HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Locked
robertmcd
Posts: 554
Joined: Tue Aug 09, 2016 9:06 am

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by robertmcd » Wed Jun 26, 2019 10:39 am

BigMoneyNoWhammies wrote:
Tue Jun 25, 2019 11:52 am
finite_difference wrote:
Sun Mar 10, 2019 12:44 pm
pezblanco wrote:
Tue Mar 05, 2019 10:43 pm
privatefarmer wrote:
Tue Mar 05, 2019 10:31 pm
One more data point, from April ‘08-present the fed fund rate has gone sideways, it started at 2.3% and currently at 2.4%. Our combo portfolio would’ve returned 18.6% CAGR, TMF returned 8.7% and VFINX 9%. So clearly In sideways markets this strategy has worked wonderfully.
So, the bull market in stocks that time saved the bacon ..... I admit that the beauty of the strategy is that if one leg (stocks) or the other leg (bonds) does well, then you're golden. I've just been pointing out that I think that the expectations that the bond leg is going to save your bacon in the future might be misplaced. I expect that it might very well give reasonable returns but ... great returns? I don't see how that could reasonably be something to bet on.
Right now rates are not very high. So if there’s a bear market there’s not much wiggle room. What will happen if treasury rates go negative? They almost did after the Great Recession. I wouldn’t expect the stock market to “save the bacon” in that type of scenario, or for stocks and treasury rates to be uncorrelated?

That is, would investors dump treasuries to see positive or 0% return elsewhere or can we expect them to pay to keep their money in treasuries?

Note: I think volatility, suppressed returns, fees, weaker correlation than expected, etc. as more of a risk than this scenario but I still think it’s worth considering.
The Fed has been adamant since the Recession that they would never do treasury instruments with negative rates like some of the European countries and japan have done. Don't expect that to change. They see it as foolish and so do most in the financial world within the US. Investors here would never be ok with locking in long term losses via negative long treasuries.
The Fed has already talked about how they may have to eventually use negative rates. Even using unconventional tools such as buying corporate bonds or stocks. Rates will continue on their downward trend to keep the party going.

User avatar
Dr. Long
Posts: 36
Joined: Sun Jun 23, 2019 2:14 pm
Location: Location, Location, Location

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by Dr. Long » Wed Jun 26, 2019 1:31 pm

Finally got around to reading the sections of this thread I skimmed, quite an informative 45 pages :) learned a good bit about backtesting data sets and also how LTTs interplay with the market.

One thing I didnt notice - and forgive me if it seems trivial - but are those of you following OPs method to the T dripping your dividends as they come (or, auto-invest is on in M1) or are you waiting until quarterly rebalances to re-invest them?

I know sometime back in the thread people posited the quarterly rebal tested best do to the possibility of the ratios following market momentum (not sure if this is the case or it is just randomness), so would dripping affect this "momentum" or better to drip as they come because early money = more money? What was done with the dividends in the backtesting?

Once again, I know it sounds trivial, just wanted to make sure I was following OP & the majority's strategy to the T, as I am going to join this train and go 5% in next Mon (start of Q3) and want to be able to accurately compare my developing data set with others down the road.

Thanks!
"(It's) the economy, stupid," - James Carville

dave_k
Posts: 401
Joined: Sat Dec 26, 2015 8:25 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by dave_k » Wed Jun 26, 2019 1:47 pm

Dr. Long wrote:
Wed Jun 26, 2019 1:31 pm
One thing I didnt notice - and forgive me if it seems trivial - but are those of you following OPs method to the T dripping your dividends as they come (or, auto-invest is on in M1) or are you waiting until quarterly rebalances to re-invest them?
I'm using Fidelity, and I reinvested the dividends along with rebalancing around the end of last quarter. According to Fidelity's info on the funds, dividends should be paid out on 7/2, and it looks like both funds pay dividends around the end of each quarter, so I'll reinvest them and rebalance again once they show up, and plan to keep doing it that way. I created a spreadsheet to track it and calculate the rebalancing along with reinvesting the dividends quarterly. Not quite as simple as M1, but I already had the IRAs at Fidelity and have plenty of free trades.

MotoTrojan
Posts: 9955
Joined: Wed Feb 01, 2017 8:39 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by MotoTrojan » Wed Jun 26, 2019 1:52 pm

dave_k wrote:
Wed Jun 26, 2019 1:47 pm
Dr. Long wrote:
Wed Jun 26, 2019 1:31 pm
One thing I didnt notice - and forgive me if it seems trivial - but are those of you following OPs method to the T dripping your dividends as they come (or, auto-invest is on in M1) or are you waiting until quarterly rebalances to re-invest them?
I'm using Fidelity, and I reinvested the dividends along with rebalancing around the end of last quarter. According to Fidelity's info on the funds, dividends should be paid out on 7/2, and it looks like both funds pay dividends around the end of each quarter, so I'll reinvest them and rebalance again once they show up, and plan to keep doing it that way. I created a spreadsheet to track it and calculate the rebalancing along with reinvesting the dividends quarterly. Not quite as simple as M1, but I already had the IRAs at Fidelity and have plenty of free trades.
I also invest them as they come. No point in leaving it in cash when it can go into the under-performing asset (rebalance boost). But I am also switching to a monthly rebalance (and occasional contribution) based on past months volatility.

dave_k
Posts: 401
Joined: Sat Dec 26, 2015 8:25 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by dave_k » Wed Jun 26, 2019 2:07 pm

MotoTrojan wrote:
Wed Jun 26, 2019 1:52 pm
I also invest them as they come. No point in leaving it in cash when it can go into the under-performing asset (rebalance boost). But I am also switching to a monthly rebalance (and occasional contribution) based on past months volatility.
I sort of followed the recent discussions about rebalancing based on volatility, and I'd consider doing this if/when I get a chance to dig further into it myself and do some back-testing, and make sure I understand how to find the correct volatility info each time. I'd need to have some confidence that it's not likely to be (much) worse in downturns, at east in terms of recovery time. I may switch to monthly if I do this, at least until the free trades run out.

How many others are confident enough in the volatility based rebalancing to give it a try?

MotoTrojan
Posts: 9955
Joined: Wed Feb 01, 2017 8:39 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by MotoTrojan » Wed Jun 26, 2019 2:21 pm

dave_k wrote:
Wed Jun 26, 2019 2:07 pm
MotoTrojan wrote:
Wed Jun 26, 2019 1:52 pm
I also invest them as they come. No point in leaving it in cash when it can go into the under-performing asset (rebalance boost). But I am also switching to a monthly rebalance (and occasional contribution) based on past months volatility.
I sort of followed the recent discussions about rebalancing based on volatility, and I'd consider doing this if/when I get a chance to dig further into it myself and do some back-testing, and make sure I understand how to find the correct volatility info each time. I'd need to have some confidence that it's not likely to be (much) worse in downturns, at east in terms of recovery time. I may switch to monthly if I do this, at least until the free trades run out.

How many others are confident enough in the volatility based rebalancing to give it a try?
For now I am using Portfolio Visualizer to output the volatility but in a pinch Excel could easily handle it and it wouldn't be THAT crazy to manually look-up price for the last 20 days (I open Google Sheets nearly daily so I could get it to auto-pull each day, daily, too).

For extremely short duration volatility events it can hurt you by reducing your equity exposure after the drop and before the rebound, but for most events of interest (including December 2018's drop, which it went into with only 25% exposure), it does quite well. It also helps because it will tend to be a bit higher equity exposure than the default 40%, during steady bulls and rebounds.

Backtest looks great, but that only describes the past. Some papers I have seen show that equity has a more favorable correlation between returns and volatility, with fixed income (treasuries) having better returns during higher volatility realms in the early days, which would be a downside to inverse volatility. You could ignore the volatility on the bonds by strictly doing volatility targeting, which worked even better (there were some insane returns the last decade since actual fund inception), but that requires picking a target volatility and I am not super comfortable with an arbitrary choice (based solely on backtesting). Thus, I like the simplicity of just using the recent inverse volatility. Speaking of arbitrary, I really disliked the sensitivity of the original portfolio to the quarterly rebalance as well.

User avatar
privatefarmer
Posts: 696
Joined: Mon Sep 08, 2014 2:45 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by privatefarmer » Wed Jun 26, 2019 3:22 pm

BigMoneyNoWhammies wrote:
Tue Jun 25, 2019 11:52 am
finite_difference wrote:
Sun Mar 10, 2019 12:44 pm
pezblanco wrote:
Tue Mar 05, 2019 10:43 pm
privatefarmer wrote:
Tue Mar 05, 2019 10:31 pm
One more data point, from April ‘08-present the fed fund rate has gone sideways, it started at 2.3% and currently at 2.4%. Our combo portfolio would’ve returned 18.6% CAGR, TMF returned 8.7% and VFINX 9%. So clearly In sideways markets this strategy has worked wonderfully.
So, the bull market in stocks that time saved the bacon ..... I admit that the beauty of the strategy is that if one leg (stocks) or the other leg (bonds) does well, then you're golden. I've just been pointing out that I think that the expectations that the bond leg is going to save your bacon in the future might be misplaced. I expect that it might very well give reasonable returns but ... great returns? I don't see how that could reasonably be something to bet on.
Right now rates are not very high. So if there’s a bear market there’s not much wiggle room. What will happen if treasury rates go negative? They almost did after the Great Recession. I wouldn’t expect the stock market to “save the bacon” in that type of scenario, or for stocks and treasury rates to be uncorrelated?

That is, would investors dump treasuries to see positive or 0% return elsewhere or can we expect them to pay to keep their money in treasuries?

Note: I think volatility, suppressed returns, fees, weaker correlation than expected, etc. as more of a risk than this scenario but I still think it’s worth considering.
The Fed has been adamant since the Recession that they would never do treasury instruments with negative rates like some of the European countries and japan have done. Don't expect that to change. They see it as foolish and so do most in the financial world within the US. Investors here would never be ok with locking in long term losses via negative long treasuries.
LOL this post nearly made me fall out of my seat. If the fed could so easily control the economy they would. They can’t. If economic growth is poor, deflation is of concern, the fed absolutely will do what they must to keep employment up and inflation within its target of 2-3%. Interest rates and the feds monetary policy are extremely complex. To think that we could simply assume rates will bottom out and zero and then the fed will just “shrug its shoulders” and let the economy plummet/deflation run wild I think is misguided. I don’t understand all the tools the fed has to “stimulate” the economy. I know there is more they can do than simply lowering rates. What I DO believe is that no investor in his right mind would buy long term treasuries over short term if he didn’t demand a premium for doing so. The entire basis of this strategy is to 1) diversify not only the different equities you own but the different RISK PREMIUMS you’re exposed to. In this case, equity risk and interest rate risk. And 2) own two assets that have historically had negative correlation with each other especially during major market crashes.

BigMoneyNoWhammies
Posts: 222
Joined: Tue Jul 11, 2017 11:58 am

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by BigMoneyNoWhammies » Thu Jun 27, 2019 8:57 am

market timer wrote:
Fri May 31, 2019 10:54 pm
HEDGEFUNDIE wrote:
Fri May 31, 2019 10:25 pm
Here are the Boglehead principles. Please point out which ones I am violating:

...

Invest with simplicity
I'd say people would take most issue with the simplicity point. However, philosophically, I see no contradiction with what you are doing and Bogleheads principles--the market just hasn't yet provided a simple product (that I'm aware of) to allow risk parity investing at low cost. I liken what you are doing to those who might have tried to invest in the S&P 500 before index funds existed. That would have seemed complicated and unorthodox at the time. Now everybody just dumps their savings into VTI without a second thought.
Based on reading through the entirety of the thread, I think the majority on here who make the stink face when seeing this thread are probably most averse to the leverage aspect (although I think this also plays into the simplicity point you make as well). The portfolio is 2 ETFs with quarterly rebalancing. That's as simple or simpler than many of the "vanilla" 3/4 fund or value tilt porfolios you see vociferously advocated on here.

Great point about the investing in the S&P pre-index fund comparison; I think that's 100% on point
Last edited by BigMoneyNoWhammies on Thu Jun 27, 2019 9:16 am, edited 2 times in total.

BigMoneyNoWhammies
Posts: 222
Joined: Tue Jul 11, 2017 11:58 am

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by BigMoneyNoWhammies » Thu Jun 27, 2019 9:13 am

privatefarmer wrote:
Wed Jun 26, 2019 3:22 pm
BigMoneyNoWhammies wrote:
Tue Jun 25, 2019 11:52 am
finite_difference wrote:
Sun Mar 10, 2019 12:44 pm
pezblanco wrote:
Tue Mar 05, 2019 10:43 pm
privatefarmer wrote:
Tue Mar 05, 2019 10:31 pm
One more data point, from April ‘08-present the fed fund rate has gone sideways, it started at 2.3% and currently at 2.4%. Our combo portfolio would’ve returned 18.6% CAGR, TMF returned 8.7% and VFINX 9%. So clearly In sideways markets this strategy has worked wonderfully.
So, the bull market in stocks that time saved the bacon ..... I admit that the beauty of the strategy is that if one leg (stocks) or the other leg (bonds) does well, then you're golden. I've just been pointing out that I think that the expectations that the bond leg is going to save your bacon in the future might be misplaced. I expect that it might very well give reasonable returns but ... great returns? I don't see how that could reasonably be something to bet on.
Right now rates are not very high. So if there’s a bear market there’s not much wiggle room. What will happen if treasury rates go negative? They almost did after the Great Recession. I wouldn’t expect the stock market to “save the bacon” in that type of scenario, or for stocks and treasury rates to be uncorrelated?

That is, would investors dump treasuries to see positive or 0% return elsewhere or can we expect them to pay to keep their money in treasuries?

