Larry Swedroe: Time Series Momentum (aka Trend Following): A Good Time For A Refresh

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
Random Walker
Posts: 2706
Joined: Fri Feb 23, 2007 8:21 pm

Larry Swedroe: Time Series Momentum (aka Trend Following): A Good Time For A Refresh

Post by Random Walker » Fri Feb 09, 2018 8:29 am

https://alphaarchitect.com/2018/02/08/t ... -evidence/

Excellent review of TS Momentum. It meets all of Larry’s criteria for a factor: persistent, pervasive, robust, intuitive, investable. It’s definitly behavioral; perhaps that’s what makes it powerful. The “smile curve” is what convinced me to invest in TS Momentum across asset classes.

Dave

kagantx
Posts: 18
Joined: Sun Feb 04, 2018 1:39 am

Re: Larry Swedroe: Time Series Momentum (aka Trend Following): A Good Time For A Refresh

Post by kagantx » Fri Feb 09, 2018 9:24 am

Has Larry looked at the returns on the AQRIX and AQRMX funds? They're awful over the past 10 years. I can't figure out how he can recommend them or what resemblance they have to the graph in his main article.

Random Walker
Posts: 2706
Joined: Fri Feb 23, 2007 8:21 pm

Re: Larry Swedroe: Time Series Momentum (aka Trend Following): A Good Time For A Refresh

Post by Random Walker » Fri Feb 09, 2018 9:37 am

Couldn’t find AQRMX, but did find AQRIX. I can’t speak for Larry. Nonetheless I see AQRIX appears to be a long only risk parity fund, investing more in less volatile/less risky/lower expected return assets and less in more volatile/more risky/higher expected return assets. I do think Larry is a fan of risk parity in general, but I don’t know that he or his firm recommends this fund.

Dave
Last edited by Random Walker on Fri Feb 09, 2018 3:21 pm, edited 1 time in total.

Theoretical
Posts: 1262
Joined: Tue Aug 19, 2014 10:09 pm

Re: Larry Swedroe: Time Series Momentum (aka Trend Following): A Good Time For A Refresh

Post by Theoretical » Fri Feb 09, 2018 2:40 pm

kagantx wrote:
Fri Feb 09, 2018 9:24 am
Has Larry looked at the returns on the AQRIX and AQRMX funds? They're awful over the past 10 years. I can't figure out how he can recommend them or what resemblance they have to the graph in his main article.
Do you mean the Managed Futures funds or the Risk parity ones? AQR uses elements of risk parity for the MF and style premia funds - mostly as a risk balancing strategy - i.e. They may go 5X leverage on t-bills or 2 year treasuries but may only be 1-2x leverage in equities.

User avatar
nisiprius
Advisory Board
Posts: 35896
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: Larry Swedroe: Time Series Momentum (aka Trend Following): A Good Time For A Refresh

Post by nisiprius » Fri Feb 09, 2018 4:59 pm

It seems fairly clear that he is identifying this strategy with the phrase "managed futures," since he writes
Strategies that attempt to capture the return premium offered by time-series momentum are often called, “managed futures,” as they take long and short positions in assets via futures markets — ideally in a multitude of futures markets around the globe.
My cynical guess is that, like many trendy new strategies, it is going to turn out that there simply are no managed futures fund with a long track record, let's say 15 years, that we could look at to find out how this strategy has worked in the real world... just as, when "commodifies" (CCF) funds were being widely recommended circa 2010 or so, it was hard to find any older than PCRIX, which started in 2008.

First question, since he cites an AQR paper, are there any AQR "managed futures" (as opposed to risk parity) funds? And, since one of his charts compares managed futures to "diversified passive," how has it compared to the moderate (60/40) model portfolio Larry recommended in his 1998 book, which can be described as "diversified passive."

There is one such AQR managed futures fund, the AQR Managed Futures Strategy Fund, AQMIX.

As I feared, it is only eight years old. And, here it is, (blue), compared to the diversified passive model portfolio Larry suggested in 1998 (red), implemented with Vanguard funds as much as possible, for maximum passivity.

Source

Image

But, someone might say, it shouldn't be considered in isolation. Although figure 3 in the paper is implicitly comparing 100% "time series momentum" to a "diversified passive" portfolio, perhaps that's too extreme. We see that in our illustration above, AQMIX had low--slightly negative--correlation with stocks. Would adding a moderate allocation to AQMIX, say 15%, improve the "diversified passive" portfolio?

