play. In March of 2020, the 4 months of options in play would have been Mar, Apr, May, Jun which were purchased in Nov, Dec, Jan, and Feb. The Net Liq of HFEA in Feb would have been around 880k and you would have spent 2.6k for each of the months of VIX options (just under 11k).

Like you, I'm very busy and building the full model is a fair amount of work using multiple tools. I'm fairly confident in my numbers, but I'll take some time and build all this out in Excel using PV and TOS when I have some time.
The approach with VIX hedge is very interesting. I am playing with thinkorswim papermoney and initiated some HFEA positions today without the hedge leaving some funds for VIX hedge. I am an options newbie and looking at the VIX options chain for the next 30 to 120 days. VIX is around 22 this week. What would be the euqivalent 10 delta VIX strick price given where it is now? Looking at your 11/2019 posts it seems about 3.5x times VIX. Is that how one would get there? Also when you intiate such hedge do you recommend only starting with one 0.3 allocation that is 120 days out and then adding one each month later on versus initiating 0.3 % for each of 30, 60, and 90 days of trenches?

TIA.
Last edited by socoolme on Fri Nov 20, 2020 4:18 pm, edited 1 time in total.
Crmck26
Posts: 29
Joined: Thu May 14, 2020 9:28 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Crmck26 »

I'm sure someone has brought something similar up before, but this thread has been going on for so long I can't remember.

I started taking my "travel savings" account that was just in a savings account and since I won't be traveling for a bit put it into M1, rebalancing monthly with contributions to:

SSO (2x S&P) 33%
QLD (2x QQQ) 33%
UBT (2x 20+ Tres) 34%

I've been happy with it and ironically outperformed the 55/45 UPRO/TMF this year. Just a thought to consider.
Semantics
Posts: 164
Joined: Tue Mar 10, 2020 1:42 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Semantics »

perfectuncertainty wrote: Mon Oct 26, 2020 1:02 pm
stockmaster wrote: Sun Oct 25, 2020 1:41 am
perfectuncertainty wrote: Fri Oct 23, 2020 9:36 pm
stockmaster wrote: Fri Oct 23, 2020 12:30 pm
perfectuncertainty wrote: Fri Oct 23, 2020 11:13 am You will stopped out with the 20% stop loss.

I'd consider a laddered VIX call hedge.

I'd hedge my portfolios with VIX 10 delta calls, usually 25-40 basis points of net liquid account worth (0.0025 - 0.0034 x NLW) depending on where the VIX is when I am buying them. Purchase them every third Wednesday of the month, 4 months out, and let them expire worthless. In a cash-secured account, your maximum theoretical drawdown occurs somewhere around a 20% decline in the S&P where you should see an unrealized 10-15% loss. At 2008 level drawdowns you would see a profit. In March 2020, they returned a substantial profit due to the speed of the decline in the S&P.
Why not allocate to a VIX ETF instead?
It's a black swan hedge and you don't want to tie up a large amount of your money in a VIX ETF that may not provide the necessary hedge if it is underfunded. The nature and leverage in VIX call options provide safety at a lower cost. Back test it and see.
But VIX ETFs do provide good hedging at very low allocation amounts. See this for a (quickly-made) example:

Image

So adding a VIX ETF behaves exactly how you want hedging to behave... You trade a lower CAGR for higher ratios and lower maximum drawdowns. And keep in mind some VIX ETFs, like XVZ, will not just shoot to zero with time (if you bought $10,000 in XVZ when it came out in September 2011 you'd still have $7,000 now) so ETF investing is not a total loss, unlike your constantly-expiring call options.

How much exactly do your options decrease your max drawdown and increase your ratios in comparison?
Baseline - If we use 55/45 UPRO / TMF from 10/2011 till now the CAGR is 31.25%. If we use 48/44/8 UPRO/TMF/XVZ the CAGR is 27.89%. So the hedge is costing us 3.36% CAGR. In dollar terms on a 100k account, the cost of allocating 8% to UVX since 10/2011 is $248k. That's a significant cost but if it saves us in a Black Swan it might be worth it.

