HEDGEFUNDIE's excellent adventure Part II: The next journey

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
perfectuncertainty
Posts: 386
Joined: Sun Feb 04, 2018 6:44 pm

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

Post by perfectuncertainty »

tradri wrote: Sat Apr 17, 2021 5:03 am
dont_know_mind wrote: Sat Apr 17, 2021 4:38 am Have you read the book “the man who solved the market” (Zuckerberg). The strategies discussed in this thread were explored by rentech and Bridgewater since the 80’s. From my interpretation of the book, Simons suggested that a large part of the returns they were generating were coming from retail - and not buy and hold Bogleheads (Simonds speculated “dentists” in the 90’s, in the 2020’s maybe Hedgefundies ?).

I think it is naive to think that technical analysis or the backtesting suggested here can generate alpha. What you are doing is 30 years behind the hedge funds. The only reason I can think of that this would generate a better than market return is the extra beta taken (through leverage) or luck.

I don’t see how doing the statistical analysis proposed or sticking it in portfolio visualiser has any hope of beating Rentech & others over time. It’s sort of delusional. What do you think your edge is with the strategy you propose?
I agree that trying to beat the professionals, that are doing this full-time and with much more resources, is pointless.

By itself, these leveraged ETFs don't generate a higher return. Over long periods of time, they perform about the same as their unlevered index. While they give you 3x the (daily) beta, this extra volatility isn't rewarded by itself.

The only reason this works, in my opinion, is because it is combined with a leveraged, uncorrelated asset, and rebalanced from time to time. This creates a natural market-timing effect.
You don't think it’s a little arrogant to call those engaging in technical analysis or other strategies naive? Many have beaten the market quite radically and with controlled risk. They just aren't using the strategies that are described here. No matter what approach you use, even purge BH Stock/Bonds AA, you are lying on probability but not on certainty.
dont_know_mind
Posts: 94
Joined: Sun Nov 22, 2020 7:14 am

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

Post by dont_know_mind »

tradri wrote: Sat Apr 17, 2021 10:45 am
dont_know_mind wrote: Sat Apr 17, 2021 10:17 am I'm paying 1.8% on non-callable debt currently, fixed for 2 years.
I was thinking of getting a margin loan at around 50% LTV and the rate on that would be around 0.75%.
I am not sure about how reliable the bank line of credit would be in severe crisis though, so I'm not sure if the lower margin rate would be worth the risk of margin call.
I've had this on my to-do list for a few months now. As the dividends in the indexes I have are paying 3-4%, and the cost of capital is half that, I haven't had a great rush to look at the lower rate margin option. But I probably will be more motivated if rates rise.
Ok.

Even at 1.5x direct leverage on the stock market, the returns are less than what a 70/30 UPRO/TMF backtest shows. (assuming you are after the highest returns)
The direct leverage was probably 2-3. The 1.5 is what I am thinking of having margin LTV if I replace some of the non-callable debt with margin debt.

I put in 0.9 in equity and bought 2.5 in index fund (with 1.6 in debt).
The profit so far is 0.54. So a 60% return on invested equity since April 2020.
Has 70/30 UPRO/TMF returned 60% since April 2020 ? It's hard to compare as different entry points.

The main thing though for me is that I felt comfortable enough in my strategy to continue to add. I bought 0.9 in April 2020, 1.0 from May-July 20, 0.6 from Aug 20 to Feb 21.
My ramblings are probably rubbish, biased by my position and talking my own book. No-one should invest based on my views. Some days I wonder whether I've had some skill or just been a lucky gambler.
dont_know_mind
Posts: 94
Joined: Sun Nov 22, 2020 7:14 am

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

Post by dont_know_mind »

perfectuncertainty wrote: Sat Apr 17, 2021 11:21 am
tradri wrote: Sat Apr 17, 2021 5:03 am
dont_know_mind wrote: Sat Apr 17, 2021 4:38 am Have you read the book “the man who solved the market” (Zuckerberg). The strategies discussed in this thread were explored by rentech and Bridgewater since the 80’s. From my interpretation of the book, Simons suggested that a large part of the returns they were generating were coming from retail - and not buy and hold Bogleheads (Simonds speculated “dentists” in the 90’s, in the 2020’s maybe Hedgefundies ?).

I think it is naive to think that technical analysis or the backtesting suggested here can generate alpha. What you are doing is 30 years behind the hedge funds. The only reason I can think of that this would generate a better than market return is the extra beta taken (through leverage) or luck.

I don’t see how doing the statistical analysis proposed or sticking it in portfolio visualiser has any hope of beating Rentech & others over time. It’s sort of delusional. What do you think your edge is with the strategy you propose?
I agree that trying to beat the professionals, that are doing this full-time and with much more resources, is pointless.

By itself, these leveraged ETFs don't generate a higher return. Over long periods of time, they perform about the same as their unlevered index. While they give you 3x the (daily) beta, this extra volatility isn't rewarded by itself.

The only reason this works, in my opinion, is because it is combined with a leveraged, uncorrelated asset, and rebalanced from time to time. This creates a natural market-timing effect.
You don't think it’s a little arrogant to call those engaging in technical analysis or other strategies naive? Many have beaten the market quite radically and with controlled risk. They just aren't using the strategies that are described here. No matter what approach you use, even purge BH Stock/Bonds AA, you are lying on probability but not on certainty.
Yes, I think you are naive if you think you can generate excess returns from technical analysis or looking up things on portfolio visualiser. You might get excess returns through luck though and think it was skill and blow up sometime in the future.
My ramblings are probably rubbish, biased by my position and talking my own book. No-one should invest based on my views. Some days I wonder whether I've had some skill or just been a lucky gambler.
tradri
Posts: 606
Joined: Tue Mar 23, 2021 5:42 pm

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

Post by tradri »

perfectuncertainty wrote: Sat Apr 17, 2021 11:21 am You don't think it’s a little arrogant to call those engaging in technical analysis or other strategies naive? Many have beaten the market quite radically and with controlled risk. They just aren't using the strategies that are described here. No matter what approach you use, even purge BH Stock/Bonds AA, you are lying on probability but not on certainty.
Show me a fool-proof strategy to beat the market and I will gladly take it. :wink:
tradri
Posts: 606
Joined: Tue Mar 23, 2021 5:42 pm

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

Post by tradri »

dont_know_mind wrote: Sat Apr 17, 2021 11:34 am
The direct leverage was probably 2-3. The 1.5 is what I am thinking of having margin LTV if I replace some of the non-callable debt with margin debt.

I put in 0.9 in equity and bought 2.5 in index fund (with 1.6 in debt).
The profit so far is 0.54. So a 60% return on invested equity since April 2020.
Has 70/30 UPRO/TMF returned 60% since April 2020 ? It's hard to compare as different entry points.

The main thing though for me is that I felt comfortable enough in my strategy to continue to add. I bought 0.9 in April 2020, 1.0 from May-July 20, 0.6 from Aug 20 to Feb 21.
70/30 UPRO/TMF returned more than 100% since April last year. https://www.portfoliovisualizer.com/bac ... tion2_1=30

And 70/30 UPRO/TYD did even better.
User avatar
OohLaLa
Posts: 459
Joined: Tue Mar 09, 2021 7:26 pm

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

Post by OohLaLa »

This is starting to veer off into some weird debates... almost strawman-like. Who here is looking to beat hedge funds at their own game? Who here thinks that HFEA is "beating the market"?

Am I, suddenly, the only one who got into this because they?:
- thought going in 100% into a stock index ETF seemed too concentrated
- wanted to take on more risk (i.e. not "beating the market") than an equivalent non-leveraged stock/ bond portfolo, while indeed expecting some additional reward
- did not intend to hold 3x leverage for decades, but instead have plans to de-leverage/ de-risk in certain ways
User avatar
OohLaLa
Posts: 459
Joined: Tue Mar 09, 2021 7:26 pm

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

Post by OohLaLa »

tradri wrote: Sat Apr 17, 2021 12:03 pm And 70/30 UPRO/TYD did even better.

Just a note for anybody looking into leveraged Intermediate Term Treasuries:
keep in mind that volume is extremely low if you are someone pumping significant money into this scheme or if it grows to a significant amount. Even having 10% allocated to something like TYD or UST could become problematic. It's really unfortunate, but I don't think there are any popular ETF options for long-term holding of that asset class. :(
tradri
Posts: 606
Joined: Tue Mar 23, 2021 5:42 pm

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

Post by tradri »

OohLaLa wrote: Sat Apr 17, 2021 1:42 pm This is starting to veer off into some weird debates... almost strawman-like. Who here is looking to beat hedge funds at their own game? Who here thinks that HFEA is "beating the market"?