Portfolio 3, orange, is a modified version of the "diversified passive" portfolio, with about a 15% allocation to AQMIX, and the rest reduced proportionately.

Source
Image

It was not an improvement.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.

User avatar
Lauretta
Posts: 656
Joined: Wed Jul 05, 2017 6:27 am

Re: Larry Swedroe: Time Series Momentum (aka Trend Following): A Good Time For A Refresh

Post by Lauretta » Fri Feb 09, 2018 5:14 pm

For what it's worth to Bogleheads in the US, I have been told there's a very good firm in France, which is run mainly by scientists, and which uses quantitative strategies which include TSM, (though I don't know whether their funds are available in the US):
https://www.cfm.fr/who-we-are/#Our%20vision
In this paper they document the robustness of trend following in different asset classes going back to 1800
https://arxiv.org/pdf/1404.3274.pdf
When everyone is thinking the same, no one is thinking at all

User avatar
nisiprius
Advisory Board
Posts: 35896
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: Larry Swedroe: Time Series Momentum (aka Trend Following): A Good Time For A Refresh

Post by nisiprius » Fri Feb 09, 2018 5:24 pm

So, the next question. Are there any managed futures fund with a much longer track record? One that takes us back before 2008-2009 for at least a few years? I went to Morningstar's basic fund screener, asked for "managed futures" funds, and as a quick screen for older funds, I asked for funds with a manager tenure of more than five years. It found me 17, but most are different share classes of just five funds:

AQR Managed Futures Strat Managed Futures
Equinox MutualHedge Futur Managed Futures
Guggenheim Managed Future Managed Futures
Salient Trend I Managed Futures
SFG Futures Strategy I Managed Futures

I'll use I class for each of them and list their inception dates.

AQMIX, 1/6/2010
MHFIX, 12/31/2009
RYIFX, 02/09/2008
SPTIX, 1/2/2013
EFSIX, 12/2011

There's only fund that's just barely, more than ten years old, and thus none that gives a good view that includes periods of time before and after 2008-2009. Three, AQMIX, MHFIX, and RYIFX are about the same age.

Just as a quick reality check, let's compare MHFIX and RYIFX to AQMIX to see whether there were any huge, obvious differences, and we'll also include Morningstar's two benchmarks... and, finally, the Vanguard LifeStrategy Moderate fund (a one-fund "diversified passive" portfolio)

The dark green line above all the others is LifeStrategy Moderate. The cluster of three more or less together is the AQR fund, the Equinox fund, and one of Morningstar's benchmarks.

Source
Image


Over the time period during which Managed Futures funds have existed, it is hard to see that they had anything in particular to offer to an investor in a "diversified passive" portfolio.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.

Random Walker
Posts: 2706
Joined: Fri Feb 23, 2007 8:21 pm

Re: Larry Swedroe: Time Series Momentum (aka Trend Following): A Good Time For A Refresh

Post by Random Walker » Fri Feb 09, 2018 5:30 pm

Nisiprius,
I agree that comparing QMHRX to 60/40 is senseless. The issue is whether it improves a stock and bond portfolio. 60/40 seems like a good comparison from that standpoint. You critique that the fund only has 8 year track record, but you then make a comparison between 60/40 and 60/40+AQMIX. Doesn’t seem fair. We’ve basically had consistent bull market for equities over the entire time period. I like TS Momentum for the potential behavior in an extended bear market for equities. I’m eager to see its effect on a 60/40 portfolio when we can incorporate that into the data.

Dave

Random Walker
Posts: 2706
Joined: Fri Feb 23, 2007 8:21 pm

Re: Larry Swedroe: Time Series Momentum (aka Trend Following): A Good Time For A Refresh

Post by Random Walker » Fri Feb 09, 2018 5:31 pm

Nisiprius,
I see you addressed my concern before I got the post up. Nice work :-)

Dave

User avatar
nisiprius
Advisory Board
Posts: 35896
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: Larry Swedroe: Time Series Momentum (aka Trend Following): A Good Time For A Refresh

Post by nisiprius » Fri Feb 09, 2018 9:17 pm

Larry Swedroe sent me a PM objecting to the comparisons I made. It gets a little weird... I don't feel I can, or should be expected to, do an effective job of presenting another person's rebuttals.