Note: I think volatility, suppressed returns, fees, weaker correlation than expected, etc. as more of a risk than this scenario but I still think it’s worth considering.
The Fed has been adamant since the Recession that they would never do treasury instruments with negative rates like some of the European countries and japan have done. Don't expect that to change. They see it as foolish and so do most in the financial world within the US. Investors here would never be ok with locking in long term losses via negative long treasuries.
LOL this post nearly made me fall out of my seat. If the fed could so easily control the economy they would. They can’t. If economic growth is poor, deflation is of concern, the fed absolutely will do what they must to keep employment up and inflation within its target of 2-3%. Interest rates and the feds monetary policy are extremely complex. To think that we could simply assume rates will bottom out and zero and then the fed will just “shrug its shoulders” and let the economy plummet/deflation run wild I think is misguided. I don’t understand all the tools the fed has to “stimulate” the economy. I know there is more they can do than simply lowering rates. What I DO believe is that no investor in his right mind would buy long term treasuries over short term if he didn’t demand a premium for doing so. The entire basis of this strategy is to 1) diversify not only the different equities you own but the different RISK PREMIUMS you’re exposed to. In this case, equity risk and interest rate risk. And 2) own two assets that have historically had negative correlation with each other especially during major market crashes.
If 2008 wasn't an instance of when the fed would utilize negative rates, I don't know what is. The fact that they didn't and chose to pursue a combination of QE and zero/near zero rates for several years instead tells me they're highly resistant to negative rates. What do you think part of the calculus has been in the fed fund rate increases over the last few quarters? It's in part so the rate has somewhere to go downward during the next recession without going negative. And look at the countries that have implemented negative rates. How has that worked for them? Their recoveries have been slower and anemic by comparison, and the Fed is well aware of that fact. They won't ever go the negative rate route, take it to the bank (pun absolutely intended).

robertmcd
Posts: 554
Joined: Tue Aug 09, 2016 9:06 am

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by robertmcd » Thu Jun 27, 2019 9:30 am

BigMoneyNoWhammies wrote:
Thu Jun 27, 2019 9:13 am
privatefarmer wrote:
Wed Jun 26, 2019 3:22 pm
BigMoneyNoWhammies wrote:
Tue Jun 25, 2019 11:52 am
finite_difference wrote:
Sun Mar 10, 2019 12:44 pm
pezblanco wrote:
Tue Mar 05, 2019 10:43 pm


So, the bull market in stocks that time saved the bacon ..... I admit that the beauty of the strategy is that if one leg (stocks) or the other leg (bonds) does well, then you're golden. I've just been pointing out that I think that the expectations that the bond leg is going to save your bacon in the future might be misplaced. I expect that it might very well give reasonable returns but ... great returns? I don't see how that could reasonably be something to bet on.
Right now rates are not very high. So if there’s a bear market there’s not much wiggle room. What will happen if treasury rates go negative? They almost did after the Great Recession. I wouldn’t expect the stock market to “save the bacon” in that type of scenario, or for stocks and treasury rates to be uncorrelated?

That is, would investors dump treasuries to see positive or 0% return elsewhere or can we expect them to pay to keep their money in treasuries?

Note: I think volatility, suppressed returns, fees, weaker correlation than expected, etc. as more of a risk than this scenario but I still think it’s worth considering.
The Fed has been adamant since the Recession that they would never do treasury instruments with negative rates like some of the European countries and japan have done. Don't expect that to change. They see it as foolish and so do most in the financial world within the US. Investors here would never be ok with locking in long term losses via negative long treasuries.
LOL this post nearly made me fall out of my seat. If the fed could so easily control the economy they would. They can’t. If economic growth is poor, deflation is of concern, the fed absolutely will do what they must to keep employment up and inflation within its target of 2-3%. Interest rates and the feds monetary policy are extremely complex. To think that we could simply assume rates will bottom out and zero and then the fed will just “shrug its shoulders” and let the economy plummet/deflation run wild I think is misguided. I don’t understand all the tools the fed has to “stimulate” the economy. I know there is more they can do than simply lowering rates. What I DO believe is that no investor in his right mind would buy long term treasuries over short term if he didn’t demand a premium for doing so. The entire basis of this strategy is to 1) diversify not only the different equities you own but the different RISK PREMIUMS you’re exposed to. In this case, equity risk and interest rate risk. And 2) own two assets that have historically had negative correlation with each other especially during major market crashes.
If 2008 wasn't an instance of when the fed would utilize negative rates, I don't know what is. The fact that they didn't and chose to pursue a combination of QE and zero/near zero rates for several years instead tells me they're highly resistant to negative rates. What do you think part of the calculus has been in the fed fund rate increases over the last few quarters? It's in part so the rate has somewhere to go downward during the next recession without going negative. And look at the countries that have implemented negative rates. How has that worked for them? Their recoveries have been slower and anemic by comparison, and the Fed is well aware of that fact. They won't ever go the negative rate route, take it to the bank (pun absolutely intended).
I will make sure to save this quote because in the next recession, we are going negative. The Fed has already said they are not opposed to negative interest rates if they are necessary for a "recovery". And it will be deemed necessary to stop asset price deflation. This is an issue that is 30-40 yrs in the making, and no longer solvable because of how far the can was kicked. How do you get growth when you are being crushed by debt? You make debt even cheaper so you can load up on more of it.

BigMoneyNoWhammies
Posts: 222
Joined: Tue Jul 11, 2017 11:58 am

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by BigMoneyNoWhammies » Thu Jun 27, 2019 10:12 am

gw wrote:
Tue Jun 11, 2019 9:34 pm
HEDGEFUNDIE wrote:
Tue Jun 11, 2019 7:38 pm
From the beginning, I have framed this strategy as an alternative to a 100% equity portfolio. That is the benchmark by which we should judge this strategy:
In your longest available backtest, it has twice the std dev of 100% equities, a 1.5x worse max drawdown (75%), and essentially the same returns (10% vs. 11%), so a pretty big fail.

But worse than that, the strategy essentially *only* made money during the post-1981 period of steadily falling rates - it steadily *lost* money throughout the previous period, when rates were rising.

https://i.imgur.com/9D8QjKf.png

Your response is to simply discard the (long!) period where your strategy underperformed, while keeping all the returns from the period where it outperformed, under an argument that rising rates are a thing of the past. You may be right, but I say it's a big bet.

Again, I would point out that your strategy is roughly equivalent to (100% equities) + 2*(LTT-CASH). It's inherently riskier than 100% equities under any reasonable expectations, has a huge exposure to rising interest rates, and it seems pretty obvious to me why it happened to perform better in the past than one could reasonably expect going forward.
gw

I don't have any beef with your arguments vis a vis the data; it's clear from the 1955-present backtest that if this strategy had been implemented starting in 1955 that the performance would have been marginal compared to a 100% S&P 500 portfolio and not worth the risk given the std dev and max drawdown comparisons and the economic realities of the first 30ish years of the data set. He's mentioned it in previous posts (which I don't fault you at all for possibly missing given how ridiculously long this thread is), but it seems to me that part of the rationale of this portfolio configuration is a fairly substantial bet that the broader economic trends on how the fed functions and the market/government reacts to adverse events going forward in the future will be substantially similar to how they've been post-1981 and that the likelihood of the economic environment being similar to the conditions experienced from 1955-1981 is relatively unlikely. I personally tend to agree with Hedefundie on that particular point, but it's a fair point to disagree on.

I'd think it foolish to totally discount the earlier period backetest data, but if you're going to implement this strategy with the assumption of market conditions going forward being relatively unchanged being baked into the rationale/implementation as I believe Hedgefundie is, the earlier data probably carries less weight for you than does the 1982-present data. That's certainly a risk, but one that seems to be acknowledged by most of those interested in pursuing the strategy with part of their portfolios.

BigMoneyNoWhammies
Posts: 222
Joined: Tue Jul 11, 2017 11:58 am

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by BigMoneyNoWhammies » Thu Jun 27, 2019 10:20 am

robertmcd wrote:
Thu Jun 27, 2019 9:30 am
BigMoneyNoWhammies wrote:
Thu Jun 27, 2019 9:13 am
privatefarmer wrote:
Wed Jun 26, 2019 3:22 pm
BigMoneyNoWhammies wrote:
Tue Jun 25, 2019 11:52 am
finite_difference wrote:
Sun Mar 10, 2019 12:44 pm


Right now rates are not very high. So if there’s a bear market there’s not much wiggle room. What will happen if treasury rates go negative? They almost did after the Great Recession. I wouldn’t expect the stock market to “save the bacon” in that type of scenario, or for stocks and treasury rates to be uncorrelated?

That is, would investors dump treasuries to see positive or 0% return elsewhere or can we expect them to pay to keep their money in treasuries?

Note: I think volatility, suppressed returns, fees, weaker correlation than expected, etc. as more of a risk than this scenario but I still think it’s worth considering.
The Fed has been adamant since the Recession that they would never do treasury instruments with negative rates like some of the European countries and japan have done. Don't expect that to change. They see it as foolish and so do most in the financial world within the US. Investors here would never be ok with locking in long term losses via negative long treasuries.
LOL this post nearly made me fall out of my seat. If the fed could so easily control the economy they would. They can’t. If economic growth is poor, deflation is of concern, the fed absolutely will do what they must to keep employment up and inflation within its target of 2-3%. Interest rates and the feds monetary policy are extremely complex. To think that we could simply assume rates will bottom out and zero and then the fed will just “shrug its shoulders” and let the economy plummet/deflation run wild I think is misguided. I don’t understand all the tools the fed has to “stimulate” the economy. I know there is more they can do than simply lowering rates. What I DO believe is that no investor in his right mind would buy long term treasuries over short term if he didn’t demand a premium for doing so. The entire basis of this strategy is to 1) diversify not only the different equities you own but the different RISK PREMIUMS you’re exposed to. In this case, equity risk and interest rate risk. And 2) own two assets that have historically had negative correlation with each other especially during major market crashes.
If 2008 wasn't an instance of when the fed would utilize negative rates, I don't know what is. The fact that they didn't and chose to pursue a combination of QE and zero/near zero rates for several years instead tells me they're highly resistant to negative rates. What do you think part of the calculus has been in the fed fund rate increases over the last few quarters? It's in part so the rate has somewhere to go downward during the next recession without going negative. And look at the countries that have implemented negative rates. How has that worked for them? Their recoveries have been slower and anemic by comparison, and the Fed is well aware of that fact. They won't ever go the negative rate route, take it to the bank (pun absolutely intended).
I will make sure to save this quote because in the next recession, we are going negative. The Fed has already said they are not opposed to negative interest rates if they are necessary for a "recovery". And it will be deemed necessary to stop asset price deflation. This is an issue that is 30-40 yrs in the making, and no longer solvable because of how far the can was kicked. How do you get growth when you are being crushed by debt? You make debt even cheaper so you can load up on more of it.
The Fed has also said in the past that they didn't have any interest in negative rates or see a need for them (Yellen made this point multiple times over the years). The way you get out from being crushed by excessive debt in the market is doing what they did last time: slash the fed fund rate to zero and buy up gobs of debt via QE, which increases the monetary supply and frees up $ for lending and spending (growth). I don't think that solution will be as easy to implement for the next recession given that the Fed hasn't yet shed all its QE holdings from the last one, and no recession is identical to the last, but if 2008 wasn't enough of a impetus to cause negative rates, I just can't envision a scenario where they pull that trigger. Clearly opinions vary. Time will tell.

PluckyDucky
Posts: 250
Joined: Tue Jan 15, 2019 8:29 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by PluckyDucky » Thu Jun 27, 2019 1:36 pm

BigMoneyNoWhammies wrote:
Thu Jun 27, 2019 10:12 am
...
gw

I don't have any beef with your arguments vis a vis the data; it's clear from the 1955-present backtest that if this strategy had been implemented starting in 1955 that the performance would have been marginal compared to a 100% S&P 500 portfolio and not worth the risk given the std dev and max drawdown comparisons and the economic realities of the first 30ish years of the data set. He's mentioned it in previous posts (which I don't fault you at all for possibly missing given how ridiculously long this thread is), but it seems to me that part of the rationale of this portfolio configuration is a fairly substantial bet that the broader economic trends on how the fed functions and the market/government reacts to adverse events going forward in the future will be substantially similar to how they've been post-1981 and that the likelihood of the economic environment being similar to the conditions experienced from 1955-1981 is relatively unlikely. I personally tend to agree with Hedefundie on that particular point, but it's a fair point to disagree on.

I'd think it foolish to totally discount the earlier period backetest data, but if you're going to implement this strategy with the assumption of market conditions going forward being relatively unchanged being baked into the rationale/implementation as I believe Hedgefundie is, the earlier data probably carries less weight for you than does the 1982-present data. That's certainly a risk, but one that seems to be acknowledged by most of those interested in pursuing the strategy with part of their portfolios.
Forgive me if I'm misremembering, but I thought we only had monthly data from 1955-1981. So we still can't backtest that period correctly for the 3x daily leverage funds.

MotoTrojan
Posts: 9955
Joined: Wed Feb 01, 2017 8:39 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by MotoTrojan » Thu Jun 27, 2019 2:20 pm

PluckyDucky wrote:
Thu Jun 27, 2019 1:36 pm
BigMoneyNoWhammies wrote:
Thu Jun 27, 2019 10:12 am
...
gw

I don't have any beef with your arguments vis a vis the data; it's clear from the 1955-present backtest that if this strategy had been implemented starting in 1955 that the performance would have been marginal compared to a 100% S&P 500 portfolio and not worth the risk given the std dev and max drawdown comparisons and the economic realities of the first 30ish years of the data set. He's mentioned it in previous posts (which I don't fault you at all for possibly missing given how ridiculously long this thread is), but it seems to me that part of the rationale of this portfolio configuration is a fairly substantial bet that the broader economic trends on how the fed functions and the market/government reacts to adverse events going forward in the future will be substantially similar to how they've been post-1981 and that the likelihood of the economic environment being similar to the conditions experienced from 1955-1981 is relatively unlikely. I personally tend to agree with Hedefundie on that particular point, but it's a fair point to disagree on.

I'd think it foolish to totally discount the earlier period backetest data, but if you're going to implement this strategy with the assumption of market conditions going forward being relatively unchanged being baked into the rationale/implementation as I believe Hedgefundie is, the earlier data probably carries less weight for you than does the 1982-present data. That's certainly a risk, but one that seems to be acknowledged by most of those interested in pursuing the strategy with part of their portfolios.
Forgive me if I'm misremembering, but I thought we only had monthly data from 1955-1981. So we still can't backtest that period correctly for the 3x daily leverage funds.
Historical monthly volatility was used to develop monthly data back to 1955. This data tracks well for the 1980's-present and the actual funds so it seems to be a robust method of determining the impact of volatility.

The OP doesn't have this shown but there are plots floating around in this thread. Funny enough the portfolio more or less meets up with the standard S&P500, which of course highlights how poor the 1955-1982 performance must've been to allow that given the amazing 1982-present performance.

samsdad
Posts: 758
Joined: Sat Jan 02, 2016 6:20 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by samsdad » Thu Jun 27, 2019 10:41 pm

MotoTrojan wrote:
Thu Jun 27, 2019 2:20 pm
PluckyDucky wrote:
Thu Jun 27, 2019 1:36 pm
BigMoneyNoWhammies wrote:
Thu Jun 27, 2019 10:12 am
...
gw

I don't have any beef with your arguments vis a vis the data; it's clear from the 1955-present backtest that if this strategy had been implemented starting in 1955 that the performance would have been marginal compared to a 100% S&P 500 portfolio and not worth the risk given the std dev and max drawdown comparisons and the economic realities of the first 30ish years of the data set. He's mentioned it in previous posts (which I don't fault you at all for possibly missing given how ridiculously long this thread is), but it seems to me that part of the rationale of this portfolio configuration is a fairly substantial bet that the broader economic trends on how the fed functions and the market/government reacts to adverse events going forward in the future will be substantially similar to how they've been post-1981 and that the likelihood of the economic environment being similar to the conditions experienced from 1955-1981 is relatively unlikely. I personally tend to agree with Hedefundie on that particular point, but it's a fair point to disagree on.

