What is wrong with AQR?

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
User avatar
Topic Author
hdas
Posts: 876
Joined: Thu Jun 11, 2015 8:24 am

What is wrong with AQR?

Post by hdas » Wed Jun 12, 2019 9:17 am

RCM alternatives has a series looking at the performance of AQR managed futures

Part 1
Part 2
Part 3

Image

And finally, perhaps most importantly of all, it hasn’t been able to outperform its peers. Our proprietary ranking model measures alternative investments in both private fund and mutual fund wrappers, and has AQR’s managed futures mutual fund AQMIX with a ranking of just 1 star out of 5, finding it in the bottom quintile of our rankings (which consider risk adjust returns and time weighted statistics). Said another way, when comparing AQMIX across multiple time frames, return metrics, and risk measures – 4 out of 5 programs in our database rank better than it. Here’s a YCharts chart again showing that graphically, with AQR plotted versus Goldman, Natixis, and Pimco’s managed futures offerings. They are all losing the war, as well, with the overall asset class suffering as volatility has gone on an extended vacation, but they are winning the battle vs AQR in terms of relative out performance, with double digit returns over the past five years while AQR is in the red.
Image
But Why?

Why is it under performing its peers? Well, again, that size factor is definitely an issue, with AQR unable to participate in a meaningful way in markets that have moved this year (like Hogs and Corn). Their footprint also can make market impact an issue, bringing to light the need for everyone to use automated execution algorithms as a new best practice. And it can’t be easy to have to unwind $9 Billion of positions without negatively impacting the one’s you still hold. What we also see, however, is a weird sort of volatility profile where AQR tends to mirror the vol of its peers on the downside, but have less volatility on the upside. Meaning they all tend to lose about the same, but when they are all winning, the others are winning more. Everyone else tended to do better in the slow crawl upwards of 2017, perhaps pointing to a long bias among others? But if that’s the case, then why was there another divergence during the big stock sell off in December 2018, where AQR didn’t capture as much of the down move in US stocks as the others? Maybe/probably a little bit of all of the above is happening.
"whenever there is a randomized way of doing something, then there is a nonrandomized way that delivers better performance but requires more thought" ET Jaynes

User avatar
nisiprius
Advisory Board
Posts: 38287
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: What is wrong with AQR?

Post by nisiprius » Wed Jun 12, 2019 4:55 pm

The question to me is: who cares?

Image

As you can see, starting at inception of AQMIX, the average managed futures fund (orange) lost money, and not a trivial amount, either. It's a total 8.2% loss spread over nine-and-a-half years, so that's something like -0.8% per year... together with a lot of volatility.

So, hey, AQMIX (blue) did much better than average, yay. But all that means is that it almost, but not quite matched the return of a money market fund (yellow).

Money market returns, with almost-stock-like volatility. (11.71% for Vanguard 500 Index, 9.04% for AQMIX, Jan 2011 - May 2019).

Now I've been hiding something, which is that Morningstar also can show us an index, "Credit Suisse Managed Futures Liquid TR" (green) and AQMIX did underperform that index.

And while that index did, at least, outperform a money market fund, Total Bond far outperformed the index and did so with far lower volatility.

Source

Image

I just don't get it. Managed Futures just look like a bad joke to me. I certainly don't want the category average, nor do I have confidence that I can pick an above-average fund.

I don't see any reason to want AQMIX.

And even if the Credit Suisse Managed Futures Strategy Fund, is able to track its index, I still don't see any reason to want it when I can get Total Bond instead.

(Yes, I know, correlations, but low correlation in this case doesn't come close to overweighing the issues of low return and high volatility).
Last edited by nisiprius on Wed Jun 12, 2019 5:03 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.

robertmcd
Posts: 392
Joined: Tue Aug 09, 2016 9:06 am

Re: What is wrong with AQR?

Post by robertmcd » Wed Jun 12, 2019 5:00 pm

The problem with managed futures is exactly this. They look great on paper and in backtests, but the real world results are so random that I can't even understand how it is possible that these funds are all using the same strategy.

larryswedroe
Posts: 15835
Joined: Thu Feb 22, 2007 8:28 am
Location: St Louis MO

Re: What is wrong with AQR?

Post by larryswedroe » Wed Jun 12, 2019 6:42 pm

Before you draw any conclusions about TS momentum and thus managed futures suggest you read this great piece by Dr. Jack Vogel of Alpha Architect (best site IMO for those interested in the academic research) which shows very clearly the danger of the mistake so many make of recency.
https://alphaarchitect.com/2019/02/20/t ... rformance/

The site has some other very good articles on TS MOM and managed futures.
Now it is true that AQR has underperformed peers in this period, with reason being their multi-metric approach has recently underperformed, which has happened before and will happen again. Of course it outperforms in other periods (:-)) AQR has had two other periods of really bad performance as a firm --one during LTCM crisis and another in the GFC era when things tended to perform against historical evidence. All risk assets go through such periods or there would be no risk.