He did not name any actual mutual fund I could look at that goes back much farther than AQMIX, MHFIX, or RYIFX, but he gives a link to a Barclay's index going back to 1980. I won't comment beyond reproducing his link to the web page and the data that is presented.

https://www.barclayhedge.com/research/i ... b/cta.html
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.

User avatar
willthrill81
Posts: 4171
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: Larry Swedroe: Time Series Momentum (aka Trend Following): A Good Time For A Refresh

Post by willthrill81 » Fri Feb 09, 2018 10:25 pm

I'm glad to see I'm not the only Boglehead who thinks that trend following is a viable strategy!

I will point out that not all trend following involves shorting assets.

Also, it's worth pointing out that, historically, trend following has had a historical tendency to lag buy-and-hold during bull markets. Where it shines is in bear markets with a generally much lower drawdown than buy-and-hold. By reducing volatility drag, this has helped it to have similar long-term absolute returns and higher risk-adjusted returns.
“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
nisiprius
Advisory Board
Posts: 35896
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: Larry Swedroe: Time Series Momentum (aka Trend Following): A Good Time For A Refresh

Post by nisiprius » Sat Feb 10, 2018 8:00 am

1) To those saying that 2008-present is not a fair test because it was a rising market, I thought the whole point of the "smile" was that the strategy performs well during both falling and rising markets. Obviously it's more exciting if it performs well during falling markets, but if that is just at the expense of bad performance in rising markets, what's the big deal? Certainly, a +10% or +12% kicker in 2008-2009 is wonderful, but is it worth the misery of eight years of subsequent underperformance?

2) With regard to the four terms "CTA," "managed futures," "time-series momentum," and "trend-following," are these four different terms for the same thing, and, if not, would someone care to clarify?

3) Is this "commodities redux," or are there mutual funds that implement this strategy (other than by active manager's personal intuition) with stocks? What are some examples of "managed futures stock funds?"
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.

Theoretical
Posts: 1262
Joined: Tue Aug 19, 2014 10:09 pm

Re: Larry Swedroe: Time Series Momentum (aka Trend Following): A Good Time For A Refresh

Post by Theoretical » Sat Feb 10, 2018 8:57 am

nisiprius wrote:
Sat Feb 10, 2018 8:00 am
1) To those saying that 2008-present is not a fair test because it was a rising market, I thought the whole point of the "smile" was that the strategy performs well during both falling and rising markets. Obviously it's more exciting if it performs well during falling markets, but if that is just at the expense of bad performance in rising markets, what's the big deal? Certainly, a +10% or +12% kicker in 2008-2009 is wonderful, but is it worth the misery of eight years of subsequent underperformance?

2) With regard to the four terms "CTA," "managed futures," "time-series momentum," and "trend-following," are these four different terms for the same thing, and, if not, would someone care to clarify?

3) Is this "commodities redux," or are there mutual funds that implement this strategy (other than by active manager's personal intuition) with stocks? What are some examples of "managed futures stock funds?"
1. Managed Futures funds tended to capitalize well on 2013 and 2014, got whipped around in 2015 and 2016, and performed flat or poorly in 2017, but have banked some gains in 2018. Likely some of this is a function of the futures contracts being "future", meaning there may be a time delay of a few months in a given year. However, I agree that it's a bit concerning that the funds haven't done that well in 2017 especially.

2. CTA (Commodities Trading Advisors) are manager/trader types who run the Managed futures Strategies, and the main difference is their sandbox is the Chicago Board of Trade and like entities, rather than the NYSE or Nikkei. Managed Futures is a term to describe funds that go long-short on one or more asset classes using the futures markets rather than trading the actual assets, such as stocks, bonds, commodities or currencies. Time-Series Momentum and trend-following I believe are cousins, with TS Mom being the academic factor (like value or quality) that asks "how is this particular asset class/sub-class" performing compared to the universe of all assets? It's distinguished from the more common (Cross-sectional) momentum by being Momentum across different asset classes than comparing performance of certain stocks with other stocks in the same index and picking the best (or worst) performers to go long (short). Trend-following relates to the technical/quantitative/mechanical trading techniques that attempt to capture time series momentum. These include things like the Simple Moving Average or other funds.

3. No, it's not commodities redux, as most of the managed futures funds are quantitative/technical analysis based and are multi-asset. That doesn't mean there's not risk attached to them, only the risk is programming or modeling errors rather than discretionary active mis-management. There's a couple of long-short commodities funds, but the vast majority are multi-asset. Most include equity index futures, interest rate futures, currency futures, and commodities futures. There's even a quasi-index fund, run by Credit Suisse, which tries to mimic aggregate managed futures moves and tracks under its expense ratio.