Am I, suddenly, the only one who got into this because they?:
- thought going in 100% into a stock index ETF seemed too concentrated
- wanted to take on more risk (i.e. not "beating the market") than an equivalent non-leveraged stock/ bond portfolo, while indeed expecting some additional reward
- did not intend to hold 3x leverage for decades, but instead have plans to de-leverage/ de-risk in certain ways
I agree with your second point. :beer
tradri
Posts: 606
Joined: Tue Mar 23, 2021 5:42 pm

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

Post by tradri »

OohLaLa wrote: Sat Apr 17, 2021 1:46 pm
tradri wrote: Sat Apr 17, 2021 12:03 pm And 70/30 UPRO/TYD did even better.

Just a note for anybody looking into leveraged Intermediate Term Treasuries:
keep in mind that volume is extremely low if you are someone pumping significant money into this scheme or if it grows to a significant amount. Even having 10% allocated to something like TYD or UST could become problematic. It's really unfortunate, but I don't think there are any popular ETF options for long-term holding of that asset class. :(
Thx for the disclaimer.
DMoogle
Posts: 549
Joined: Sat Oct 31, 2020 10:24 pm

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

Post by DMoogle »

tradri wrote: Sat Apr 17, 2021 11:56 am
perfectuncertainty wrote: Sat Apr 17, 2021 11:21 am You don't think it’s a little arrogant to call those engaging in technical analysis or other strategies naive? Many have beaten the market quite radically and with controlled risk. They just aren't using the strategies that are described here. No matter what approach you use, even purge BH Stock/Bonds AA, you are lying on probability but not on certainty.
Show me a fool-proof strategy to beat the market and I will gladly take it. :wink:
How do you define "beating the market"? Many people have different definitions.
tradri
Posts: 606
Joined: Tue Mar 23, 2021 5:42 pm

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

Post by tradri »

DMoogle wrote: Sat Apr 17, 2021 2:53 pm How do you define "beating the market"? Many people have different definitions.
Isn't "beating the market" defined as generating alpha?

If you can generate excess returns that aren't explained by known risk factors, like market risk, value, size, etc...you know something that others don't, right?

Regarding this risk parity approach we are pursuing here, I wouldn't call it "beating the market", but simply the free lunch of diversifying into uncorrelated assets.
DMoogle
Posts: 549
Joined: Sat Oct 31, 2020 10:24 pm

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

Post by DMoogle »

tradri wrote: Sat Apr 17, 2021 2:58 pmIsn't "beating the market" defined as generating alpha?

If you can generate excess returns that aren't explained by known risk factors, like market risk, value, size, etc...you know something that others don't, right?

Regarding this risk parity approach we are pursuing here, I wouldn't call it "beating the market", but simply the free lunch of diversifying into uncorrelated assets.
Even generating alpha has multiple definitions, and is generally a fuzzy concept. If you define it simply as "generating returns in excess of a benchmark portfolio," then the question becomes: what is that benchmark?

If you alter it instead to be "risk-adjusted returns," then how do we define risk? Standard deviation (ala Sharpe Ratio) only goes so far, and does not adequately represent market returns (which are not, historically, normally distributed - you'd want to look at skewness and kurtosis, at a minimum).

The CAPM pricing model [r = Rf + Beta * (Rm - Rf) + Alpha] would imply that using leverage itself, by definition, won't beat the market (because you can never borrow below the risk-free rate) unless the underlying portfolio is already generating alpha, because all you're really doing is raising the Beta.

So let me revise my question: what would be sufficient evidence to demonstrate that a portfolio has generated alpha?
tradri
Posts: 606
Joined: Tue Mar 23, 2021 5:42 pm

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

Post by tradri »

DMoogle wrote: Sat Apr 17, 2021 4:11 pm Even generating alpha has multiple definitions, and is generally a fuzzy concept. If you define it simply as "generating returns in excess of a benchmark portfolio," then the question becomes: what is that benchmark?

If you alter it instead to be "risk-adjusted returns," then how do we define risk? Standard deviation (ala Sharpe Ratio) only goes so far, and does not adequately represent market returns (which are not, historically, normally distributed - you'd want to look at skewness and kurtosis, at a minimum).

The CAPM pricing model [r = Rf + Beta * (Rm - Rf) + Alpha] would imply that using leverage itself, by definition, won't beat the market (because you can never borrow below the risk-free rate) unless the underlying portfolio is already generating alpha, because all you're really doing is raising the Beta.

So let me revise my question: what would be sufficient evidence to demonstrate that a portfolio has generated alpha?
I think if there isn't a logical explanation for why a portfolio is generating alpha (not yet explained beta), there is no reason to believe that this should persist in the future.

I agree that using leverage doesn't generate alpha. It raises the volatility and the return (minus borrowing costs, fees, volatility decay), so the risk-adjusted return is worse.

Apart from having a clear advantage (like the data advantage Renaissance Technologies had), I don't see a reason why one should believe to be able to outsmart all the other rational actors consistently.
Semantics
Posts: 352
Joined: Tue Mar 10, 2020 1:42 pm

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

Post by Semantics »

perfectuncertainty wrote: Sat Apr 17, 2021 11:21 am
tradri wrote: Sat Apr 17, 2021 5:03 am
dont_know_mind wrote: Sat Apr 17, 2021 4:38 am Have you read the book “the man who solved the market” (Zuckerberg). The strategies discussed in this thread were explored by rentech and Bridgewater since the 80’s. From my interpretation of the book, Simons suggested that a large part of the returns they were generating were coming from retail - and not buy and hold Bogleheads (Simonds speculated “dentists” in the 90’s, in the 2020’s maybe Hedgefundies ?).

I think it is naive to think that technical analysis or the backtesting suggested here can generate alpha. What you are doing is 30 years behind the hedge funds. The only reason I can think of that this would generate a better than market return is the extra beta taken (through leverage) or luck.

I don’t see how doing the statistical analysis proposed or sticking it in portfolio visualiser has any hope of beating Rentech & others over time. It’s sort of delusional. What do you think your edge is with the strategy you propose?
I agree that trying to beat the professionals, that are doing this full-time and with much more resources, is pointless.

By itself, these leveraged ETFs don't generate a higher return. Over long periods of time, they perform about the same as their unlevered index. While they give you 3x the (daily) beta, this extra volatility isn't rewarded by itself.

The only reason this works, in my opinion, is because it is combined with a leveraged, uncorrelated asset, and rebalanced from time to time. This creates a natural market-timing effect.
You don't think it’s a little arrogant to call those engaging in technical analysis or other strategies naive? Many have beaten the market quite radically and with controlled risk. They just aren't using the strategies that are described here. No matter what approach you use, even purge BH Stock/Bonds AA, you are lying on probability but not on certainty.
Other strategies maybe, but there are millions of people who research and employ TA. If it worked and could be reliably reproduced it would have minted more than a few billionaires and I am not aware of a single one other than maybe Jim Simons, but I doubt what they do could be reasonably called TA.
dont_know_mind
Posts: 94
Joined: Sun Nov 22, 2020 7:14 am

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

Post by dont_know_mind »

Hydromod wrote: Fri Apr 16, 2021 11:40 am I've been doing a little experimenting to verify the unpredictability of forecasting based on recent history. For example, does recent volatility give any predictive capability on the time scales that I can trade?

It's pretty clear that daily variation is not normally distributed in any of the funds I've looked at, as is well known. Even the Laplace distribution appears to underestimate the tails by a wide margin, although it's pretty reasonable for the smaller moves that occur frequently enough to happen within most quarters.

I haven't yet seen a statistical bias in either the size or direction of future moves over various periods ranging from a day on up, given lookbacks from a week to a quarter. I've looked at indexes, long-term ETFs, gold, commodities. There is some hint of predictability for longer periods, maybe, but I'd be hard pressed to thing that any signal would be large enough to be actionable. I haven't looked at momentum yet, but I don't think volatility gives an actionable estimate of future returns.

This isn't to say that there aren't short-term patterns, but any pattern is destined to disappear.

Without significant predictability, trading cannot be expected to add value, and slippage during trading costs value.

The only predictive measure I've seen that consistently seems to have added value is the unemployment rate, but that wasn't useful for COVID, recent estimates are always inaccurate, and the predictability only operates on the longer time scales.
Where VIX has been above 70 and the market has been more than 30% below the prior peak, is there an above average, statistically significant 2 year return ?
My ramblings are probably rubbish, biased by my position and talking my own book. No-one should invest based on my views. Some days I wonder whether I've had some skill or just been a lucky gambler.
reln
Posts: 718
Joined: Fri Apr 19, 2019 4:01 pm

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

Post by reln »

Semantics wrote: Sat Apr 17, 2021 5:20 pm
perfectuncertainty wrote: Sat Apr 17, 2021 11:21 am
tradri wrote: Sat Apr 17, 2021 5:03 am
dont_know_mind wrote: Sat Apr 17, 2021 4:38 am Have you read the book “the man who solved the market” (Zuckerberg). The strategies discussed in this thread were explored by rentech and Bridgewater since the 80’s. From my interpretation of the book, Simons suggested that a large part of the returns they were generating were coming from retail - and not buy and hold Bogleheads (Simonds speculated “dentists” in the 90’s, in the 2020’s maybe Hedgefundies ?).