Best wishes
Larry

afan
Posts: 4173
Joined: Sun Jul 25, 2010 4:01 pm

Re: What is wrong with AQR?

Post by afan » Thu Jun 13, 2019 4:34 am

Since it is not a cap weighted index fund I would never invest in it. The highway robbery level of expenses is just a bonus.

That said, if one likes to gamble and has to have some secret, special, only the pros know about this sophisticated strategy, way of playing with money, I can think of worse. If you add a 10% allocation to AQMIX to a total stock and total bond portfolio you get slightly lower volatility with slightly lower Sharpe ratio. Depending on how you feel about volatility, maybe that is worth it?

I know that reducing the Sharpe ratio is not supposed to be desirable. But we are talking about a behavioral need to hold exotic portfolio, not a simple desire for increased risk adjusted returns. You just decide to ignore the increased negative skewness.

At least it gives you something to talk about at parties and the deep satisfaction of knowing that you are among the expert investors whose 1% advisors have steered you into managed futures.

Plus, given its volatility, there will be times when holding this fund would be profitable. All you have to do is ignore the times when you would be better off without it.

Of course, you could have lowered volatility more and INCREASED Sharpe ratio by going to 50/50 stock and bond, skipping the futures entirely.

But where is the fun in that? You don't need an advisor to help you and AQR does not benefit.
We don't know how to beat the market on a risk-adjusted basis, and we don't know anyone that does know either | --Swedroe | We assume that markets are efficient, that prices are right | --Fama

Forester
Posts: 130
Joined: Sat Jan 19, 2019 2:50 pm
Location: UK

Re: What is wrong with AQR?

Post by Forester » Thu Jun 13, 2019 5:18 am

I can't see the point of managed futures as an investment. Profits and losses equal zero. If the stock market becomes more efficient I still get market beta, if the futures markets become more efficient (smarter hedgers, competition among speculators), there's less profit to be had for the trend follower working for me.

columbia
Posts: 1567
Joined: Tue Aug 27, 2013 5:30 am

Re: What is wrong with AQR?

Post by columbia » Thu Jun 13, 2019 5:50 am

How much of the goal is low/no correlation? I can get that with cash.

User avatar
nisiprius
Advisory Board
Posts: 38287
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: What is wrong with AQR?

Post by nisiprius » Thu Jun 13, 2019 6:20 am

larryswedroe wrote:
Wed Jun 12, 2019 6:42 pm
...AQR has had two other periods of really bad performance as a firm --one during LTCM crisis and another in the GFC era when things tended to perform against historical evidence...
Any asset, strategy, or firm has had great performance if you excuse the periods when it had bad performance. An actual investor in the fund can't do that (unless, of course, they know how to time the market and get out whenever the market is "tending to perform against historical evidence.") Money lost "against historical evidence" is still gone.

Dumb Boglehead non-stock investments--like the Vanguard Total Bond Market Index Fund (and similar funds from Vanguard's competitors) did not collapse during those time periods.

Source

Image

That wasn't because Vanguard was special--I didn't want to clutter the chart but if you add PTTRX (PIMCO Total Return), DFGBX (DFA Five-Year Global Fixed Income), you don't see anything alarming during the LTCM or GFC eras. I attribute this to sticking to simplicity, as opposed to using sophisticated strategies and non-traditional assets to gain an edge. Sometimes that edge is a fragile, and closely-calculated balance, based on quantitative expectations of how assets are going to move in relation to each other--a balance that fails whenever the market "tends to perform against historical evidence."
Last edited by nisiprius on Thu Jun 13, 2019 11:25 am, edited 1 time 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.

Forester
Posts: 130
Joined: Sat Jan 19, 2019 2:50 pm
Location: UK

Re: What is wrong with AQR?

Post by Forester » Thu Jun 13, 2019 7:24 am

Correct me if I'm wrong but these futures markets are the same as betting exchanges which exist for sports and politics. It's an insulated system where all the profits & losses add up to zero, minus the trading costs.

afan
Posts: 4173
Joined: Sun Jul 25, 2010 4:01 pm

Re: What is wrong with AQR?

Post by afan » Thu Jun 13, 2019 8:58 am

On the subject of academic studies:

"A comprehensive look at the empirical performance of equity risk premium prediction"
Welch and Goyal
Review of Financial Studies, 2008, vol 21, no 4, 1455-1508

Careful analysis of a wide variety of rules that purport to predict equity risk premia and therefore guide movea into and out of stocks.