On the stock side, there's a long-flat S&P 500 ETF that's come out recently that switches to t-bills if set market signals go negative. There's also the extremely transparent Alpha Architect Value Momentum Trend ETF VMOT, which uses 2 simple signals on a 50-50 basis to determine whether to short a major stock index (S&P or EAFE) while keeping the holdings of their 2 value and 2 stock funds intact.

long_gamma
Posts: 226
Joined: Mon Sep 14, 2015 4:13 pm

Re: Larry Swedroe: Time Series Momentum (aka Trend Following): A Good Time For A Refresh

Post by long_gamma » Sat Feb 10, 2018 9:00 am

"Everyone has a plan 'till they get punched in the mouth." --Mike Tyson

scone
Posts: 1447
Joined: Wed Jul 11, 2012 4:46 pm

Re: Larry Swedroe: Time Series Momentum (aka Trend Following): A Good Time For A Refresh

Post by scone » Sat Feb 10, 2018 9:08 am

nisiprius wrote:
Fri Feb 09, 2018 9:17 pm
Larry Swedroe sent me a PM objecting to the comparisons I made. It gets a little weird... I don't feel I can, or should be expected to, do an effective job of presenting another person's rebuttals.

He did not name any actual mutual fund I could look at that goes back much farther than AQMIX, MHFIX, or RYIFX, but he gives a link to a Barclay's index going back to 1980. I won't comment beyond reproducing his link to the web page and the data that is presented.

https://www.barclayhedge.com/research/i ... b/cta.html
At least Larry is still reading this forum. Thank you for all your help over the years, Larry! :beer
"My bond allocation is the amount of money that I cannot afford to lose." -- Taylor Larimore

long_gamma
Posts: 226
Joined: Mon Sep 14, 2015 4:13 pm

Re: Larry Swedroe: Time Series Momentum (aka Trend Following): A Good Time For A Refresh

Post by long_gamma » Sat Feb 10, 2018 9:09 am

Soceity Generalle trend index with signal disclosed

https://cib.societegenerale.com/fileadm ... index.html

Image
"Everyone has a plan 'till they get punched in the mouth." --Mike Tyson

User avatar
nisiprius
Advisory Board
Posts: 35896
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: Larry Swedroe: Time Series Momentum (aka Trend Following): A Good Time For A Refresh

Post by nisiprius » Sat Feb 10, 2018 5:46 pm

But, quite apart from the question of whether there's any way to invest in a CTA index, are CTA indexes, like the BarclayHedge, any good?

In Swedroe and Kizer (2009), The Only Guide to Alternative Investments You'll Ever Need: The Good, the Flawed, the Bad, and the Ugly, "commodities" are included under "good," but CTAs come in for some harsh criticism as a way to access them. In a discussion of them on pp. 55-6, in part, it says:
Light regulation of CTA funds results in limited information on the performance as well as the riskiness of these funds... As of October 1996, there were 304 CTAs in operation. Since the end of 1986, 597 had been dissolved... [in another time period there was] a failure rate of 71% in just seven years.... [In a 2000 study] the probability of a commodity fund dropping out of the database was 19% per year... for the period 1989-1997, once reported returns were corrected for survivorship bias, realized return were reduced by 3.6% per year.

In summary, there is a significant body of evidence demonstrating that the only ones likely to create wealth from CTAs are the fund sponsors and the commission-driven salespeople or advisers who market them.... the evidence is that they are highly risky investments with poor risk/return characteristics.
Now, clearly, if the BarclayHedge CTA index were a valid representation of what investors with CTAs were actually getting, they would seem to be at least decent investments. To make such a harsh judgement, the authors must believe that there are indeed serious questions about the validity and about survivorship bias in CTA databases.
Last edited by nisiprius on Sat Feb 10, 2018 5:49 pm, edited 2 times in total.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.

Random Walker
Posts: 2706
Joined: Fri Feb 23, 2007 8:21 pm

Re: Larry Swedroe: Time Series Momentum (aka Trend Following): A Good Time For A Refresh

Post by Random Walker » Sat Feb 10, 2018 5:47 pm

Time Series Momentum versus Cross Sectional Momentum

CS Momentum is relative momentum. Independent of whether investments in an asset class go up or down, CS Momentum will go long the top performers and short the worst performers. Thus it remains net no exposure to the asset class. For example, in stocks, CS Momentum might buy the top third performing S&P 500 stocks and sell the bottom third. The best performers might have gone up the most or gone down the least in the prior period.