I think it is naive to think that technical analysis or the backtesting suggested here can generate alpha. What you are doing is 30 years behind the hedge funds. The only reason I can think of that this would generate a better than market return is the extra beta taken (through leverage) or luck.

I don’t see how doing the statistical analysis proposed or sticking it in portfolio visualiser has any hope of beating Rentech & others over time. It’s sort of delusional. What do you think your edge is with the strategy you propose?
I agree that trying to beat the professionals, that are doing this full-time and with much more resources, is pointless.

By itself, these leveraged ETFs don't generate a higher return. Over long periods of time, they perform about the same as their unlevered index. While they give you 3x the (daily) beta, this extra volatility isn't rewarded by itself.

The only reason this works, in my opinion, is because it is combined with a leveraged, uncorrelated asset, and rebalanced from time to time. This creates a natural market-timing effect.
You don't think it’s a little arrogant to call those engaging in technical analysis or other strategies naive? Many have beaten the market quite radically and with controlled risk. They just aren't using the strategies that are described here. No matter what approach you use, even purge BH Stock/Bonds AA, you are lying on probability but not on certainty.
Other strategies maybe, but there are millions of people who research and employ TA. If it worked and could be reliably reproduced it would have minted more than a few billionaires and I am not aware of a single one other than maybe Jim Simons, but I doubt what they do could be reasonably called TA.
From the very little that has be known of what Jim Simons and his team have done, it seems to have included leverage. Once I learned that, I was less impressed.
dont_know_mind
Posts: 94
Joined: Sun Nov 22, 2020 7:14 am

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

Post by dont_know_mind »

OohLaLa wrote: Sat Apr 17, 2021 1:42 pm This is starting to veer off into some weird debates... almost strawman-like. Who here is looking to beat hedge funds at their own game? Who here thinks that HFEA is "beating the market"?

Am I, suddenly, the only one who got into this because they?:
- thought going in 100% into a stock index ETF seemed too concentrated
- wanted to take on more risk (i.e. not "beating the market") than an equivalent non-leveraged stock/ bond portfolo, while indeed expecting some additional reward
- did not intend to hold 3x leverage for decades, but instead have plans to de-leverage/ de-risk in certain ways
Sorry if you feel the comments are straw man trolling or not adding to the discussion.
I find the HFEA strategy and particularly UPRO very interesting.

I’m here to challenge my views.

I have an opposing position to the HFEA position.
I am levered risk assets but short 2 year fixed debt, long non-US, tilted away from tech, tilted towards value/resources/EM.

I always like to know who is on the other side of my position and why. Hopefully they both can do well.
Good luck.
dont_know_mind
Posts: 94
Joined: Sun Nov 22, 2020 7:14 am

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

Post by dont_know_mind »

tradri wrote: Sat Apr 17, 2021 12:03 pm
dont_know_mind wrote: Sat Apr 17, 2021 11:34 am
The direct leverage was probably 2-3. The 1.5 is what I am thinking of having margin LTV if I replace some of the non-callable debt with margin debt.

I put in 0.9 in equity and bought 2.5 in index fund (with 1.6 in debt).
The profit so far is 0.54. So a 60% return on invested equity since April 2020.
Has 70/30 UPRO/TMF returned 60% since April 2020 ? It's hard to compare as different entry points.

The main thing though for me is that I felt comfortable enough in my strategy to continue to add. I bought 0.9 in April 2020, 1.0 from May-July 20, 0.6 from Aug 20 to Feb 21.
70/30 UPRO/TMF returned more than 100% since April last year. https://www.portfoliovisualizer.com/bac ... tion2_1=30

And 70/30 UPRO/TYD did even better.
That is great and better than my returns in the last year.
It does seem to generate excess return. Where does this really come from? What is the underlying basis for it and why does it persist ?

Under what conditions do you think this would fail?

Maybe this has been discussed previously and please refer me to the previous comment if so.
My ramblings are probably rubbish, biased by my position and talking my own book. No-one should invest based on my views. Some days I wonder whether I've had some skill or just been a lucky gambler.
perfectuncertainty
Posts: 386
Joined: Sun Feb 04, 2018 6:44 pm

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

Post by perfectuncertainty »

dont_know_mind wrote: Sat Apr 17, 2021 11:43 am
perfectuncertainty wrote: Sat Apr 17, 2021 11:21 am
tradri wrote: Sat Apr 17, 2021 5:03 am
dont_know_mind wrote: Sat Apr 17, 2021 4:38 am Have you read the book “the man who solved the market” (Zuckerberg). The strategies discussed in this thread were explored by rentech and Bridgewater since the 80’s. From my interpretation of the book, Simons suggested that a large part of the returns they were generating were coming from retail - and not buy and hold Bogleheads (Simonds speculated “dentists” in the 90’s, in the 2020’s maybe Hedgefundies ?).

I think it is naive to think that technical analysis or the backtesting suggested here can generate alpha. What you are doing is 30 years behind the hedge funds. The only reason I can think of that this would generate a better than market return is the extra beta taken (through leverage) or luck.

I don’t see how doing the statistical analysis proposed or sticking it in portfolio visualiser has any hope of beating Rentech & others over time. It’s sort of delusional. What do you think your edge is with the strategy you propose?
I agree that trying to beat the professionals, that are doing this full-time and with much more resources, is pointless.

By itself, these leveraged ETFs don't generate a higher return. Over long periods of time, they perform about the same as their unlevered index. While they give you 3x the (daily) beta, this extra volatility isn't rewarded by itself.

The only reason this works, in my opinion, is because it is combined with a leveraged, uncorrelated asset, and rebalanced from time to time. This creates a natural market-timing effect.
You don't think it’s a little arrogant to call those engaging in technical analysis or other strategies naive? Many have beaten the market quite radically and with controlled risk. They just aren't using the strategies that are described here. No matter what approach you use, even purge BH Stock/Bonds AA, you are lying on probability but not on certainty.
Yes, I think you are naive if you think you can generate excess returns from technical analysis or looking up things on portfolio visualiser. You might get excess returns through luck though and think it was skill and blow up sometime in the future.
So yes to my question (arrogant). Please disregard statistics too then, since many technical studies have a statistical basis. Indexing can blow up pretty bad too - we saw it in 2008, 2018, and 2020. Are people who follow indexing naive too?
Last edited by perfectuncertainty on Sat Apr 17, 2021 9:29 pm, edited 1 time in total.
perfectuncertainty
Posts: 386
Joined: Sun Feb 04, 2018 6:44 pm

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

Post by perfectuncertainty »

tradri wrote: Sat Apr 17, 2021 11:56 am
perfectuncertainty wrote: Sat Apr 17, 2021 11:21 am You don't think it’s a little arrogant to call those engaging in technical analysis or other strategies naive? Many have beaten the market quite radically and with controlled risk. They just aren't using the strategies that are described here. No matter what approach you use, even purge BH Stock/Bonds AA, you are lying on probability but not on certainty.
Show me a fool-proof strategy to beat the market and I will gladly take it. :wink:
So according to your logic, the market can't be beat. Berkshire Hathaway performance all luck? And if you feel it can't be beat, why are you participating in HFEA? Just to be a naysayer?
perfectuncertainty
Posts: 386
Joined: Sun Feb 04, 2018 6:44 pm

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

Post by perfectuncertainty »

Semantics wrote: Sat Apr 17, 2021 5:20 pm
perfectuncertainty wrote: Sat Apr 17, 2021 11:21 am
tradri wrote: Sat Apr 17, 2021 5:03 am
dont_know_mind wrote: Sat Apr 17, 2021 4:38 am Have you read the book “the man who solved the market” (Zuckerberg). The strategies discussed in this thread were explored by rentech and Bridgewater since the 80’s. From my interpretation of the book, Simons suggested that a large part of the returns they were generating were coming from retail - and not buy and hold Bogleheads (Simonds speculated “dentists” in the 90’s, in the 2020’s maybe Hedgefundies ?).

I think it is naive to think that technical analysis or the backtesting suggested here can generate alpha. What you are doing is 30 years behind the hedge funds. The only reason I can think of that this would generate a better than market return is the extra beta taken (through leverage) or luck.

I don’t see how doing the statistical analysis proposed or sticking it in portfolio visualiser has any hope of beating Rentech & others over time. It’s sort of delusional. What do you think your edge is with the strategy you propose?
I agree that trying to beat the professionals, that are doing this full-time and with much more resources, is pointless.

By itself, these leveraged ETFs don't generate a higher return. Over long periods of time, they perform about the same as their unlevered index. While they give you 3x the (daily) beta, this extra volatility isn't rewarded by itself.