Excellent paper and worth a read.
Oversimplified conclusion: the rules and strategies do not work.
Some worked, sort of, in sample but none worked out of sample.
We don't know how to beat the market on a risk-adjusted basis, and we don't know anyone that does know either | --Swedroe | We assume that markets are efficient, that prices are right | --Fama

afan
Posts: 4173
Joined: Sun Jul 25, 2010 4:01 pm

Re: What is wrong with AQR?

Post by afan » Thu Jun 13, 2019 9:08 am

Forester wrote:
Thu Jun 13, 2019 7:24 am
Correct me if I'm wrong but these futures markets are the same as betting exchanges which exist for sports and politics. It's an insulated system where all the profits & losses add up to zero, minus the trading costs.
It would add up to zero, minus trading costs, if you were not paying an investment manager to trade for you, an advisor to help you pick investment managers and an advisor advisor to help you pick advisors. Then of course, you share any gains with the IRS.

Once you add up the costs, it would take a long string of good luck to come out ahead. For a while. Then start losing again.

As long as you are happy to ignore the losses and brag about the intermittent gains, sounds like a great investment.
We don't know how to beat the market on a risk-adjusted basis, and we don't know anyone that does know either | --Swedroe | We assume that markets are efficient, that prices are right | --Fama

eigenperson
Posts: 16
Joined: Mon Nov 09, 2015 7:16 pm

Re: What is wrong with AQR?

Post by eigenperson » Thu Jun 13, 2019 9:32 am

nisiprius wrote:
Wed Jun 12, 2019 4:55 pm
The question to me is: who cares?
There are reviews for psychics on Yelp. You and I may laugh at that, since there's no point in differentiating between a good psychic and a bad one, but still -- who wants to go to a bad psychic? There's no possible upside of that.

Some psychics, I'm sure, are able to predict bad events but struggle to predict the good ones. Similarly (at least to my mind), some funds have "a weird sort of volatility profile where [they tend] to mirror the vol of [their] peers on the downside, but have less volatility on the upside." I'm pretty sure both of these are superstitions, but if you believe in the respective superstitions, neither is information you can afford to ignore.

skeptic42
Posts: 64
Joined: Mon Feb 11, 2019 5:27 pm

Re: What is wrong with AQR?

Post by skeptic42 » Thu Jun 13, 2019 10:25 am

Shouldn't a long position in a futures contract return almost the same as an investment in the underlying asset?
Then why was the performance of managed futures so bad?
How could they miss all the returns of stocks and bonds in the last years?

User avatar
nisiprius
Advisory Board
Posts: 38287
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: What is wrong with AQR?

Post by nisiprius » Thu Jun 13, 2019 10:31 am

[Inaccurate comment retracted, see hdas' post immediately below].
Last edited by nisiprius on Thu Jun 13, 2019 11:29 am, 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.

User avatar
Topic Author
hdas
Posts: 876
Joined: Thu Jun 11, 2015 8:24 am

Re: What is wrong with AQR?

Post by hdas » Thu Jun 13, 2019 10:38 am

nisiprius wrote:
Thu Jun 13, 2019 10:31 am
skeptic42 wrote:
Thu Jun 13, 2019 10:25 am
Shouldn't a long position in a futures contract return almost the same as an investment in the underlying asset?
Then why was the performance of managed futures so bad?
How could they miss all the returns of stocks and bonds in the last years?
I don't think "managed futures" invest in stock and bond futures. Or not typically. I think are typically investing in something, somehow, related to "commodities."
You should enforce your traditional evidence approach across the board when commenting and don't spread misinformation.
Managed Futures ppl trade everything, including bonds and stocks. In fact, most of the performance of this year for a lot of them has been driven by being long fixed income futures. They were very long stocks late 2017 as well. :greedy
"whenever there is a randomized way of doing something, then there is a nonrandomized way that delivers better performance but requires more thought" ET Jaynes

azanon
Posts: 2368
Joined: Mon Nov 07, 2011 10:34 am
Location: Little Rock, AR
Contact:

Re: What is wrong with AQR?

Post by azanon » Thu Jun 13, 2019 10:39 am

I never personally bought the book, but wasn't managed futures on the "bad or ugly" list in "The Only Guide to Alternative Investments You'll Ever Need: The Good, the Flawed, the Bad, and the Ugly"? If so, when did that change?

I could only find the "good" list quick searching the internet, and I did confirm, it wasn't one of those.

User avatar
Topic Author
hdas
Posts: 876
Joined: Thu Jun 11, 2015 8:24 am

Re: What is wrong with AQR?