First some observations about XVZ
1. Look at the XVZ volume - it's almost nonexistent and has a spread of 0.25.
2. When everything was getting killed in 3/20 the high was only 20% higher than it is now (so not much protection). Is that a sufficient enough hedge? Let's see what the hedge would be worth on 3/16/2020.
How it follow from the bolded that XVZ doesn't offer much protection? It more than doubled between late-Feb and early-March. If anything, I would think the fact that it stabilized only 20% lower than its peak would be a point in its favor, since holding onto its gains like that limits the timing element of rebalancing.

My issues with XVZ are more the fact it's unlikely to perform well in a longer more gradual downturn like 2018 or the early 2000s, and as you said the low volume (only $30M AUM).

Your approach seems similar to the VIX tail hedge index (^VXTH), which did pretty well in both the GFC and pandemic, but I think this type of strategy is much more valuable in a leveraged portfolio, since SPX will recover from a downturn, but UPRO may not.
leenaerts
Posts: 28
Joined: Sun Sep 02, 2018 2:15 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by leenaerts »

Basically I like leverage, but it looks like you can overleverage on 1 thing, lose it all and never recover. If you spread it out to un/very low correlated assets your chances of losing all become way less and make it way more likely to recover.

Currently I'm having 100k in UPRO, 100k loaned out on crypto.com for 12% interest a year before taxes (I know not without risk, but imo low risk). 300k in bitcoin that I am planning on selling at least half of it.

My plan after reading this: I don't like TMF with low interest rates, but like to keep my investments in crypto.com-loans and bitcoin so I was thinking of a 33/33/33 upro/crypto/btc, which should be earning similar (besides the decay) to 60/20/20 spy/crypto/btc. btc is probably as volatile as upro or not more so dont feel like putting more than 1/3 of my equity in.

What do you think about this? Any other advice?

What about putting some extra margin on upro (and maybe btc?) and investing some more in crypto @ 12%?
Way more profitable, just how much leverage can i put on upro without risking margin calls? 33% (yearly rebalanced?)? so starting with 4x leverage on spy?

I am thinking yearly rebalancing for tax reasons to make profits count long term.
Thereum
Posts: 58
Joined: Sun Jun 14, 2020 9:05 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Thereum »

Not to beat a dead horse, but short SQQQ is absolutely killing long TQQQ. Long TQQQ has an 18% drawdown from its peak in September, while short SQQQ is up 1% in the same time frame.

I believe shorting SQQQ is not only much better than going long on TQQQ, but is among the best strategies in the world. You don't even need to go long on LTTs to get great absolute and risk-adjusted returns with this strategy.
keith6014
Posts: 381
Joined: Thu Jan 02, 2014 11:58 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by keith6014 »

Thereum wrote: Sat Nov 21, 2020 12:50 am Not to beat a dead horse, but short SQQQ is absolutely killing long TQQQ. Long TQQQ has an 18% drawdown from its peak in September, while short SQQQ is up 1% in the same time frame.

I believe shorting SQQQ is not only much better than going long on TQQQ, but is among the best strategies in the world. You don't even need to go long on LTTs to get great absolute and risk-adjusted returns with this strategy.
How often do you rebalance?
Thereum
Posts: 58
Joined: Sun Jun 14, 2020 9:05 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Thereum »

keith6014 wrote: Sat Nov 21, 2020 8:14 am
Thereum wrote: Sat Nov 21, 2020 12:50 am Not to beat a dead horse, but short SQQQ is absolutely killing long TQQQ. Long TQQQ has an 18% drawdown from its peak in September, while short SQQQ is up 1% in the same time frame.

I believe shorting SQQQ is not only much better than going long on TQQQ, but is among the best strategies in the world. You don't even need to go long on LTTs to get great absolute and risk-adjusted returns with this strategy.
How often do you rebalance?
The long term plan is monthly. However, it's possible that better results would be achieved with weekly re-balancing. Re-balancing with this strategy is nice because you never realize gains. You just short SQQQ into oblivion.

I am also looking at hedges for this strategy. It seems like you want to short TQQQ and TMV as well.
User avatar
privatefarmer
Posts: 710
Joined: Mon Sep 08, 2014 2:45 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by privatefarmer »

Thereum wrote: Sat Nov 21, 2020 5:35 pm
keith6014 wrote: Sat Nov 21, 2020 8:14 am
Thereum wrote: Sat Nov 21, 2020 12:50 am Not to beat a dead horse, but short SQQQ is absolutely killing long TQQQ. Long TQQQ has an 18% drawdown from its peak in September, while short SQQQ is up 1% in the same time frame.

