HEDGEFUNDIE's excellent adventure Part II: The next journey

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
MotoTrojan
Posts: 10778
Joined: Wed Feb 01, 2017 8:39 pm

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

Post by MotoTrojan »

305pelusa wrote: Wed Dec 18, 2019 7:25 pm
HEDGEFUNDIE wrote: Wed Dec 18, 2019 7:17 pm
305pelusa wrote: Wed Dec 18, 2019 11:50 am
MotoTrojan wrote: Wed Dec 18, 2019 11:35 am
As to the other point. A Lifecycle investor may hold 2x leverage early on, and eventually back off as they get closer to decumulation where they’ll be entirely unleveraged. My point is that this isn’t very different than a more typical “age in bonds” approach where someone holds nearly 100% equity early in life and moves towards bonds later on. Both investors follow a similar glide path but the lifecycle investor just leveraged that glidepath. Basically I’m saying most of us use or suggest a detuned lifecycle process, it’s not an uncommon method.
Actually, it is a very different allocation, and it leads to very different results. Historically, people who use true Lifecycle Investing manage to produce ~1.6 times final wealth than those who use conventional TDFs, with the same risk (uncertainty in final wealth). It's like getting paid the same, but in Euros (back when the Euro was 1.6 to the dollar). Every retirement cohort in the US since 1871 would've come out ahead. Same for studies in Britain and Japan. Lifecycle Investing is massively supported by theory and historical results. And it's a free lunch.

In fact, results from correct temporal diversification dwarf results from asset diversification.

IMO, it makes little sense to get into complicated Multi-factor investing and tilting, which might or might not produce benefits, when you could apply Lifecycle Investing, which is widely supported by literature, theory and practice.
Would it still be a very different allocation if you include a mid-six figure mortgage in the asset calculation for the "normal AA"-holding young person?
I believe so. Provided your cash flow is good, someone with a mortgage and a house, given all other finances equal, should probably invest similarly to someone who rents for his whole life. Your "rent money" could either go into rent, or a house purchase.

That's just a fancy way of saying that the "negative bond" of a large mortgage is more or less offset by the "positive bond" from a house that pays you dividends (in the form of not paying rent).

That's how I see it any ways.
The paper says this technique is meant for someone with stable income so I think that’s a reasonable conclusion.
User avatar
czeckers
Posts: 1049
Joined: Thu May 17, 2007 3:49 pm
Location: USA

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

Post by czeckers »

The approach of using leverage early on in life sounds great on paper, but is harder to do in real life.

The problem with leverage is that it can wipe you out if you face a crash early on. Look at past posts by market timer: viewtopic.php?t=5934. He did exactly that and got killed in the 2009 crash because of margin calls. In theory, he may have been fine had he been able to ride it out, but didn't have enough cash to cover the margin calls and ended up being forced to sell at a loss.

It's a long but fascinating thread and gives you a real feel for what leverage feels like in a down market.
The Espresso portfolio: | | 20% US TSM, 20% Small Value, 10% US REIT, 10% Dev Int'l, 10% EM, 10% Commodities, 20% Inter-term US Treas | | "A journey of a thousand miles begins with a single step."
UberGrub
Posts: 141
Joined: Wed Aug 21, 2019 3:47 pm

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

Post by UberGrub »

czeckers wrote: Wed Dec 18, 2019 7:58 pm The approach of using leverage early on in life sounds great on paper, but is harder to do in real life.

The problem with leverage is that it can wipe you out if you face a crash early on. Look at past posts by market timer: viewtopic.php?t=5934. He did exactly that and got killed in the 2009 crash because of margin calls. In theory, he may have been fine had he been able to ride it out, but didn't have enough cash to cover the margin calls and ended up being forced to sell at a loss.

It's a long but fascinating thread and gives you a real feel for what leverage feels like in a down market.
My understanding is that when backtested historically, wipeouts never occurred because you reconstitute your leverage monthly. And never has there been a 50% loss over one month. But even if a wipeout occurred, you just keep going, diversifying temporally. They tested higher margin requirements and higher leverage (3:1) where wipeouts did occur and it still came out ahead.

MT's and 305pelusa's approach of not rebalancing leverage monthly, however, could lead to trouble during a prolonged Bear market. And that's not what was backtested. That said, I think MT used far higher leverage and was not even contributing in savings. 305 seems to use lower leverage, and is frequently contributing. I think he really believes the approach is superior and is trying to implement it in a very careful and reasonable way. I think he's giving the approach a good name :mrgreen:
RayKeynes
Posts: 259
Joined: Mon Nov 11, 2019 2:14 am

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

Post by RayKeynes »

siamond wrote: Wed Dec 18, 2019 5:23 pm
RayKeynes wrote: Wed Dec 18, 2019 2:44 am @ HedgeFundie

There is LTT data available since 1928 (on a yearly basis by SIAMOND), which don't you try to simulate TMF from 1929? I would be very interested to see the backtesting results for TMF since 1929.

I know that there is no LIBOR data available since 1929, but you may take the average spread for UPRO (around 0.48%) and add it to the effective federal funds rate to simulate the borrowing costs?!
Actually, the LTT data in Simba before 1942 is VERY CRUDE. The trouble is there are no long-term rates from FRED until then (nor from other sources). So we had to solely rely on 10yrs rates to derive bond returns, which some view as somewhat long-term, but most would not agree. TMF is based on a 20+ years treasury index, as a case in point. Plus we only have annual data and no decent proxy for daily volatility by then.

Also the EFFR only goes back to 1955, this is why we started the simulated model at this point in time. One could use T-Bills as a proxy, but we know this is a rather poor proxy and this would be adding even more noise to the model. Finally we know that during WW-II, the government played all sorts of games with treasuries, which lead to a dramatic drop in value of treasuries in real terms.

Frankly, I'd be happy to go farther back in time than 1955, there is always a lot to learn from history, but in this case, it definitely sounds like garbage in, garbage out...
If there is data available which is faily accurate since 1955 - why don't we backtest from 1955 onwards to estimate TMF returns? I would be VERY interested to see how TMF would have performed from 1955 - 1980 and it would definitely change my view, as from 1955 - 1980 interest rates experienced a sharp rise and I want to make sure that such a scenario is also reflected in a backtest.

If TMF does perform rather "OK" in such a scenario as well, I'd consider to invest a small amount in such a strategy.
MotoTrojan
Posts: 10778
Joined: Wed Feb 01, 2017 8:39 pm

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

Post by MotoTrojan »

RayKeynes wrote: Thu Dec 19, 2019 3:11 am
siamond wrote: Wed Dec 18, 2019 5:23 pm
RayKeynes wrote: Wed Dec 18, 2019 2:44 am @ HedgeFundie

There is LTT data available since 1928 (on a yearly basis by SIAMOND), which don't you try to simulate TMF from 1929? I would be very interested to see the backtesting results for TMF since 1929.

I know that there is no LIBOR data available since 1929, but you may take the average spread for UPRO (around 0.48%) and add it to the effective federal funds rate to simulate the borrowing costs?!
Actually, the LTT data in Simba before 1942 is VERY CRUDE. The trouble is there are no long-term rates from FRED until then (nor from other sources). So we had to solely rely on 10yrs rates to derive bond returns, which some view as somewhat long-term, but most would not agree. TMF is based on a 20+ years treasury index, as a case in point. Plus we only have annual data and no decent proxy for daily volatility by then.

Also the EFFR only goes back to 1955, this is why we started the simulated model at this point in time. One could use T-Bills as a proxy, but we know this is a rather poor proxy and this would be adding even more noise to the model. Finally we know that during WW-II, the government played all sorts of games with treasuries, which lead to a dramatic drop in value of treasuries in real terms.

Frankly, I'd be happy to go farther back in time than 1955, there is always a lot to learn from history, but in this case, it definitely sounds like garbage in, garbage out...
If there is data available which is faily accurate since 1955 - why don't we backtest from 1955 onwards to estimate TMF returns? I would be VERY interested to see how TMF would have performed from 1955 - 1980 and it would definitely change my view, as from 1955 - 1980 interest rates experienced a sharp rise and I want to make sure that such a scenario is also reflected in a backtest.

If TMF does perform rather "OK" in such a scenario as well, I'd consider to invest a small amount in such a strategy.
Have you not seen the OP with TMF and TMF/UPRO data from 1955-2018? TMF was pummeled.
RayKeynes
Posts: 259
Joined: Mon Nov 11, 2019 2:14 am

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

Post by RayKeynes »

MotoTrojan wrote: Thu Dec 19, 2019 7:59 am
RayKeynes wrote: Thu Dec 19, 2019 3:11 am
siamond wrote: Wed Dec 18, 2019 5:23 pm
RayKeynes wrote: Wed Dec 18, 2019 2:44 am @ HedgeFundie

There is LTT data available since 1928 (on a yearly basis by SIAMOND), which don't you try to simulate TMF from 1929? I would be very interested to see the backtesting results for TMF since 1929.

I know that there is no LIBOR data available since 1929, but you may take the average spread for UPRO (around 0.48%) and add it to the effective federal funds rate to simulate the borrowing costs?!
Actually, the LTT data in Simba before 1942 is VERY CRUDE. The trouble is there are no long-term rates from FRED until then (nor from other sources). So we had to solely rely on 10yrs rates to derive bond returns, which some view as somewhat long-term, but most would not agree. TMF is based on a 20+ years treasury index, as a case in point. Plus we only have annual data and no decent proxy for daily volatility by then.