Post by hdas » Thu Jun 13, 2019 10:47 am

Forester wrote:
Thu Jun 13, 2019 7:24 am
Correct me if I'm wrong but these futures markets are the same as betting exchanges which exist for sports and politics. It's an insulated system where all the profits & losses add up to zero, minus the trading costs.
Futures Exchanges are an important part of the financial system. See, humans have learn about the benefits for civilization of efficient transfer of risk. You need a variety of players with different incentives to interact in a regulated and relative transparent environment. An important part of this ecosystem is the speculator. See Speculator as a Hero
Granted, speculators am not angels; many are motivated by gambling and greed, and when given the chance will take advantage of the public as much as the next person. But the efficiency of a competitive marketplace helps to ferret out and reduce unscrupulous conduct.

The intellectual raises his eyebrows at the economic and historical analysis and contemptuously says, “Man cannot live by bread alone.” To this I respond that without us, there would be no bread.

I am proud to be a speculator. I am proud that my humble attempts to predict Tuesday’s prices on Monday are an indispensable component of our society. By buying low and selling high, I create harmony and freedom.
"whenever there is a randomized way of doing something, then there is a nonrandomized way that delivers better performance but requires more thought" ET Jaynes

User avatar
nisiprius
Advisory Board
Posts: 38287
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: What is wrong with AQR?

Post by nisiprius » Thu Jun 13, 2019 10:50 am

azanon wrote:
Thu Jun 13, 2019 10:39 am
I never personally bought the book, but wasn't managed futures on the "bad or ugly" list in "The Only Guide to Alternative Investments You'll Ever Need: The Good, the Flawed, the Bad, and the Ugly"? If so, when did that change?

I could only find the "good" list quick searching the internet, and I did confirm, it wasn't one of those.
"Managed futures" isn't mentioned by name one way or the other, which tells you something right there: managed futures funds were not familiar enough in 2008 for Swedroe and Kizer to think they were worth mentioning.

The full list:

The Good: Real estate, Inflation-protected securities, commodities [including "CTAs], international equities, fixed annuities, stable-value funds.

The Flawed: High-yield (junk) bonds, private equity, covered calls, socially responsible mutual funds [now called ESG], precious metals equities, preferred stocks, convertible bonds, emerging markets bonds.

The Bad: Hedge funds, leveraged buyouts, variable annuities.

The Ugly: Equity-indexed annuities [now called 'fixed indexed annuities'], structured investment products, leveraged funds.
Last edited by nisiprius on Thu Jun 13, 2019 11:47 am, edited 5 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.

User avatar
nisiprius
Advisory Board
Posts: 38287
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: What is wrong with AQR?

Post by nisiprius » Thu Jun 13, 2019 11:16 am

hdas wrote:
Thu Jun 13, 2019 10:47 am
...You need a variety of players with different incentives to interact in a regulated and relative transparent environment. An important part of this ecosystem is the speculator. See Speculator as a Hero ...
Whether or not speculators are "heroes" in their own minds,

...speculation has nothing to do with "Investing Advice Inspired by Jack Bogle."

John C. Bogle took sides on this. He wrote a book about it:

Image

In Common Sense on Mutual Funds, he wrote--p. 54 of the 2010 "tenth anniversary edition:"
As Table 2.1 demonstrates, investment, or enterprise, has prevailed over speculation in the long run.... In the very long run, speculation has proved to be a neutral factor in the shaping of returns. Speculation cannot feed on itself forever.
To Bogle, speculation was at best "neutral," not heroic.

Even more directly: (Alas, this quotation is "all over the Internet" but I haven't found the source)
Ask yourself: Am I an investor, or am I a speculator? An investor is a person who owns business and holds it forever and enjoys the returns that U.S. businesses, and to some extent global businesses, have earned since the beginning of time. Speculation is betting on price. Speculation has no place in the portfolio or the kit of the typical investor.
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
Topic Author
hdas
Posts: 876
Joined: Thu Jun 11, 2015 8:24 am

Re: What is wrong with AQR?

Post by hdas » Thu Jun 13, 2019 11:34 am

nisiprius wrote:
Thu Jun 13, 2019 11:16 am
hdas wrote:
Thu Jun 13, 2019 10:47 am
...You need a variety of players with different incentives to interact in a regulated and relative transparent environment. An important part of this ecosystem is the speculator. See Speculator as a Hero
Whether or not speculators are "heroes" in their own minds,

...speculation has nothing to do with "Investing Advice Inspired by Jack Bogle."