TS Momentum is absolute Momentum. Simply go long what has gone up in an asset class and short what has gone down. If everything has gone up in the prior period, then nothing gets shorted. So TS Momentum necessarily implies a shifting allocation. For example in equities, net equity exposure will be greater after a rising market in the prior period than it would be after a falling market in the prior period.

Both CS Momentum and TS Momentum can be applied to multiple asset classes, and this is indeed what the AQR funds do. They take advantage of the lack of momentum correlation between various asset classes.

Dave


User avatar
Taylor Larimore
Advisory Board
Posts: 26999
Joined: Tue Feb 27, 2007 8:09 pm
Location: Miami FL

Past Performance

Post by Taylor Larimore » Sat Feb 10, 2018 8:21 pm

Bogleheads:

There is one fact that is so important that the government requires all mutual fund companies to warn us:

Past performance does not predict future performance.

Best wishes.
Taylor
"Simplicity is the master key to financial success." -- Jack Bogle

stlutz
Posts: 4487
Joined: Fri Jan 02, 2009 1:08 am

Re: Larry Swedroe: Time Series Momentum (aka Trend Following): A Good Time For A Refresh

Post by stlutz » Sat Feb 10, 2018 8:35 pm

Nisi--Some of the concerns about the CTA indexes are also discussed in this article: http://www.etf.com/publications/journal ... nopaging=1 . Their issues are:
1. Returns may be biased upward: The returns for indexes of CTA managers tend to be biased upward as a result of the voluntary nature of self-reporting performance. A CTA with poor performance for a period of time is less likely to report unfavorable returns to these types of databases, resulting in an index that includes mostly favorable performance.
2. Lack of a natural measuring stick: Other asset classes like equities or bonds have a natural benchmark for performance reporting. Market-capitalization-weighted benchmarks for equities or bonds mathematically represent the average return in aggregate to investors in the equity or bond markets. These are meaningful performance measures. How would one apply that to the managed futures space?
The article ends up proposing what they view as a more truly systematic approach. Wisdom Tree has an ETF following this approach, trading under the ticker WTMF. I'll let others look up how the fund has performed.

long_gamma
Posts: 226
Joined: Mon Sep 14, 2015 4:13 pm

Re: Larry Swedroe: Time Series Momentum (aka Trend Following): A Good Time For A Refresh

Post by long_gamma » Sat Feb 10, 2018 9:09 pm

stlutz wrote:
Sat Feb 10, 2018 8:35 pm
Wisdom Tree has an ETF following this approach, trading under the ticker WTMF. I'll let others look up how the fund has performed.
WTMF has some glaring flaws.

* It uses only 24 future components. Most of the CTA funds has lot more diversity. Even basic trend indicator, which i posted upthread has 50+ futures. Funds like Winton and AQR can do trend following in spreads also.
* Fund uses some arbitrary rules like, not shorting energy futures. So it is long-short in some and long only in energy

Image

* Other CTA's combine time-series with other factors like carry and cross-sectional, which offers better risk adjusted return
* It does not include any equity futures or foreign bond futures
"Everyone has a plan 'till they get punched in the mouth." --Mike Tyson

User avatar
nisiprius
Advisory Board
Posts: 35896
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: Larry Swedroe: Time Series Momentum (aka Trend Following): A Good Time For A Refresh

Post by nisiprius » Sat Feb 10, 2018 9:51 pm

long_gamma wrote:
Sat Feb 10, 2018 9:09 pm
stlutz wrote:
Sat Feb 10, 2018 8:35 pm
Wisdom Tree has an ETF following this approach, trading under the ticker WTMF. I'll let others look up how the fund has performed.
WTMF has some glaring flaws.

* It uses only 24 future components. Most of the CTA funds has lot more diversity. Even basic trend indicator, which i posted upthread has 50+ futures. Funds like Winton and AQR can do trend following in spreads also.
* Fund uses some arbitrary rules like, not shorting energy futures. So it is long-short in some and long only in energy.
* Other CTA's combine time-series with other factors like carry and cross-sectional, which offers better risk adjusted return
* It does not include any equity futures or foreign bond futures
You can't blame people for looking at what's actually out there. If WTMF isn't a fair thing to look at because of its "glaring flaws," and if AQMIX isn't a fair thing to look at because it wasn't launched until 2010, then please suggest a fund or ETF that would be a fair thing to look at.