Also the EFFR only goes back to 1955, this is why we started the simulated model at this point in time. One could use T-Bills as a proxy, but we know this is a rather poor proxy and this would be adding even more noise to the model. Finally we know that during WW-II, the government played all sorts of games with treasuries, which lead to a dramatic drop in value of treasuries in real terms.

Frankly, I'd be happy to go farther back in time than 1955, there is always a lot to learn from history, but in this case, it definitely sounds like garbage in, garbage out...
If there is data available which is faily accurate since 1955 - why don't we backtest from 1955 onwards to estimate TMF returns? I would be VERY interested to see how TMF would have performed from 1955 - 1980 and it would definitely change my view, as from 1955 - 1980 interest rates experienced a sharp rise and I want to make sure that such a scenario is also reflected in a backtest.

If TMF does perform rather "OK" in such a scenario as well, I'd consider to invest a small amount in such a strategy.
Have you not seen the OP with TMF and TMF/UPRO data from 1955-2018? TMF was pummeled.
Can you share the daily/monthly data for TMF from 1955 - 1985? I will then integrate it into my backtesting. Thank you
MotoTrojan
Posts: 10778
Joined: Wed Feb 01, 2017 8:39 pm

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

Post by MotoTrojan »

RayKeynes wrote: Thu Dec 19, 2019 8:02 am
MotoTrojan wrote: Thu Dec 19, 2019 7:59 am
RayKeynes wrote: Thu Dec 19, 2019 3:11 am
siamond wrote: Wed Dec 18, 2019 5:23 pm
RayKeynes wrote: Wed Dec 18, 2019 2:44 am @ HedgeFundie

There is LTT data available since 1928 (on a yearly basis by SIAMOND), which don't you try to simulate TMF from 1929? I would be very interested to see the backtesting results for TMF since 1929.

I know that there is no LIBOR data available since 1929, but you may take the average spread for UPRO (around 0.48%) and add it to the effective federal funds rate to simulate the borrowing costs?!
Actually, the LTT data in Simba before 1942 is VERY CRUDE. The trouble is there are no long-term rates from FRED until then (nor from other sources). So we had to solely rely on 10yrs rates to derive bond returns, which some view as somewhat long-term, but most would not agree. TMF is based on a 20+ years treasury index, as a case in point. Plus we only have annual data and no decent proxy for daily volatility by then.

Also the EFFR only goes back to 1955, this is why we started the simulated model at this point in time. One could use T-Bills as a proxy, but we know this is a rather poor proxy and this would be adding even more noise to the model. Finally we know that during WW-II, the government played all sorts of games with treasuries, which lead to a dramatic drop in value of treasuries in real terms.

Frankly, I'd be happy to go farther back in time than 1955, there is always a lot to learn from history, but in this case, it definitely sounds like garbage in, garbage out...
If there is data available which is faily accurate since 1955 - why don't we backtest from 1955 onwards to estimate TMF returns? I would be VERY interested to see how TMF would have performed from 1955 - 1980 and it would definitely change my view, as from 1955 - 1980 interest rates experienced a sharp rise and I want to make sure that such a scenario is also reflected in a backtest.

If TMF does perform rather "OK" in such a scenario as well, I'd consider to invest a small amount in such a strategy.
Have you not seen the OP with TMF and TMF/UPRO data from 1955-2018? TMF was pummeled.
Can you share the daily/monthly data for TMF from 1955 - 1985? I will then integrate it into my backtesting. Thank you
I don’t have it handy (was monthly) but the plots are there. It wasn’t pretty.
User avatar
czeckers
Posts: 1049
Joined: Thu May 17, 2007 3:49 pm
Location: USA

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

Post by czeckers »

I'm not sure back testing is very useful for anything other than seeing how a portfolio will perform under certain specific market conditions that occurred in the past. For example, testing a portfolio while looking at specific time periods of inflation or deflation.

The type of back tests you are talking about are heavily skewed by recent market performance which happens to be very favorable right now. If you end the back test say in 2010, you will get very different results. This is the big danger of back testing, it can convince you to chase recent performance. Right now, a back test will convince you to tilt your portfolio to large growth and that market risks are minimal. However, moving forward, large growth may just as well under-perform the rest of the market and risks may be quite higher. In 2010, the same back test would have convinced you that risks were terribly high and you should be in long-term bonds and that stocks were a terrible investment. In 2008, a back test would have convinced you that international diversification and small value were the best choices and there was no place for large or small growth in a portfolio. Anyone adjusting their portfolio based on back test results would have bought the latest outperforming asset class and subsequently missed out on the next outperforming asset class. Right now, using leverage to invest in US large cap stocks sounds great, but it's just chasing past performance. There is a probability that domestic large caps will underperform for the next decade.
The Espresso portfolio: | | 20% US TSM, 20% Small Value, 10% US REIT, 10% Dev Int'l, 10% EM, 10% Commodities, 20% Inter-term US Treas | | "A journey of a thousand miles begins with a single step."
dspencer
Posts: 225
Joined: Wed Jul 06, 2016 11:29 am

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

Post by dspencer »

305pelusa wrote: Wed Dec 18, 2019 1:08 pm
I would probably start with the paper:
https://papers.ssrn.com/sol3/papers.cfm ... id=1149340

I was sold just on the theory and logic of it. It's nice to see the backtest overwhelmingly favor it too.

The thread I started to document the process is (I hope) a good resource:
viewtopic.php?f=10&t=274390

Finally, they came out with a book. It's on Amazon:
https://www.amazon.com/Lifecycle-Invest ... 449&sr=8-3

If you sign up for Audible, you can use the credit to get the audio book for free. I liked it so much, I bought the book so they'd get some royalties. It has completely changed my investing mindset.
The theory/logic definitely makes sense to me assuming the cost of the leverage is reasonable. There are two main reservations I have:

1. Behavior/psychology - Many people believe they can "keep the faith" through a steep loss. In practice, not everyone can. It's sort of like statistics on divorce. Not many people say "til death do us part" while thinking "but really it's 50/50." Yet that turns out to be the case. What percentage of people who think "I can stomach a 70% loss without changing course" can actually do so? Is it possible to truly know in advance?

2. Limits of backtesting - We live in a rapidly changing economic environment. How relevant is a backtest starting in 1871? Recent tests are more useful but no amount of backtesting should ever be mistaken for a crystal ball. On the plus side, this strategy doesn't rely on any particular theory that seems likely to change with technology. Nor does it seem likely to fail without warning in a way that would spare Age-20 in bonds type investors.

The combination of the two is the most dangerous. The next crash is always "different" and surely there will be people saying that the entire financial system is collapsing.
User avatar
Steve Reading
Posts: 2839
Joined: Fri Nov 16, 2018 10:20 pm

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

Post by Steve Reading »

dspencer wrote: Thu Dec 19, 2019 12:24 pm
305pelusa wrote: Wed Dec 18, 2019 1:08 pm
I would probably start with the paper:
https://papers.ssrn.com/sol3/papers.cfm ... id=1149340

I was sold just on the theory and logic of it. It's nice to see the backtest overwhelmingly favor it too.

The thread I started to document the process is (I hope) a good resource:
viewtopic.php?f=10&t=274390

Finally, they came out with a book. It's on Amazon:
https://www.amazon.com/Lifecycle-Invest ... 449&sr=8-3

If you sign up for Audible, you can use the credit to get the audio book for free. I liked it so much, I bought the book so they'd get some royalties. It has completely changed my investing mindset.
The theory/logic definitely makes sense to me assuming the cost of the leverage is reasonable. There are two main reservations I have:

1. Behavior/psychology - Many people believe they can "keep the faith" through a steep loss. In practice, not everyone can. It's sort of like statistics on divorce. Not many people say "til death do us part" while thinking "but really it's 50/50." Yet that turns out to be the case. What percentage of people who think "I can stomach a 70% loss without changing course" can actually do so? Is it possible to truly know in advance?
That’s certainly true. I can’t argue with that. If you’re not committed to the plan, it won’t work.
dspencer wrote: Thu Dec 19, 2019 12:24 pm 2. Limits of backtesting - We live in a rapidly changing economic environment. How relevant is a backtest starting in 1871? Recent tests are more useful but no amount of backtesting should ever be mistaken for a crystal ball. On the plus side, this strategy doesn't rely on any particular theory that seems likely to change with technology. Nor does it seem likely to fail without warning in a way that would spare Age-20 in bonds type investors.
I honestly couldn’t care less about the backtest. The theory is mathematically and theoretical sound. It was developed by Merton and Samuelson decades ago. And if you read academic asset allocation literature (like Campbell’s Strategic Asset Allocation), the formulas are all there.

All that paper did was put the theory to the test. They tested in the US since 1871. In Britain. In Japan. They redid the simulations with increased borrowing costs, lower Equity Risk Premiums and higher volatility. It came out ahead every single time. Even when they put it against a Monte Carlos simulator, to see how it does against random data (not just past data) it keeps coming ahead.