John C. Bogle took sides on this. He wrote a book about it:

Image

In Common Sense on Mutual Funds, he wrote--p. 54 of the 2010 "tenth anniversary edition:"
As Table 2.1 demonstrates, investment, or enterprise, has prevailed over speculation in the long run.... In the very long run, speculation has proved to be a neutral factor in the shaping of returns. Speculation cannot feed on itself forever.
To Bogle, speculation was at best "neutral," not heroic.

Even more directly: (Alas, this quotation is "all over the Internet" but I haven't found the source)
Ask yourself: Am I an investor, or am I a speculator? An investor is a person who owns business and holds it forever and enjoys the returns that U.S. businesses, and to some extent global businesses, have earned since the beginning of time. Speculation is betting on price. Speculation has no place in the portfolio or the kit of the typical investor.
Nisi,

Yes this is and old debate, more philosophical than practical. Nobody is advocating for other people to become full time speculators. Most people don't have the gumption for it anyways. The point is to stop demonizing the indispensable market participants and embrace the necessary diversity in the market ecosystem. Next time the market crashes and you are scared about the safety your financial nest, you surely will be grateful that some of us will be awake overnight bidding the sp futures betting in the imminent next rise. :greedy
"whenever there is a randomized way of doing something, then there is a nonrandomized way that delivers better performance but requires more thought" ET Jaynes

User avatar
corn18
Posts: 1393
Joined: Fri May 22, 2015 6:24 am

Re: What is wrong with AQR?

Post by corn18 » Thu Jun 13, 2019 12:22 pm

I am super happy that there are people out there trying out their great ideas to beat my average returns. They generate my returns. What a beautiful system.
Don't do something, just stand there!

User avatar
Taylor Larimore
Advisory Board
Posts: 28279
Joined: Tue Feb 27, 2007 8:09 pm
Location: Miami FL

Re: What is wrong with AQR?

Post by Taylor Larimore » Thu Jun 13, 2019 1:03 pm

Bogleheads:

One of the things "wrong with AQR" is that investors using their funds are almost certainly paying the added expense of "revenue sharing" (an industry euphemism for "kick-back.") Vanguard has never participated in Revenue Sharing.

Revenue Sharing Fund Families.

An Alternative Approach to Alternatives: Investing With Even More Style by Larry Swedroe

Best wishes.
Taylor
Last edited by Taylor Larimore on Thu Jun 13, 2019 2:11 pm, edited 1 time in total.
"Simplicity is the master key to financial success." -- Jack Bogle

User avatar
HomerJ
Posts: 12868
Joined: Fri Jun 06, 2008 12:50 pm

Re: What is wrong with AQR?

Post by HomerJ » Thu Jun 13, 2019 2:03 pm

Nothing is wrong with AQR.

The owners of AQR went from not being on the billionaire list to being on the billionaire list (plus some).

From 2017:

https://www.forbes.com/sites/nathanvard ... b352e723a4
At 50, Asness is today also a billionaire. His net worth is estimated at $3 billion. Two of his fellow AQR co-founders, David Kabiller and John Liew, are each worth $1 billion, making AQR one of the few Wall Street firms to mint three billionaires. All three men appear on Forbes’ list of billionaires for the first time. The way Asness and his partners got there says a lot about what is happening on Wall Street these days.
It does indeed say a lot about what is happening on Wall Street. AQR is probably working even better than the owners intended.
The J stands for Jay

columbia
Posts: 1567
Joined: Tue Aug 27, 2013 5:30 am

Re: What is wrong with AQR?

Post by columbia » Thu Jun 13, 2019 5:44 pm

eigenperson wrote:
Thu Jun 13, 2019 9:32 am
nisiprius wrote:
Wed Jun 12, 2019 4:55 pm
The question to me is: who cares?
There are reviews for psychics on Yelp. You and I may laugh at that, since there's no point in differentiating between a good psychic and a bad one, but still -- who wants to go to a bad psychic? There's no possible upside of that.

Some psychics, I'm sure, are able to predict bad events but struggle to predict the good ones. Similarly (at least to my mind), some funds have "a weird sort of volatility profile where [they tend] to mirror the vol of [their] peers on the downside, but have less volatility on the upside." I'm pretty sure both of these are superstitions, but if you believe in the respective superstitions, neither is information you can afford to ignore.
I once walked past pyschic’s storefront. A sign on the door indicated that said person had to go out for coffee

Presumably, a real psychic would have foreseen running out of it....

larryswedroe
Posts: 15835
Joined: Thu Feb 22, 2007 8:28 am
Location: St Louis MO

Re: What is wrong with AQR?