If it is CTAs themselves that are the proof of the pudding, how do you respond to Swedroe and Kizer's (2009) statement that
there is a significant body of evidence demonstrating that the only ones likely to create wealth from CTAs are the fund sponsors and the commission-driven salespeople or advisers who market them.... the evidence is that they are highly risky investments with poor risk/return characteristics.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.

long_gamma
Posts: 226
Joined: Mon Sep 14, 2015 4:13 pm

Re: Larry Swedroe: Time Series Momentum (aka Trend Following): A Good Time For A Refresh

Post by long_gamma » Sat Feb 10, 2018 11:05 pm

nisiprius wrote:
Sat Feb 10, 2018 9:51 pm
You can't blame people for looking at what's actually out there. If WTMF isn't a fair thing to look at because of its "glaring flaws," and if AQMIX isn't a fair thing to look at because it wasn't launched until 2010, then please suggest a fund or ETF that would be a fair thing to look at.

If it is CTAs themselves that are the proof of the pudding, how do you respond to Swedroe and Kizer's (2009) statement that
there is a significant body of evidence demonstrating that the only ones likely to create wealth from CTAs are the fund sponsors and the commission-driven salespeople or advisers who market them.... the evidence is that they are highly risky investments with poor risk/return characteristics.
I am not blaming anybody, just pointing out flaws. Missing out entire equity asset class in 2017 or oil bear market was very costly for WTMF.

Unfortunately, it is difficult to structure in ETF form with all disclosed signals, like they do in market cap ETF's. It can be front run like USO, UNG etf's when it came out. JP morgan ETF seems to be interesting.

If one wants to invest in funds one would go with fund which uses most diversified lineup and very mechanical (systematic, rather than discretionary CTA's.) in their approach. AQR, Winton, Man alpha all fits that bill.

Regarding the other statement, you have to ask the authors of that quote. CTA's can come in many flavor's. I would focus on trend following quant CTA's. There was (and probably still is) some truth to that statement atleast in 2009. Now days I am not so sure. If JP Morgan can deliver trend following for 85 basis points, which is probably less than average retail equity mutual funds charged in early 2000, then that statement is no longer holds true. AQR has been game changer in this space, sucking up all the money from other CTA's.

Regarding risk/return characteristics, with quant CTA's investors already know risk/return characteristics based on the backtest of their signal generating process. Poor risk/return characteristics is mostly with discretionary, non-systematic , short term trading CTA's or star CTA's based on their prediction skills.

Personally, I don't invest in any of those funds or ETF's. I trade my own futures account.
"Everyone has a plan 'till they get punched in the mouth." --Mike Tyson

User avatar
nisiprius
Advisory Board
Posts: 35896
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: Larry Swedroe: Time Series Momentum (aka Trend Following): A Good Time For A Refresh

Post by nisiprius » Sun Feb 11, 2018 10:34 am

long_gamma wrote:
Sat Feb 10, 2018 11:05 pm
nisiprius wrote:
Sat Feb 10, 2018 9:51 pm
... If WTMF isn't a fair thing to look at because of its "glaring flaws," and if AQMIX isn't a fair thing to look at because it wasn't launched until 2010, then please suggest a fund or ETF that would be a fair thing to look at...
... JP morgan ETF seems to be interesting...
This one?

JPMF, JPMorgan Managed Futures Strategy ETF

Source

Image

(Inception 12/5/2017)
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.

User avatar
Lauretta
Posts: 656
Joined: Wed Jul 05, 2017 6:27 am

Re: Larry Swedroe: Time Series Momentum (aka Trend Following): A Good Time For A Refresh

Post by Lauretta » Sun Feb 11, 2018 12:06 pm

nisiprius wrote:
Sun Feb 11, 2018 10:34 am
long_gamma wrote:
Sat Feb 10, 2018 11:05 pm
nisiprius wrote:
Sat Feb 10, 2018 9:51 pm
... If WTMF isn't a fair thing to look at because of its "glaring flaws," and if AQMIX isn't a fair thing to look at because it wasn't launched until 2010, then please suggest a fund or ETF that would be a fair thing to look at...
... JP morgan ETF seems to be interesting...
This one?