But is it surprising? Not at all. It makes perfect sense. Regardless of how the future looks, you will, on average, come out ahead by spreading whatever lifetime exposure to stocks you’ll have across as many years as possible. It’s just diversification, from a temporal perspective.
"... 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
drock
Posts: 28
Joined: Sat Mar 23, 2019 4:47 pm

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

Post by drock »

So if diversification is a free lunch, and life-cycle investing style leverage is a free lunch would combining the two be a REALLY free lunch? Why not use a diverse set of things like buy futures on bonds, s&p index, nasdaq100, r2k, gold, sell 4-7 month out vix contracts, etc in combination? For extra credit why wouldn't I then reduce exposure/leverage when the spy is below its (insert an SMA length here)? As reddit would say...it literally couldn't go tits up right?
no simpler
Posts: 109
Joined: Sat Jul 13, 2019 4:54 pm

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

Post by no simpler »

305pelusa wrote: Thu Dec 19, 2019 12:43 pm
dspencer wrote: Thu Dec 19, 2019 12:24 pm
305pelusa wrote: Wed Dec 18, 2019 1:08 pm
I would probably start with the paper:
https://papers.ssrn.com/sol3/papers.cfm ... id=1149340

I was sold just on the theory and logic of it. It's nice to see the backtest overwhelmingly favor it too.

The thread I started to document the process is (I hope) a good resource:
viewtopic.php?f=10&t=274390

Finally, they came out with a book. It's on Amazon:
https://www.amazon.com/Lifecycle-Invest ... 449&sr=8-3

If you sign up for Audible, you can use the credit to get the audio book for free. I liked it so much, I bought the book so they'd get some royalties. It has completely changed my investing mindset.
The theory/logic definitely makes sense to me assuming the cost of the leverage is reasonable. There are two main reservations I have:

1. Behavior/psychology - Many people believe they can "keep the faith" through a steep loss. In practice, not everyone can. It's sort of like statistics on divorce. Not many people say "til death do us part" while thinking "but really it's 50/50." Yet that turns out to be the case. What percentage of people who think "I can stomach a 70% loss without changing course" can actually do so? Is it possible to truly know in advance?
That’s certainly true. I can’t argue with that. If you’re not committed to the plan, it won’t work.
dspencer wrote: Thu Dec 19, 2019 12:24 pm 2. Limits of backtesting - We live in a rapidly changing economic environment. How relevant is a backtest starting in 1871? Recent tests are more useful but no amount of backtesting should ever be mistaken for a crystal ball. On the plus side, this strategy doesn't rely on any particular theory that seems likely to change with technology. Nor does it seem likely to fail without warning in a way that would spare Age-20 in bonds type investors.
I honestly couldn’t care less about the backtest. The theory is mathematically and theoretical sound. It was developed by Merton and Samuelson decades ago. And if you read academic asset allocation literature (like Campbell’s Strategic Asset Allocation), the formulas are all there.

All that paper did was put the theory to the test. They tested in the US since 1871. In Britain. In Japan. They redid the simulations with increased borrowing costs, lower Equity Risk Premiums and higher volatility. It came out ahead every single time. Even when they put it against a Monte Carlos simulator, to see how it does against random data (not just past data) it keeps coming ahead.

But is it surprising? Not at all. It makes perfect sense. Regardless of how the future looks, you will, on average, come out ahead by spreading whatever lifetime exposure to stocks you’ll have across as many years as possible. It’s just diversification, from a temporal perspective.
I would agree that there are pretty strong reasons for temporal diversification. There is a parallel here to VC backed startups. A VC backed startup will take on a lot of risk when it is young and target very high levels of growth, often compressing % margins in order to take more market share. Why? Because the goal is to maximize the discount of all future cashflows in $, not optimize for a % margin. When the company reaches a more stable growth rate, it will focus on consistent profitable growth, but ideally it will be at such a large run rate that this amounts to a shitload of free cashflow in total $ terms.

One of the big insights of lifecycle investing is to think in terms of total dollars, which is what ultimately matters.
no simpler
Posts: 109
Joined: Sat Jul 13, 2019 4:54 pm

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

Post by no simpler »

drock wrote: Fri Dec 20, 2019 6:31 am So if diversification is a free lunch, and life-cycle investing style leverage is a free lunch would combining the two be a REALLY free lunch? Why not use a diverse set of things like buy futures on bonds, s&p index, nasdaq100, r2k, gold, sell 4-7 month out vix contracts, etc in combination? For extra credit why wouldn't I then reduce exposure/leverage when the spy is below its (insert an SMA length here)? As reddit would say...it literally couldn't go tits up right?
Yep, if there is an argument for temporal diversification, then there is definitely an argument for temporal shifts in asset allocation, whether based on total market cap weights of different asset classes or volatility based weights. A fixed allocation (60:40, etc) is just to reduce cognitive burden, but cannot possibly be the most efficient or optimal weighting scheme for a given point in time, since the optimal weights change continuously through time.
User avatar
Steve Reading
Posts: 2839
Joined: Fri Nov 16, 2018 10:20 pm

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

Post by Steve Reading »

drock wrote: Fri Dec 20, 2019 6:31 am So if diversification is a free lunch, and life-cycle investing style leverage is a free lunch would combining the two be a REALLY free lunch? Why not use a diverse set of things like buy futures on bonds, s&p index, nasdaq100, r2k, gold, sell 4-7 month out vix contracts, etc in combination? For extra credit why wouldn't I then reduce exposure/leverage when the spy is below its (insert an SMA length here)? As reddit would say...it literally couldn't go tits up right?
I couldn’t care less about gold or making a volatility stance or concentrating on a sector (nasdaq). Idk what r2k is. I don’t market time and I’m suspicious others can. If looking at the SMA actually helped, everyone would do it and the market would drop today. It’s not consistent with the random walk hypothesis (while asset and temporal diversification absolutely are). I don’t invest in fixed income because my salary is already a giant fixed income “bond” that pays me coupons every 2 weeks. That said, to the extent you will temporally diversity, you should also consider keeping your position well diversified. I am 60/40 in domestic/Int, I don’t skimp on my EM, and I have diversified into a few factors.
"... 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: 2839
Joined: Fri Nov 16, 2018 10:20 pm

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

Post by Steve Reading »

no simpler wrote: Fri Dec 20, 2019 9:42 am Yep, if there is an argument for temporal diversification, then there is definitely an argument for temporal shifts in asset allocation, whether based on total market cap weights of different asset classes or volatility based weights. A fixed allocation (60:40, etc) is just to reduce cognitive burden, but cannot possibly be the most efficient or optimal weighting scheme for a given point in time, since the optimal weights change continuously through time.
This isn’t necessarily true. That temporal diversification is recommended doesn’t mean tactical asset allocation is too.

Temporal diversification works because given a constant volatility and ERP, the optimal long term asset allocation is a fixed percent allocation. That’s because the long term risk-adjusted returns are maximized when you maximize short term. This is aka Myopic Portfolio choice. Samuelson decades ago proved investment horizon is irrelevant to the optimal portfolio.

Given that a constant % allocation is optimal, then a person should try to keep a constant % in stocks of their whole wealth through their life. If your salary is very risk less, that means leverage early on to offset that human capital.

That is ALL temporal diversification requires. It’s simple and straightforward.

To make an argument for tactical asset allocation, reducing leverage when expected volatility is high, ERP has changed or trying to guess what the optimal weight is, you need to take a step further and claim these things change and you can predict them better than a 50/50 chance.

This very well might be an argument you can make and I’m all ears to learn. But you cannot claim that “if there’s an argument for Lifecycle Investing, there’s one for shifting an allocation”. That one is recommended does not imply the other is too. The former is a theoretical certainty, the latter can only be proven via backtests to show it can reliably be done and can persist in the future. Very different.

Hope that makes sense
"... 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
Lee_WSP
Posts: 3987
Joined: Fri Apr 19, 2019 5:15 pm
Location: Arizona

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

Post by Lee_WSP »

no simpler wrote: Fri Dec 20, 2019 9:28 am
305pelusa wrote: Thu Dec 19, 2019 12:43 pm
dspencer wrote: Thu Dec 19, 2019 12:24 pm
305pelusa wrote: Wed Dec 18, 2019 1:08 pm
I would probably start with the paper:
https://papers.ssrn.com/sol3/papers.cfm ... id=1149340

I was sold just on the theory and logic of it. It's nice to see the backtest overwhelmingly favor it too.

The thread I started to document the process is (I hope) a good resource:
viewtopic.php?f=10&t=274390

Finally, they came out with a book. It's on Amazon:
https://www.amazon.com/Lifecycle-Invest ... 449&sr=8-3

If you sign up for Audible, you can use the credit to get the audio book for free. I liked it so much, I bought the book so they'd get some royalties. It has completely changed my investing mindset.
The theory/logic definitely makes sense to me assuming the cost of the leverage is reasonable. There are two main reservations I have:

1. Behavior/psychology - Many people believe they can "keep the faith" through a steep loss. In practice, not everyone can. It's sort of like statistics on divorce. Not many people say "til death do us part" while thinking "but really it's 50/50." Yet that turns out to be the case. What percentage of people who think "I can stomach a 70% loss without changing course" can actually do so? Is it possible to truly know in advance?
That’s certainly true. I can’t argue with that. If you’re not committed to the plan, it won’t work.
dspencer wrote: Thu Dec 19, 2019 12:24 pm 2. Limits of backtesting - We live in a rapidly changing economic environment. How relevant is a backtest starting in 1871? Recent tests are more useful but no amount of backtesting should ever be mistaken for a crystal ball. On the plus side, this strategy doesn't rely on any particular theory that seems likely to change with technology. Nor does it seem likely to fail without warning in a way that would spare Age-20 in bonds type investors.
I honestly couldn’t care less about the backtest. The theory is mathematically and theoretical sound. It was developed by Merton and Samuelson decades ago. And if you read academic asset allocation literature (like Campbell’s Strategic Asset Allocation), the formulas are all there.