I'd think it foolish to totally discount the earlier period backetest data, but if you're going to implement this strategy with the assumption of market conditions going forward being relatively unchanged being baked into the rationale/implementation as I believe Hedgefundie is, the earlier data probably carries less weight for you than does the 1982-present data. That's certainly a risk, but one that seems to be acknowledged by most of those interested in pursuing the strategy with part of their portfolios.
Forgive me if I'm misremembering, but I thought we only had monthly data from 1955-1981. So we still can't backtest that period correctly for the 3x daily leverage funds.
Historical monthly volatility was used to develop monthly data back to 1955. This data tracks well for the 1980's-present and the actual funds so it seems to be a robust method of determining the impact of volatility.

The OP doesn't have this shown but there are plots floating around in this thread. Funny enough the portfolio more or less meets up with the standard S&P500, which of course highlights how poor the 1955-1982 performance must've been to allow that given the amazing 1982-present performance.
I don’t have access to a computer, so this is a mobile screen shot. Forgive me if this is potatoed on your screens.

Image

A couple of thoughts. First, I’m thoroughly convinced that the simulated data back to 1955 developed previously is as accurate as anything that is going to be available. Second, I’m also thoroughly convinced that pre-1982ish data as it relates to long-term treasuries is irrelevant. As HEDGEFUNDIE and others, such as vineviz, have noted, there was a sea change with Volcker and to ignore that and point excitedly at the graph above as indisputable proof that this strategy is doomed to failure is being willfully ignorant, imho.

For completeness, here’s 1982 till the end of last year:

Image

These are, of course, quarterly-rebalanced results.

samsdad
Posts: 758
Joined: Sat Jan 02, 2016 6:20 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by samsdad » Thu Jun 27, 2019 11:01 pm

Here’s long-term treasuries compared with the simulated TMF fund from 1955-1981, keeping the 500 in for reference:

Image

And, a slightly happier picture, beginning 1982 through last year:

Image

samsdad
Posts: 758
Joined: Sat Jan 02, 2016 6:20 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by samsdad » Thu Jun 27, 2019 11:13 pm

And, for the why-not-100%-UPRO? crowd, here’s what 1955-1981 would’ve looked like:

Image

And 1982-2018:

Image

I hope these graphs ends that particular discussion, at least for a few hundred posts.

MotoTrojan
Posts: 9955
Joined: Wed Feb 01, 2017 8:39 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by MotoTrojan » Thu Jun 27, 2019 11:17 pm

samsdad wrote:
Thu Jun 27, 2019 11:13 pm
And, for the why-not-100%-UPRO? crowd, here’s what 1955-1981 would’ve looked like:

I hope these graphs ends that particular discussion, at least for a few hundred posts.
I don't know, there are some nice looking periods :twisted: . Honestly, during a time of equity distress would you rather throw a little fun money at 100% UPRO or Bitcoin?

I am mostly kidding.

samsdad
Posts: 758
Joined: Sat Jan 02, 2016 6:20 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by samsdad » Fri Jun 28, 2019 12:09 am

MotoTrojan wrote:
Thu Jun 27, 2019 11:17 pm
samsdad wrote:
Thu Jun 27, 2019 11:13 pm
And, for the why-not-100%-UPRO? crowd, here’s what 1955-1981 would’ve looked like:

I hope these graphs ends that particular discussion, at least for a few hundred posts.
I don't know, there are some nice looking periods :twisted: . Honestly, during a time of equity distress would you rather throw a little fun money at 100% UPRO or Bitcoin?

I am mostly kidding.
I’d mostly be eating antacids by the handful if those were my only two choices.

comeinvest
Posts: 291
Joined: Mon Mar 12, 2012 6:57 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by comeinvest » Fri Jun 28, 2019 5:26 am

privatefarmer wrote:
Thu Jun 20, 2019 4:28 pm
MotoTrojan wrote:
Thu Jun 20, 2019 4:03 pm
privatefarmer wrote:
Thu Jun 20, 2019 4:00 pm
reformed.trader wrote:
Thu Jun 20, 2019 10:30 am
privatefarmer wrote:
Thu Jun 20, 2019 12:11 am


right now I am borrowing 600k on a 650k portfolio, so i'm basically 2x levered. My goal is a risk parity portfolio of 1.25mil, I hope to have that unlevered in 5-10yrs.

The beauty of this is that before I discovered this thread I was severely limited in how much of a margin loan I could have bc of fear of getting hit w/ a margin call. Now, if my account comes even close to a margin call, I simply swap out unlevered ETFs w/ levered ETFs, allowing me to pay off some of the margin loan while maintaining the same amount of leverage.
Keep in mind your margin requirements will go up as you move to 2x and 3x funds compared to unlevered funds.
Yes. So because of “portfolio margin” at interactive brokers, I’m able to use a margin loan on unlevered ETFs for nearly 100% of my leverage. I then use a very small amount of LETFs within my Roth IRA to get the rest of the exposure up to my target portfolio size (1.25mil). IB will let me borrow up to 80 cents for every 20 cents invested with a balanced, unleveraged portfolio. If the market dips, I will buy more LETFs in my iras and pay down the margin loan to prevent margin call.

Basically, what I’m thankful to HEDGEDUNDIE for, is just opening my eyes to risk parity in general. Over the last couple of months I have fine tuned the methods I use to obtain adequate leverage but whether you’re using margin loan, futures or LETFs the philosophy is the same.
Curious what return you expect or assume to get from your unleveraged risk parity portfolio? What withdrawal rate do you plan to use once you reach $1.25M and presumably start drawing down?
I am a pessimist. However, PV shows an 11.3% CAGR since 1978 using a 50/50 mid cap value/LTT with a sharpe ratio of 0.68. Since I’m using mostly a margin loan for my leverage, I do not have to worry as much about volatility decay. So I am able to tilt strongly towards small/value. When I do use LETFs it’ll be LCB/LTTs as SIamond and hedgefundie have shown that you’re better off with a lower volatile asset like LCB to reduce volatility decay. So, I’m HOPING for maybe a 10% CAGR but if it’s less then it’s less.

I plan on using a variable w/d strategy instead of a fixed w/d as this portfolio is more to supplement my income vs fund a complete retirement, so I can adjust my lifestyle with variable distributions. According to PV, I should be able to expect anywhere from 6-8%/year with a 50/50 MCV/LTT portfolio as a variable w/d that would give me, historically, a 50% chance of portfolio increasing in value and a 50% chance of decreasing in value. So that’s my “ceiling” as far as withdrawals. Hopefully, after I’ve hit 1.25mil, I won’t be withdrawing anymore than like 3-4% variable but I figure my ceiling will be in the 6-8%. The more SCV tilt I have the closer it’d be to 8%.

This is all based on data from 1978 which is not long of a time period and I am well aware of dropping rates since then. My larger focus rather on what return I’ll get is what portfolio offers the highest expected sharpe ratio, select that portfolio and then titrate with leverage to get an adequate return. I BELIEVE a 50/50 portfolio of SCV/LTT would have the highest sharpe ratio but if someone has better data please do let me know.

I also have 20% in emerging markets SCV (DGS) just bc I am not comfortable completely ignoring INTL equities. I do not know if this will help or hinder the sharpe ratio but it’ll help me sleep better so there’s that.
I read your posts with interested, as I too have been thinking about combining risk parity, leverage, factor tilts, and a life cycle approach - which I understand is what you are attempting, right? I personally however believe that your return assumptions are far too optimistic. If you have 50/50 equities / LTT, the LTT portion i.e. 50% of your portfolio will likely contribute approximately nothing to your real returns during the next few decades; even if long term interest rates were to drop from 2-2.5% to zero, this would result in a one-time boost of maybe 40%, but zero nominal thereafter, so however you slice and dice it, the real return will be close to nothing, in striking contrast with the CAGR of LTTs from 1978 to now which I would say is not repeatable from current levels with 100% mathematical likelihood, so that for the bonds portion even attempting to extrapolate the past seems meaningless to me.

However, with your assumptions I am interested what your life cycle glide path algorithm for deleveraging is to get you from 2 x leverage to no leverage within 5-10 years? In other words, do you have a specific algorithm in mind for applying dividends, interest, and capital gains to your margin loan?

jaj2276
Posts: 493
Joined: Sat Apr 16, 2011 5:13 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by jaj2276 » Fri Jun 28, 2019 10:44 am

Did my second (first full quarter) rebalance today. My portfolio is up 24% since Feb 15, 14% for this quarter. My portfolio had drifted to 62% TMF, 38%UPRO. My rebalance was a bit more complex because I decided to split the portfolio in to three slices: 1) naive risk parity (40/60, qtrly rebalance), 2) 16 target vol (25%, monthly rebal), 3) adaptive allocation (25%, monthly rebal). The vol lookback will be 21 days for slice 2 and 3. It's been a "fun" ride so far, will be interesting to see how it fares when the market has a December 2018 repeat.

MotoTrojan
Posts: 9955
Joined: Wed Feb 01, 2017 8:39 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by MotoTrojan » Fri Jun 28, 2019 11:04 am

jaj2276 wrote:
Fri Jun 28, 2019 10:44 am
Did my second (first full quarter) rebalance today. My portfolio is up 24% since Feb 15, 14% for this quarter. My portfolio had drifted to 62% TMF, 38%UPRO. My rebalance was a bit more complex because I decided to split the portfolio in to three slices: 1) naive risk parity (40/60, qtrly rebalance), 2) 16 target vol (25%, monthly rebal), 3) adaptive allocation (25%, monthly rebal). The vol lookback will be 21 days for slice 2 and 3. It's been a "fun" ride so far, will be interesting to see how it fares when the market has a December 2018 repeat.
Curious how you picked the 25% target vol? It did great since inception of these funds but using simulated data the sharpe was a good bit off from optimized. Will you be letting these three options drift independently?

jaj2276
Posts: 493
Joined: Sat Apr 16, 2011 5:13 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by jaj2276 » Fri Jun 28, 2019 12:26 pm

MotoTrojan wrote:
Fri Jun 28, 2019 11:04 am
jaj2276 wrote:
Fri Jun 28, 2019 10:44 am
Did my second (first full quarter) rebalance today. My portfolio is up 24% since Feb 15, 14% for this quarter. My portfolio had drifted to 62% TMF, 38%UPRO. My rebalance was a bit more complex because I decided to split the portfolio in to three slices: 1) naive risk parity (40/60, qtrly rebalance), 2) 16 target vol (25%, monthly rebal), 3) adaptive allocation (25%, monthly rebal). The vol lookback will be 21 days for slice 2 and 3. It's been a "fun" ride so far, will be interesting to see how it fares when the market has a December 2018 repeat.
Curious how you picked the 25% target vol? It did great since inception of these funds but using simulated data the sharpe was a good bit off from optimized. Will you be letting these three options drift independently?
It's a 16 target vol, not 25 (the 25 was the percentage of portfolio allocated to the 16-vol slice). In PV I simply chose a vol that was greater than the 40/60 qtr rebalance but had a lower mkt correlation than the 40/60.

MotoTrojan
Posts: 9955
Joined: Wed Feb 01, 2017 8:39 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by MotoTrojan » Fri Jun 28, 2019 1:50 pm

1st monthly 20 day look-back volatility rebalance today; 47% UPRO, up 19.6% with a 109% XIRR.

User avatar
privatefarmer
Posts: 696
Joined: Mon Sep 08, 2014 2:45 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by privatefarmer » Fri Jun 28, 2019 10:44 pm

comeinvest wrote:
Fri Jun 28, 2019 5:26 am
privatefarmer wrote:
Thu Jun 20, 2019 4:28 pm
MotoTrojan wrote:
Thu Jun 20, 2019 4:03 pm
privatefarmer wrote:
Thu Jun 20, 2019 4:00 pm
reformed.trader wrote:
Thu Jun 20, 2019 10:30 am


Keep in mind your margin requirements will go up as you move to 2x and 3x funds compared to unlevered funds.
Yes. So because of “portfolio margin” at interactive brokers, I’m able to use a margin loan on unlevered ETFs for nearly 100% of my leverage. I then use a very small amount of LETFs within my Roth IRA to get the rest of the exposure up to my target portfolio size (1.25mil). IB will let me borrow up to 80 cents for every 20 cents invested with a balanced, unleveraged portfolio. If the market dips, I will buy more LETFs in my iras and pay down the margin loan to prevent margin call.

Basically, what I’m thankful to HEDGEDUNDIE for, is just opening my eyes to risk parity in general. Over the last couple of months I have fine tuned the methods I use to obtain adequate leverage but whether you’re using margin loan, futures or LETFs the philosophy is the same.
Curious what return you expect or assume to get from your unleveraged risk parity portfolio? What withdrawal rate do you plan to use once you reach $1.25M and presumably start drawing down?
I am a pessimist. However, PV shows an 11.3% CAGR since 1978 using a 50/50 mid cap value/LTT with a sharpe ratio of 0.68. Since I’m using mostly a margin loan for my leverage, I do not have to worry as much about volatility decay. So I am able to tilt strongly towards small/value. When I do use LETFs it’ll be LCB/LTTs as SIamond and hedgefundie have shown that you’re better off with a lower volatile asset like LCB to reduce volatility decay. So, I’m HOPING for maybe a 10% CAGR but if it’s less then it’s less.

I plan on using a variable w/d strategy instead of a fixed w/d as this portfolio is more to supplement my income vs fund a complete retirement, so I can adjust my lifestyle with variable distributions. According to PV, I should be able to expect anywhere from 6-8%/year with a 50/50 MCV/LTT portfolio as a variable w/d that would give me, historically, a 50% chance of portfolio increasing in value and a 50% chance of decreasing in value. So that’s my “ceiling” as far as withdrawals. Hopefully, after I’ve hit 1.25mil, I won’t be withdrawing anymore than like 3-4% variable but I figure my ceiling will be in the 6-8%. The more SCV tilt I have the closer it’d be to 8%.

This is all based on data from 1978 which is not long of a time period and I am well aware of dropping rates since then. My larger focus rather on what return I’ll get is what portfolio offers the highest expected sharpe ratio, select that portfolio and then titrate with leverage to get an adequate return. I BELIEVE a 50/50 portfolio of SCV/LTT would have the highest sharpe ratio but if someone has better data please do let me know.

I also have 20% in emerging markets SCV (DGS) just bc I am not comfortable completely ignoring INTL equities. I do not know if this will help or hinder the sharpe ratio but it’ll help me sleep better so there’s that.
I read your posts with interested, as I too have been thinking about combining risk parity, leverage, factor tilts, and a life cycle approach - which I understand is what you are attempting, right? I personally however believe that your return assumptions are far too optimistic. If you have 50/50 equities / LTT, the LTT portion i.e. 50% of your portfolio will likely contribute approximately nothing to your real returns during the next few decades; even if long term interest rates were to drop from 2-2.5% to zero, this would result in a one-time boost of maybe 40%, but zero nominal thereafter, so however you slice and dice it, the real return will be close to nothing, in striking contrast with the CAGR of LTTs from 1978 to now which I would say is not repeatable from current levels with 100% mathematical likelihood, so that for the bonds portion even attempting to extrapolate the past seems meaningless to me.