Post by larryswedroe » Thu Jun 13, 2019 7:37 pm

Taylor
Hate to say that but your comments are not only totally false but disgraceful to make such accusations when you don't know the facts. That's fake news at its worst. You should know better.

While you can certainly feel the fees are high or too high, there is no revenue sharing nor any kick backs of any kind by AQR, nor Stone Ridge, nor DFA, nor Bridgeway, nor any other firm I would recommend. None, nada, zilch.
In fact it is illegal for any RIA to take kickbacks or share revenue from any money manager as that violates the fiduciary rule of only accepting payments from the money managers they place client assets with.
Larry

User avatar
Taylor Larimore
Advisory Board
Posts: 28279
Joined: Tue Feb 27, 2007 8:09 pm
Location: Miami FL

Re: What is wrong with AQR?

Post by Taylor Larimore » Thu Jun 13, 2019 7:56 pm

While you can certainly feel the fees are high or too high, there is no revenue sharing nor any kick backs of any kind by AQR, nor Stone Ridge, nor DFA, nor Bridgeway, nor any other firm I would recommend.
Larry:

I am very happy to read this.

Thank you and best wishes.
Taylor
"Simplicity is the master key to financial success." -- Jack Bogle

User avatar
nedsaid
Posts: 11573
Joined: Fri Nov 23, 2012 12:33 pm

Re: What is wrong with AQR?

Post by nedsaid » Thu Jun 13, 2019 9:10 pm

larryswedroe wrote:
Thu Jun 13, 2019 7:37 pm
Taylor
Hate to say that but your comments are not only totally false but disgraceful to make such accusations when you don't know the facts. That's fake news at its worst. You should know better.

While you can certainly feel the fees are high or too high, there is no revenue sharing nor any kick backs of any kind by AQR, nor Stone Ridge, nor DFA, nor Bridgeway, nor any other firm I would recommend. None, nada, zilch.
In fact it is illegal for any RIA to take kickbacks or share revenue from any money manager as that violates the fiduciary rule of only accepting payments from the money managers they place client assets with.
Larry
Larry, I know that Ameriprise had a preferred list of providers who did precisely what Taylor is talking about. Ameriprise did disclose this. The funds that the advisors recommended just happened to be on that preferred list which is the classic conflict of interest. I also know that there was a lawsuit regarding American Funds and the buying of so-called shelf space, my very foggy memory thinks this might be Edward Jones.

Not aware of the rules for Registered Investment Advisors and doubtless you are correct here but this practice has happened in the industry in the past.

I think Fiduciary status has a lot to do with this. I remember years ago I asked one of Ray Lucia's advisors if they were fiduciaries and you would have thought I threw water on the Wicked Witch of the West. He sputtered out an answer but thought it to be very unsatisfactory.

Larry if you could elaborate on this, it would be much appreciated. Not sure if Ameriprise is still doing this. Not sure how the newer Fiduciary rule affected all of this. I know Edward Jones changed their model, they charged a 1.35% AUM Fee coupled with cheap proprietary funds instead of the old model of commission sales. Don't think they sell American Funds anymore.
A fool and his money are good for business.

garlandwhizzer
Posts: 2282
Joined: Fri Aug 06, 2010 3:42 pm

Re: What is wrong with AQR?

Post by garlandwhizzer » Thu Jun 13, 2019 9:50 pm

What's wrong with AQR? High costs, lousy returns in the recent past. Most investors find that combination hard to tolerate for a significant span of years. AQR is fully up to date on all the academic investing literature and it is run by brilliant people who base portfolio selection on a careful analysis of factor models. In recent years that has not produced the anticipated investment success and investors are performance chasing, fleeing and going elsewhere. The question it raises is: does intelligent analysis of and insight into factor models produce improved returns after costs? It's too short a time frame to decide for certain but clearly in recent years the market clearly has not rewarded their sophisticated and sometimes complex investment products. Most investors find underperformance hard to tolerate especially when it costs more.

Garland Whizzer

User avatar
nedsaid
Posts: 11573
Joined: Fri Nov 23, 2012 12:33 pm

Re: What is wrong with AQR?

Post by nedsaid » Thu Jun 13, 2019 11:49 pm

garlandwhizzer wrote:
Thu Jun 13, 2019 9:50 pm
What's wrong with AQR? High costs, lousy returns in the recent past. Most investors find that combination hard to tolerate for a significant span of years. AQR is fully up to date on all the academic investing literature and it is run by brilliant people who base portfolio selection on a careful analysis of factor models. In recent years that has not produced the anticipated investment success and investors are performance chasing, fleeing and going elsewhere. The question it raises is: does intelligent analysis of and insight into factor models produce improved returns after costs? It's too short a time frame to decide for certain but clearly in recent years the market clearly has not rewarded their sophisticated and sometimes complex investment products. Most investors find underperformance hard to tolerate especially when it costs more.