All that paper did was put the theory to the test. They tested in the US since 1871. In Britain. In Japan. They redid the simulations with increased borrowing costs, lower Equity Risk Premiums and higher volatility. It came out ahead every single time. Even when they put it against a Monte Carlos simulator, to see how it does against random data (not just past data) it keeps coming ahead.

But is it surprising? Not at all. It makes perfect sense. Regardless of how the future looks, you will, on average, come out ahead by spreading whatever lifetime exposure to stocks you’ll have across as many years as possible. It’s just diversification, from a temporal perspective.
I would agree that there are pretty strong reasons for temporal diversification. There is a parallel here to VC backed startups. A VC backed startup will take on a lot of risk when it is young and target very high levels of growth, often compressing % margins in order to take more market share. Why? Because the goal is to maximize the discount of all future cashflows in $, not optimize for a % margin. When the company reaches a more stable growth rate, it will focus on consistent profitable growth, but ideally it will be at such a large run rate that this amounts to a shitload of free cashflow in total $ terms.

One of the big insights of lifecycle investing is to think in terms of total dollars, which is what ultimately matters.
That's a really spot on analogy. The party keeps on going as long as VC/investor keeps on pumping money in until either the venture runs out of credit or becomes solvent on its own.

A law review goes over the pitfalls of lifecycle investing pretty well. I guess even the authors acknowledge only maybe half the workforce could even implement it and of that small cohort even less have the discipline and financial means to stick with the strategy.

https://lawreview.vermontlaw.edu/wp-con ... ol.-36.pdf
User avatar
Steve Reading
Posts: 2839
Joined: Fri Nov 16, 2018 10:20 pm

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

Post by Steve Reading »

Lee_WSP wrote: Fri Dec 20, 2019 1:07 pm
no simpler wrote: Fri Dec 20, 2019 9:28 am
305pelusa wrote: Thu Dec 19, 2019 12:43 pm
dspencer wrote: Thu Dec 19, 2019 12:24 pm
305pelusa wrote: Wed Dec 18, 2019 1:08 pm
I would probably start with the paper:
https://papers.ssrn.com/sol3/papers.cfm ... id=1149340

I was sold just on the theory and logic of it. It's nice to see the backtest overwhelmingly favor it too.

The thread I started to document the process is (I hope) a good resource:
viewtopic.php?f=10&t=274390

Finally, they came out with a book. It's on Amazon:
https://www.amazon.com/Lifecycle-Invest ... 449&sr=8-3

If you sign up for Audible, you can use the credit to get the audio book for free. I liked it so much, I bought the book so they'd get some royalties. It has completely changed my investing mindset.
The theory/logic definitely makes sense to me assuming the cost of the leverage is reasonable. There are two main reservations I have:

1. Behavior/psychology - Many people believe they can "keep the faith" through a steep loss. In practice, not everyone can. It's sort of like statistics on divorce. Not many people say "til death do us part" while thinking "but really it's 50/50." Yet that turns out to be the case. What percentage of people who think "I can stomach a 70% loss without changing course" can actually do so? Is it possible to truly know in advance?
That’s certainly true. I can’t argue with that. If you’re not committed to the plan, it won’t work.
dspencer wrote: Thu Dec 19, 2019 12:24 pm 2. Limits of backtesting - We live in a rapidly changing economic environment. How relevant is a backtest starting in 1871? Recent tests are more useful but no amount of backtesting should ever be mistaken for a crystal ball. On the plus side, this strategy doesn't rely on any particular theory that seems likely to change with technology. Nor does it seem likely to fail without warning in a way that would spare Age-20 in bonds type investors.
I honestly couldn’t care less about the backtest. The theory is mathematically and theoretical sound. It was developed by Merton and Samuelson decades ago. And if you read academic asset allocation literature (like Campbell’s Strategic Asset Allocation), the formulas are all there.

All that paper did was put the theory to the test. They tested in the US since 1871. In Britain. In Japan. They redid the simulations with increased borrowing costs, lower Equity Risk Premiums and higher volatility. It came out ahead every single time. Even when they put it against a Monte Carlos simulator, to see how it does against random data (not just past data) it keeps coming ahead.

But is it surprising? Not at all. It makes perfect sense. Regardless of how the future looks, you will, on average, come out ahead by spreading whatever lifetime exposure to stocks you’ll have across as many years as possible. It’s just diversification, from a temporal perspective.
I would agree that there are pretty strong reasons for temporal diversification. There is a parallel here to VC backed startups. A VC backed startup will take on a lot of risk when it is young and target very high levels of growth, often compressing % margins in order to take more market share. Why? Because the goal is to maximize the discount of all future cashflows in $, not optimize for a % margin. When the company reaches a more stable growth rate, it will focus on consistent profitable growth, but ideally it will be at such a large run rate that this amounts to a shitload of free cashflow in total $ terms.

One of the big insights of lifecycle investing is to think in terms of total dollars, which is what ultimately matters.
That's a really spot on analogy. The party keeps on going as long as VC/investor keeps on pumping money in until either the venture runs out of credit or becomes solvent on its own.

A law review goes over the pitfalls of lifecycle investing pretty well. I guess even the authors acknowledge only maybe half the workforce could even implement it and of that small cohort even less have the discipline and financial means to stick with the strategy.

https://lawreview.vermontlaw.edu/wp-con ... ol.-36.pdf
Thanks for the link, I hadn’t seen it before. I really like and agree with it.

The one thing is that they mention the returns on stocks must be higher than the borrowing costs, otherwise it doesn’t work. That’s true but the strategy technically accounts for that. When you figure out tour allocation, you substract equity expected returns from expected borrowing costs. If that number is negative, it would tell you not to put anything in stocks as long as you carried that debt. So if you’re following the book correctly, you should never end up in this pitfall.

What if you’re already leveraged and borrowing costs go up? Just liquidate. They recommend everything in tax advantaged space any ways.

I agree strongly that it’s complicated. Perhaps some day mutual funds can automate it for everyone else. For now, it’s very DIY and I only support it in places like Bogleheads where I think people are willing to put the extra effort.
"... 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
Forester
Posts: 1835
Joined: Sat Jan 19, 2019 2:50 pm
Location: UK

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

Post by Forester »

Summer 2019 might be the "generational buying opportunity" 2009/10 period for ex-US assets. I'm reminded of Reddit posts where someone decided to put $10k into a leveraged ETF which they held onto until 2016-19 when they get cold feet.

RAFI estimate about 180% real return for ex-US equities in the 2020s vs 104% for US large cap.
no simpler
Posts: 109
Joined: Sat Jul 13, 2019 4:54 pm

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

Post by no simpler »

305pelusa wrote: Fri Dec 20, 2019 10:05 am
no simpler wrote: Fri Dec 20, 2019 9:42 am Yep, if there is an argument for temporal diversification, then there is definitely an argument for temporal shifts in asset allocation, whether based on total market cap weights of different asset classes or volatility based weights. A fixed allocation (60:40, etc) is just to reduce cognitive burden, but cannot possibly be the most efficient or optimal weighting scheme for a given point in time, since the optimal weights change continuously through time.
This isn’t necessarily true. That temporal diversification is recommended doesn’t mean tactical asset allocation is too.

Temporal diversification works because given a constant volatility and ERP, the optimal long term asset allocation is a fixed percent allocation. That’s because the long term risk-adjusted returns are maximized when you maximize short term. This is aka Myopic Portfolio choice. Samuelson decades ago proved investment horizon is irrelevant to the optimal portfolio.

Given that a constant % allocation is optimal, then a person should try to keep a constant % in stocks of their whole wealth through their life. If your salary is very risk less, that means leverage early on to offset that human capital.

That is ALL temporal diversification requires. It’s simple and straightforward.

To make an argument for tactical asset allocation, reducing leverage when expected volatility is high, ERP has changed or trying to guess what the optimal weight is, you need to take a step further and claim these things change and you can predict them better than a 50/50 chance.

This very well might be an argument you can make and I’m all ears to learn. But you cannot claim that “if there’s an argument for Lifecycle Investing, there’s one for shifting an allocation”. That one is recommended does not imply the other is too. The former is a theoretical certainty, the latter can only be proven via backtests to show it can reliably be done and can persist in the future. Very different.

Hope that makes sense
Can you share the Samuelson paper? There's a reason physicists tend to do a whole hell of a lot better than economists when it comes to investing...and I say this as someone who studied econ, not physics. There's a reason LTCM blew up and Rentech thrived. From doing time series analysis, I can say with absolute certainty that portfolio co-moments, and by extension the optimal portfolio, are NOT scale invariant. Did he use real data, or just a whole bunch of theoretical assumptions? Keep in mind there have been nobel prizes in economics awarded to folks for their work that assumes stock market returns are normally distributed, so treat even celebrated economists with suspicion.