JPMF, JPMorgan Managed Futures Strategy ETF

Source

Image

(Inception 12/5/2017)
this one looks better though unfortunately is very recent; but the people running it are very smart so it will be interesting to see how it goes in future
https://www.bloomberg.com/quote/PIM1966:AU
When everyone is thinking the same, no one is thinking at all

long_gamma
Posts: 226
Joined: Mon Sep 14, 2015 4:13 pm

Re: Larry Swedroe: Time Series Momentum (aka Trend Following): A Good Time For A Refresh

Post by long_gamma » Sun Feb 11, 2018 12:14 pm

Yes.

If you are alluding to drop in recent prices, then it is pretty normal. Trend following does not do well in sudden reversal (momentum crashes). It does well in slow grind bear or bull markets.

Prospectus does not say, what volatility they are targeting for. Looks like much less vol. targeting compared to AQR. AQR has about 8.5% drawdown from recent high watermark.
"Everyone has a plan 'till they get punched in the mouth." --Mike Tyson

pkcrafter
Posts: 12749
Joined: Sun Mar 04, 2007 12:19 pm
Location: CA
Contact:

Re: Larry Swedroe: Time Series Momentum (aka Trend Following): A Good Time For A Refresh

Post by pkcrafter » Sun Feb 11, 2018 1:03 pm

Well, Taylor has nailed it once again. If "trend following is go good, why are there (at least) two versions.

Here's an explanation of trend following.

https://www.trendfollowing.com/trend/

I would especially like to hear Nisiprius' comments on it. Also, what would be another term for trend following? Hint. a term which we are all familiar with and don't recommend.


Paul
When times are good, investors tend to forget about risk and focus on opportunity. When times are bad, investors tend to forget about opportunity and focus on risk.

long_gamma
Posts: 226
Joined: Mon Sep 14, 2015 4:13 pm

Re: Larry Swedroe: Time Series Momentum (aka Trend Following): A Good Time For A Refresh

Post by long_gamma » Sun Feb 11, 2018 1:53 pm

Interview of David harding (Winton group) on Bloomberg last year.

https://www.youtube.com/watch?v=lu92IYyk2Sw
"Everyone has a plan 'till they get punched in the mouth." --Mike Tyson

cashcows
Posts: 9
Joined: Sun Nov 01, 2015 3:06 pm

Re: Larry Swedroe: Time Series Momentum (aka Trend Following): A Good Time For A Refresh

Post by cashcows » Sat Mar 10, 2018 9:07 pm

Theoretical wrote:
Sat Feb 10, 2018 8:57 am
nisiprius wrote:
Sat Feb 10, 2018 8:00 am
1) To those saying that 2008-present is not a fair test because it was a rising market, I thought the whole point of the "smile" was that the strategy performs well during both falling and rising markets. Obviously it's more exciting if it performs well during falling markets, but if that is just at the expense of bad performance in rising markets, what's the big deal? Certainly, a +10% or +12% kicker in 2008-2009 is wonderful, but is it worth the misery of eight years of subsequent underperformance?

2) With regard to the four terms "CTA," "managed futures," "time-series momentum," and "trend-following," are these four different terms for the same thing, and, if not, would someone care to clarify?

3) Is this "commodities redux," or are there mutual funds that implement this strategy (other than by active manager's personal intuition) with stocks? What are some examples of "managed futures stock funds?"
1. Managed Futures funds tended to capitalize well on 2013 and 2014, got whipped around in 2015 and 2016, and performed flat or poorly in 2017, but have banked some gains in 2018. Likely some of this is a function of the futures contracts being "future", meaning there may be a time delay of a few months in a given year. However, I agree that it's a bit concerning that the funds haven't done that well in 2017 especially.

2. CTA (Commodities Trading Advisors) are manager/trader types who run the Managed futures Strategies, and the main difference is their sandbox is the Chicago Board of Trade and like entities, rather than the NYSE or Nikkei. Managed Futures is a term to describe funds that go long-short on one or more asset classes using the futures markets rather than trading the actual assets, such as stocks, bonds, commodities or currencies. Time-Series Momentum and trend-following I believe are cousins, with TS Mom being the academic factor (like value or quality) that asks "how is this particular asset class/sub-class" performing compared to the universe of all assets? It's distinguished from the more common (Cross-sectional) momentum by being Momentum across different asset classes than comparing performance of certain stocks with other stocks in the same index and picking the best (or worst) performers to go long (short). Trend-following relates to the technical/quantitative/mechanical trading techniques that attempt to capture time series momentum. These include things like the Simple Moving Average or other funds.