However, with your assumptions I am interested what your life cycle glide path algorithm for deleveraging is to get you from 2 x leverage to no leverage within 5-10 years? In other words, do you have a specific algorithm in mind for applying dividends, interest, and capital gains to your margin loan?
So I think the first point is that a 50/50 portfolio, unleveraged, historically has had a high sharpe ratio and I don’t foresee that changing in the future. The reason is not because LTTs have provided significant return rather they have had negative correlation with equities particularly during market crashes. You are diversifying your exposure to different risk premiums (equity risk and interest rate risk) which have low to negative correlation with each other thus reducing your overall volatility and hopefully increasing your sharpe ratio. So I wouldn’t focus just on LTTs rather look at the portfolio as a whole.

So I am using the life cycle method where I have my target portfolio which I hope to essentially “retire” with and have used leverage to gain exposure to that exact portfolio today. This means I am essentially 2:1 levered today using a combo of LETFs and a margin loan from IB. The margin loan is my preference as it is cheap and you don’t have to worry about volatility decay (unless you had a margin call). So, I have the margin loan essentially “maxed out” and then I use LETFs to gain the rest of my needed leverage. I am currently able to do this with just 2x LETFs in part of my Roth IRA, if the market dips I’ll need to sell off some stocks in my margin account to prevent a margin call and then increase the amount of LETFs in my Roth to keep the portfolio size constant. Over time, as I rebalance quarterly, as my portfolio grows in size I will convert LETFs to unlevered ETFs and then after that is complete I will sell stocks to pay down my margin loan until I am 100% unlevered. I hope to have this completed within 10 years but not in a rush, the market will do whatever it’s going to do. The good news is that as the portfolio grows I become less and less leveraged.

comeinvest
Posts: 291
Joined: Mon Mar 12, 2012 6:57 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by comeinvest » Sat Jun 29, 2019 5:01 am

privatefarmer wrote:
Fri Jun 28, 2019 10:44 pm
comeinvest wrote:
Fri Jun 28, 2019 5:26 am

I read your posts with interested, as I too have been thinking about combining risk parity, leverage, factor tilts, and a life cycle approach - which I understand is what you are attempting, right? I personally however believe that your return assumptions are far too optimistic. If you have 50/50 equities / LTT, the LTT portion i.e. 50% of your portfolio will likely contribute approximately nothing to your real returns during the next few decades; even if long term interest rates were to drop from 2-2.5% to zero, this would result in a one-time boost of maybe 40%, but zero nominal thereafter, so however you slice and dice it, the real return will be close to nothing, in striking contrast with the CAGR of LTTs from 1978 to now which I would say is not repeatable from current levels with 100% mathematical likelihood, so that for the bonds portion even attempting to extrapolate the past seems meaningless to me.

However, with your assumptions I am interested what your life cycle glide path algorithm for deleveraging is to get you from 2 x leverage to no leverage within 5-10 years? In other words, do you have a specific algorithm in mind for applying dividends, interest, and capital gains to your margin loan?
So I think the first point is that a 50/50 portfolio, unleveraged, historically has had a high sharpe ratio and I don’t foresee that changing in the future. The reason is not because LTTs have provided significant return rather they have had negative correlation with equities particularly during market crashes. You are diversifying your exposure to different risk premiums (equity risk and interest rate risk) which have low to negative correlation with each other thus reducing your overall volatility and hopefully increasing your sharpe ratio. So I wouldn’t focus just on LTTs rather look at the portfolio as a whole.

So I am using the life cycle method where I have my target portfolio which I hope to essentially “retire” with and have used leverage to gain exposure to that exact portfolio today. This means I am essentially 2:1 levered today using a combo of LETFs and a margin loan from IB. The margin loan is my preference as it is cheap and you don’t have to worry about volatility decay (unless you had a margin call). So, I have the margin loan essentially “maxed out” and then I use LETFs to gain the rest of my needed leverage. I am currently able to do this with just 2x LETFs in part of my Roth IRA, if the market dips I’ll need to sell off some stocks in my margin account to prevent a margin call and then increase the amount of LETFs in my Roth to keep the portfolio size constant. Over time, as I rebalance quarterly, as my portfolio grows in size I will convert LETFs to unlevered ETFs and then after that is complete I will sell stocks to pay down my margin loan until I am 100% unlevered. I hope to have this completed within 10 years but not in a rush, the market will do whatever it’s going to do. The good news is that as the portfolio grows I become less and less leveraged.
Thank you. So your strategy is to keep the nominal portfolio size constant (1.25 mil) by selling securities (or converting the leveraged products to non-leveraged) whenever the combined portfolio has risen, but do nothing (in terms of portfolio size; i.e. other than switching between margin and leveraged products in various accounts if necessary) and "sit it out" whenever the market drops, i.e. no redemption until the asset size is back at 1.25 mil? If yes, have you calculated the maximum equity drawdown that this algorithm could sustain (with reasonable assumptions on how the LTTs behave during the equity drawdown) without forced liquidations? Sorry if this has been discussed earlier as I did not read the entire 45 pages and 2226 posts. The most critical periods are probably the first few years with the highest leverage.
On another note, you probably realized that the 1.25 mil will not be worth 1.25 mil in real terms due to inflation, I assume you adjusted your target number accordingly.

User avatar
privatefarmer
Posts: 696
Joined: Mon Sep 08, 2014 2:45 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by privatefarmer » Sat Jun 29, 2019 5:37 am

comeinvest wrote:
Sat Jun 29, 2019 5:01 am
privatefarmer wrote:
Fri Jun 28, 2019 10:44 pm
comeinvest wrote:
Fri Jun 28, 2019 5:26 am

I read your posts with interested, as I too have been thinking about combining risk parity, leverage, factor tilts, and a life cycle approach - which I understand is what you are attempting, right? I personally however believe that your return assumptions are far too optimistic. If you have 50/50 equities / LTT, the LTT portion i.e. 50% of your portfolio will likely contribute approximately nothing to your real returns during the next few decades; even if long term interest rates were to drop from 2-2.5% to zero, this would result in a one-time boost of maybe 40%, but zero nominal thereafter, so however you slice and dice it, the real return will be close to nothing, in striking contrast with the CAGR of LTTs from 1978 to now which I would say is not repeatable from current levels with 100% mathematical likelihood, so that for the bonds portion even attempting to extrapolate the past seems meaningless to me.

However, with your assumptions I am interested what your life cycle glide path algorithm for deleveraging is to get you from 2 x leverage to no leverage within 5-10 years? In other words, do you have a specific algorithm in mind for applying dividends, interest, and capital gains to your margin loan?
So I think the first point is that a 50/50 portfolio, unleveraged, historically has had a high sharpe ratio and I don’t foresee that changing in the future. The reason is not because LTTs have provided significant return rather they have had negative correlation with equities particularly during market crashes. You are diversifying your exposure to different risk premiums (equity risk and interest rate risk) which have low to negative correlation with each other thus reducing your overall volatility and hopefully increasing your sharpe ratio. So I wouldn’t focus just on LTTs rather look at the portfolio as a whole.

So I am using the life cycle method where I have my target portfolio which I hope to essentially “retire” with and have used leverage to gain exposure to that exact portfolio today. This means I am essentially 2:1 levered today using a combo of LETFs and a margin loan from IB. The margin loan is my preference as it is cheap and you don’t have to worry about volatility decay (unless you had a margin call). So, I have the margin loan essentially “maxed out” and then I use LETFs to gain the rest of my needed leverage. I am currently able to do this with just 2x LETFs in part of my Roth IRA, if the market dips I’ll need to sell off some stocks in my margin account to prevent a margin call and then increase the amount of LETFs in my Roth to keep the portfolio size constant. Over time, as I rebalance quarterly, as my portfolio grows in size I will convert LETFs to unlevered ETFs and then after that is complete I will sell stocks to pay down my margin loan until I am 100% unlevered. I hope to have this completed within 10 years but not in a rush, the market will do whatever it’s going to do. The good news is that as the portfolio grows I become less and less leveraged.
Thank you. So your strategy is to keep the nominal portfolio size constant (1.25 mil) by selling securities (or converting the leveraged products to non-leveraged) whenever the combined portfolio has risen, but do nothing (in terms of portfolio size; i.e. other than switching between margin and leveraged products in various accounts if necessary) and "sit it out" whenever the market drops, i.e. no redemption until the asset size is back at 1.25 mil? If yes, have you calculated the maximum equity drawdown that this algorithm could sustain (with reasonable assumptions on how the LTTs behave during the equity drawdown) without forced liquidations? Sorry if this has been discussed earlier as I did not read the entire 45 pages and 2226 posts. The most critical periods are probably the first few years with the highest leverage.
On another note, you probably realized that the 1.25 mil will not be worth 1.25 mil in real terms due to inflation, I assume you adjusted your target number accordingly.
yes that is basically what I am doing. I will first convert LETFs--> unlevered ETFs and then when thats complete I will start to sell funds/pay down my margin as the portfolio grows.

I currently have enough "cushion" in my margin account to allow for a 15% drawdown before hitting a margin call. if it comes close to that, I will sell off some funds/pay down some margin and then, in my roth IRA, exchange some unlevered ETFs for LETFs (or exchange my 2x LETFs for 3x LETFs) to maintain a constant exposure of 1.25mil. The risk of a margin call is essentially zero since I have a significant amount of unlevered funds that could be converted to 3x funds if needed, so I can easily sell off/pay down my margin loan without losing leverage.

And, yes, I will increase that 1.25mil target if my spending increases. I came up w/ 1.25mil based on my monthly spending requirements and a monte carlo simulation, so if my spending goes up w/ inflation then the portfolio size needed will also increase.

This is very similar to the recommendations of the "lifecycle investing" book however they recommend : start with 200% equities --> 100% equities --> target portfolio of stocks/bonds.

comeinvest
Posts: 291
Joined: Mon Mar 12, 2012 6:57 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by comeinvest » Sat Jun 29, 2019 6:15 am

privatefarmer wrote:
Sat Jun 29, 2019 5:37 am
comeinvest wrote:
Sat Jun 29, 2019 5:01 am
privatefarmer wrote:
Fri Jun 28, 2019 10:44 pm
comeinvest wrote:
Fri Jun 28, 2019 5:26 am

I read your posts with interested, as I too have been thinking about combining risk parity, leverage, factor tilts, and a life cycle approach - which I understand is what you are attempting, right? I personally however believe that your return assumptions are far too optimistic. If you have 50/50 equities / LTT, the LTT portion i.e. 50% of your portfolio will likely contribute approximately nothing to your real returns during the next few decades; even if long term interest rates were to drop from 2-2.5% to zero, this would result in a one-time boost of maybe 40%, but zero nominal thereafter, so however you slice and dice it, the real return will be close to nothing, in striking contrast with the CAGR of LTTs from 1978 to now which I would say is not repeatable from current levels with 100% mathematical likelihood, so that for the bonds portion even attempting to extrapolate the past seems meaningless to me.

However, with your assumptions I am interested what your life cycle glide path algorithm for deleveraging is to get you from 2 x leverage to no leverage within 5-10 years? In other words, do you have a specific algorithm in mind for applying dividends, interest, and capital gains to your margin loan?
So I think the first point is that a 50/50 portfolio, unleveraged, historically has had a high sharpe ratio and I don’t foresee that changing in the future. The reason is not because LTTs have provided significant return rather they have had negative correlation with equities particularly during market crashes. You are diversifying your exposure to different risk premiums (equity risk and interest rate risk) which have low to negative correlation with each other thus reducing your overall volatility and hopefully increasing your sharpe ratio. So I wouldn’t focus just on LTTs rather look at the portfolio as a whole.

So I am using the life cycle method where I have my target portfolio which I hope to essentially “retire” with and have used leverage to gain exposure to that exact portfolio today. This means I am essentially 2:1 levered today using a combo of LETFs and a margin loan from IB. The margin loan is my preference as it is cheap and you don’t have to worry about volatility decay (unless you had a margin call). So, I have the margin loan essentially “maxed out” and then I use LETFs to gain the rest of my needed leverage. I am currently able to do this with just 2x LETFs in part of my Roth IRA, if the market dips I’ll need to sell off some stocks in my margin account to prevent a margin call and then increase the amount of LETFs in my Roth to keep the portfolio size constant. Over time, as I rebalance quarterly, as my portfolio grows in size I will convert LETFs to unlevered ETFs and then after that is complete I will sell stocks to pay down my margin loan until I am 100% unlevered. I hope to have this completed within 10 years but not in a rush, the market will do whatever it’s going to do. The good news is that as the portfolio grows I become less and less leveraged.
Thank you. So your strategy is to keep the nominal portfolio size constant (1.25 mil) by selling securities (or converting the leveraged products to non-leveraged) whenever the combined portfolio has risen, but do nothing (in terms of portfolio size; i.e. other than switching between margin and leveraged products in various accounts if necessary) and "sit it out" whenever the market drops, i.e. no redemption until the asset size is back at 1.25 mil? If yes, have you calculated the maximum equity drawdown that this algorithm could sustain (with reasonable assumptions on how the LTTs behave during the equity drawdown) without forced liquidations? Sorry if this has been discussed earlier as I did not read the entire 45 pages and 2226 posts. The most critical periods are probably the first few years with the highest leverage.
On another note, you probably realized that the 1.25 mil will not be worth 1.25 mil in real terms due to inflation, I assume you adjusted your target number accordingly.
yes that is basically what I am doing. I will first convert LETFs--> unlevered ETFs and then when thats complete I will start to sell funds/pay down my margin as the portfolio grows.

I currently have enough "cushion" in my margin account to allow for a 15% drawdown before hitting a margin call. if it comes close to that, I will sell off some funds/pay down some margin and then, in my roth IRA, exchange some unlevered ETFs for LETFs (or exchange my 2x LETFs for 3x LETFs) to maintain a constant exposure of 1.25mil. The risk of a margin call is essentially zero since I have a significant amount of unlevered funds that could be converted to 3x funds if needed, so I can easily sell off/pay down my margin loan without losing leverage.

And, yes, I will increase that 1.25mil target if my spending increases. I came up w/ 1.25mil based on my monthly spending requirements and a monte carlo simulation, so if my spending goes up w/ inflation then the portfolio size needed will also increase.