I believe shorting SQQQ is not only much better than going long on TQQQ, but is among the best strategies in the world. You don't even need to go long on LTTs to get great absolute and risk-adjusted returns with this strategy.
How often do you rebalance?
The long term plan is monthly. However, it's possible that better results would be achieved with weekly re-balancing. Re-balancing with this strategy is nice because you never realize gains. You just short SQQQ into oblivion.

I am also looking at hedges for this strategy. It seems like you want to short TQQQ and TMV as well.
Do you use interactive brokers? What’s the cost to borrow the shares of SQQQ?
langlands
Posts: 600
Joined: Wed Apr 03, 2019 10:05 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by langlands »

Thereum wrote: Sat Nov 21, 2020 5:35 pm
keith6014 wrote: Sat Nov 21, 2020 8:14 am
Thereum wrote: Sat Nov 21, 2020 12:50 am Not to beat a dead horse, but short SQQQ is absolutely killing long TQQQ. Long TQQQ has an 18% drawdown from its peak in September, while short SQQQ is up 1% in the same time frame.

I believe shorting SQQQ is not only much better than going long on TQQQ, but is among the best strategies in the world. You don't even need to go long on LTTs to get great absolute and risk-adjusted returns with this strategy.
How often do you rebalance?
The long term plan is monthly. However, it's possible that better results would be achieved with weekly re-balancing. Re-balancing with this strategy is nice because you never realize gains. You just short SQQQ into oblivion.

I am also looking at hedges for this strategy. It seems like you want to short TQQQ and TMV as well.
I've asked you this once before, but what do you think happens if you rebalance daily?
Thereum
Posts: 58
Joined: Sun Jun 14, 2020 9:05 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Thereum »

langlands wrote: Sat Nov 21, 2020 8:26 pm I've asked you this once before, but what do you think happens if you rebalance daily?
You are just long TQQQ in that case.

I've done some thinking about short SQQQ in terms of greeks. I believe it is long delta, long theta, short gamma, and long vega. Long TQQQ would be long delta, short theta, long gamma, and short vega. Since SQQQ is short gamma, the position size must frequently be increased as the underlying rallies.

This greeks analysis might explain why long TQQQ is problematic -- you are long gamma and short theta, but you are short vega. A call option would give you long vega exposure with the other greeks being the same. You are much better off buying index ETF call options with leverage than buying leveraged ETFs.
langlands
Posts: 600
Joined: Wed Apr 03, 2019 10:05 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by langlands »

Thereum wrote: Sat Nov 21, 2020 8:56 pm
langlands wrote: Sat Nov 21, 2020 8:26 pm I've asked you this once before, but what do you think happens if you rebalance daily?
You are just long TQQQ in that case.

Yes exactly.

I've done some thinking about short SQQQ in terms of greeks. I believe it is long delta, long theta, short gamma, and long vega. Long TQQQ would be long delta, short theta, long gamma, and short vega. Since SQQQ is short gamma, the position size must frequently be increased as the underlying rallies.

This greeks analysis might explain why long TQQQ is problematic -- you are long gamma and short theta, but you are short vega. A call option would give you long vega exposure with the other greeks being the same. You are much better off buying index ETF call options with leverage than buying leveraged ETFs.

Interesting to apply greeks to non-option securities. I assume the theta you're talking about is essentially the borrow cost.
Besides that, technically I think all of them except the delta are 0 if looked at on an instantaneous single-point basis, which is how they're defined (they are just leveraged contracts, so the delta is 3x). However, if you look at it from a buy and hold (non-rebalancing) perspective over the course of say a month, I kind of see where you're coming from. You have to be careful about what exactly you're talking about though and we run into this re-occurring median vs. mean issue aka picking up pennies in front of a steam roller. For instance, in the median case short SQQQ is long vega and long TQQQ is short vega (held unbalanced over the course of a month say). But in expected value, SQQQ and TQQQ both have 0 vega. I hope that's clear. The detailed math is too much for me to explain in this type of forum, but it's all to do with exactly how volatility decay works (key point: volatility decay IS NOT free money).