My argument is simply that the temporal component of investing is generally ignored, or addressed with very sloppy heuristics.
langlands
Posts: 849
Joined: Wed Apr 03, 2019 10:05 pm

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

Post by langlands »

no simpler wrote: Fri Dec 20, 2019 5:40 pm
305pelusa wrote: Fri Dec 20, 2019 10:05 am
no simpler wrote: Fri Dec 20, 2019 9:42 am Yep, if there is an argument for temporal diversification, then there is definitely an argument for temporal shifts in asset allocation, whether based on total market cap weights of different asset classes or volatility based weights. A fixed allocation (60:40, etc) is just to reduce cognitive burden, but cannot possibly be the most efficient or optimal weighting scheme for a given point in time, since the optimal weights change continuously through time.
This isn’t necessarily true. That temporal diversification is recommended doesn’t mean tactical asset allocation is too.

Temporal diversification works because given a constant volatility and ERP, the optimal long term asset allocation is a fixed percent allocation. That’s because the long term risk-adjusted returns are maximized when you maximize short term. This is aka Myopic Portfolio choice. Samuelson decades ago proved investment horizon is irrelevant to the optimal portfolio.

Given that a constant % allocation is optimal, then a person should try to keep a constant % in stocks of their whole wealth through their life. If your salary is very risk less, that means leverage early on to offset that human capital.

That is ALL temporal diversification requires. It’s simple and straightforward.

To make an argument for tactical asset allocation, reducing leverage when expected volatility is high, ERP has changed or trying to guess what the optimal weight is, you need to take a step further and claim these things change and you can predict them better than a 50/50 chance.

This very well might be an argument you can make and I’m all ears to learn. But you cannot claim that “if there’s an argument for Lifecycle Investing, there’s one for shifting an allocation”. That one is recommended does not imply the other is too. The former is a theoretical certainty, the latter can only be proven via backtests to show it can reliably be done and can persist in the future. Very different.

Hope that makes sense
Can you share the Samuelson paper? There's a reason physicists tend to do a whole hell of a lot better than economists when it comes to investing...and I say this as someone who studied econ, not physics. There's a reason LTCM blew up and Rentech thrived. From doing time series analysis, I can say with absolute certainty that portfolio co-moments, and by extension the optimal portfolio, are NOT scale invariant. Did he use real data, or just a whole bunch of theoretical assumptions? Keep in mind there have been nobel prizes in economics awarded to folks for their work that assumes stock market returns are normally distributed, so treat even celebrated economists with suspicion.

My argument is simply that the temporal component of investing is generally ignored, or addressed with very sloppy heuristics.
I'm pretty sure he's referring to Samuelson's Fallacy of Large Numbers. Unfortunately, the result makes very strong assumptions on the utility function. Ross in the paper https://www.cambridge.org/core/journal ... 2B27CE740 makes a pretty strong case that for most reasonable utility functions, Samuelson's argument doesn't apply.
User avatar
Steve Reading
Posts: 2839
Joined: Fri Nov 16, 2018 10:20 pm

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

Post by Steve Reading »

no simpler wrote: Fri Dec 20, 2019 5:40 pm
305pelusa wrote: Fri Dec 20, 2019 10:05 am
no simpler wrote: Fri Dec 20, 2019 9:42 am Yep, if there is an argument for temporal diversification, then there is definitely an argument for temporal shifts in asset allocation, whether based on total market cap weights of different asset classes or volatility based weights. A fixed allocation (60:40, etc) is just to reduce cognitive burden, but cannot possibly be the most efficient or optimal weighting scheme for a given point in time, since the optimal weights change continuously through time.
This isn’t necessarily true. That temporal diversification is recommended doesn’t mean tactical asset allocation is too.

Temporal diversification works because given a constant volatility and ERP, the optimal long term asset allocation is a fixed percent allocation. That’s because the long term risk-adjusted returns are maximized when you maximize short term. This is aka Myopic Portfolio choice. Samuelson decades ago proved investment horizon is irrelevant to the optimal portfolio.

Given that a constant % allocation is optimal, then a person should try to keep a constant % in stocks of their whole wealth through their life. If your salary is very risk less, that means leverage early on to offset that human capital.

That is ALL temporal diversification requires. It’s simple and straightforward.

To make an argument for tactical asset allocation, reducing leverage when expected volatility is high, ERP has changed or trying to guess what the optimal weight is, you need to take a step further and claim these things change and you can predict them better than a 50/50 chance.

This very well might be an argument you can make and I’m all ears to learn. But you cannot claim that “if there’s an argument for Lifecycle Investing, there’s one for shifting an allocation”. That one is recommended does not imply the other is too. The former is a theoretical certainty, the latter can only be proven via backtests to show it can reliably be done and can persist in the future. Very different.

Hope that makes sense
Can you share the Samuelson paper? There's a reason physicists tend to do a whole hell of a lot better than economists when it comes to investing...and I say this as someone who studied econ, not physics. There's a reason LTCM blew up and Rentech thrived. From doing time series analysis, I can say with absolute certainty that portfolio co-moments, and by extension the optimal portfolio, are NOT scale invariant. Did he use real data, or just a whole bunch of theoretical assumptions? Keep in mind there have been nobel prizes in economics awarded to folks for their work that assumes stock market returns are normally distributed, so treat even celebrated economists with suspicion.

My argument is simply that the temporal component of investing is generally ignored, or addressed with very sloppy heuristics.
I listed his assumptions already. Given that volatility and ERP are constant, with normal distributions, the optimal asset allocation for a given utility function is a constant % allocation:
http://www.wiwi.uni-muenster.de/05/down ... on1969.pdf

This is a theoretical result. It depends on no data (and hence, suffers from no data mining). It forms the basis for Lifecycle Investing.

If you’re unconvinced by the assumptions, disagree with them or whatever else, that’s cool. But you cannot sit here and claim that Samuelson’s results (and Lifecycle Investing), which are mathematical facts given some assumptions, support the idea of tactical asset allocation. They absolutely do not.

If you want to prove some argument for tactical asset allocation then, like I already said, you need to take additional steps. You know to show that ERP does vary and you can somehow predict it to a reasonable extent. Ditto for volatility. These are, by definition, data proofs; you’d need to show it worked on data and then make a further proof that it can continue in the future at a given confidence interval. Similar to factors. This is completely different than Samuelson’s result.

Personally, I know volatility and ERP vary but I do not believe I can predict it better than a coin flip. So I picked some long term averages, made an allocation using Samuelson’s equation, and just bought/hold/followed the course from them.

In this way, I’m just a regular BH who believes they cannot tactically shift allocations profitably and that the market is a random walk. I’m just using leverage to more evenly spread the stock risk I’m planning on taking over decades.
"... 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
get_g0ing
Posts: 750
Joined: Sat Dec 09, 2017 11:09 am

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

Post by get_g0ing »

So in M1 Finance, what's a good way to keep track of the actual return if you are also making contributions?
I mean the initial amount grows to a certain value and M1 shows you the return - so far so good - but if you make additional contributions, M1 also includes that in the gain/return.
MotoTrojan
Posts: 10778
Joined: Wed Feb 01, 2017 8:39 pm

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

Post by MotoTrojan »

get_g0ing wrote: Sat Dec 21, 2019 12:24 pm So in M1 Finance, what's a good way to keep track of the actual return if you are also making contributions?
I mean the initial amount grows to a certain value and M1 shows you the return - so far so good - but if you make additional contributions, M1 also includes that in the gain/return.
XIRR, google sheets/finance is how I did it. Allowed me to compare to a total return S&P500 index too, for sand contribution history.
assetalloc
Posts: 74
Joined: Tue Nov 13, 2018 8:45 pm

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

Post by assetalloc »

RayKeynes wrote: Thu Dec 19, 2019 3:11 am
If there is data available which is faily accurate since 1955 - why don't we backtest from 1955 onwards to estimate TMF returns? I would be VERY interested to see how TMF would have performed from 1955 - 1980 and it would definitely change my view, as from 1955 - 1980 interest rates experienced a sharp rise and I want to make sure that such a scenario is also reflected in a backtest.

If TMF does perform rather "OK" in such a scenario as well, I'd consider to invest a small amount in such a strategy.
I hate to admit, but I have a speculative theory that keeps me worried:
I divide recent financial markets in 2 chunks, one Pre-Allan greenspan/Clinton & one Post. The reason I do that is that I believe that in recent history, US government started to artificially control financial markets through FED. When I apply this theory, I see very different returns before 2000, and 2000-2019.

2000, it was irrational exuberance due to interest rates manipulation blamed on tech industry
2008, it was the housing crisis,
2021, it could be Bush/Bernanke/Obama $10T printing ..... We have HUGE debt bubble growing exponentially with leveraged instruments

The crash would have come sooner, but POTUS is doing well to keep pumping more money in the pot: Tax cut, Space Force, Oil???
langlands
Posts: 849
Joined: Wed Apr 03, 2019 10:05 pm

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

Post by langlands »

305pelusa wrote: Sat Dec 21, 2019 12:03 pm
no simpler wrote: Fri Dec 20, 2019 5:40 pm
305pelusa wrote: Fri Dec 20, 2019 10:05 am
no simpler wrote: Fri Dec 20, 2019 9:42 am Yep, if there is an argument for temporal diversification, then there is definitely an argument for temporal shifts in asset allocation, whether based on total market cap weights of different asset classes or volatility based weights. A fixed allocation (60:40, etc) is just to reduce cognitive burden, but cannot possibly be the most efficient or optimal weighting scheme for a given point in time, since the optimal weights change continuously through time.
This isn’t necessarily true. That temporal diversification is recommended doesn’t mean tactical asset allocation is too.