The only reason this works, in my opinion, is because it is combined with a leveraged, uncorrelated asset, and rebalanced from time to time. This creates a natural market-timing effect.
You don't think it’s a little arrogant to call those engaging in technical analysis or other strategies naive? Many have beaten the market quite radically and with controlled risk. They just aren't using the strategies that are described here. No matter what approach you use, even purge BH Stock/Bonds AA, you are lying on probability but not on certainty.
Other strategies maybe, but there are millions of people who research and employ TA. If it worked and could be reliably reproduced it would have minted more than a few billionaires and I am not aware of a single one other than maybe Jim Simons, but I doubt what they do could be reasonably called TA.
Simons uses some
Soros
Tudor Jones
Steve Cohen
Dailio uses technical analysis on occasion

I'm not arguing that technical analysis is some holy grail. I'm merely countering those who arrogantly state that it has no value and successful traders don't use it.
Marseille07
Posts: 16054
Joined: Fri Nov 06, 2020 12:41 pm

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

Post by Marseille07 »

perfectuncertainty wrote: Sat Apr 17, 2021 9:19 pm So yes to my question (arrogant). Please disregard statistics too then, since many technical studies have a statistical basis. Indexing can blow up pretty bad too - we saw it in 2008, 2018, and 2020. Are people who follow indexing naive too?
There is no point arguing with naysayers. If your TA underperforms, they say TA doesn't work. If your TA overperforms, they call you lucky. Just ignore them.

Calling someone lucky because they haven't blown up is nonsensical, anyway. By that logic, everyone's lucky...including Buffett, Simons, Dalio, Ackman, Munger and Wood.
taojaxx
Posts: 311
Joined: Wed Jul 18, 2012 8:25 pm

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

Post by taojaxx »

dont_know_mind wrote: Sat Apr 17, 2021 8:25 pm
tradri wrote: Sat Apr 17, 2021 12:03 pm
dont_know_mind wrote: Sat Apr 17, 2021 11:34 am
The direct leverage was probably 2-3. The 1.5 is what I am thinking of having margin LTV if I replace some of the non-callable debt with margin debt.

I put in 0.9 in equity and bought 2.5 in index fund (with 1.6 in debt).
The profit so far is 0.54. So a 60% return on invested equity since April 2020.
Has 70/30 UPRO/TMF returned 60% since April 2020 ? It's hard to compare as different entry points.

The main thing though for me is that I felt comfortable enough in my strategy to continue to add. I bought 0.9 in April 2020, 1.0 from May-July 20, 0.6 from Aug 20 to Feb 21.
70/30 UPRO/TMF returned more than 100% since April last year. https://www.portfoliovisualizer.com/bac ... tion2_1=30

And 70/30 UPRO/TYD did even better.
That is great and better than my returns in the last year.
It does seem to generate excess return. Where does this really come from? What is the underlying basis for it and why does it persist ?

Under what conditions do you think this would fail?

Maybe this has been discussed previously and please refer me to the previous comment if so.
I certainly don't think that dont_know_mind trolls or brings strawman arguments. Matter of fact I found his questioning healthy and challenging. The thread has been extending beyond proportions but few have questioned the HFEA tenets and he's one of the very few who does. All of us on board HFEA benefit from such challenges.
I think his approach is perfectly legit and works fine. It is also a workaround for rebalancing issues but it does in no way invalidate HFEA in my mind. It's not the other side of the HFEA trade, just a different trade. He has a leveraged position in one asset -total market with some tilts- and his is a fixed term leverage. Contrary to how he presents it, he is not any more short fixed income than we are: HFEA leverage is being short fixed income as well. We just apply it to capturing the ERP -as he does- only on US stocks. But we add a long position on long Treasuries as insurance against market crashes.
As to what can make HFEA crash, we all know it: stagflation with positive stocks/bonds correlation on the down side.
No one here thinks we reinvented the wheel and plans to beat Jim Simons -who, by the way, lately did poorly on his clients fund and splendidly on his employee only fund (kind of embarassing)-
Just trying to make good money and survive crashes. HFEA delivers under most foreseeable market conditions. And the asset mix can be adjusted for other environments (UPRO/UGL or something else). What matters is leverage and uncorrelated assets.
Better lucky than smart.
User avatar
cos
Posts: 506
Joined: Fri Aug 23, 2019 7:34 pm
Location: Boston
Contact:

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

Post by cos »

perfectuncertainty wrote: Sat Apr 17, 2021 9:20 pm
tradri wrote: Sat Apr 17, 2021 11:56 am Show me a fool-proof strategy to beat the market and I will gladly take it. :wink:
So according to your logic, the market can't be beat. Berkshire Hathaway performance all luck? And if you feel it can't be beat, why are you participating in HFEA? Just to be a naysayer?
I think you two are talking past each other. After accounting for established risk factors like market, value, size, and term, Berkshire Hathaway's unexplained alpha becomes statistically insignificant (regression). As with HFEA, Berkshire Hathaway achieves its outperformance not by "beating the market" but by taking on more risk than the market via factor tilts and leverage.

To be clear, I'm defining "unexplained alpha" as "outperformance not explained by common equity and fixed income factor models," and I'm defining "beating the market" as "outperforming the market on a risk-adjusted basis."
tradri
Posts: 606
Joined: Tue Mar 23, 2021 5:42 pm

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

Post by tradri »

dont_know_mind wrote: Sat Apr 17, 2021 8:25 pm
That is great and better than my returns in the last year.
It does seem to generate excess return. Where does this really come from? What is the underlying basis for it and why does it persist ?

Under what conditions do you think this would fail?

Maybe this has been discussed previously and please refer me to the previous comment if so.
Well, the stock market had an incredible rally since April last year, so of course any strategy that includes a 3x S&P 500 would have done very well.

But this isn't the point. The point is, that including an uncorrelated asset like treasuries allows you to rebalance into a stock market crash and benefit from the rally once the stock market recovers. The idea behind this is risk parity. When one asset fails the other one saves the day.

The market environment under which this would fail is stagflation, as both stocks and bonds do poorly in such an environment. Adding gold is a way to hedge against that, but after reading up on why stagflation happened in the 1970s, I am willing to take the bet that this won't happen again anytime soon. https://www.thebalance.com/what-is-stagflation-3305964

In the original HFEA post, he outlined many other concerns as well.
tradri
Posts: 606
Joined: Tue Mar 23, 2021 5:42 pm

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

Post by tradri »

perfectuncertainty wrote: Sat Apr 17, 2021 9:20 pm
So according to your logic, the market can't be beat. Berkshire Hathaway performance all luck? And if you feel it can't be beat, why are you participating in HFEA? Just to be a naysayer?
Thank you for bringing up Warren Buffet. As far as I know, he is as opposed to technical analysis/trading, as one can be.

From what I have heard, the reason for Buffet's outperformance was, that he discovered certain factors like value, decades before the academic literature caught up. In recent years he failed to outperform the market, because value also underperformed the market.

HFEA isn't beating the market. A 3x leveraged stock/bond portfolio has a worse risk-adjusted return than the unlevered stock/bond portfolio. One can only expect higher returns with much higher volatility.
perfectuncertainty
Posts: 386
Joined: Sun Feb 04, 2018 6:44 pm

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

Post by perfectuncertainty »

Marseille07 wrote: Sat Apr 17, 2021 9:40 pm
perfectuncertainty wrote: Sat Apr 17, 2021 9:19 pm So yes to my question (arrogant). Please disregard statistics too then, since many technical studies have a statistical basis. Indexing can blow up pretty bad too - we saw it in 2008, 2018, and 2020. Are people who follow indexing naive too?
There is no point arguing with naysayers. If your TA underperforms, they say TA doesn't work. If your TA overperforms, they call you lucky. Just ignore them.

Calling someone lucky because they haven't blown up is nonsensical, anyway. By that logic, everyone's lucky...including Buffett, Simons, Dalio, Ackman, Munger and Wood.
Thank you for your response. There are many approaches to investing / trading that aren't simply indexed Buy and Hold. Some involve TA, some don't. Some use charts to gauge sentiment, some use statistical averages to look at probability, some use market psychology, some use PE ranges. I actually think S&P indexing makes a ton of sense for the vast majority of investors. But those who scoff at others using different or blended approaches are arrogant.

For instance, I bought a large amount of QLD and COST on March 5th 2021. I paid $102.71 for QLD and $310.49 based on factors in charts. Collectively up 24.7% in 42 days.

This past week I traded SPX options actively but conservatively and made 7.6% (no losers). It's one the easiest ways to make money with options and with limited risk and high probability.

You will call me naive and/or lucky. I actually don't hang my ego on what anyone calls me at all. A lesson I learned a long time ago. I'm simply sharing that I do not regard something as some holy grail. From early 2000 to early 2013 the SP500 returned ZERO!! During that same time LTT didn’t fare too much better (from June 2003 to August 2011, TLT returned ZERO percent). Indexing sucked for a decade. Were the indexers and 60/40 BH naive? Just unlucky? How about those retiring in late 2002? Naive?