I think you are overcomplicating the analysis a little and putting TQQQ in an unfairly negative light by not thinking of TQQQ in the context of a rebalancing strategy (which you do consider for SQQQ). At the end of the day, with either SQQQ or TQQQ, all there is is your leverage ratio. With TQQQ, if you don't rebalance, it's 3x every day. With short SQQQ, it's 3x today and can go up or down depending on what happens to the underlying. But no one's stopping you from adjusting the ratio however you want. In a tax advantaged account, there aren't tax issues to consider either and you can rebalance at will. So the core decision at the end of the day is just how much leverage do you want on a daily basis.

User avatar
Steve Reading
Posts: 2498
Joined: Fri Nov 16, 2018 10:20 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Steve Reading »

Thereum wrote: Sat Nov 21, 2020 8:56 pm Long TQQQ would be long delta, short theta, long gamma, and short vega.
This isn't quite right. Long TQQQ is theta-neutral, it is gamma neutral from day to day (and actually short gamma during the day) and long vega. At first glance, it might appear that TQQQ does poorly in high volatility so it must be betting on low volatility ("short vega") but that's not quite right.

In Options as a Strategic Investment, McMillan has an excellent chapter where he works through the math of a buy-and-hold and a buy-and-hold-and-rebalance investment in index funds, and shows the latter is effectively short vega (and in fact, identical to a position formulated with options). That is because the idea that you'd sell an asset after a rise in price is fundamentally what shorting volatility is. B&H with no rebalancing, OTOH, is vega-neutral.

Hence, an investment like TQQQ, which buys more exposure after a rise in price, is long volatility and long vega.

Here's a more intuitive explanation: Volatility just means the likelihood of an underlying finishing farther away from its current price. Long TQQQ does better with positions that have a high chance of ending up very far from their current price, either on the upside (by constantly re-leveraging up) or downside (by selling exposure on the way down).

Put another way, if I told you SPY will finish exactly where it is right now in a year's time, with no possibility of a higher or lower return (no volatility), then strategies that sell highs and buy dips (rebalancing) will make money (because they're short vega, as McMillan says) and strategies that buys highs and sells dips will lose money (as they are long vega). Strategies that just buy-and-hold (like a non-rebalanced 60/40, or just 100% stocks) are vega-neutral and make no money.
"... so high a present discounted value of wealth, it is only prudent for him to put more into common stocks compared to his present tangible wealth, borrowing if necessary" - Paul Samuelson
tbfanman
Posts: 12
Joined: Sun Apr 19, 2020 1:44 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by tbfanman »

Steve Reading wrote: Sat Nov 21, 2020 9:43 pm
Thereum wrote: Sat Nov 21, 2020 8:56 pm Long TQQQ would be long delta, short theta, long gamma, and short vega.
This isn't quite right. Long TQQQ is theta-neutral, it is gamma neutral from day to day (and actually short gamma during the day) and long vega. At first glance, it might appear that TQQQ does poorly in high volatility so it must be betting on low volatility ("short vega") but that's not quite right.

In Options as a Strategic Investment, McMillan has an excellent chapter where he works through the math of a buy-and-hold and a buy-and-hold-and-rebalance investment in index funds, and shows the latter is effectively short vega (and in fact, identical to a position formulated with options). That is because the idea that you'd sell an asset after a rise in price is fundamentally what shorting volatility is. B&H with no rebalancing, OTOH, is vega-neutral.

Hence, an investment like TQQQ, which buys more exposure after a rise in price, is long volatility and long vega.

Here's a more intuitive explanation: Volatility just means the likelihood of an underlying finishing farther away from its current price. Long TQQQ does better with positions that have a high chance of ending up very far from their current price, either on the upside (by constantly re-leveraging up) or downside (by selling exposure on the way down).