Temporal diversification works because given a constant volatility and ERP, the optimal long term asset allocation is a fixed percent allocation. That’s because the long term risk-adjusted returns are maximized when you maximize short term. This is aka Myopic Portfolio choice. Samuelson decades ago proved investment horizon is irrelevant to the optimal portfolio.

Given that a constant % allocation is optimal, then a person should try to keep a constant % in stocks of their whole wealth through their life. If your salary is very risk less, that means leverage early on to offset that human capital.

That is ALL temporal diversification requires. It’s simple and straightforward.

To make an argument for tactical asset allocation, reducing leverage when expected volatility is high, ERP has changed or trying to guess what the optimal weight is, you need to take a step further and claim these things change and you can predict them better than a 50/50 chance.

This very well might be an argument you can make and I’m all ears to learn. But you cannot claim that “if there’s an argument for Lifecycle Investing, there’s one for shifting an allocation”. That one is recommended does not imply the other is too. The former is a theoretical certainty, the latter can only be proven via backtests to show it can reliably be done and can persist in the future. Very different.

Hope that makes sense
Can you share the Samuelson paper? There's a reason physicists tend to do a whole hell of a lot better than economists when it comes to investing...and I say this as someone who studied econ, not physics. There's a reason LTCM blew up and Rentech thrived. From doing time series analysis, I can say with absolute certainty that portfolio co-moments, and by extension the optimal portfolio, are NOT scale invariant. Did he use real data, or just a whole bunch of theoretical assumptions? Keep in mind there have been nobel prizes in economics awarded to folks for their work that assumes stock market returns are normally distributed, so treat even celebrated economists with suspicion.

My argument is simply that the temporal component of investing is generally ignored, or addressed with very sloppy heuristics.
I listed his assumptions already. Given that volatility and ERP are constant, with normal distributions, the optimal asset allocation for a given utility function is a constant % allocation:
http://www.wiwi.uni-muenster.de/05/down ... on1969.pdf

This is a theoretical result. It depends on no data (and hence, suffers from no data mining). It forms the basis for Lifecycle Investing.

If you’re unconvinced by the assumptions, disagree with them or whatever else, that’s cool. But you cannot sit here and claim that Samuelson’s results (and Lifecycle Investing), which are mathematical facts given some assumptions, support the idea of tactical asset allocation. They absolutely do not.

If you want to prove some argument for tactical asset allocation then, like I already said, you need to take additional steps. You know to show that ERP does vary and you can somehow predict it to a reasonable extent. Ditto for volatility. These are, by definition, data proofs; you’d need to show it worked on data and then make a further proof that it can continue in the future at a given confidence interval. Similar to factors. This is completely different than Samuelson’s result.

Personally, I know volatility and ERP vary but I do not believe I can predict it better than a coin flip. So I picked some long term averages, made an allocation using Samuelson’s equation, and just bought/hold/followed the course from them.

In this way, I’m just a regular BH who believes they cannot tactically shift allocations profitably and that the market is a random walk. I’m just using leverage to more evenly spread the stock risk I’m planning on taking over decades.
Samuelson, in that paper, makes quite a few assumptions relating to the objective function. First of all, his main result only applies to isoelastic utilities, which admittedly is not a big deal and does include the biggie, namely log wealth utility. To me, the main assumption is the form of the objective itself: he is trying to maximize a functional of the form int_0^T U(C(t))dt where U(x) is the utility you derive from an amount of consumption x and C(t) is the consumption at time t.

While this might be a good objective from an economic theory perspective for the "representative agent", it's certainly not very close to my own objective, and probably many other people on this forum. Without getting too moralistic, note that this is quite a hedonistic metric and is trying to maximize your overall "happiness" in life. In fact, Samuelson even makes the assumption throughout most of his paper that you want your wealth at time T+1 to be 0 since you can't take money with you into the grave (he does provide results for other assumptions about the final wealth). The objective function I personally try to maximize is much closer to max U(x(T)) where x(t) is my portfolio size at time t and T is about 30 years from now. I anticipate my C(t), i.e. materialistic needs to stay very constant throughout this entire period and I will not derive significantly greater utility through greater consumption. For this kind of objective function and under the typical isoelastic utilities, as Ross shows in https://www.cambridge.org/core/journals ... 42B27CE740 the time period under consideration absolutely can affect your percent allocation.
Last edited by langlands on Sat Dec 21, 2019 2:37 pm, edited 2 times in total.
User avatar
lock.that.stock
Posts: 136
Joined: Sun Apr 28, 2019 2:01 pm

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

Post by lock.that.stock »

get_g0ing wrote: Sat Dec 21, 2019 12:24 pm So in M1 Finance, what's a good way to keep track of the actual return if you are also making contributions?
I mean the initial amount grows to a certain value and M1 shows you the return - so far so good - but if you make additional contributions, M1 also includes that in the gain/return.
There is a spreadsheet that longinvest created on this forum along with contributions from others that I found useful for calculating XIRR.

viewtopic.php?f=10&t=150025
User avatar
Steve Reading
Posts: 2839
Joined: Fri Nov 16, 2018 10:20 pm

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

Post by Steve Reading »

langlands wrote: Sat Dec 21, 2019 2:26 pm
305pelusa wrote: Sat Dec 21, 2019 12:03 pm
no simpler wrote: Fri Dec 20, 2019 5:40 pm
305pelusa wrote: Fri Dec 20, 2019 10:05 am
no simpler wrote: Fri Dec 20, 2019 9:42 am Yep, if there is an argument for temporal diversification, then there is definitely an argument for temporal shifts in asset allocation, whether based on total market cap weights of different asset classes or volatility based weights. A fixed allocation (60:40, etc) is just to reduce cognitive burden, but cannot possibly be the most efficient or optimal weighting scheme for a given point in time, since the optimal weights change continuously through time.
This isn’t necessarily true. That temporal diversification is recommended doesn’t mean tactical asset allocation is too.

Temporal diversification works because given a constant volatility and ERP, the optimal long term asset allocation is a fixed percent allocation. That’s because the long term risk-adjusted returns are maximized when you maximize short term. This is aka Myopic Portfolio choice. Samuelson decades ago proved investment horizon is irrelevant to the optimal portfolio.

Given that a constant % allocation is optimal, then a person should try to keep a constant % in stocks of their whole wealth through their life. If your salary is very risk less, that means leverage early on to offset that human capital.

That is ALL temporal diversification requires. It’s simple and straightforward.

To make an argument for tactical asset allocation, reducing leverage when expected volatility is high, ERP has changed or trying to guess what the optimal weight is, you need to take a step further and claim these things change and you can predict them better than a 50/50 chance.

This very well might be an argument you can make and I’m all ears to learn. But you cannot claim that “if there’s an argument for Lifecycle Investing, there’s one for shifting an allocation”. That one is recommended does not imply the other is too. The former is a theoretical certainty, the latter can only be proven via backtests to show it can reliably be done and can persist in the future. Very different.

Hope that makes sense
Can you share the Samuelson paper? There's a reason physicists tend to do a whole hell of a lot better than economists when it comes to investing...and I say this as someone who studied econ, not physics. There's a reason LTCM blew up and Rentech thrived. From doing time series analysis, I can say with absolute certainty that portfolio co-moments, and by extension the optimal portfolio, are NOT scale invariant. Did he use real data, or just a whole bunch of theoretical assumptions? Keep in mind there have been nobel prizes in economics awarded to folks for their work that assumes stock market returns are normally distributed, so treat even celebrated economists with suspicion.

My argument is simply that the temporal component of investing is generally ignored, or addressed with very sloppy heuristics.
I listed his assumptions already. Given that volatility and ERP are constant, with normal distributions, the optimal asset allocation for a given utility function is a constant % allocation:
http://www.wiwi.uni-muenster.de/05/down ... on1969.pdf

This is a theoretical result. It depends on no data (and hence, suffers from no data mining). It forms the basis for Lifecycle Investing.

If you’re unconvinced by the assumptions, disagree with them or whatever else, that’s cool. But you cannot sit here and claim that Samuelson’s results (and Lifecycle Investing), which are mathematical facts given some assumptions, support the idea of tactical asset allocation. They absolutely do not.

If you want to prove some argument for tactical asset allocation then, like I already said, you need to take additional steps. You know to show that ERP does vary and you can somehow predict it to a reasonable extent. Ditto for volatility. These are, by definition, data proofs; you’d need to show it worked on data and then make a further proof that it can continue in the future at a given confidence interval. Similar to factors. This is completely different than Samuelson’s result.

Personally, I know volatility and ERP vary but I do not believe I can predict it better than a coin flip. So I picked some long term averages, made an allocation using Samuelson’s equation, and just bought/hold/followed the course from them.

In this way, I’m just a regular BH who believes they cannot tactically shift allocations profitably and that the market is a random walk. I’m just using leverage to more evenly spread the stock risk I’m planning on taking over decades.
Samuelson, in that paper, makes quite a few assumptions relating to the objective function. First of all, his main result only applies to isoelastic utilities, which admittedly is not a big deal and does include the biggie, namely log wealth utility. To me, the main assumption is the form of the objective itself: he is trying to maximize a functional of the form int_0^T U(C(t))dt where U(x) is the utility you derive from an amount of consumption x and C(t) is the consumption at time t.