My point for disclosing the above is actually to point out that all of this, including HFEA, is an exercise in probability and risk management. There is no "free lunch" even in diversification - else you would allocate each allocate with equal percentages. If you change them you are trying to gain an edge based on history or on risk perception (black magic?). TLT or TMF can sink much lower and drag your return into the mud - so far 55/45 HFEA is up 4.75% YTD whereas the SP500 is up over 11%. 60/40 BH is up 2.52%. There are methods/systems/approaches that provide different probability and risk/return of capital appreciation. BH, 60/40, everything else is an exercise in probability and risk/return.
Marseille07
Posts: 16054
Joined: Fri Nov 06, 2020 12:41 pm

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

Post by Marseille07 »

perfectuncertainty wrote: Sun Apr 18, 2021 2:17 pm
Marseille07 wrote: Sat Apr 17, 2021 9:40 pm
perfectuncertainty wrote: Sat Apr 17, 2021 9:19 pm So yes to my question (arrogant). Please disregard statistics too then, since many technical studies have a statistical basis. Indexing can blow up pretty bad too - we saw it in 2008, 2018, and 2020. Are people who follow indexing naive too?
There is no point arguing with naysayers. If your TA underperforms, they say TA doesn't work. If your TA overperforms, they call you lucky. Just ignore them.

Calling someone lucky because they haven't blown up is nonsensical, anyway. By that logic, everyone's lucky...including Buffett, Simons, Dalio, Ackman, Munger and Wood.
Thank you for your response. There are many approaches to investing / trading that aren't simply indexed Buy and Hold. Some involve TA, some don't. Some use charts to gauge sentiment, some use statistical averages to look at probability, some use market psychology, some use PE ranges. I actually think S&P indexing makes a ton of sense for the vast majority of investors. But those who scoff at others using different or blended approaches are arrogant.

For instance, I bought a large amount of QLD and COST on March 5th 2021. I paid $102.71 for QLD and $310.49 based on factors in charts. Collectively up 24.7% in 42 days.

This past week I traded SPX options actively but conservatively and made 7.6% (no losers). It's one the easiest ways to make money with options and with limited risk and high probability.

You will call me naive and/or lucky. I actually don't hang my ego on what anyone calls me at all. A lesson I learned a long time ago. I'm simply sharing that I do not regard something as some holy grail. From early 2000 to early 2013 the SP500 returned ZERO!! During that same time LTT didn’t fare too much better (from June 2003 to August 2011, TLT returned ZERO percent). Indexing sucked for a decade. Were the indexers and 60/40 BH naive? Just unlucky? How about those retiring in late 2002? Naive?

My point for disclosing the above is actually to point out that all of this, including HFEA, is an exercise in probability and risk management. There is no "free lunch" even in diversification - else you would allocate each allocate with equal percentages. If you change them you are trying to gain an edge based on history or on risk perception (black magic?). TLT or TMF can sink much lower and drag your return into the mud - so far 55/45 HFEA is up 4.75% YTD whereas the SP500 is up over 11%. 60/40 BH is up 2.52%. There are methods/systems/approaches that provide different probability and risk/return of capital appreciation. BH, 60/40, everything else is an exercise in probability and risk/return.
You're right, there's no free lunch in leverage or diversification. As far as HFEA, the risk parity idea isn't a terrible one. However, the outperformance was coming from the 4-decade-long yields decline, which may have ended last August.
tradri
Posts: 606
Joined: Tue Mar 23, 2021 5:42 pm

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

Post by tradri »

I had a quick look at the PIMCO StocksPLUS® Long Duration Fund (PSLDX). Since the fund's inception, the performance has been relatively flat.

Comparing this fund to 50/50 UPRO/TMF rebalanced quarterly, PSLDX significantly underperformed.

What is the reason for this? Is it because PSLDX rebalances its stocks/bonds daily, and therefore the stocks and bonds don't have time to build up momentum before they are rebalanced?

P.S. My mistake. :oops: PSLDX uses 2x leverage, not 3x.
Last edited by tradri on Sun Apr 18, 2021 3:04 pm, edited 2 times in total.
User avatar
cos
Posts: 506
Joined: Fri Aug 23, 2019 7:34 pm
Location: Boston
Contact:

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

Post by cos »

tradri wrote: Sun Apr 18, 2021 2:58 pm I had a quick look at the PIMCO StocksPLUS® Long Duration Fund (PSLDX). Since the fund's inception, the performance has been relatively flat.

Comparing this fund to 50/50 UPRO/TMF rebalanced quarterly, PSLDX significantly underperformed.

What is the reason for this? Is it because PSLDX rebalances its stocks/bonds daily, and therefore the stocks and bonds don't have time to build up momentum before they are rebalanced?
What are you talking about? Keep in mind that PSLDX throws off most of its growth in the form of dividends. If you don't reinvest them in a backtest, you're going to get inaccurate results.
tradri
Posts: 606
Joined: Tue Mar 23, 2021 5:42 pm

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

Post by tradri »

cos wrote: Sun Apr 18, 2021 3:03 pm
tradri wrote: Sun Apr 18, 2021 2:58 pm I had a quick look at the PIMCO StocksPLUS® Long Duration Fund (PSLDX). Since the fund's inception, the performance has been relatively flat.

Comparing this fund to 50/50 UPRO/TMF rebalanced quarterly, PSLDX significantly underperformed.

What is the reason for this? Is it because PSLDX rebalances its stocks/bonds daily, and therefore the stocks and bonds don't have time to build up momentum before they are rebalanced?
What are you talking about? Keep in mind that PSLDX throws off most of its growth in the form of dividends. If you don't reinvest them in a backtest, you're going to get inaccurate results.
Already figured it out. PSLDX is 2x leveraged, not 3x. :oops:
tradri
Posts: 606
Joined: Tue Mar 23, 2021 5:42 pm

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

Post by tradri »

cos wrote: Sun Apr 18, 2021 3:03 pm Keep in mind that PSLDX throws off most of its growth in the form of dividends. If you don't reinvest them in a backtest, you're going to get inaccurate results.
Is this the reason why PSLDX looks so much worse on the chart from Google? https://www.google.com/search?q=psldx&o ... e&ie=UTF-8
User avatar
cos
Posts: 506
Joined: Fri Aug 23, 2019 7:34 pm
Location: Boston
Contact:

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

Post by cos »

tradri wrote: Sun Apr 18, 2021 3:19 pm
cos wrote: Sun Apr 18, 2021 3:03 pm Keep in mind that PSLDX throws off most of its growth in the form of dividends. If you don't reinvest them in a backtest, you're going to get inaccurate results.
Is this the reason why PSLDX looks so much worse on the chart from Google? https://www.google.com/search?q=psldx&o ... e&ie=UTF-8
Bingo.
occambogle
Posts: 953
Joined: Thu Dec 12, 2019 3:58 am

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

Post by occambogle »

I don’t own PSLDX but it’s what I use as a test to verify if any graphing site is actually showing re-invested dividends or not... with PSLDX it’s instantly obvious.
corp_sharecropper
Posts: 590
Joined: Thu Nov 07, 2013 1:36 pm

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

Post by corp_sharecropper »

tradri wrote: Sun Apr 18, 2021 3:19 pm
cos wrote: Sun Apr 18, 2021 3:03 pm Keep in mind that PSLDX throws off most of its growth in the form of dividends. If you don't reinvest them in a backtest, you're going to get inaccurate results.
Is this the reason why PSLDX looks so much worse on the chart from Google? https://www.google.com/search?q=psldx&o ... e&ie=UTF-8
Yes
Hydromod
Posts: 1052
Joined: Tue Mar 26, 2019 10:21 pm

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

Post by Hydromod »

perfectuncertainty wrote: Sun Apr 18, 2021 2:17 pm You will call me naive and/or lucky. I actually don't hang my ego on what anyone calls me at all. A lesson I learned a long time ago. I'm simply sharing that I do not regard something as some holy grail. From early 2000 to early 2013 the SP500 returned ZERO!! During that same time LTT didn’t fare too much better (from June 2003 to August 2011, TLT returned ZERO percent). Indexing sucked for a decade. Were the indexers and 60/40 BH naive? Just unlucky? How about those retiring in late 2002? Naive?

My point for disclosing the above is actually to point out that all of this, including HFEA, is an exercise in probability and risk management. There is no "free lunch" even in diversification - else you would allocate each allocate with equal percentages. If you change them you are trying to gain an edge based on history or on risk perception (black magic?). TLT or TMF can sink much lower and drag your return into the mud - so far 55/45 HFEA is up 4.75% YTD whereas the SP500 is up over 11%. 60/40 BH is up 2.52%. There are methods/systems/approaches that provide different probability and risk/return of capital appreciation. BH, 60/40, everything else is an exercise in probability and risk/return.
It appears that 2000 to 2013 would have been a pretty good time for the HFEA, even though UPRO and TMF would have sucked individually.

Assuming I've properly accounted for expenses etc., and my software is doing the calculations correctly, I calculate roughly 40% CAGR over that period (worst 3-yr rolling period briefly dipped below 20% CAGR to 17 or 18% maybe three times for a month or two at a clip) using monthly rebalancing with risk parity (assigning a risk fraction of 60% to UPRO). Almost certainly an overestimate that doesn't account for all expenses, but still a reasonably good return.