Put another way, if I told you SPY will finish exactly where it is right now in a year's time, with no possibility of a higher or lower return (no volatility), then strategies that sell highs and buy dips (rebalancing) will make money (because they're short vega, as McMillan says) and strategies that buys highs and sells dips will lose money (as they are long vega). Strategies that just buy-and-hold (like a non-rebalanced 60/40, or just 100% stocks) are vega-neutral and make no money.
Suppose, then, that we wanted to mimic the exposure to volatility that shorting SPXS (or Sqqq) has, but wanted to avoid the margin call potential. Would holding upro (or tqqq) and buying the dips (perhaps even buying enough that you increase nominal exposure as the market drops) effectively accomplish this?
User avatar
Uncorrelated
Posts: 1105
Joined: Sun Oct 13, 2019 3:16 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Uncorrelated »

langlands wrote: Sat Nov 21, 2020 8:26 pm
Thereum wrote: Sat Nov 21, 2020 5:35 pm
keith6014 wrote: Sat Nov 21, 2020 8:14 am
Thereum wrote: Sat Nov 21, 2020 12:50 am Not to beat a dead horse, but short SQQQ is absolutely killing long TQQQ. Long TQQQ has an 18% drawdown from its peak in September, while short SQQQ is up 1% in the same time frame.

I believe shorting SQQQ is not only much better than going long on TQQQ, but is among the best strategies in the world. You don't even need to go long on LTTs to get great absolute and risk-adjusted returns with this strategy.
How often do you rebalance?
The long term plan is monthly. However, it's possible that better results would be achieved with weekly re-balancing. Re-balancing with this strategy is nice because you never realize gains. You just short SQQQ into oblivion.

I am also looking at hedges for this strategy. It seems like you want to short TQQQ and TMV as well.
I've asked you this once before, but what do you think happens if you rebalance daily?
To add, daily rebalancing is theoretically optimal (before taxes). Any difference between TQQQ or short SQQQ must be based on a difference in expense ratio or random noise. I don't know any theoretical arguments why shorting SQQQ would lead to better expected performance than TQQQ.

I reproduced my analysis on the CER of options and LETF under different rebalance intervals and found that shorting a short LETF consistently results in the worst certainty equivalent return across all range of parameters except when the position is daily rebalanced. With daily rebalancing all investments are equal.

Image

The number on the Y axis is the certainty equivalent amount ($$$) of the strategy (over 21 days) starting with $10000. Higher is better.

The CER loss of using short-short-letf instead of letf is about 0.22% per month, or 2.64% per year. For both investments to be equal, the expense ratio difference must be that big.
Thereum
Posts: 58
Joined: Sun Jun 14, 2020 9:05 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Thereum »

Uncorrelated wrote: Sun Nov 22, 2020 3:22 am
To add, daily rebalancing is theoretically optimal (before taxes). Any difference between TQQQ or short SQQQ must be based on a difference in expense ratio or random noise. I don't know any theoretical arguments why shorting SQQQ would lead to better expected performance than TQQQ.
You haven't seen the numerous arguments about volatility decay of leveraged ETFs? Why not just do a simple backtest and obverse how short SQQQ crushes TQQQ in absolute and risk-adjusted returns.

Long TQQQ vs short SQQQ:

Image

Also, here's what it looks like if you short SQQQ and TQQQ with a beta target of zero:

Image

Shorting leveraged ETFs is better than going long leveraged ETFs. It is really that simple. I challenge you to find one case where going long would return more than shorting the inverse.
Last edited by Thereum on Sun Nov 22, 2020 4:44 am, edited 2 times in total.
Thereum
Posts: 58
Joined: Sun Jun 14, 2020 9:05 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Thereum »

Another example, with a different underlying. I am comparing shorting FAZ, long FAS, and long XLF (unleveraged equivalent):

Image

Yes, that's not a glitch -- you make significantly more holding the unleveraged ETF than the 3x. Shorting the inverse, of course, blows away the alternatives and produces unbelievable returns. As I said earlier, I have a very hard time finding better investment strategies than shorting leveraged inverse ETFs.
User avatar
Uncorrelated
Posts: 1105
Joined: Sun Oct 13, 2019 3:16 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Uncorrelated »

Thereum wrote: Sun Nov 22, 2020 4:19 am
Uncorrelated wrote: Sun Nov 22, 2020 3:22 am
To add, daily rebalancing is theoretically optimal (before taxes). Any difference between TQQQ or short SQQQ must be based on a difference in expense ratio or random noise. I don't know any theoretical arguments why shorting SQQQ would lead to better expected performance than TQQQ.
Seriously? You haven't seen the numerous arguments about volatility decay of leveraged ETFs? Why not just do a simple backtest and obverse how short SQQQ crushes TQQQ in absolute and risk-adjusted returns.