This is very similar to the recommendations of the "lifecycle investing" book however they recommend : start with 200% equities --> 100% equities --> target portfolio of stocks/bonds.
Thank you. However I still think that with your return assumptions you are "hiding" behind the concept sharpe ratio. With your starting allocation 50% stocks 50% bonds, leveraged 2x, you effectively have (just differently combined, and all percentage numbers expressed as percentage of initial portfolio equity of $625k): 100% equity market exposure, plus 100% bonds, minus 100% cash.
Mathematically we can set parentheses as we like, for example: (100% equity market) + (100% bonds - minus 100% cash).
With current, - relatively - mainstream long-term forecasts of nominal 5% for equities, 2% long term treasuries (based on current yields) and 2% cash (based on current rates assuming fed funds rate stays constant), and 1% spread vs. benchmark rates for your IB loan of 625k and/or similar effective cost of other methods of leverage, your effective cost of leverage is ca. 3%, i.e. the performance of the term in the second pair of parentheses becomes negative, and your grand total expected return on *equity* can be expressed as follows:
5% + 2% - 3%, or (equity market return - 1%).
You would IMHO be better off with putting 100% in an equity index fund (or factor fund of your choice) and be done, and outperform your complex strategy by probably more than 1% annually (and save a whole lot of time implementing your complex life cycle path across multiple accounts). This is simple math and the complex concept of "sharpe ratio" doesn't make this simple math prettier. Granted, you would have some of the dips between now and retirement not buffered by treasuries temporarily moving in the opposite direction; but how much do you care what's happening between now and your target date, as long as the end result is right... and is it worth 1% per annum detraction from performance to you?
To add insult to injury, I think a 100% equity strategy is by magnitudes more tax efficient than your proposed strategy.

I am just realizing that we discussed the same topic a while ago either in this or another thread. It's basically really hard (say impossible) to benefit from leveraging an asset class with cash that yields less (especially after implementation cost) than the cash you are leveraging with. Mentally combining asset classes and leveraging the whole thing doesn't help, as a simple mathematical decomposition into components unveils the negative arbitrage. There are many strategies that worked in the past, but kind of mathematically fell apart since real interest rates and real expected returns from treasuries became negative or close to zero.

User avatar
Steve Reading
Posts: 1906
Joined: Fri Nov 16, 2018 10:20 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by Steve Reading » Sat Jun 29, 2019 9:14 am

privatefarmer wrote:
Sat Jun 29, 2019 5:37 am

yes that is basically what I am doing. I will first convert LETFs--> unlevered ETFs and then when thats complete I will start to sell funds/pay down my margin as the portfolio grows.

I currently have enough "cushion" in my margin account to allow for a 15% drawdown before hitting a margin call. if it comes close to that, I will sell off some funds/pay down some margin and then, in my roth IRA, exchange some unlevered ETFs for LETFs (or exchange my 2x LETFs for 3x LETFs) to maintain a constant exposure of 1.25mil. The risk of a margin call is essentially zero since I have a significant amount of unlevered funds that could be converted to 3x funds if needed, so I can easily sell off/pay down my margin loan without losing leverage.

And, yes, I will increase that 1.25mil target if my spending increases. I came up w/ 1.25mil based on my monthly spending requirements and a monte carlo simulation, so if my spending goes up w/ inflation then the portfolio size needed will also increase.

This is very similar to the recommendations of the "lifecycle investing" book however they recommend : start with 200% equities --> 100% equities --> target portfolio of stocks/bonds.
It looks similar (leverage a portfolio when young and delever as you age) but it's actually very different. Lifecycle Investing operates under the assumption that your income must be accounted for as your portfolio. And for most people, the income is uncorrelated and even negatively correlated from equities (i.e. it's a large bond). So leveraging stocks (and only stocks) is just a way for overcompensating to hit your desired asset allocation. Leveraging isn't technically required either, it just helps hit your desired numbers earlier.

If you worked in the finance industry, where your job might in fact depend on the market, you'd do the opposite. You might be 100% in bonds or even use leveraged bonds.

Leveraging a balanced or risk-paired portfolio is a fine idea (although comeinvest makes an excellent point). But it's missing the whole theory of Lifecycle Investing where you take the nature of your salary into account and discount every future payment to the present to understand how much it's worth. That's really the fundamental part of the strategy.

MotoTrojan
Posts: 9955
Joined: Wed Feb 01, 2017 8:39 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by MotoTrojan » Sat Jun 29, 2019 9:18 am

comeinvest wrote:
Sat Jun 29, 2019 6:15 am
privatefarmer wrote:
Sat Jun 29, 2019 5:37 am
comeinvest wrote:
Sat Jun 29, 2019 5:01 am
privatefarmer wrote:
Fri Jun 28, 2019 10:44 pm
comeinvest wrote:
Fri Jun 28, 2019 5:26 am

I read your posts with interested, as I too have been thinking about combining risk parity, leverage, factor tilts, and a life cycle approach - which I understand is what you are attempting, right? I personally however believe that your return assumptions are far too optimistic. If you have 50/50 equities / LTT, the LTT portion i.e. 50% of your portfolio will likely contribute approximately nothing to your real returns during the next few decades; even if long term interest rates were to drop from 2-2.5% to zero, this would result in a one-time boost of maybe 40%, but zero nominal thereafter, so however you slice and dice it, the real return will be close to nothing, in striking contrast with the CAGR of LTTs from 1978 to now which I would say is not repeatable from current levels with 100% mathematical likelihood, so that for the bonds portion even attempting to extrapolate the past seems meaningless to me.

However, with your assumptions I am interested what your life cycle glide path algorithm for deleveraging is to get you from 2 x leverage to no leverage within 5-10 years? In other words, do you have a specific algorithm in mind for applying dividends, interest, and capital gains to your margin loan?
So I think the first point is that a 50/50 portfolio, unleveraged, historically has had a high sharpe ratio and I don’t foresee that changing in the future. The reason is not because LTTs have provided significant return rather they have had negative correlation with equities particularly during market crashes. You are diversifying your exposure to different risk premiums (equity risk and interest rate risk) which have low to negative correlation with each other thus reducing your overall volatility and hopefully increasing your sharpe ratio. So I wouldn’t focus just on LTTs rather look at the portfolio as a whole.

So I am using the life cycle method where I have my target portfolio which I hope to essentially “retire” with and have used leverage to gain exposure to that exact portfolio today. This means I am essentially 2:1 levered today using a combo of LETFs and a margin loan from IB. The margin loan is my preference as it is cheap and you don’t have to worry about volatility decay (unless you had a margin call). So, I have the margin loan essentially “maxed out” and then I use LETFs to gain the rest of my needed leverage. I am currently able to do this with just 2x LETFs in part of my Roth IRA, if the market dips I’ll need to sell off some stocks in my margin account to prevent a margin call and then increase the amount of LETFs in my Roth to keep the portfolio size constant. Over time, as I rebalance quarterly, as my portfolio grows in size I will convert LETFs to unlevered ETFs and then after that is complete I will sell stocks to pay down my margin loan until I am 100% unlevered. I hope to have this completed within 10 years but not in a rush, the market will do whatever it’s going to do. The good news is that as the portfolio grows I become less and less leveraged.
Thank you. So your strategy is to keep the nominal portfolio size constant (1.25 mil) by selling securities (or converting the leveraged products to non-leveraged) whenever the combined portfolio has risen, but do nothing (in terms of portfolio size; i.e. other than switching between margin and leveraged products in various accounts if necessary) and "sit it out" whenever the market drops, i.e. no redemption until the asset size is back at 1.25 mil? If yes, have you calculated the maximum equity drawdown that this algorithm could sustain (with reasonable assumptions on how the LTTs behave during the equity drawdown) without forced liquidations? Sorry if this has been discussed earlier as I did not read the entire 45 pages and 2226 posts. The most critical periods are probably the first few years with the highest leverage.
On another note, you probably realized that the 1.25 mil will not be worth 1.25 mil in real terms due to inflation, I assume you adjusted your target number accordingly.
yes that is basically what I am doing. I will first convert LETFs--> unlevered ETFs and then when thats complete I will start to sell funds/pay down my margin as the portfolio grows.

I currently have enough "cushion" in my margin account to allow for a 15% drawdown before hitting a margin call. if it comes close to that, I will sell off some funds/pay down some margin and then, in my roth IRA, exchange some unlevered ETFs for LETFs (or exchange my 2x LETFs for 3x LETFs) to maintain a constant exposure of 1.25mil. The risk of a margin call is essentially zero since I have a significant amount of unlevered funds that could be converted to 3x funds if needed, so I can easily sell off/pay down my margin loan without losing leverage.

And, yes, I will increase that 1.25mil target if my spending increases. I came up w/ 1.25mil based on my monthly spending requirements and a monte carlo simulation, so if my spending goes up w/ inflation then the portfolio size needed will also increase.

This is very similar to the recommendations of the "lifecycle investing" book however they recommend : start with 200% equities --> 100% equities --> target portfolio of stocks/bonds.
Thank you. However I still think that with your return assumptions you are "hiding" behind the concept sharpe ratio. With your starting allocation 50% stocks 50% bonds, leveraged 2x, you effectively have (just differently combined, and all percentage numbers expressed as percentage of initial portfolio equity of $625k): 100% equity market exposure, plus 100% bonds, minus 100% cash.
Mathematically we can set parentheses as we like, for example: (100% equity market) + (100% bonds - minus 100% cash).
With current, - relatively - mainstream long-term forecasts of nominal 5% for equities, 2% long term treasuries (based on current yields) and 2% cash (based on current rates assuming fed funds rate stays constant), and 1% spread vs. benchmark rates for your IB loan of 625k and/or similar effective cost of other methods of leverage, your effective cost of leverage is ca. 3%, i.e. the performance of the term in the second pair of parentheses becomes negative, and your grand total expected return on *equity* can be expressed as follows:
5% + 2% - 3%, or (equity market return - 1%).
You would IMHO be better off with putting 100% in an equity index fund (or factor fund of your choice) and be done, and outperform your complex strategy by probably more than 1% annually (and save a whole lot of time implementing your complex life cycle path across multiple accounts). This is simple math and the complex concept of "sharpe ratio" doesn't make this simple math prettier. Granted, you would have some of the dips between now and retirement not buffered by treasuries temporarily moving in the opposite direction; but how much do you care what's happening between now and your target date, as long as the end result is right... and is it worth 1% per annum detraction from performance to you?
To add insult to injury, I think a 100% equity strategy is by magnitudes more tax efficient than your proposed strategy.

I am just realizing that we discussed the same topic a while ago either in this or another thread. It's basically really hard (say impossible) to benefit from leveraging an asset class with cash that yields less (especially after implementation cost) than the cash you are leveraging with. Mentally combining asset classes and leveraging the whole thing doesn't help, as a simple mathematical decomposition into components unveils the negative arbitrage. There are many strategies that worked in the past, but kind of mathematically fell apart since real interest rates and real expected returns from treasuries became negative or close to zero.
To be fair it’s more like 300% bond - 200% cash so even if the bond yield is equal to cash there is a positive expected return. Also I have previously shown that regardless of market period (even 1955-1982) the combined portfolio had a higher CAGR than the weighted average of its components returns, so you are ignoring an efficiency boost which is related to the higher sharpe.

comeinvest
Posts: 291
Joined: Mon Mar 12, 2012 6:57 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by comeinvest » Sat Jun 29, 2019 5:12 pm

MotoTrojan wrote:
Sat Jun 29, 2019 9:18 am
To be fair it’s more like 300% bond - 200% cash so even if the bond yield is equal to cash there is a positive expected return. Also I have previously shown that regardless of market period (even 1955-1982) the combined portfolio had a higher CAGR than the weighted average of its components returns, so you are ignoring an efficiency boost which is related to the higher sharpe.
Not sure where you get your numbers of 300%; I think my numbers were correct, referring to privatefarmer's strategy.
I think interest rates fell since somewhere around 1980. So the interesting backtest period is before 1980. For a static portfolio, the return is the weighted average of its initial components. So the only way the combined portfolio can have a higher CAGR (before leverage) must be from the rebalancing effect. Can the return from rebalancing be more than 1%? I don't have numbers, but I think it would be unlikely and require volatile equity and bond markets.
What is historically the return from rebalancing, a.k.a. the implied "market timing" effect of a rebalancing strategy based on the assumption of reversion to the mean? (I know the intent of rebalancing is to rebalance the risk, but as a side effect it also has a market timing effect.)
I know there were periods of rising interest rates before around 1980; but do we have any historical precedence of zero or negative real interest rates of cash and treasuries? If not, any simple extrapolation or backtesting without adjustments is IMHO meaningless.
I have yet to see one study, simulation, or successful backtest of a risk parity style strategy or any asset allocation for that matter, where one component has negative expected real returns (or returns below the margin cost that the investor can obtain). I would be glad to see one.

User avatar
Steve Reading
Posts: 1906
Joined: Fri Nov 16, 2018 10:20 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by Steve Reading » Sat Jun 29, 2019 7:05 pm

comeinvest wrote:
Sat Jun 29, 2019 5:12 pm
I have yet to see one study, simulation, or successful backtest of a risk parity style strategy or any asset allocation for that matter, where one component has negative expected real returns (or returns below the margin cost that the investor can obtain). I would be glad to see one.
Well to be fair, commodities have an expected real return of 0% before fees. After fund fees, they have a negative expected real return. And they are used in most traditional risk-parity portfolios. So any study/simulation/backtest of a risk-parity portfolio with commodities (i.e. most risk-parity portfolios) has "a component with negative expected real returns".

Clearly, this rebalancing bonus due to the volatility and uncorrelated assets has worked really well in the past, no doubt about it. Up to you how you interpret the information though.

comeinvest
Posts: 291
Joined: Mon Mar 12, 2012 6:57 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by comeinvest » Sat Jun 29, 2019 7:33 pm

305pelusa wrote:
Sat Jun 29, 2019 7:05 pm
comeinvest wrote:
Sat Jun 29, 2019 5:12 pm
I have yet to see one study, simulation, or successful backtest of a risk parity style strategy or any asset allocation for that matter, where one component has negative expected real returns (or returns below the margin cost that the investor can obtain). I would be glad to see one.
Well to be fair, commodities have an expected real return of 0% before fees. After fund fees, they have a negative expected real return. And they are used in most traditional risk-parity portfolios. So any study/simulation/backtest of a risk-parity portfolio with commodities (i.e. most risk-parity portfolios) has "a component with negative expected real returns".

Clearly, this rebalancing bonus due to the volatility and uncorrelated assets has worked really well in the past, no doubt about it. Up to you how you interpret the information though.
... what was the magnitude of the rebalancing bonus in the past? Can we quantify this?

I don't mean to categorically dismiss the validity of the approach, but I'm questioning the utility of backtesting over time periods where parameters were fundamentally different, if not contrary, to today's, without intelligent adjustments or simulations.

Frankly, I did not study the effect of commodities as an asset component with negative real returns; but my gut feeling tells me that it might make a minor contribution to a rebalancing bonus, and/or "smooth out" volatility of the combined asset base, but at the expense of expected total return or expected terminal value of your portfolio. I personally question the utility of "volatility" as a measure of risk for long-term investors, and the utility of smoothing out drops in the market. Investors as sophisticated as those reading 45 pages and more than 2000 posts of this thread trying to optimize their risk-adjusted returns using 6x leverage on risk parity portfolios or similar strategies, will be rational enough and have the mental strength to sit out market drops until markets recover. The risk measure for a rational, long-term investor should be shortfall probability of terminal value.