Sure, there are no guarantees and I don't think we'll see such returns going forward with the TMF tailwind fizzling out, but this suggests that there can indeed be a free lunch when money is sloshing back and forth between two funds, one has both funds, and one rebalances.
perfectuncertainty
Posts: 386
Joined: Sun Feb 04, 2018 6:44 pm

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

Post by perfectuncertainty »

Hydromod wrote: Sun Apr 18, 2021 6:27 pm
perfectuncertainty wrote: Sun Apr 18, 2021 2:17 pm You will call me naive and/or lucky. I actually don't hang my ego on what anyone calls me at all. A lesson I learned a long time ago. I'm simply sharing that I do not regard something as some holy grail. From early 2000 to early 2013 the SP500 returned ZERO!! During that same time LTT didn’t fare too much better (from June 2003 to August 2011, TLT returned ZERO percent). Indexing sucked for a decade. Were the indexers and 60/40 BH naive? Just unlucky? How about those retiring in late 2002? Naive?

My point for disclosing the above is actually to point out that all of this, including HFEA, is an exercise in probability and risk management. There is no "free lunch" even in diversification - else you would allocate each allocate with equal percentages. If you change them you are trying to gain an edge based on history or on risk perception (black magic?). TLT or TMF can sink much lower and drag your return into the mud - so far 55/45 HFEA is up 4.75% YTD whereas the SP500 is up over 11%. 60/40 BH is up 2.52%. There are methods/systems/approaches that provide different probability and risk/return of capital appreciation. BH, 60/40, everything else is an exercise in probability and risk/return.
It appears that 2000 to 2013 would have been a pretty good time for the HFEA, even though UPRO and TMF would have sucked individually.

Assuming I've properly accounted for expenses etc., and my software is doing the calculations correctly, I calculate roughly 40% CAGR over that period (worst 3-yr rolling period briefly dipped below 20% CAGR to 17 or 18% maybe three times for a month or two at a clip) using monthly rebalancing with risk parity (assigning a risk fraction of 60% to UPRO). Almost certainly an overestimate that doesn't account for all expenses, but still a reasonably good return.

Sure, there are no guarantees and I don't think we'll see such returns going forward with the TMF tailwind fizzling out, but this suggests that there can indeed be a free lunch when money is sloshing back and forth between two funds, one has both funds, and one rebalances.
March 2000 to March 2013 the 55/45 VIFNX/VUSTX had a CAGR of 6.17%. The SP500 had a return of 2.89%. HFEA would have been well below 18% at best. LTT returned 8.19%. LINK
tradri
Posts: 606
Joined: Tue Mar 23, 2021 5:42 pm

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

Post by tradri »

Hydromod wrote: Sun Apr 18, 2021 6:27 pm It appears that 2000 to 2013 would have been a pretty good time for the HFEA, even though UPRO and TMF would have sucked individually.

Assuming I've properly accounted for expenses etc., and my software is doing the calculations correctly, I calculate roughly 40% CAGR over that period (worst 3-yr rolling period briefly dipped below 20% CAGR to 17 or 18% maybe three times for a month or two at a clip) using monthly rebalancing with risk parity (assigning a risk fraction of 60% to UPRO). Almost certainly an overestimate that doesn't account for all expenses, but still a reasonably good return.

Sure, there are no guarantees and I don't think we'll see such returns going forward with the TMF tailwind fizzling out, but this suggests that there can indeed be a free lunch when money is sloshing back and forth between two funds, one has both funds, and one rebalances.
40% CAGR? Even during the strong bull run of the last 12 years HFEA didn't achieve a 40% CAGR.

According to the Simba backtesting spreadsheet the CAGR for HFEA was 12.86% from 2000 to 2013, which is still fantastic, but 40% would have been... :moneybag
Hydromod
Posts: 1052
Joined: Tue Mar 26, 2019 10:21 pm

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

Post by Hydromod »

It's certainly possible I'm doing something incorrect.

For begin 2000 to end 2013, I get
  • ~8% CAGR with SPY/TLT
  • ~18% with SSO/UBT
  • ~39% with UPRO/TMF
  • ~18% with UOPIX/UBT and ~25% with QLD/UBT (cross-checking the synthetic LETF implementation with NASDAQ)
I used the actual funds after they were available, and spliced in synthetic funds before.

The comparison between UOPIX/UBT and QLD/UBT has synthetic QLD before 2006 (UOPIX started 12/1997). This suggests that either the synthetic QLD LETF is quite optimistically constructed (which we suspected already) or the UOPIX implementation had quite a bit of slippage. If it's the first case, the synthetic UPRO LETF implementation will be quite optimistic too.

UPRO is synthetic before 6/2009, SSO before 6/2006.
TLT is synthetic before 7/2002, UBT before 2010, TMF before 4/2009.

This doesn't have much trade slippage prior to the actual fund being available though.

I used a risk-budget form of inverse volatility with two-month lookback, with the equities given 61% and treasuries 39% of the volatility budget (straight inverse volatility has a 50/50 split).

This risk budget is more aggressive than straight inverse volatility (which you can do with rolling optimization in Portfolio Visualizer, although with a minimum of 12-month lookback).

The weights range from 15/85 equity/treasury to 75/25 over time with the given risk budget, rather than staying fixed. Really not very practical outside of tax-protected space.

For comparison, the calculations from 2010-present with the UPRO/TMF sequence (with actual LETFs) results in about 31% CAGR with the same risk budget approach.

I've been checking the implementation against the efficient frontier calculation in PV, and it seems to give weights that are on or near the frontier (although some of the funds have different weights in PV) and close to the minimum volatility point even with several funds.

But the point of this is not so much what the CAGR would have been, it's that there has been real synergy with the equity/treasury working together even when they aren't necessarily doing so well separately.
perfectuncertainty
Posts: 386
Joined: Sun Feb 04, 2018 6:44 pm

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

Post by perfectuncertainty »

Hydromod wrote: Sun Apr 18, 2021 8:56 pm It's certainly possible I'm doing something incorrect.

For begin 2000 to end 2013, I get
  • ~8% CAGR with SPY/TLT
  • ~18% with SSO/UBT
  • ~39% with UPRO/TMF
  • ~18% with UOPIX/UBT and ~25% with QLD/UBT (cross-checking the synthetic LETF implementation with NASDAQ)
I used the actual funds after they were available, and spliced in synthetic funds before.

The comparison between UOPIX/UBT and QLD/UBT has synthetic QLD before 2006 (UOPIX started 12/1997). This suggests that either the synthetic QLD LETF is quite optimistically constructed (which we suspected already) or the UOPIX implementation had quite a bit of slippage. If it's the first case, the synthetic UPRO LETF implementation will be quite optimistic too.

UPRO is synthetic before 6/2009, SSO before 6/2006.
TLT is synthetic before 7/2002, UBT before 2010, TMF before 4/2009.

This doesn't have much trade slippage prior to the actual fund being available though.

I used a risk-budget form of inverse volatility with two-month lookback, with the equities given 61% and treasuries 39% of the volatility budget (straight inverse volatility has a 50/50 split).

This risk budget is more aggressive than straight inverse volatility (which you can do with rolling optimization in Portfolio Visualizer, although with a minimum of 12-month lookback).

The weights range from 15/85 equity/treasury to 75/25 over time with the given risk budget, rather than staying fixed. Really not very practical outside of tax-protected space.

For comparison, the calculations from 2010-present with the UPRO/TMF sequence (with actual LETFs) results in about 31% CAGR with the same risk budget approach.

I've been checking the implementation against the efficient frontier calculation in PV, and it seems to give weights that are on or near the frontier (although some of the funds have different weights in PV) and close to the minimum volatility point even with several funds.

But the point of this is not so much what the CAGR would have been, it's that there has been real synergy with the equity/treasury working together even when they aren't necessarily doing so well separately.
1. You aren't using the dates I posted so I'm, not sure what you are trying to get at with this. The dates are March 2000 to March 2013 (13 years exactly).
2. At best the LETF combo will achieve 3x the unlevered combo.
3. During the period in question (March 2000 to March 2013), 55/45 VIFNX/VUSTX had a CAGR of 6.17%. So it makes no sense that 55/45 UPRO/TMF would perform better than 18.5%. No statistics needed.
Hydromod
Posts: 1052
Joined: Tue Mar 26, 2019 10:21 pm

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

Post by Hydromod »

perfectuncertainty wrote: Sun Apr 18, 2021 10:38 pm
Hydromod wrote: Sun Apr 18, 2021 8:56 pm It's certainly possible I'm doing something incorrect.

For begin 2000 to end 2013, I get
  • ~8% CAGR with SPY/TLT
  • ~18% with SSO/UBT
  • ~39% with UPRO/TMF
  • ~18% with UOPIX/UBT and ~25% with QLD/UBT (cross-checking the synthetic LETF implementation with NASDAQ)
I used the actual funds after they were available, and spliced in synthetic funds before.