Yes I have. I was one of the most vocal opponents of LETF's until I realized that the math doesn't actually work out that way. How can volatility decay be bad if a continuously rebalanced asset allocation is optimal1?

It turns out that the backtests were misleading. When viewed through an expected utility optimization lens and using all available stock market data, daily rebalancing is superior. This can clearly be observed by comparing the CER of options and LETF's. Despite the existence of "volatility decay", LETF's outperform options over all rebalance intervals (ignoring expense ratio).

Your simple backtest has numerous errors, such as not performing margin calls, forgetting borrowing costs, no out-of-sample tests, not specifying an utility function, etc etc.

1 https://cs.uwaterloo.ca/~paforsyt/agon.pdf theorem 1, page 119.
Thereum wrote: Sun Nov 22, 2020 4:19 am Shorting leveraged ETFs is better than going long leveraged ETFs. It is really that simple. I challenge you to find one case where going long would return more than shorting the inverse.
The image in my previous post shows exactly that, based on simulated market data going back to 1932. Although the probability distribution of outcomes for LETF is strictly superior to short short LETF, it is not guaranteed you will get a better outcome. Just as betting on black in the casino is not guaranteed to result in a loss.
langlands
Posts: 600
Joined: Wed Apr 03, 2019 10:05 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by langlands »

This recent post by Semantics in a different thread is I think also quite relevant to this discussion and summarizes the situation well:

viewtopic.php?p=5612084#p5612084
corp_sharecropper
Posts: 333
Joined: Thu Nov 07, 2013 2:36 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by corp_sharecropper »

Learned something interesting on Friday. Apparently, the Small Exchange is/has launched a new 10 year treasury futures contract with a notional value of 100x the index (currently notional value is ~$830/contract). The big problem w/futures, imo, has been size of treasury futures (>$100K). Looks like this contract is/will be available at IB (I didn't try to find out as the market was closed, it doesn't trade the same hours, 24/5, that CME/NYMEX/etc trade). I have to imagine it won't be too easy to trade at the beginning but I really hope it catches on and other treasury yield futures follow. Sort of wish they went for a slightly higher notional value but oh well.

Ticker is 10YSME
User avatar
Ramjet
Posts: 509
Joined: Thu Feb 06, 2020 11:45 am

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Ramjet »

Someone made a separate thread a while back where people participating in the Excellent Adventure could post their results, allocation, and when they entered into it. I can't seem to locate it. Help anyone?
User avatar
Steve Reading
Posts: 2498
Joined: Fri Nov 16, 2018 10:20 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Steve Reading »

Uncorrelated wrote: Sun Nov 22, 2020 3:22 am Any difference between TQQQ or short SQQQ must be based on a difference in expense ratio or random noise.
The difference is due to stock exposure differences, and we should expect those differences to be quite large. Ex:

1) Say your portfolio is $100K of UPRO. This portfolio has 3x exposure to the market on day one ($300K worth of SPY exposure). Say the market dips 10% so you end the day with $70K. Day 2, leverage resets and you're still effectively 300% stocks ($270K worth of SPY exposure).
2) Say your portfolio instead is a short of $100K of SPXS shares. That's also 300% exposure to the market on day one. But say the market dips 10% on day one here too. SPXS shares appreciate 30%, so your SPXS short position becomes a $130K worth of shares. Your equity has dropped to 70K like in the previous example. So your actual stock leverage becomes 390K/70K = 5.57x.