Having that said, if you can demonstrate, or cite a study, that adding an asset with negative expected return reduces my shortfall risk, or similar measure of risk/return characteristics of my expected terminal value over long time periods, I'm happy to hear about it. Unlike today, I think in the past most asset classes had positive expected return, and I think commodities were typically a small percentage of the asset base in model portfolios.

User avatar
willthrill81
Posts: 19257
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by willthrill81 » Sat Jun 29, 2019 8:09 pm

MotoTrojan wrote:
Sat Jun 22, 2019 11:58 am
willthrill81 wrote:
Sat Jun 22, 2019 11:05 am
willthrill81 wrote:
Sat Jun 22, 2019 12:14 am
Hydromod wrote:
Sat Jun 22, 2019 12:05 am
HEDGEFUNDIE wrote:
Fri Jun 21, 2019 10:53 pm
Glad to see the experimentation, but just wanted to point out that the 1 month inverse volatility strategy has only returned 15% YTD while the original strategy has returned 29%.

Just like with plain ol’ index funds, tracking error is a thing.
Cagr since Jan 2018 was 10.68% for inverse volatility versus 10.01% for the original. Can't catch every zig. We just hope for a larger fraction of good guesses...
As I noted above, a very good argument could be made to diversify your implementation strategy, maybe putting 1/3 in the OP's fixed 40/60 split, 1/3 in inverse volatility, and 1/3 in target volatility. When I get a little time, I crunch the numbers to see how that would have impacted the results; I'm pretty sure it would have smoothed out things a bit.
I ran the results through PV and analyzed them in Excel. The original strategy (i.e. 40/60 with quarterly rebalancing) and the inverse volatility approach with a 20 day timing period were very strongly correlated, with an r of .953, giving an r-square of .908. Hence, there's very little diversification benefit there. However, using a target volatility approach with a 25% target, the correlation between its returns and the original strategy was just .601, with an r-square of .361. So I used a 50/50 split between them since that produces much more diversification. Below are the annual returns, with the results through May, 2019.

2010: 40.10%
2011: 67.74%
2012: 35.87%
2013: 24.44%
2014: 60.79%
2015: -6.61%
2016: 27.70%
2017: 53.28%
2018: -3.29%
2019: 16.82%

This improved significantly on the 'worst year' of both strategies independently (both = -6.61%, original = -16.67%, target volatility = -7.64%) and improves the returns of the original strategy significantly (CAGR both = 29.43%, original = 24.19%, target volatility = 36.61%). Also, the returns of the combined strategies were weakly correlated with the S&P 500 (VFINX) with an r of .331, with an r-square of just .11. So the S&P 500's returns only account for 11% of the variance in the returns of the combined strategies.

Thoughts?
I don’t have the numbers handy but I think your 25% target is on the high end and is biased towards the bull market since fund inception. I believe 18% was more optimal for the longer daily simulated data and doesn’t generate the same outperformance. As to blending, I doubt I’d independently run the different variants (although this would be quite easy if M1 allows splitting one account into 2 pies?) but I pondered splitting the difference by rebalancing monthly to a mid-point between the suggested inverse volatility AA and the long run 40/60 which would net a similar effect and reduce tracking error regret in both directions.
While the results of the 25% target volatility approach were not strongly correlated with stocks, I've noticed a different problem with it: at times over the last decade, you would have been 100% in UPRO. There's no way that I would ever want that to happen as it would leave you wide open for massive sudden declines (i.e. intra-month) with no protection. With your inverse volatility/risk parity approach, however, over the last decade, the monthly allocation for UPRO would never have been higher than 73% and usually between 40% and 60%. This would have sacrificed returns compared to the 25% target volatility approach, but without the constant downside protection of TMF, it seems like too big of a gamble to me.
“It's a dangerous business, Frodo, going out your door. You step onto the road, and if you don't keep your feet, there's no knowing where you might be swept off to.” J.R.R. Tolkien,The Lord of the Rings

User avatar
Steve Reading
Posts: 1906
Joined: Fri Nov 16, 2018 10:20 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by Steve Reading » Sat Jun 29, 2019 8:39 pm

comeinvest wrote:
Sat Jun 29, 2019 7:33 pm

... what was the magnitude of the rebalancing bonus in the past? Can we quantify this?
I'm sure there are ways. Here's a crude example that occurs to me. Using Simba's backtest spreadsheet, gold returned 5.66% from 1970 to 2018. A portfolio of 53% US stocks/47% cash returned the same thing. I added cash to the stocks to "dilute" the returns and achieve the same returns as the gold. If you now add them (portfolio of 50% gold, 26.5% US stocks, 23.5% cash) had a return of 6.52%. So that's right under a full percent just from rebalancing bonus in this very specific scenario.

Please don't take it as "see? Not enough to make up the additional 1% from margin fees in privatefarmer's case". The above is just an example to prove that it existed (the rebalancing bonus).

If you're more interested in the magnitude/ways to quantify it, as well as how it fares with other definitions of risk (like terminal value risk), I suggest you delve into the literature more. You could also use a Monte Carlos simulator and play around with how negative the asset return needs to get, or high positive the correlation can get, before you eliminate the rebalancing bonus completely.

User avatar
privatefarmer
Posts: 696
Joined: Mon Sep 08, 2014 2:45 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by privatefarmer » Sat Jun 29, 2019 9:57 pm

Why would we assume the return of LTTs would be equal to cash? If that were the case why would anyone tie their $$ up for 30 years in a LTT? Because of the added risk, of course the expected return of LTTs is > cash. So you are borrowing at the risk free rate + 1% (at IB) and investing in a risk parity portfolio which historically has had returns slightly less than 100% equities but with far lower drawdowns. I understand the argument that you are borrowing and then essentially investing at least some of that into LTTs which to many seems illogical. However, you have to look at the portfolio from the top down, as a whole. There absolutely is a premium you get from having negatively correlated assets as HEDGEFUNDIE has pointed out in several posts.

There have been numerous periods (for example 6/2003-9/2018) where interest rates were essentially flat that the risk parity portfolio did very well precisely because of the negative correlation btwn LTTs and equities.

User avatar
privatefarmer
Posts: 696
Joined: Mon Sep 08, 2014 2:45 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by privatefarmer » Sun Jun 30, 2019 12:23 am

305pelusa wrote:
Sat Jun 29, 2019 9:14 am
privatefarmer wrote:
Sat Jun 29, 2019 5:37 am

yes that is basically what I am doing. I will first convert LETFs--> unlevered ETFs and then when thats complete I will start to sell funds/pay down my margin as the portfolio grows.

I currently have enough "cushion" in my margin account to allow for a 15% drawdown before hitting a margin call. if it comes close to that, I will sell off some funds/pay down some margin and then, in my roth IRA, exchange some unlevered ETFs for LETFs (or exchange my 2x LETFs for 3x LETFs) to maintain a constant exposure of 1.25mil. The risk of a margin call is essentially zero since I have a significant amount of unlevered funds that could be converted to 3x funds if needed, so I can easily sell off/pay down my margin loan without losing leverage.

And, yes, I will increase that 1.25mil target if my spending increases. I came up w/ 1.25mil based on my monthly spending requirements and a monte carlo simulation, so if my spending goes up w/ inflation then the portfolio size needed will also increase.

This is very similar to the recommendations of the "lifecycle investing" book however they recommend : start with 200% equities --> 100% equities --> target portfolio of stocks/bonds.
It looks similar (leverage a portfolio when young and delever as you age) but it's actually very different. Lifecycle Investing operates under the assumption that your income must be accounted for as your portfolio. And for most people, the income is uncorrelated and even negatively correlated from equities (i.e. it's a large bond). So leveraging stocks (and only stocks) is just a way for overcompensating to hit your desired asset allocation. Leveraging isn't technically required either, it just helps hit your desired numbers earlier.

If you worked in the finance industry, where your job might in fact depend on the market, you'd do the opposite. You might be 100% in bonds or even use leveraged bonds.

Leveraging a balanced or risk-paired portfolio is a fine idea (although comeinvest makes an excellent point). But it's missing the whole theory of Lifecycle Investing where you take the nature of your salary into account and discount every future payment to the present to understand how much it's worth. That's really the fundamental part of the strategy.
Fair points. I guess, in my view, a risk parity portfolio is superior to 100% equities. I understand the lifecycle strategy is to start with 200% equities and then eventually transition to a non-levered balanced portfolio. I would prefer to pick the portfolio with the highest expected sharpe ratio (a risk parity portfolio) and then titrate my leverage to an acceptable amount. So, in theory, I may never completely pay down my leverage, I may get it down to a point that is low enough where I am okay with the volatility and then just keep it constant from there.

But, I think the general idea is that I am picking roughly what my ideal portfolio would be and using leverage to gain exposure to that today and then over the next several years pay down that leverage either to zero or closer to zero, depending on my risk tolerance at that point in time.

comeinvest
Posts: 291
Joined: Mon Mar 12, 2012 6:57 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by comeinvest » Sun Jun 30, 2019 3:05 am

privatefarmer wrote:
Sat Jun 29, 2019 9:57 pm
Why would we assume the return of LTTs would be equal to cash? If that were the case why would anyone tie their $$ up for 30 years in a LTT? Because of the added risk, of course the expected return of LTTs is > cash.
I strongly believe this argument is wrong, assuming that something must have higher returns because it carries risk. First off, whether long term bonds carry more or less risk than cash or short term bonds is debatable and has actually been debated in academic literature, and probably depends on the viewpoint - a long term bond provides a guaranteed stream of income, while a cash investor's outcome over the duration of the bond is dependent on unknown future interest developments over that time frame. Some say this is the reason why long term bonds have historically provided little excess returns over cash.
Secondly, even if we assume that LTTs are more risky - the return of an LTT over its duration between now and its maturity is 100% predictable and known in advance. Interest rates and inflation are not known in advance, but current assumptions made by the market can be derived from spreads between treasuries and inflation protected securities (TIPS), and similar data. TIPS currently yield about half a percent (real yield). With that set of data, the real return of current treasuries until maturity will be close to nothing, and you will lose money if you finance with the predicted cash rates. I guess to make money with LTTs, you would have to assume sustained periods like decades of minimal inflation or deflation, which is inconsistent with current TIPS spreads.
Germany, Japan, Switzerland and many other countries' 10 year government bonds currently have negative nominal yields and are ca. 2% below expected inflation. Why anyone invests in those is beyond me, but I know German life insurances for example are legally required to have most of their assets in government bonds. Most of those life insurers are effectively losing money on the policies they wrote, and would be bankrupt if they were not on life support by their holding companies. Would you still use assets with a real loss of -2% annually for portfolio construction, based on backtesting from the past when they yielded a few percentage points above cash and inflation?
privatefarmer wrote:
Sat Jun 29, 2019 9:57 pm
So you are borrowing at the risk free rate + 1% (at IB) and investing in a risk parity portfolio which historically has had returns slightly less than 100% equities but with far lower drawdowns. I understand the argument that you are borrowing and then essentially investing at least some of that into LTTs which to many seems illogical. However, you have to look at the portfolio from the top down, as a whole. There absolutely is a premium you get from having negatively correlated assets as HEDGEFUNDIE has pointed out in several posts.

There have been numerous periods (for example 6/2003-9/2018) where interest rates were essentially flat that the risk parity portfolio did very well precisely because of the negative correlation btwn LTTs and equities.
In 2003, 10 year treasuries yielded about 4%, and 30 years treasuries about 5.3%. Both several percentage points above inflation and future inflation expectations at that time, and above cash rates. Surely a great constituent of a risk parity portfolio back then, and a great contributor to performance as well as diversification between 2003 and 2018. I would be jumping all in. 2019 completely different story.

"you have to look at the portfolio from the top down, as a whole" - agreed, sometimes this makes sense, depending on what you are analyzing. However, decomposition and pairing as I demonstrated in my earlier post can unveil negative arbitrage (or negative quasi-arbitrage) between pairs with current valuations, which would then have to be logically cancelled out or they would act as nothing but a drag on overall performance, something that is hidden in the holistic view. I do admit that the rebalancing bonus would have to be accounted for, so again, I am not categorically dismissing the approach, just the method of justification by backtesting without intelligent adjustments for current macro data that are fundamentally different from the backtesting period. Let me just say I am very skeptical that the effect of the rebalancing bonus on my risk-adjusted expected terminal value is higher than the loss from the (treasuries minus cash) pair, but happy to see a study, *realistic* backtest, or simulation that proves me wrong. I would be so happy in fact that I would be jumping all in. I'm trying hard to justify it, but I'm not convinced based on the evidence presented so far. With current macroeconomic data and "mainstream" long-term return forecasts, I am seeing less advantageous risk/return characteristics of my terminal portfolio value than e.g. with a passive 100% equity portfolio.
Last edited by comeinvest on Sun Jun 30, 2019 3:48 am, edited 1 time in total.

User avatar
privatefarmer
Posts: 696
Joined: Mon Sep 08, 2014 2:45 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by privatefarmer » Sun Jun 30, 2019 3:46 am

comeinvest wrote:
Sun Jun 30, 2019 3:05 am
privatefarmer wrote:
Sat Jun 29, 2019 9:57 pm
Why would we assume the return of LTTs would be equal to cash? If that were the case why would anyone tie their $$ up for 30 years in a LTT? Because of the added risk, of course the expected return of LTTs is > cash.
I strongly believe this argument is wrong, assuming that something must have higher returns because it carries risk. First off, whether long term bonds carry more or less risk than cash or short term bonds is debatable and has actually been debated in academic literature, and probably depends on the viewpoint - a long term bond provides a guaranteed stream of income, while a cash investor's outcome over the duration of the bond is dependent on unknown future interest developments over that time frame. Some say this is the reason why long term bonds have historically provided little excess returns over cash.
Secondly, even if we assume that LTTs are more risky - the return of an LTT over its duration between now and its maturity is 100% predictable and known in advance. Interest rates and inflation are not known in advance, but current assumptions made by the market can be derived from spreads between treasuries and inflation protected securities (TIPS), and similar data. TIPS currently yield about half a percent (real yield). With that set of data, the real return of current treasuries until maturity will be close to nothing, and you will lose money if you finance with the predicted cash rates. I guess to make money with LTTs, you would have to assume sustained periods like decades of minimal inflation or deflation, which is inconsistent with current TIPS spreads.
Germany, Japan, Switzerland and many other countries' 10 year government bonds currently have negative nominal yields and are ca. 2% below expected inflation. Why anyone invests in those is beyond me, but I know German life insurances for example are legally required to have most of their assets in government bonds. Most of those life insurers are effectively losing money on the policies they wrote, and would be bankrupt if they were not on life support by their holding companies. Would you still use assets with a real loss of -2% annually for portfolio construction, based on backtesting from the past when they yielded a few percentage points above cash and inflation?
privatefarmer wrote:
Sat Jun 29, 2019 9:57 pm
So you are borrowing at the risk free rate + 1% (at IB) and investing in a risk parity portfolio which historically has had returns slightly less than 100% equities but with far lower drawdowns. I understand the argument that you are borrowing and then essentially investing at least some of that into LTTs which to many seems illogical. However, you have to look at the portfolio from the top down, as a whole. There absolutely is a premium you get from having negatively correlated assets as HEDGEFUNDIE has pointed out in several posts.