While this might be a good objective from an economic theory perspective for the "representative agent", it's certainly not very close to my own objective, and probably many other people on this forum. Without getting too moralistic, note that this is quite a hedonistic metric and is trying to maximize your overall "happiness" in life. In fact, Samuelson even makes the assumption throughout most of his paper that you want your wealth at time T+1 to be 0 since you can't take money with you into the grave (he does provide results for other assumptions about the final wealth). The objective function I personally try to maximize is much closer to max U(x(T)) where x(t) is my portfolio size at time t and T is about 30 years from now. I anticipate my C(t), i.e. materialistic needs to stay very constant throughout this entire period and I will not derive significantly greater utility through greater consumption. For this kind of objective function and under the typical isoelastic utilities, as Ross shows in https://www.cambridge.org/core/journals ... 42B27CE740 the time period under consideration absolutely can affect your percent allocation.
BHs are notorious for suggesting % asset allocations independent of your wealth. Be it 100k, 1M, 10M, etc. This is consistent with a log wealth utility. And it's the essence of "stay the course". Even Bogle, without any need for any more money, stayed 50/50 if I recall correctly. So I think it's fair to say many posters would derive ever-increasing utilities from higher wealth, but at a decreasing pace (similar to log wealth utility once again). Even if they don't spend it, getting to donate or leave it behind brings them joy. So they're willing to keep risking approximately the same, %-wise.

It sounds like this just isn't for you. Presumably, once you're hit more than you need to maintain your constant, steady desired consumption, you'd put it all in TIPs. That doesn't resonate with me but it does with others like Bernstein. That's cool. :happy

I couldn't read the paper you posted. How do I access it?

Cheers mate.
"... 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
ltdshred
Posts: 65
Joined: Sun Jul 07, 2019 9:41 pm

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

Post by ltdshred »

Ran a few simulation strategies just now and took a look at the distributions and found some interesting results. Compared 3 portfolios: 43-57 TQQQ-TMF; 55-45 UPRO-TMF; 100 SP500. Shouldn't be surprised, but I noticed that 55-45 had a fatter tail than 43-57, but limited upside. My parameters were a 100,000 starting balance and a look through period in the next 5 years.


TQQQ-TMF 43-57 Strategy:
Image

UPRO-TMF 55-45 Strategy:
Image

SP500 Strategy:
Image
danielfp
Posts: 35
Joined: Sat Dec 21, 2019 3:30 pm

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

Post by danielfp »

Hi guys,

New here. This is one of the few sources talking about this sort of long term leveraged strategy (excellent work btw). I work in the hedge fund industry and have invested 90%+ of my net worth using long term holding of 3x leveraged etf portfolios for some time now. It will take me a good time to read the entire thread but I just wanted to congratulate HEDGEFUNDIE for his efforts :happy Look forward to talking about this further here!

Daniel
Topic Author
HEDGEFUNDIE
Posts: 4801
Joined: Sun Oct 22, 2017 2:06 pm

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

Post by HEDGEFUNDIE »

danielfp wrote: Sat Dec 21, 2019 3:37 pm Hi guys,

New here. This is one of the few sources talking about this sort of long term leveraged strategy (excellent work btw). I work in the hedge fund industry and have invested 90%+ of my net worth using long term holding of 3x leveraged etf portfolios for some time now. It will take me a good time to read the entire thread but I just wanted to congratulate HEDGEFUNDIE for his efforts :happy Look forward to talking about this further here!

Daniel
Welcome! What LETFs do you hold and in what proportions?
danielfp
Posts: 35
Joined: Sat Dec 21, 2019 3:30 pm

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

Post by danielfp »

I personally hold 34/33/33 MIDU/TMF/TYD, up 45% so far this year. I also time new investments when the portfolio goes into drawdowns.
MotoTrojan
Posts: 10778
Joined: Wed Feb 01, 2017 8:39 pm

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

Post by MotoTrojan »

danielfp wrote: Sat Dec 21, 2019 5:11 pm I personally hold 34/33/33 MIDU/TMF/TYD, up 45% so far this year. I also time new investments when the portfolio goes into drawdowns.
No concerns with the tracking error of TYD? Why not hold TMF and EDV to get your target effective duration?

Why mid caps?
danielfp
Posts: 35
Joined: Sat Dec 21, 2019 3:30 pm

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

Post by danielfp »

The correlation of returns of the 7-10 year just plays better with the risk to return profile than EDV exposure. A leveraged short end is not equivalent to just a long end in my evaluation process, the short end just behaves differently. I am not seeking just a target duration but a correlation of returns that plays a certain way. About tracking error, it has not been a problem or really of any large significance. The only issue I see with TYD is liquidity, reason why for a large allocation you might prefer to negotiate creation shares with direxion rather than try to buy them at market
danielfp
Posts: 35
Joined: Sat Dec 21, 2019 3:30 pm

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

Post by danielfp »

About midcaps, I like their consistent outperformance over large caps and I prefer their risk to return profile. I think they are a sweet spot in the size factor (large enough to have a lot of safety while small enough to grow aggressively).
langlands
Posts: 849
Joined: Wed Apr 03, 2019 10:05 pm

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

Post by langlands »

305pelusa wrote: Sat Dec 21, 2019 2:50 pm BHs are notorious for suggesting % asset allocations independent of your wealth. Be it 100k, 1M, 10M, etc. This is consistent with a log wealth utility. And it's the essence of "stay the course". Even Bogle, without any need for any more money, stayed 50/50 if I recall correctly. So I think it's fair to say many posters would derive ever-increasing utilities from higher wealth, but at a decreasing pace (similar to log wealth utility once again). Even if they don't spend it, getting to donate or leave it behind brings them joy. So they're willing to keep risking approximately the same, %-wise.

It sounds like this just isn't for you. Presumably, once you're hit more than you need to maintain your constant, steady desired consumption, you'd put it all in TIPs. That doesn't resonate with me but it does with others like Bernstein. That's cool. :happy

I couldn't read the paper you posted. How do I access it?

Cheers mate.
Hm, I think you got the wrong impression. I didn't mean to criticize lifecycle investing. I just never found Samuelson's fallacy of large numbers argument particularly compelling. But I also don't find it particularly relevant to lifecycle investing. In fact, I'm a little confused because Samuelson is famously against time diversification so it seems strange to invoke his name in support of it.

Once I hit the amount needed to maintain my constant desired consumption, I definitely wouldn't put it all in TIPS, just like Warren Buffet doesn't put all his money into TIPS even though he consumes about 0.000001% of his net worth. My entire point was that my utility function is essentially uncorrelated to my consumption needs, which might seem very strange to most, but probably resonates with quite a few as well. By applying lifecycle investing and other principles, I hope to achieve a net worth by the time I'm nearing retirement so that retirement itself is no longer an issue for me. In other words, I hope that my time horizon is effectively always very long exactly so that time diversification is always on my side. I think that short time horizon is a disadvantage and forces more conservation allocation.

I think one way to access is through JStor where you get something like 5 free articles per month. Another way is just to google for something called sci-hub where you can access almost all papers...
MotoTrojan
Posts: 10778
Joined: Wed Feb 01, 2017 8:39 pm

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

Post by MotoTrojan »

danielfp wrote: Sat Dec 21, 2019 6:00 pm The correlation of returns of the 7-10 year just plays better with the risk to return profile than EDV exposure. A leveraged short end is not equivalent to just a long end in my evaluation process, the short end just behaves differently. I am not seeking just a target duration but a correlation of returns that plays a certain way. About tracking error, it has not been a problem or really of any large significance. The only issue I see with TYD is liquidity, reason why for a large allocation you might prefer to negotiate creation shares with direxion rather than try to buy them at market
The tell-tale plots I saw for TYD showed pretty poor ability to track it's index, while TMF was doing much better. Otherwise thanks for the insight, makes sense.
MotoTrojan
Posts: 10778
Joined: Wed Feb 01, 2017 8:39 pm

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

Post by MotoTrojan »

danielfp wrote: Sat Dec 21, 2019 6:04 pm About midcaps, I like their consistent outperformance over large caps and I prefer their risk to return profile. I think they are a sweet spot in the size factor (large enough to have a lot of safety while small enough to grow aggressively).
While the unleveraged ones have outperformed large caps, I would think the increased volatility (decay) and trading costs would eat into that advantage.
User avatar
Steve Reading
Posts: 2839
Joined: Fri Nov 16, 2018 10:20 pm

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

Post by Steve Reading »

langlands wrote: Sat Dec 21, 2019 6:40 pm
Hm, I think you got the wrong impression. I didn't mean to criticize lifecycle investing. I just never found Samuelson's fallacy of large numbers argument particularly compelling. But I also don't find it particularly relevant to lifecycle investing. In fact, I'm a little confused because Samuelson is famously against time diversification so it seems strange to invoke his name in support of it.

Once I hit the amount needed to maintain my constant desired consumption, I definitely wouldn't put it all in TIPS, just like Warren Buffet doesn't put all his money into TIPS even though he consumes about 0.000001% of his net worth. My entire point was that my utility function is essentially uncorrelated to my consumption needs, which might seem very strange to most, but probably resonates with quite a few as well. By applying lifecycle investing and other principles, I hope to achieve a net worth by the time I'm nearing retirement so that retirement itself is no longer an issue for me. In other words, I hope that my time horizon is effectively always very long exactly so that time diversification is always on my side. I think that short time horizon is a disadvantage and forces more conservation allocation.