Now, the above is hand-wavy in terms of collateral but the point remains. Buy and hold UPRO is the only portfolio that remains 300% stocks day after day. Anything else (buy and hold futures, short SPXS, etc) does not remain 300% day after day. And so you should expect a different (in fact VERY different) performance after one day. It's not expense or noise, it's just asset allocation.
tbfanman wrote: Sun Nov 22, 2020 12:02 am Suppose, then, that we wanted to mimic the exposure to volatility that shorting SPXS (or Sqqq) has, but wanted to avoid the margin call potential. Would holding upro (or tqqq) and buying the dips (perhaps even buying enough that you increase nominal exposure as the market drops) effectively accomplish this?
You can do some of it, but how much of it you can accomplish depends on the depth of the dip and your starting UPRO/cash exposure. You can imagine a dip large enough where you simply cannot keep tracking the short SPXS position. In my example above, say you started with $100K cash and $100K UPRO. If the market drops 10%, you'd have to dump $60K of the cash into UPRO (to end up with $130K UPRO and $390K nominal total exposure, like the short SPXS portfolio). After another 10% dip, you'd be out of cash and cannot keep tracking short SPXS. If you start with more cash, you can track for longer, but then your starting exposure is lower.
"... so high a present discounted value of wealth, it is only prudent for him to put more into common stocks compared to his present tangible wealth, borrowing if necessary" - Paul Samuelson
User avatar
Steve Reading
Posts: 2498
Joined: Fri Nov 16, 2018 10:20 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Steve Reading »

Thereum wrote: Sun Nov 22, 2020 4:22 am Another example, with a different underlying. I am comparing shorting FAZ, long FAS, and long XLF (unleveraged equivalent):

Image

Yes, that's not a glitch -- you make significantly more holding the unleveraged ETF than the 3x. Shorting the inverse, of course, blows away the alternatives and produces unbelievable returns. As I said earlier, I have a very hard time finding better investment strategies than shorting leveraged inverse ETFs.
Didn't we already talk about this?

IBKR requires a 90% maintenance margin on a FAZ short. So if you start with 200% cash, -100% FAZ, a measly 5% rise in FAZ already triggers a margin call. Have you seen the graph of FAZ? Its a LETF, it gains 5% all the time. It tripled in March. It gained 50% in Dec 2018 and June 2020. You wouldn't have maintained exposure, the graphs above are all fairy tales. I suspect PV doesn't actually close a position until it reaches $0 dollars and I don't think it does intra-month (correct me if I'm wrong though).
I don't know for certain but I'll guess it's not much better with options.

But you know that already, your short LETF portfolios always carry protective puts for this very reason IIRC. I don't think you accounted for those in the graphs above.

Or maybe you do this in a tiny portion of your account (say <5%) to ensure you have overwhelming collateral. Obviously, that solves the issue. If you have effectively infinite collateral and know markets will recover (as they have thus far in the graphs above), then it's a no-brainer to just increase exposure as the market drops. It's free money at that point. But a bit misleading IMO.
"... so high a present discounted value of wealth, it is only prudent for him to put more into common stocks compared to his present tangible wealth, borrowing if necessary" - Paul Samuelson
Semantics
Posts: 164
Joined: Tue Mar 10, 2020 1:42 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Semantics »

Thereum wrote: Sun Nov 22, 2020 4:19 am
Uncorrelated wrote: Sun Nov 22, 2020 3:22 am
To add, daily rebalancing is theoretically optimal (before taxes). Any difference between TQQQ or short SQQQ must be based on a difference in expense ratio or random noise. I don't know any theoretical arguments why shorting SQQQ would lead to better expected performance than TQQQ.
You haven't seen the numerous arguments about volatility decay of leveraged ETFs? Why not just do a simple backtest and obverse how short SQQQ crushes TQQQ in absolute and risk-adjusted returns.

Long TQQQ vs short SQQQ:

Image

Also, here's what it looks like if you short SQQQ and TQQQ with a beta target of zero:

Image

Shorting leveraged ETFs is better than going long leveraged ETFs. It is really that simple. I challenge you to find one case where going long would return more than shorting the inverse.
Short SQQQ assuming 100% collateral and not more would actually have gone to $0 mid-March. You shouldn't rely on Portfolio Visualizer, it only uses monthly returns. So the massive drop mid-March doesn't show up at all you just see where it was at the end of the month.
Thereum
Posts: 58
Joined: Sun Jun 14, 2020 9:05 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Thereum »

Steve Reading wrote: Sun Nov 22, 2020 11:51 am Didn't we already talk about this?

IBKR requires a 90% maintenance margin on a FAZ short. So if you start with 200% cash, -100% FAZ, a measly 5% rise in FAZ already triggers a margin call. Have you seen the graph of FAZ? Its a LETF, it gains 5% all the time. It tripled in March. It gained 50% in Dec 2018 and June 2020. You wouldn't have maintained exposure, the graphs above are all fairy tales. I suspect PV doesn't actually close a position until it reaches $0 dollars and I don't think it does intra-month (correct me if I'm wrong though).
I don't know for certain but I'll guess it's not much better with options.