Garland Whizzer
Hi Garland. Markets are dynamic, the problem is that markets might be even more dynamic than I thought and perhaps more dynamic than Cliff Asness and his team believed.

There are huge pools of loosely regulated investment pools called Hedge Funds, funds that make extensive use of shorts, leverage, and derivatives. From what I have read, the market value of the derivative market is many times larger than the markets of the securities upon what they are based. Since Hedge Funds are rather opaque, if even that, impossible to know what big bets are out there. Add to that the flash traders and the ground can shift under the feet of factor models at frightening speed. Sort of like trying to build a skyscraper upon a foundation of quicksand. So when a fund unwinds a very large leveraged bet, markets can experience unexpected volatility. We have seen the Dow fall by 1,000 points and recover the same day!

Asset classes act in certain ways most of the time but in the shorter term just about anything can happen. Asset classes don't have to behave the way we think they should particularly in a crisis. Add in the bets made with leverage, shorts, derivatives and volatility can strike with lightning speed. How do you model that? Markets can crash and recover before you can figure out what might have caused it.

I could pull up a couple of AQR annual reports and I suppose Cliff Asness can tell us there what happened. I suppose that is a homework assignment. If I get time, and can wade through it, I might report back. I will focus on AQR Style Premia I. He and his team have taken on a huge intellectual challenge, I commend them for trying, but it might just be too complex of a problem with too many variables. My guess is that AQR funds haven't done well because Value hasn't done well relative to the market, foggy memory seems to remember comments to this effect on the forum.

I have a Morningstar subscription and looked at what the Morningstar analyst had to say. It is still rated 3 star and the analyst seemed optimistic about the fund's future prospects. The analyst report was written on 12/5/2018 and as I suspected blamed the relatively poor performance of the Value factor. Morningstar did dock this fund on fees, saying that the fund was too expensive.
A fool and his money are good for business.

larryswedroe
Posts: 15835
Joined: Thu Feb 22, 2007 8:28 am
Location: St Louis MO

Re: What is wrong with AQR?

Post by larryswedroe » Fri Jun 14, 2019 8:49 am

nedsaid,
Re Ameriprise, First no one should work with them because they are not fiduciaries, just salespeople. Second, none of the funds we work with pay anyone anything. So if Ameriprise is recommending the funds say of AQR (don't even know if they have access) they are NOT getting any form of kickback from them, nor DFA, nor Bridgeway, nor Stone Ridge.

And on AQR what people are missing (and no fault of theirs) is that many of the strategies which have done poorly recently have done well in live funds over the longer term (just not in public form). And AQR has had two other periods like this when all things seem to go "haywire" ---like in LTCM crisis and in the Quant Crisis described in the Scott Patterson's excellent book. Note btw, that all the trades that blew up on LTCM would have ultimately made money if they could have held on (but excessive leverage prevented that).

I would note that trend following and managed futures, while having no risk story has worked across stocks and bonds and currencies and commodities for literally centuries, and there are even logical explanations for momentum at times. Just think about stocks rising and the shorts having to cover, or stocks crashing and then some investors are forced to sell (either by margin calls or stomach screaming get me out) and then others get pushed into it. But it tends to perform poorly in most periods and then do well when needed most in EXTENDED crashes, so most people cannot hold on. We have not had an extended crash since 08. So no big surprise it has not done well. Really nice piece by Jack Vogel showing how if look at MF for last 10 years the SR is bad, but go back 1 year and it looks GREAT.

Which is why I personally have a small allocation to the strategy inside the AQR alternative risk premium fund. Some downside protection and different source or risk/return.
Larry

User avatar
nedsaid
Posts: 11573
Joined: Fri Nov 23, 2012 12:33 pm

Re: What is wrong with AQR?

Post by nedsaid » Fri Jun 14, 2019 9:17 am

Hi Larry, you have answered the question about kick-backs, and as I suspected it has to do with fiduciary status. Thank you for clarifying that.

I am not down on the factor Alts that use shorts and leverage, neutral actually, but I have seen that you need shorting to isolate the factors from market beta to get a source of return that doesn't correlate to the US Stock Market. The problem with long only factor funds is that most of the return and risk still comes from market beta. So I think a good case can be made for a fund like AQR Style Premia QSPIX to have an allocation of maybe 5% of a portfolio. The light has gone off in my head.