There have been numerous periods (for example 6/2003-9/2018) where interest rates were essentially flat that the risk parity portfolio did very well precisely because of the negative correlation btwn LTTs and equities.
In 2003, 10 year treasuries yielded about 4%, and 30 years treasuries about 5.3%. Both several percentage points above inflation and future inflation expectations at that time, and above cash rates. Surely a great constituent of a risk parity portfolio back then, and a great contributor to performance as well as diversification between 2003 and 2018. I would be jumping all in. 2019 completely different story.

"you have to look at the portfolio from the top down, as a whole" - agreed, sometimes this makes sense, depending on what you are analyzing. However, decomposition and pairing as I demonstrated in my earlier post can unveil negative arbitrage (or negative quasi-arbitrage) between pairs, which would then have to be logically cancelled out or they would act as nothing but a drag on overall performance, something that is hidden in the holistic view. I do admit that the rebalancing bonus would have to be accounted for, so again, I am not categorically dismissing the approach, just the method of justification by backtesting without intelligent adjustments for current macro data that are fundamentally different from the backtesting period. Let me just say I am very skeptical that the effect of the rebalancing bonus on my expected risk-adjusted terminal value is higher than the loss from the (treasuries minus cash) pair, but happy to see a study, *realistic* backtest, or simulation that proves me wrong. I would be so happy in fact that I would be jumping all in. I'm trying hard to justify it, but I'm not convinced base on the evidence presented so far.
I can understand your reasoning however I think the benefit you would have gotten from the negative correlation between LTTs and equities is not being fully appreciated.

From 6/2003-9/2018, LTT yields started and ended at ~3%.
a 60/40 portfolio had a CAGR of 8.4%, STD Dev 8% sharpe=0.91
100% s/p500 had a CAGR of 9.6%, STD Dev 13% sharpe = 0.68
100% LTTs had a CAGR of 4.95% (why this is higher than the yield I don't know, considering there should not have been any price appreciation), STD Dev 11%

So while the CAGR was reduced from 100% equities it was only reduced slightly and the sharpe increased from 0.68-->0.91. The max DD of 100% equities was 51% vs 29% in the 60/40. If you were to leverage the 60/40 portfolio 2:1 you would have had roughly the same STD Dev but a much higher CAGR vs 100% equities.

comeinvest
Posts: 291
Joined: Mon Mar 12, 2012 6:57 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by comeinvest » Sun Jun 30, 2019 4:54 am

privatefarmer wrote:
Sun Jun 30, 2019 3:46 am
comeinvest wrote:
Sun Jun 30, 2019 3:05 am
privatefarmer wrote:
Sat Jun 29, 2019 9:57 pm
Why would we assume the return of LTTs would be equal to cash? If that were the case why would anyone tie their $$ up for 30 years in a LTT? Because of the added risk, of course the expected return of LTTs is > cash.
I strongly believe this argument is wrong, assuming that something must have higher returns because it carries risk. First off, whether long term bonds carry more or less risk than cash or short term bonds is debatable and has actually been debated in academic literature, and probably depends on the viewpoint - a long term bond provides a guaranteed stream of income, while a cash investor's outcome over the duration of the bond is dependent on unknown future interest developments over that time frame. Some say this is the reason why long term bonds have historically provided little excess returns over cash.
Secondly, even if we assume that LTTs are more risky - the return of an LTT over its duration between now and its maturity is 100% predictable and known in advance. Interest rates and inflation are not known in advance, but current assumptions made by the market can be derived from spreads between treasuries and inflation protected securities (TIPS), and similar data. TIPS currently yield about half a percent (real yield). With that set of data, the real return of current treasuries until maturity will be close to nothing, and you will lose money if you finance with the predicted cash rates. I guess to make money with LTTs, you would have to assume sustained periods like decades of minimal inflation or deflation, which is inconsistent with current TIPS spreads.
Germany, Japan, Switzerland and many other countries' 10 year government bonds currently have negative nominal yields and are ca. 2% below expected inflation. Why anyone invests in those is beyond me, but I know German life insurances for example are legally required to have most of their assets in government bonds. Most of those life insurers are effectively losing money on the policies they wrote, and would be bankrupt if they were not on life support by their holding companies. Would you still use assets with a real loss of -2% annually for portfolio construction, based on backtesting from the past when they yielded a few percentage points above cash and inflation?
privatefarmer wrote:
Sat Jun 29, 2019 9:57 pm
So you are borrowing at the risk free rate + 1% (at IB) and investing in a risk parity portfolio which historically has had returns slightly less than 100% equities but with far lower drawdowns. I understand the argument that you are borrowing and then essentially investing at least some of that into LTTs which to many seems illogical. However, you have to look at the portfolio from the top down, as a whole. There absolutely is a premium you get from having negatively correlated assets as HEDGEFUNDIE has pointed out in several posts.

There have been numerous periods (for example 6/2003-9/2018) where interest rates were essentially flat that the risk parity portfolio did very well precisely because of the negative correlation btwn LTTs and equities.
In 2003, 10 year treasuries yielded about 4%, and 30 years treasuries about 5.3%. Both several percentage points above inflation and future inflation expectations at that time, and above cash rates. Surely a great constituent of a risk parity portfolio back then, and a great contributor to performance as well as diversification between 2003 and 2018. I would be jumping all in. 2019 completely different story.

"you have to look at the portfolio from the top down, as a whole" - agreed, sometimes this makes sense, depending on what you are analyzing. However, decomposition and pairing as I demonstrated in my earlier post can unveil negative arbitrage (or negative quasi-arbitrage) between pairs, which would then have to be logically cancelled out or they would act as nothing but a drag on overall performance, something that is hidden in the holistic view. I do admit that the rebalancing bonus would have to be accounted for, so again, I am not categorically dismissing the approach, just the method of justification by backtesting without intelligent adjustments for current macro data that are fundamentally different from the backtesting period. Let me just say I am very skeptical that the effect of the rebalancing bonus on my expected risk-adjusted terminal value is higher than the loss from the (treasuries minus cash) pair, but happy to see a study, *realistic* backtest, or simulation that proves me wrong. I would be so happy in fact that I would be jumping all in. I'm trying hard to justify it, but I'm not convinced base on the evidence presented so far.
I can understand your reasoning however I think the benefit you would have gotten from the negative correlation between LTTs and equities is not being fully appreciated.

From 6/2003-9/2018, LTT yields started and ended at ~3%.
a 60/40 portfolio had a CAGR of 8.4%, STD Dev 8% sharpe=0.91
100% s/p500 had a CAGR of 9.6%, STD Dev 13% sharpe = 0.68
100% LTTs had a CAGR of 4.95% (why this is higher than the yield I don't know, considering there should not have been any price appreciation), STD Dev 11%

So while the CAGR was reduced from 100% equities it was only reduced slightly and the sharpe increased from 0.68-->0.91. The max DD of 100% equities was 51% vs 29% in the 60/40. If you were to leverage the 60/40 portfolio 2:1 you would have had roughly the same STD Dev but a much higher CAGR vs 100% equities.
"Rebalancing bonus" in your 60/40 example would have been 7.6% - 4.95% = 2.605%.
A data graph that I looked at before had 4% and 5.3% yields in 2003 for 10-year and 30-year treasuries, which might explain the higher CAGR. I think a blend of treasuries currently yields closer to 2% on average, which is the expected return until ("average") maturity with reasonable accuracy. That would be about 3% annually less than 2003-2018, more than the rebalancing bonus in 2003-2018.

Admittedly more meaningful is maybe the spread in current expected returns of the constituent asset classes vs. the realized spread between 2003-2018, in an attempt to compare the "drag" of the lesser performing asset class. Realized spread was 9.6% - 4.95% = 4.65%. Current spread based on asset return forecasts is ca.5% - 2% = 3%, less than in 2003-2018. This would speak in favor of the risk parity strategies.
However, treasury yields are naturally limited on the low end. Although previously assumed limits have been tested and probably broken in the case of European government bonds, they can only go so far below zero before everybody would start hoarding dollar bills. As a result, with treasury rates as low as they are, the buffering effect of treasuries on equity market drawdowns might be a lot more limited than in the past. Conceivably, the expected rebalancing bonus might be reduced likewise.

The rebalancing bonus in 2003-2018 might be exceptionally high because of the large stock market swings up to 2007 and the crash of 2007-2009 and subsequent recovery. I think I heard of excess returns of around 1% due to rebalancing over longer timeframes.

Still not dismissing, but still hesitant to jump on this strategy, for lack meaningful backtesting periods with conditions similar to current conditions. A lot of reasoning that was valid in the past either fell apart, or entails completely different results with current parameters. As nobody knows better than bogleheads, it's too easy to outsmart yourself with complex strategies.

User avatar
Steve Reading
Posts: 1906
Joined: Fri Nov 16, 2018 10:20 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by Steve Reading » Sun Jun 30, 2019 7:35 am

privatefarmer wrote:
Sun Jun 30, 2019 12:23 am
305pelusa wrote:
Sat Jun 29, 2019 9:14 am
privatefarmer wrote:
Sat Jun 29, 2019 5:37 am

yes that is basically what I am doing. I will first convert LETFs--> unlevered ETFs and then when thats complete I will start to sell funds/pay down my margin as the portfolio grows.

I currently have enough "cushion" in my margin account to allow for a 15% drawdown before hitting a margin call. if it comes close to that, I will sell off some funds/pay down some margin and then, in my roth IRA, exchange some unlevered ETFs for LETFs (or exchange my 2x LETFs for 3x LETFs) to maintain a constant exposure of 1.25mil. The risk of a margin call is essentially zero since I have a significant amount of unlevered funds that could be converted to 3x funds if needed, so I can easily sell off/pay down my margin loan without losing leverage.

And, yes, I will increase that 1.25mil target if my spending increases. I came up w/ 1.25mil based on my monthly spending requirements and a monte carlo simulation, so if my spending goes up w/ inflation then the portfolio size needed will also increase.

This is very similar to the recommendations of the "lifecycle investing" book however they recommend : start with 200% equities --> 100% equities --> target portfolio of stocks/bonds.
It looks similar (leverage a portfolio when young and delever as you age) but it's actually very different. Lifecycle Investing operates under the assumption that your income must be accounted for as your portfolio. And for most people, the income is uncorrelated and even negatively correlated from equities (i.e. it's a large bond). So leveraging stocks (and only stocks) is just a way for overcompensating to hit your desired asset allocation. Leveraging isn't technically required either, it just helps hit your desired numbers earlier.

If you worked in the finance industry, where your job might in fact depend on the market, you'd do the opposite. You might be 100% in bonds or even use leveraged bonds.

Leveraging a balanced or risk-paired portfolio is a fine idea (although comeinvest makes an excellent point). But it's missing the whole theory of Lifecycle Investing where you take the nature of your salary into account and discount every future payment to the present to understand how much it's worth. That's really the fundamental part of the strategy.
Fair points. I guess, in my view, a risk parity portfolio is superior to 100% equities. I understand the lifecycle strategy is to start with 200% equities and then eventually transition to a non-levered balanced portfolio. I would prefer to pick the portfolio with the highest expected sharpe ratio (a risk parity portfolio) and then titrate my leverage to an acceptable amount. So, in theory, I may never completely pay down my leverage, I may get it down to a point that is low enough where I am okay with the volatility and then just keep it constant from there.

But, I think the general idea is that I am picking roughly what my ideal portfolio would be and using leverage to gain exposure to that today and then over the next several years pay down that leverage either to zero or closer to zero, depending on my risk tolerance at that point in time.
You know how comeinvest keeps saying "look at the cost of leverage and the LT bonds... that's a negative return" and you tell him to look at the entire portfolio? I am asking you the same thing here. Don't just look at the portfolio; look at all of your sources of revenue as your portfolio (including salary). So even though risk parity seems superior to 200% equities, the question is whether it's superior when a person has a salary with bond-like characteristics.

The reason you like LT bonds is to have something that goes up when equities go down. If your salary keeps providing a stream during recessions, then you have a crude risk-parity type of portfolio when you're 200% in equities. In fact, when starting, 200% equities will not even be enough.


The point I'm trying to make is the following: You wouldn't make a portfolio by just adding the assets with the highest Sharpe. What matters is that the portfolio has the highest Sharpe, even if made up by suboptimal assets. But take that further. I'm claiming you also shouldn't just choose a portfolio because it has the highest Sharpe; what matters is a more holistic approach where you take income, mortgage, horizon, leverage, spouse, etc into account to create the financial plan with the highest Sharpe. That's what Lifecycle Investing is about. This takes much more work and effort, but I believe is overall superior. And for younger people with jobs uncorrelated with the market, it will almost certainly lead to 100-200% in equities.

User avatar
privatefarmer
Posts: 696
Joined: Mon Sep 08, 2014 2:45 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by privatefarmer » Sun Jun 30, 2019 7:40 am

305pelusa wrote:
Sun Jun 30, 2019 7:35 am
privatefarmer wrote:
Sun Jun 30, 2019 12:23 am
305pelusa wrote:
Sat Jun 29, 2019 9:14 am
privatefarmer wrote:
Sat Jun 29, 2019 5:37 am

yes that is basically what I am doing. I will first convert LETFs--> unlevered ETFs and then when thats complete I will start to sell funds/pay down my margin as the portfolio grows.

I currently have enough "cushion" in my margin account to allow for a 15% drawdown before hitting a margin call. if it comes close to that, I will sell off some funds/pay down some margin and then, in my roth IRA, exchange some unlevered ETFs for LETFs (or exchange my 2x LETFs for 3x LETFs) to maintain a constant exposure of 1.25mil. The risk of a margin call is essentially zero since I have a significant amount of unlevered funds that could be converted to 3x funds if needed, so I can easily sell off/pay down my margin loan without losing leverage.

And, yes, I will increase that 1.25mil target if my spending increases. I came up w/ 1.25mil based on my monthly spending requirements and a monte carlo simulation, so if my spending goes up w/ inflation then the portfolio size needed will also increase.

This is very similar to the recommendations of the "lifecycle investing" book however they recommend : start with 200% equities --> 100% equities --> target portfolio of stocks/bonds.
It looks similar (leverage a portfolio when young and delever as you age) but it's actually very different. Lifecycle Investing operates under the assumption that your income must be accounted for as your portfolio. And for most people, the income is uncorrelated and even negatively correlated from equities (i.e. it's a large bond). So leveraging stocks (and only stocks) is just a way for overcompensating to hit your desired asset allocation. Leveraging isn't technically required either, it just helps hit your desired numbers earlier.