I think one way to access is through JStor where you get something like 5 free articles per month. Another way is just to google for something called sci-hub where you can access almost all papers...
The argument against the fallacy of time diversification is very compelling IMO. Mathematically and historically, we know stocks do get riskier, not less, over longer periods of time.

Don’t get time diversification (a fallacy) confused with temporal diversification (aka Lifecycle Investing). The latter says that given a certain number of dollar*years in stocks, you’re better off spreading it in as many years as possible (instead of just concentrated in a few years). This is very different than the (fallacious) idea that stocks are less risky over time. The latter is what some call time diversification, the idea that good years will balance bad years so if held for long enough, stocks actually get safer. That’s not true.
"... 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
danielfp
Posts: 35
Joined: Sat Dec 21, 2019 3:30 pm

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

Post by danielfp »

MotoTrojan wrote: Sat Dec 21, 2019 6:57 pm
danielfp wrote: Sat Dec 21, 2019 6:04 pm About midcaps, I like their consistent outperformance over large caps and I prefer their risk to return profile. I think they are a sweet spot in the size factor (large enough to have a lot of safety while small enough to grow aggressively).
While the unleveraged ones have outperformed large caps, I would think the increased volatility (decay) and trading costs would eat into that advantage.
The advantage holds up well enough for me to make this choice.
MotoTrojan
Posts: 10778
Joined: Wed Feb 01, 2017 8:39 pm

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

Post by MotoTrojan »

danielfp wrote: Sat Dec 21, 2019 7:12 pm
MotoTrojan wrote: Sat Dec 21, 2019 6:57 pm
danielfp wrote: Sat Dec 21, 2019 6:04 pm About midcaps, I like their consistent outperformance over large caps and I prefer their risk to return profile. I think they are a sweet spot in the size factor (large enough to have a lot of safety while small enough to grow aggressively).
While the unleveraged ones have outperformed large caps, I would think the increased volatility (decay) and trading costs would eat into that advantage.
The advantage holds up well enough for me to make this choice.
If you say so...

https://www.portfoliovisualizer.com/bac ... ion2_2=100

I know mid-caps have not out-performed in this timeframe, but man is that MIDU volatility scary.
danielfp
Posts: 35
Joined: Sat Dec 21, 2019 3:30 pm

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

Post by danielfp »

MotoTrojan wrote: Sat Dec 21, 2019 7:16 pm
danielfp wrote: Sat Dec 21, 2019 7:12 pm
MotoTrojan wrote: Sat Dec 21, 2019 6:57 pm
danielfp wrote: Sat Dec 21, 2019 6:04 pm About midcaps, I like their consistent outperformance over large caps and I prefer their risk to return profile. I think they are a sweet spot in the size factor (large enough to have a lot of safety while small enough to grow aggressively).
While the unleveraged ones have outperformed large caps, I would think the increased volatility (decay) and trading costs would eat into that advantage.
The advantage holds up well enough for me to make this choice.
If you say so...

https://www.portfoliovisualizer.com/bac ... ion2_2=100

I know mid-caps have not out-performed in this timeframe, but man is that MIDU volatility scary.
I know, it does not look great in the recent past but I've ran tests on a 3x leveraged midcap synthetic from the 70s to make the assessment. Sadly not data I can share but I am happier with this choice vs large caps.
danielfp
Posts: 35
Joined: Sat Dec 21, 2019 3:30 pm

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

Post by danielfp »

It is worth noting that my midcap allocation is only 34% so its contribution to my overall volatility is not as awful as if I was doing the larger equity allocations on this thread.
MotoTrojan
Posts: 10778
Joined: Wed Feb 01, 2017 8:39 pm

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

Post by MotoTrojan »

danielfp wrote: Sat Dec 21, 2019 7:25 pm It is worth noting that my midcap allocation is only 34% so its contribution to my overall volatility is not as awful as if I was doing the larger equity allocations on this thread.
How long have you been in this portfolio?
danielfp
Posts: 35
Joined: Sat Dec 21, 2019 3:30 pm

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

Post by danielfp »

MotoTrojan wrote: Sat Dec 21, 2019 8:01 pm
danielfp wrote: Sat Dec 21, 2019 7:25 pm It is worth noting that my midcap allocation is only 34% so its contribution to my overall volatility is not as awful as if I was doing the larger equity allocations on this thread.
How long have you been in this portfolio?
A couple of years now in this exact setup. Although longer in other similar leveraged approaches.
stipeman
Posts: 75
Joined: Tue Sep 18, 2007 5:26 pm

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

Post by stipeman »

danielfp wrote: Sun Dec 22, 2019 1:17 am
MotoTrojan wrote: Sat Dec 21, 2019 8:01 pm
danielfp wrote: Sat Dec 21, 2019 7:25 pm It is worth noting that my midcap allocation is only 34% so its contribution to my overall volatility is not as awful as if I was doing the larger equity allocations on this thread.
How long have you been in this portfolio?
A couple of years now in this exact setup. Although longer in other similar leveraged approaches.
Why not futures? Much lower cost and more flexible. You can use /EMD for the midcaps or unlevered + treasury futures.
danielfp
Posts: 35
Joined: Sat Dec 21, 2019 3:30 pm

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

Post by danielfp »

stipeman wrote: Sun Dec 22, 2019 12:22 pm
danielfp wrote: Sun Dec 22, 2019 1:17 am
MotoTrojan wrote: Sat Dec 21, 2019 8:01 pm
danielfp wrote: Sat Dec 21, 2019 7:25 pm It is worth noting that my midcap allocation is only 34% so its contribution to my overall volatility is not as awful as if I was doing the larger equity allocations on this thread.
How long have you been in this portfolio?
A couple of years now in this exact setup. Although longer in other similar leveraged approaches.
Why not futures? Much lower cost and more flexible. You can use /EMD for the midcaps or unlevered + treasury futures.
I cannot trade futures due to conflicts of interest with the asset management firm I work for, I can only hold LETFs long term.
danielfp
Posts: 35
Joined: Sat Dec 21, 2019 3:30 pm

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

Post by danielfp »

The cost difference is also not that big considering all the taxable events generating when rolling over futures contracts, plus the 60/40 tax structure vs long term capital gains tax in LETFs only incurred when selling.
get_g0ing
Posts: 750
Joined: Sat Dec 09, 2017 11:09 am

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

Post by get_g0ing »

lock.that.stock wrote: Sat Dec 21, 2019 2:31 pm
get_g0ing wrote: Sat Dec 21, 2019 12:24 pm So in M1 Finance, what's a good way to keep track of the actual return if you are also making contributions?
I mean the initial amount grows to a certain value and M1 shows you the return - so far so good - but if you make additional contributions, M1 also includes that in the gain/return.
There is a spreadsheet that longinvest created on this forum along with contributions from others that I found useful for calculating XIRR.

viewtopic.php?f=10&t=150025
MotoTrojan wrote: Sat Dec 21, 2019 1:44 pm
get_g0ing wrote: Sat Dec 21, 2019 12:24 pm So in M1 Finance, what's a good way to keep track of the actual return if you are also making contributions?
I mean the initial amount grows to a certain value and M1 shows you the return - so far so good - but if you make additional contributions, M1 also includes that in the gain/return.
XIRR, google sheets/finance is how I did it. Allowed me to compare to a total return S&P500 index too, for sand contribution history.
Thanks guys !!
stormcrow
Posts: 103
Joined: Thu Nov 21, 2019 4:12 pm

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

Post by stormcrow »

danielfp wrote: Sat Dec 21, 2019 6:00 pm The correlation of returns of the 7-10 year just plays better with the risk to return profile than EDV exposure. A leveraged short end is not equivalent to just a long end in my evaluation process, the short end just behaves differently. I am not seeking just a target duration but a correlation of returns that plays a certain way. About tracking error, it has not been a problem or really of any large significance. The only issue I see with TYD is liquidity, reason why for a large allocation you might prefer to negotiate creation shares with direxion rather than try to buy them at market
Finally someone else looking at TYD! All the backtesting I did definitely suggested that there is a sweet spot of risk/reward that TYD helps hit.
MotoTrojan
Posts: 10778
Joined: Wed Feb 01, 2017 8:39 pm

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

Post by MotoTrojan »

stormcrow wrote: Mon Dec 23, 2019 10:47 am
danielfp wrote: Sat Dec 21, 2019 6:00 pm The correlation of returns of the 7-10 year just plays better with the risk to return profile than EDV exposure. A leveraged short end is not equivalent to just a long end in my evaluation process, the short end just behaves differently. I am not seeking just a target duration but a correlation of returns that plays a certain way. About tracking error, it has not been a problem or really of any large significance. The only issue I see with TYD is liquidity, reason why for a large allocation you might prefer to negotiate creation shares with direxion rather than try to buy them at market
Finally someone else looking at TYD! All the backtesting I did definitely suggested that there is a sweet spot of risk/reward that TYD helps hit.
Care to share that backtesting? From what I saw there is not a significant difference in return based on what part of the yield curve you target, with the main consideration being your effective duration. EDV is a powerful tool to reduce your effective duration away from TMF, while also avoiding the 1% ER and volatility decay of these leveraged products. Basically start with TMF and if that is too much duration then start diluting with EDV until you reach your target. I did not see any advantage to using TMF and IEF for example, rather than using less TMF and some EDV instead of IEF.
Post Reply