But you know that already, your short LETF portfolios always carry protective puts for this very reason IIRC. I don't think you accounted for those in the graphs above.

Or maybe you do this in a tiny portion of your account (say <5%) to ensure you have overwhelming collateral. Obviously, that solves the issue. If you have effectively infinite collateral and know markets will recover (as they have thus far in the graphs above), then it's a no-brainer to just increase exposure as the market drops. It's free money at that point. But a bit misleading IMO.
I am seeing a 50% margin requirement for shorting FAZ. SQQQ has only a 13% margin requirement for me. But yes, margin calls are an issue. There are several ways to avoid this.

1) Keep position size small -- you will still crush long leveraged and long leveraged even with only a target of 15% of the portfolio exposed

2) Shorting both leveraged ETFs. I showed earlier how you can achieve exceptional returns with a delta neutral approach. You can even get excellent returns using a delta negative approach, if you increase the short size of TQQQ.

3) Using options to define risk. I used to short SQQQ using call spreads, but I have switched to buying very short term call options, making the position more like a diagonal. These call options typically cost a few pennies but will completely protect me against a crash like we saw March 2020. In fact, I have structured my trades to actually profit if this happens.

By the way, you corrected me about the greeks. I was thinking about things a bit differently. What I was trying to suggest is that leveraged ETFs under-perform if realized volatility is high. For instance, if the SP500 is flat for an entire year, then SQQQ and TQQQ will be down. I believe this is a mathematically certainty (someone please correct me if I am wrong). The higher the volatility during this period, the more the leveraged ETFs will have fallen.

If realized volatility is low, then the market will almost certainly be higher than in the previous year, which means SQQQ will be down significantly.

The only scenario in which SQQQ would be up is if the market is down. However, markets almost always experience severe volatility when they crash, which means that SQQQ will still under-perform.

Portfolio Visualizer marks to market the portfolio every month, so you are correct that a lot of the strategies would have blown up. That's why a hedge is necessary if the target position size is 100%.
User avatar
Uncorrelated
Posts: 1105
Joined: Sun Oct 13, 2019 3:16 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Uncorrelated »

Thereum wrote: Sun Nov 22, 2020 5:15 pm 1) Keep position size small -- you will still crush long leveraged and long leveraged even with only a target of 15% of the portfolio exposed.
With a target of 15% of the portfolio exposed (that's pretty vague...), it's less risky than total stock market. Do you really think your proposed strategy has higher risk-adjusted returns than a bog-standard two-fund portfolio?

Which do you think is more likely, hedge fund managers across the world spent billions and billions on market research, yet are completely unaware of your superior strategy. Or the risks to this strategy are far greater than you estimate?
2) Shorting both leveraged ETFs. I showed earlier how you can achieve exceptional returns with a delta neutral approach. You can even get excellent returns using a delta negative approach, if you increase the short size of TQQQ.
That is not delta neutral over time horizons longer than one day. If you do rebalance daily, the only difference between TQQQ and short SQQQ is the difference in expense ratio, but if the market is even vaguely efficient it's impossible to make any profit of this. If you do not rebalance daily, the strategy is suboptimal (see proof that a constant asset allocation is optimal a few posts back) and basically amounts to market timing. This was also pointed out by Steve Reading and Semantics using slightly different arguments and terminology.

3) Using options to define risk. I used to short SQQQ using call spreads, but I have switched to buying very short term call options, making the position more like a diagonal. These call options typically cost a few pennies but will completely protect me against a crash like we saw March 2020. In fact, I have structured my trades to actually profit if this happens.
As far as I'm aware of, calculating whether such options actually results in a preferable distribution of outcomes is an open problem. According to the insurance market model, purchasing insurance (call options) is likely to result in improved expected utility if you have a lower risk tolerance than average. And selling insurance is likely to result in improved expected utility if you have a high risk tolerance. I see strong theoretical reasons why purchasing insurance in conjunction with high risk strategies results in lower expected utility.
Post Reply