If you worked in the finance industry, where your job might in fact depend on the market, you'd do the opposite. You might be 100% in bonds or even use leveraged bonds.

Leveraging a balanced or risk-paired portfolio is a fine idea (although comeinvest makes an excellent point). But it's missing the whole theory of Lifecycle Investing where you take the nature of your salary into account and discount every future payment to the present to understand how much it's worth. That's really the fundamental part of the strategy.
Fair points. I guess, in my view, a risk parity portfolio is superior to 100% equities. I understand the lifecycle strategy is to start with 200% equities and then eventually transition to a non-levered balanced portfolio. I would prefer to pick the portfolio with the highest expected sharpe ratio (a risk parity portfolio) and then titrate my leverage to an acceptable amount. So, in theory, I may never completely pay down my leverage, I may get it down to a point that is low enough where I am okay with the volatility and then just keep it constant from there.

But, I think the general idea is that I am picking roughly what my ideal portfolio would be and using leverage to gain exposure to that today and then over the next several years pay down that leverage either to zero or closer to zero, depending on my risk tolerance at that point in time.
You know how comeinvest keeps saying "look at the cost of leverage and the LT bonds... that's a negative return" and you tell him to look at the entire portfolio? I am asking you the same thing here. Don't just look at the portfolio; look at all of your sources of revenue as your portfolio (including salary). So even though risk parity seems superior to 200% equities, the question is whether it's superior when a person has a salary with bond-like characteristics.

The reason you like LT bonds is to have something that goes up when equities go down. If your salary keeps providing a stream during recessions, then you have a crude risk-parity type of portfolio when you're 200% in equities. In fact, when starting, 200% equities will not even be enough.


The point I'm trying to make is the following: You wouldn't make a portfolio by just adding the assets with the highest Sharpe. What matters is that the portfolio has the highest Sharpe, even if made up by suboptimal assets. But take that further. I'm claiming you also shouldn't just choose a portfolio because it has the highest Sharpe; what matters is a more holistic approach where you take income, mortgage, horizon, leverage, spouse, etc into account to create the financial plan with the highest Sharpe. That's what Lifecycle Investing is about. This takes much more work and effort, but I believe is overall superior. And for younger people with jobs uncorrelated with the market, it will almost certainly lead to 100-200% in equities.
Good point. I will have to sit down and do the math. I can say however that my savings now are mostly a “drop in the bucket” relative to my portfolio size. The growth of the portfolio should vastly dwarf my future savings but it should be taken into account regardless.

MotoTrojan
Posts: 9955
Joined: Wed Feb 01, 2017 8:39 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by MotoTrojan » Sun Jun 30, 2019 10:36 am

privatefarmer wrote:
Sun Jun 30, 2019 3:46 am


From 6/2003-9/2018, LTT yields started and ended at ~3%.
a 60/40 portfolio had a CAGR of 8.4%, STD Dev 8% sharpe=0.91
100% s/p500 had a CAGR of 9.6%, STD Dev 13% sharpe = 0.68
100% LTTs had a CAGR of 4.95% (why this is higher than the yield I don't know, considering there should not have been any price appreciation), STD Dev 11%

It actually makes perfect sense that CAGR is higher than yield if it smoothly increased in rates and then decreased with no price change, no? Should equal the weighted average yield.

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

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by HEDGEFUNDIE » Sun Jun 30, 2019 11:06 pm

The strategy’s 2019 return has hit 40.5%, compared to 18.5% for the S&P.

10 year return of 28.8%, compared to 14.5%.

Strategy shows superior Sharpe ratio in both cases.

https://www.portfoliovisualizer.com/bac ... 0&total3=0

User avatar
privatefarmer
Posts: 696
Joined: Mon Sep 08, 2014 2:45 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by privatefarmer » Sun Jun 30, 2019 11:58 pm

HEDGEFUNDIE wrote:
Sun Jun 30, 2019 11:06 pm
The strategy’s 2019 return has hit 40.5%, compared to 18.5% for the S&P.

10 year return of 28.8%, compared to 14.5%.

Strategy shows superior Sharpe ratio in both cases.

https://www.portfoliovisualizer.com/bac ... 0&total3=0
so what size yacht are you going to get?

comeinvest
Posts: 291
Joined: Mon Mar 12, 2012 6:57 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by comeinvest » Mon Jul 01, 2019 4:53 am

privatefarmer wrote:
Sun Jun 30, 2019 7:40 am
Good point. I will have to sit down and do the math. I can say however that my savings now are mostly a “drop in the bucket” relative to my portfolio size. The growth of the portfolio should vastly dwarf my future savings but it should be taken into account regardless.
I too am interested in an optimal life cycle, rebalanced, risk parity strategy. I hope I'm not retired by the time I'm done with the math ;) Also, fuzzy input parameters to the statistical models, backtests highly dependent on time periods chosen, make the task not easier.

JBeck
Posts: 147
Joined: Fri Apr 15, 2016 4:54 am

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by JBeck » Mon Jul 01, 2019 7:42 am

Deleted
Last edited by JBeck on Wed Jul 03, 2019 7:21 am, edited 1 time in total.

MotoTrojan
Posts: 9955
Joined: Wed Feb 01, 2017 8:39 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by MotoTrojan » Mon Jul 01, 2019 8:47 am

willthrill81 wrote:
Sat Jun 29, 2019 8:09 pm
MotoTrojan wrote:
Sat Jun 22, 2019 11:58 am
willthrill81 wrote:
Sat Jun 22, 2019 11:05 am
willthrill81 wrote:
Sat Jun 22, 2019 12:14 am
Hydromod wrote:
Sat Jun 22, 2019 12:05 am


Cagr since Jan 2018 was 10.68% for inverse volatility versus 10.01% for the original. Can't catch every zig. We just hope for a larger fraction of good guesses...
As I noted above, a very good argument could be made to diversify your implementation strategy, maybe putting 1/3 in the OP's fixed 40/60 split, 1/3 in inverse volatility, and 1/3 in target volatility. When I get a little time, I crunch the numbers to see how that would have impacted the results; I'm pretty sure it would have smoothed out things a bit.
I ran the results through PV and analyzed them in Excel. The original strategy (i.e. 40/60 with quarterly rebalancing) and the inverse volatility approach with a 20 day timing period were very strongly correlated, with an r of .953, giving an r-square of .908. Hence, there's very little diversification benefit there. However, using a target volatility approach with a 25% target, the correlation between its returns and the original strategy was just .601, with an r-square of .361. So I used a 50/50 split between them since that produces much more diversification. Below are the annual returns, with the results through May, 2019.

2010: 40.10%
2011: 67.74%
2012: 35.87%
2013: 24.44%
2014: 60.79%
2015: -6.61%
2016: 27.70%
2017: 53.28%
2018: -3.29%
2019: 16.82%

This improved significantly on the 'worst year' of both strategies independently (both = -6.61%, original = -16.67%, target volatility = -7.64%) and improves the returns of the original strategy significantly (CAGR both = 29.43%, original = 24.19%, target volatility = 36.61%). Also, the returns of the combined strategies were weakly correlated with the S&P 500 (VFINX) with an r of .331, with an r-square of just .11. So the S&P 500's returns only account for 11% of the variance in the returns of the combined strategies.

Thoughts?
I don’t have the numbers handy but I think your 25% target is on the high end and is biased towards the bull market since fund inception. I believe 18% was more optimal for the longer daily simulated data and doesn’t generate the same outperformance. As to blending, I doubt I’d independently run the different variants (although this would be quite easy if M1 allows splitting one account into 2 pies?) but I pondered splitting the difference by rebalancing monthly to a mid-point between the suggested inverse volatility AA and the long run 40/60 which would net a similar effect and reduce tracking error regret in both directions.
While the results of the 25% target volatility approach were not strongly correlated with stocks, I've noticed a different problem with it: at times over the last decade, you would have been 100% in UPRO. There's no way that I would ever want that to happen as it would leave you wide open for massive sudden declines (i.e. intra-month) with no protection. With your inverse volatility/risk parity approach, however, over the last decade, the monthly allocation for UPRO would never have been higher than 73% and usually between 40% and 60%. This would have sacrificed returns compared to the 25% target volatility approach, but without the constant downside protection of TMF, it seems like too big of a gamble to me.
Exactly. I couldn’t handle 100% UPRO exposure either. Feeling good today about the monthly rebalance to 47% though.

User avatar
willthrill81
Posts: 19257
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by willthrill81 » Mon Jul 01, 2019 8:49 am

MotoTrojan wrote:
Mon Jul 01, 2019 8:47 am
willthrill81 wrote:
Sat Jun 29, 2019 8:09 pm
MotoTrojan wrote:
Sat Jun 22, 2019 11:58 am
willthrill81 wrote:
Sat Jun 22, 2019 11:05 am
willthrill81 wrote:
Sat Jun 22, 2019 12:14 am


As I noted above, a very good argument could be made to diversify your implementation strategy, maybe putting 1/3 in the OP's fixed 40/60 split, 1/3 in inverse volatility, and 1/3 in target volatility. When I get a little time, I crunch the numbers to see how that would have impacted the results; I'm pretty sure it would have smoothed out things a bit.
I ran the results through PV and analyzed them in Excel. The original strategy (i.e. 40/60 with quarterly rebalancing) and the inverse volatility approach with a 20 day timing period were very strongly correlated, with an r of .953, giving an r-square of .908. Hence, there's very little diversification benefit there. However, using a target volatility approach with a 25% target, the correlation between its returns and the original strategy was just .601, with an r-square of .361. So I used a 50/50 split between them since that produces much more diversification. Below are the annual returns, with the results through May, 2019.

2010: 40.10%
2011: 67.74%
2012: 35.87%
2013: 24.44%
2014: 60.79%
2015: -6.61%
2016: 27.70%
2017: 53.28%
2018: -3.29%
2019: 16.82%

This improved significantly on the 'worst year' of both strategies independently (both = -6.61%, original = -16.67%, target volatility = -7.64%) and improves the returns of the original strategy significantly (CAGR both = 29.43%, original = 24.19%, target volatility = 36.61%). Also, the returns of the combined strategies were weakly correlated with the S&P 500 (VFINX) with an r of .331, with an r-square of just .11. So the S&P 500's returns only account for 11% of the variance in the returns of the combined strategies.

Thoughts?
I don’t have the numbers handy but I think your 25% target is on the high end and is biased towards the bull market since fund inception. I believe 18% was more optimal for the longer daily simulated data and doesn’t generate the same outperformance. As to blending, I doubt I’d independently run the different variants (although this would be quite easy if M1 allows splitting one account into 2 pies?) but I pondered splitting the difference by rebalancing monthly to a mid-point between the suggested inverse volatility AA and the long run 40/60 which would net a similar effect and reduce tracking error regret in both directions.
While the results of the 25% target volatility approach were not strongly correlated with stocks, I've noticed a different problem with it: at times over the last decade, you would have been 100% in UPRO. There's no way that I would ever want that to happen as it would leave you wide open for massive sudden declines (i.e. intra-month) with no protection. With your inverse volatility/risk parity approach, however, over the last decade, the monthly allocation for UPRO would never have been higher than 73% and usually between 40% and 60%. This would have sacrificed returns compared to the 25% target volatility approach, but without the constant downside protection of TMF, it seems like too big of a gamble to me.
Exactly. I couldn’t handle 100% UPRO exposure either. Feeling good today about the monthly rebalance to 47% though.
I must admit that it's a bit tempting. UPRO is up almost 25% over the last month alone.
“It's a dangerous business, Frodo, going out your door. You step onto the road, and if you don't keep your feet, there's no knowing where you might be swept off to.” J.R.R. Tolkien,The Lord of the Rings

MotoTrojan
Posts: 9955
Joined: Wed Feb 01, 2017 8:39 pm

Re: HEDGEFUNDIE's excellent adventure [risk parity strategy using 3x leveraged ETFs]

Post by MotoTrojan » Mon Jul 01, 2019 12:54 pm

willthrill81 wrote:
Mon Jul 01, 2019 8:49 am
MotoTrojan wrote:
Mon Jul 01, 2019 8:47 am
willthrill81 wrote:
Sat Jun 29, 2019 8:09 pm
MotoTrojan wrote:
Sat Jun 22, 2019 11:58 am
willthrill81 wrote:
Sat Jun 22, 2019 11:05 am


I ran the results through PV and analyzed them in Excel. The original strategy (i.e. 40/60 with quarterly rebalancing) and the inverse volatility approach with a 20 day timing period were very strongly correlated, with an r of .953, giving an r-square of .908. Hence, there's very little diversification benefit there. However, using a target volatility approach with a 25% target, the correlation between its returns and the original strategy was just .601, with an r-square of .361. So I used a 50/50 split between them since that produces much more diversification. Below are the annual returns, with the results through May, 2019.

2010: 40.10%
2011: 67.74%
2012: 35.87%
2013: 24.44%
2014: 60.79%
2015: -6.61%
2016: 27.70%
2017: 53.28%
2018: -3.29%
2019: 16.82%

This improved significantly on the 'worst year' of both strategies independently (both = -6.61%, original = -16.67%, target volatility = -7.64%) and improves the returns of the original strategy significantly (CAGR both = 29.43%, original = 24.19%, target volatility = 36.61%). Also, the returns of the combined strategies were weakly correlated with the S&P 500 (VFINX) with an r of .331, with an r-square of just .11. So the S&P 500's returns only account for 11% of the variance in the returns of the combined strategies.

Thoughts?
I don’t have the numbers handy but I think your 25% target is on the high end and is biased towards the bull market since fund inception. I believe 18% was more optimal for the longer daily simulated data and doesn’t generate the same outperformance. As to blending, I doubt I’d independently run the different variants (although this would be quite easy if M1 allows splitting one account into 2 pies?) but I pondered splitting the difference by rebalancing monthly to a mid-point between the suggested inverse volatility AA and the long run 40/60 which would net a similar effect and reduce tracking error regret in both directions.
While the results of the 25% target volatility approach were not strongly correlated with stocks, I've noticed a different problem with it: at times over the last decade, you would have been 100% in UPRO. There's no way that I would ever want that to happen as it would leave you wide open for massive sudden declines (i.e. intra-month) with no protection. With your inverse volatility/risk parity approach, however, over the last decade, the monthly allocation for UPRO would never have been higher than 73% and usually between 40% and 60%. This would have sacrificed returns compared to the 25% target volatility approach, but without the constant downside protection of TMF, it seems like too big of a gamble to me.
Exactly. I couldn’t handle 100% UPRO exposure either. Feeling good today about the monthly rebalance to 47% though.
I must admit that it's a bit tempting. UPRO is up almost 25% over the last month alone.
I’d consider it with daily rebalancing and an exponential look back period (weights recent history higher).

Locked