The comparison between UOPIX/UBT and QLD/UBT has synthetic QLD before 2006 (UOPIX started 12/1997). This suggests that either the synthetic QLD LETF is quite optimistically constructed (which we suspected already) or the UOPIX implementation had quite a bit of slippage. If it's the first case, the synthetic UPRO LETF implementation will be quite optimistic too.

UPRO is synthetic before 6/2009, SSO before 6/2006.
TLT is synthetic before 7/2002, UBT before 2010, TMF before 4/2009.

This doesn't have much trade slippage prior to the actual fund being available though.

I used a risk-budget form of inverse volatility with two-month lookback, with the equities given 61% and treasuries 39% of the volatility budget (straight inverse volatility has a 50/50 split).

This risk budget is more aggressive than straight inverse volatility (which you can do with rolling optimization in Portfolio Visualizer, although with a minimum of 12-month lookback).

The weights range from 15/85 equity/treasury to 75/25 over time with the given risk budget, rather than staying fixed. Really not very practical outside of tax-protected space.

For comparison, the calculations from 2010-present with the UPRO/TMF sequence (with actual LETFs) results in about 31% CAGR with the same risk budget approach.

I've been checking the implementation against the efficient frontier calculation in PV, and it seems to give weights that are on or near the frontier (although some of the funds have different weights in PV) and close to the minimum volatility point even with several funds.

But the point of this is not so much what the CAGR would have been, it's that there has been real synergy with the equity/treasury working together even when they aren't necessarily doing so well separately.
1. You aren't using the dates I posted so I'm, not sure what you are trying to get at with this. The dates are March 2000 to March 2013 (13 years exactly).
2. At best the LETF combo will achieve 3x the unlevered combo.
3. During the period in question (March 2000 to March 2013), 55/45 VIFNX/VUSTX had a CAGR of 6.17%. So it makes no sense that 55/45 UPRO/TMF would perform better than 18.5%. No statistics needed.
I’m just picking roughly off a graph, estimates are only good to about 2 decimal places. The curve is pretty smooth but it won’t be quite accurate. I’m consistently picking the dates I reported, even if they don’t exactly match your dates. And I’m pointing out that there is error in the synthetic LETF, likely to be optimistic, so the numbers are likely overestimated. I’m trying to match your dates as best as I can.

And I don’t disagree with the premise that 55/45 UPRO/TMF won’t do much better than 3 times VFINX/VUSTX, give or take a little rebalancing momentum.

All of my descriptions above said that I wasn’t calculating 55/45 though. You can do better if you are overweight in UPRO when UPRO is doing well and overweight in TMF when TMF is doing well. That is what the risk budget thing ended up doing in automated calculations. I’ve found before that it tends to do noticeably better than a fixed proportion comparing the same funds head to head as well, instead of 1x vs. 2x vs. 3x.

I’m impressed that the algorithm does well, so I’m sharing the details, but my main point is to illustrate that the combination of the two assets can do better than the two assets separately. It seemed that you were trying to say that wasn’t possible, so I gave a counter example. Of course it won’t always be the case.
Ramjet
Posts: 1464
Joined: Thu Feb 06, 2020 10:45 am

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

Post by Ramjet »

tradri wrote: Sat Apr 17, 2021 6:59 am
dont_know_mind wrote: Sat Apr 17, 2021 6:33 am For an uncorrelated asset, at current interest rates that would appear to be cash. Other options would be a resource company index (Vale/BHP/other profitable miners). Or a country index with heavy resource tilt (UK, Canada, Australia). Gold, like oil has idiosyncratic risk. I could imagine situations where inflation and resource prices in general increase but gold does not (eg a central bank decides to sell its gold reserves). Similarly with oil or any one specific resource (eg oil if oil use has peaked).

I think my only edge as a retail investor is that I can employ leverage when institutions are forced to liquidate. The only time I take on leverage is when TSHTF and I spend the next 5-7 years paying that down.

I would not use the HFEA as:
A) too popular
B) having significant amount of net worth in black box, untested UPRO (in that it has not been through any systemic crisis where counterparty risk was stress tested) is to me nutty.

Cleanest application of leverage in my mind is to apply non-callable debt to a vanilla index fund. Then there are no moving parts other than a) index fund expense ratio, b) expected return of the index, c) cost of capital, d) exchange rate effects.

I think some non-US indexes were great value last year. I levered up 2:1 in April 20 and the carry was positive to the tune of 1% pa. I intend to pay the leverage off over the next 5-7 years. I don’t have any rebalancing hassles.

Like buying a chunk of real estate applying leverage in this way, the risks are obvious. If you buy when it is down 30% and it falls another 30%, you’re in the hole. But in another way the risk is controlled if you have a stable job and are youngish. If your leverage is 2-3X yearly income, bad case scenario, you end up paying the debt off over 5-7 years with nothing to show. Worst case scenario you can’t work and file for bankruptcy.

I like to think of leverage in a cashflow manner rather than the statistical one advocated in this thread. To me you are just leveraging your human capital based on your risk tolerance rather than finding some holy grail that works in all situations.

And yes, if you leverage you could lose! No way around that. The HFEA is not alchemy. The bond-stock correlation could kick you in the nuts. The whole triple levered long bonds (TMF) is to me like triple levering a pegged currency assuming it can’t break. I don’t do stuff that could kill me.

I favoured leveraging non US indexes with value and resources tilt last year precisely because it has not worked in the last 20 years and is unpopular. I would not be leveraging anything that is popular and the fact this is one of the most popular threads on bogleheads concerns me. As does the Bitcoin. I am allergic to both. Something being popular means to me the runway is probably not going to be long. I could be wrong. I figure I will still be ok with my investments if I am wrong as the imbedded expectations are low.
Personally, I feel more comfortable investing in a leveraged strategy where I I know that a) I can't lose more than I invested and b) I can't face a total loss..
Enormous plus of this strategy
perfectuncertainty
Posts: 386
Joined: Sun Feb 04, 2018 6:44 pm

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

Post by perfectuncertainty »

Hydromod wrote: Sun Apr 18, 2021 11:30 pm
perfectuncertainty wrote: Sun Apr 18, 2021 10:38 pm
Hydromod wrote: Sun Apr 18, 2021 8:56 pm It's certainly possible I'm doing something incorrect.

For begin 2000 to end 2013, I get
  • ~8% CAGR with SPY/TLT
  • ~18% with SSO/UBT
  • ~39% with UPRO/TMF
  • ~18% with UOPIX/UBT and ~25% with QLD/UBT (cross-checking the synthetic LETF implementation with NASDAQ)
I used the actual funds after they were available, and spliced in synthetic funds before.

The comparison between UOPIX/UBT and QLD/UBT has synthetic QLD before 2006 (UOPIX started 12/1997). This suggests that either the synthetic QLD LETF is quite optimistically constructed (which we suspected already) or the UOPIX implementation had quite a bit of slippage. If it's the first case, the synthetic UPRO LETF implementation will be quite optimistic too.

UPRO is synthetic before 6/2009, SSO before 6/2006.
TLT is synthetic before 7/2002, UBT before 2010, TMF before 4/2009.

This doesn't have much trade slippage prior to the actual fund being available though.

I used a risk-budget form of inverse volatility with two-month lookback, with the equities given 61% and treasuries 39% of the volatility budget (straight inverse volatility has a 50/50 split).

This risk budget is more aggressive than straight inverse volatility (which you can do with rolling optimization in Portfolio Visualizer, although with a minimum of 12-month lookback).

The weights range from 15/85 equity/treasury to 75/25 over time with the given risk budget, rather than staying fixed. Really not very practical outside of tax-protected space.

For comparison, the calculations from 2010-present with the UPRO/TMF sequence (with actual LETFs) results in about 31% CAGR with the same risk budget approach.

I've been checking the implementation against the efficient frontier calculation in PV, and it seems to give weights that are on or near the frontier (although some of the funds have different weights in PV) and close to the minimum volatility point even with several funds.

But the point of this is not so much what the CAGR would have been, it's that there has been real synergy with the equity/treasury working together even when they aren't necessarily doing so well separately.
1. You aren't using the dates I posted so I'm, not sure what you are trying to get at with this. The dates are March 2000 to March 2013 (13 years exactly).
2. At best the LETF combo will achieve 3x the unlevered combo.
3. During the period in question (March 2000 to March 2013), 55/45 VIFNX/VUSTX had a CAGR of 6.17%. So it makes no sense that 55/45 UPRO/TMF would perform better than 18.5%. No statistics needed.
I’m just picking roughly off a graph, estimates are only good to about 2 decimal places. The curve is pretty smooth but it won’t be quite accurate. I’m consistently picking the dates I reported, even if they don’t exactly match your dates. And I’m pointing out that there is error in the synthetic LETF, likely to be optimistic, so the numbers are likely overestimated. I’m trying to match your dates as best as I can.