3. No, it's not commodities redux, as most of the managed futures funds are quantitative/technical analysis based and are multi-asset. That doesn't mean there's not risk attached to them, only the risk is programming or modeling errors rather than discretionary active mis-management. There's a couple of long-short commodities funds, but the vast majority are multi-asset. Most include equity index futures, interest rate futures, currency futures, and commodities futures. There's even a quasi-index fund, run by Credit Suisse, which tries to mimic aggregate managed futures moves and tracks under its expense ratio.

On the stock side, there's a long-flat S&P 500 ETF that's come out recently that switches to t-bills if set market signals go negative. There's also the extremely transparent Alpha Architect Value Momentum Trend ETF VMOT, which uses 2 simple signals on a 50-50 basis to determine whether to short a major stock index (S&P or EAFE) while keeping the holdings of their 2 value and 2 stock funds intact.
I see that Merrill Edge blocks the trading of VMOT. Any recommendations on where to purchase without commission?

golfCaddy
Posts: 504
Joined: Wed Jan 10, 2018 10:02 pm

Re: Larry Swedroe: Time Series Momentum (aka Trend Following): A Good Time For A Refresh

Post by golfCaddy » Sun Mar 11, 2018 11:36 pm

There's a lot of issues with the behavioral theory that's being ignored. First, why is the auto-correlation between monthly returns this odd parabola(from 1 to 11 months), with a maximum at 1-month and 11-months, and nearly zero at 7 months? Then, why is there this sharp reversal, at 13-months and 15-months being significant, but strongly negative, while 14-months appears to be near zero? There's no intuitive or satisfying behavioral theory as to why the chart should like this. If this was just a case of investors being unduly influenced by recent performance, you might expect a monotonically decreasing curve, until you reached zero at which point it should stabilize.

Image

Swelfie
Posts: 200
Joined: Mon Mar 14, 2016 12:54 am

Re: Larry Swedroe: Time Series Momentum (aka Trend Following): A Good Time For A Refresh

Post by Swelfie » Tue Mar 13, 2018 5:27 am

nisiprius wrote:
Fri Feb 09, 2018 4:59 pm
It seems fairly clear that he is identifying this strategy with the phrase "managed futures," since he writes
Strategies that attempt to capture the return premium offered by time-series momentum are often called, “managed futures,” as they take long and short positions in assets via futures markets — ideally in a multitude of futures markets around the globe.
My cynical guess is that, like many trendy new strategies, it is going to turn out that there simply are no managed futures fund with a long track record, let's say 15 years, that we could look at to find out how this strategy has worked in the real world... just as, when "commodifies" (CCF) funds were being widely recommended circa 2010 or so, it was hard to find any older than PCRIX, which started in 2008.

First question, since he cites an AQR paper, are there any AQR "managed futures" (as opposed to risk parity) funds? And, since one of his charts compares managed futures to "diversified passive," how has it compared to the moderate (60/40) model portfolio Larry recommended in his 1998 book, which can be described as "diversified passive."

There is one such AQR managed futures fund, the AQR Managed Futures Strategy Fund, AQMIX.

As I feared, it is only eight years old. And, here it is, (blue), compared to the diversified passive model portfolio Larry suggested in 1998 (red), implemented with Vanguard funds as much as possible, for maximum passivity.

Source

Image

But, someone might say, it shouldn't be considered in isolation. Although figure 3 in the paper is implicitly comparing 100% "time series momentum" to a "diversified passive" portfolio, perhaps that's too extreme. We see that in our illustration above, AQMIX had low--slightly negative--correlation with stocks. Would adding a moderate allocation to AQMIX, say 15%, improve the "diversified passive" portfolio?

Portfolio 3, orange, is a modified version of the "diversified passive" portfolio, with about a 15% allocation to AQMIX, and the rest reduced proportionately.

Source
Image

It was not an improvement.
How is this not a risk adjusted returns improvement? Portfolio 3 wins on both Sharpe and Sorting metrics. CAGR is down a bit but both your volatility and downside only volatility is down even more which is the whole point of adding such an asset class. It's very limited data but it seems to imply Larry is right and that adding managed futures improved the portfolio.

Post Reply