And I don’t disagree with the premise that 55/45 UPRO/TMF won’t do much better than 3 times VFINX/VUSTX, give or take a little rebalancing momentum.

All of my descriptions above said that I wasn’t calculating 55/45 though. You can do better if you are overweight in UPRO when UPRO is doing well and overweight in TMF when TMF is doing well. That is what the risk budget thing ended up doing in automated calculations. I’ve found before that it tends to do noticeably better than a fixed proportion comparing the same funds head to head as well, instead of 1x vs. 2x vs. 3x.

I’m impressed that the algorithm does well, so I’m sharing the details, but my main point is to illustrate that the combination of the two assets can do better than the two assets separately. It seemed that you were trying to say that wasn’t possible, so I gave a counter example. Of course it won’t always be the case.
Thanks for clarifying. The reason I raised the 13 years of poor performance with Indexing is because those of us who don't religiously use buy and hold indexing were being called naive and other dumb terms. 13 years is one-third of a working life for a person.
Hydromod
Posts: 1052
Joined: Tue Mar 26, 2019 10:21 pm

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

Post by Hydromod »

Hydromod wrote: Sun Apr 18, 2021 8:56 pm For begin 2000 to end 2013, I get
  • ~8% CAGR with SPY/TLT
  • ~18% with SSO/UBT
  • ~39% with UPRO/TMF
  • ~18% with UOPIX/UBT and ~25% with QLD/UBT (cross-checking the synthetic LETF implementation with NASDAQ)
I plotted 13-yr CAGR for precision and looked at start of March to end of April, to give an idea of variability.

Over this period, the actual funds I used returned
  • 2 to 3% CAGR with SPY
  • 2.2 to 4.3% CAGR with SSO
  • 18.2 to 22.2% CAGR with UPRO (clearly extremely optimistic)
  • 8 to 9% CAGR with TLT
  • 19 to 21% CAGR with UBT (a bit optimistic)
  • 29 to 32% CAGR with TMF (more optimistic)
  • -3.5 to -0.5% CAGR with QQQ
  • -20 to -14% CAGR with UOPIX
  • -4.5 to 1.5% CAGR with QLD
  • 7 to 7.8% CAGR with SPY/TLT (risk budget)
  • 5.5 to 6.5% CAGR with SPY/TLT (55/45)
  • 16.2 to 18% with SSO/UBT (risk budget)
  • 13.2 to 15% with SSO/UBT (55/45)
  • 35.2 to 38.2% with UPRO/TMF (risk budget)
  • 31.2 to 34.2% with UPRO/TMF (55/45)
  • 7.7 to 9.0% with QQQ/TLT (risk budget)
  • 3 to 5.1% with QQQ/TLT (55/45)
  • 14.5 to 17.2% with UOPIX/UBT (risk budget)
  • 1.8 to 6.2% with UOPIX/UBT (55/45)
  • 19.9 to 22.5% with QLD/UBT (risk budget)
  • 12.1 to 16.5% with QLD/UBT (55/45)
This gives the information for head to head comparisons. All of these use a strict 21-day trade interval.

We see the issue with simulating LETFs likely being optimistic.

We see that the 55/45 gives values between the two funds.

We see that risk budget weighting gave returns higher than fixed 55/45 weighting for all combinations during this period, in some cases small and others large.

We see that risk budget weighting can give returns higher than either fund by itself, but not always.

I don't think that there is nearly enough information from historical volatility to trade risk-on/risk-off. It is just a small statistical bias that accumulates over time, even with slippage from trades (maybe not enough to overcome capital gains taxes though).

And this does support the claim that one can do better than buy-and-hold, at least incrementally.
User avatar
RovenSkyfall
Posts: 267
Joined: Wed Apr 01, 2020 11:40 am

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

Post by RovenSkyfall »

tradri wrote: Sun Apr 18, 2021 3:42 am
dont_know_mind wrote: Sat Apr 17, 2021 8:25 pm
That is great and better than my returns in the last year.
It does seem to generate excess return. Where does this really come from? What is the underlying basis for it and why does it persist ?

Under what conditions do you think this would fail?

Maybe this has been discussed previously and please refer me to the previous comment if so.
Well, the stock market had an incredible rally since April last year, so of course any strategy that includes a 3x S&P 500 would have done very well.

But this isn't the point. The point is, that including an uncorrelated asset like treasuries allows you to rebalance into a stock market crash and benefit from the rally once the stock market recovers. The idea behind this is risk parity. When one asset fails the other one saves the day.

The market environment under which this would fail is stagflation, as both stocks and bonds do poorly in such an environment. Adding gold is a way to hedge against that, but after reading up on why stagflation happened in the 1970s, I am willing to take the bet that this won't happen again anytime soon. https://www.thebalance.com/what-is-stagflation-3305964

In the original HFEA post, he outlined many other concerns as well.
It seems like this type of concave rebalancing is less ideal than a convex rebalancing strategy for something like HFEA. A paper better explaining can be found here.

Convex rebalancing (or Constant Proportion Portfolio Insurance - CPPI) is theorized to be better for constantly rising or constantly declining portfolios (which HFEA is intended to be the majority of the time), where as concave rebalancing (constant mix) tends to outperform in trendless markets that oscilate. I have a hard time calling the HFEA trendless, but maybe the volatility is too much for CPPI.
I saved my money, but it can't save me | The Chariot
tradri
Posts: 606
Joined: Tue Mar 23, 2021 5:42 pm

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

Post by tradri »

RovenSkyfall wrote: Mon Apr 19, 2021 2:58 pm
It seems like this type of concave rebalancing is less ideal than a convex rebalancing strategy for something like HFEA. A paper better explaining can be found here.

Convex rebalancing (or Constant Proportion Portfolio Insurance - CPPI) is theorized to be better for constantly rising or constantly declining portfolios (which HFEA is intended to be the majority of the time), where as concave rebalancing (constant mix) tends to outperform in trendless markets that oscilate. I have a hard time calling the HFEA trendless, but maybe the volatility is too much for CPPI.
So, the idea is that one overweights 3x stocks in a bull market, and 3x treasuries in a bear market, right?

Are there any convincing backtests for such a rebalancing strategy with LETFs?

P.S. Also, I wouldn't agree that a 3x stocks/treasuries portfolio is constantly declining. It experiences sudden drops when the stock market crashes, but I don't think this can be called a constantly declining portfolio.
tradri
Posts: 606
Joined: Tue Mar 23, 2021 5:42 pm

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

Post by tradri »

I'm not sure if this has been discussed before, but is there an optimal rebalancing day?

I have seen some conflicting articles online. Some say end of month is ideal, others say the first of the month is ideal...

Should one even worry about that, or simply rebalance on the 1st of every quarter for convenience?
DMoogle
Posts: 549
Joined: Sat Oct 31, 2020 10:24 pm

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

Post by DMoogle »

If you don't believe in market timing, then it doesn't matter.
User avatar
RovenSkyfall
Posts: 267
Joined: Wed Apr 01, 2020 11:40 am

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

Post by RovenSkyfall »

tradri wrote: Mon Apr 19, 2021 3:13 pm
RovenSkyfall wrote: Mon Apr 19, 2021 2:58 pm
It seems like this type of concave rebalancing is less ideal than a convex rebalancing strategy for something like HFEA. A paper better explaining can be found here.

Convex rebalancing (or Constant Proportion Portfolio Insurance - CPPI) is theorized to be better for constantly rising or constantly declining portfolios (which HFEA is intended to be the majority of the time), where as concave rebalancing (constant mix) tends to outperform in trendless markets that oscilate. I have a hard time calling the HFEA trendless, but maybe the volatility is too much for CPPI.
So, the idea is that one overweights 3x stocks in a bull market, and 3x treasuries in a bear market, right?

Are there any convincing backtests for such a rebalancing strategy with LETFs?

P.S. Also, I wouldn't agree that a 3x stocks/treasuries portfolio is constantly declining. It experiences sudden drops when the stock market crashes, but I don't think this can be called a constantly declining portfolio.
To answer your first question it’s not that simple. There is a floor, a cushion and a multiplier all of which determine your equity allocation. The equation is in the article but it maintains a minimum of the floor and as the value of the portfolio increases you add more to equities. As equities fall in value you sell. Opposite to a constant mix allocation.

I am not aware of any backtests for LETFs using a convex rebalancing strategy. I may ultimately see if I can figure out how to code for it in R, as I am also curious.

On average I think HFEA is increasing in value.
Last edited by RovenSkyfall on Mon Apr 19, 2021 5:16 pm, edited 1 time in total.
I saved my money, but it can't save me | The Chariot
tradri
Posts: 606
Joined: Tue Mar 23, 2021 5:42 pm

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

Post by tradri »

DMoogle wrote: Mon Apr 19, 2021 3:41 pm If you don't believe in market timing, then it doesn't matter.
True.
Post Reply