Have looked on the Fidelity website, I have access to certain Alt funds, not sure there is access to a multi-factor Premia (long-short) fund similar to QSPIX. I do have access to certain AQR funds. A lot of the Market Neutral funds just don't look all that impressive, I nick-name them return neutral funds. They seem to do not much, my understanding is that such funds will do well when you have high volatility in the markets. Vanguard has such a fund and its record is not impressive to say the least.

Again, though QSPIX has been somewhat disappointing recently, the analysts at Morningstar had good things to say about it. I am not ready to pull the trigger on these. Lots of folks like me looking for a form of portfolio insurance but have held off sticking to more traditional investments.
A fool and his money are good for business.

larryswedroe
Posts: 15835
Joined: Thu Feb 22, 2007 8:28 am
Location: St Louis MO

Re: What is wrong with AQR?

Post by larryswedroe » Fri Jun 14, 2019 9:35 pm

nedsaid, very much disagree with this statement
but I have seen that you need shorting to isolate the factors from market beta to get a source of return that doesn't correlate to the US Stock Market
And easy to show it's wrong. Here's example with start date the inception of the DFA SV fund in April 93 through year end 2018.
So run a 60/40 (A) with US TSM and say intermediate Treasury using Vanguard for Portfolio A and then because own higher EXPECTED returning DFA SV fund you do 40/60 (B), with same Vanguard intermediate Treasury.

What you see is very similar returns and SD (B slightly better in both cases), but loadings for A are .6 on market beta and .16 on term. For B it's like .41 market beta, .31 on size, .26 on value, and importantly .23 on TERM (so much higher there, helping to pull in tails). So clearly with just long only funds you can diversify more toward risk parity. Then see the worst years and the SR and the best years.

And then can add other sources of risk should one desire like quality or reinsurance and carry or whatever to create even more risk parity with long only funds.

Bottom line don't need to add long-short if willing to have very high tilts. Now of course would add international exposure as well.

Best wishes
Larry

User avatar
nisiprius
Advisory Board
Posts: 38287
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: What is wrong with AQR?

Post by nisiprius » Fri Jun 14, 2019 10:11 pm

larryswedroe wrote:
Fri Jun 14, 2019 8:49 am
...I would note that trend following and managed futures...tend to perform poorly in most periods and then do well when needed most in EXTENDED crashes, so most people cannot hold on. We have not had an extended crash since 08. So no big surprise it has not done well. Really nice piece by Jack Vogel showing how if look at MF for last 10 years the SR is bad, but go back 1 year and it looks GREAT.
]
Image
The grey line is the Credit Suisse managed futures index. The blue "managed futures" is the Morningstar category average for managed futures funds. I'm not seeing either of them "looking great" or "doing well when needed most" during 2008-2009. They didn't crash by -50%, but neither did my Vanguard Total Bond fund.

I do not see any place at all in the blue line (managed futures category average) that could be described as "doing great."
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: 3902
Joined: Fri Feb 23, 2007 8:21 pm

Re: What is wrong with AQR?

Post by Random Walker » Sat Jun 15, 2019 12:18 am

nedsaid wrote:
Fri Jun 14, 2019 9:17 am
I am not down on the factor Alts that use shorts and leverage, neutral actually, but I have seen that you need shorting to isolate the factors from market beta to get a source of return that doesn't correlate to the US Stock Market. The problem with long only factor funds is that most of the return and risk still comes from market beta. So I think a good case can be made for a fund like AQR Style Premia QSPIX to have an allocation of maybe 5% of a portfolio. The light has gone off in my head.
This brings up the issue of cost per unit value added versus cost alone. The long only asset class funds are generally more expensive than the total market or core funds. If one chooses an asset class fund with deeper exposure to the desired factors, then he will need less of the more expensive fund to achieve his factor targets. Moreover, he will need to take on less market factor risk (the risk he is trying to diversify away from) to achieve the exposure he wants to the other factors.

Dave

User avatar
nedsaid
Posts: 11573
Joined: Fri Nov 23, 2012 12:33 pm

Re: What is wrong with AQR?

Post by nedsaid » Sat Jun 15, 2019 1:21 am

larryswedroe wrote:
Fri Jun 14, 2019 9:35 pm
nedsaid, very much disagree with this statement
but I have seen that you need shorting to isolate the factors from market beta to get a source of return that doesn't correlate to the US Stock Market
And easy to show it's wrong. Here's example with start date the inception of the DFA SV fund in April 93 through year end 2018.
So run a 60/40 (A) with US TSM and say intermediate Treasury using Vanguard for Portfolio A and then because own higher EXPECTED returning DFA SV fund you do 40/60 (B), with same Vanguard intermediate Treasury.

What you see is very similar returns and SD (B slightly better in both cases), but loadings for A are .6 on market beta and .16 on term. For B it's like .41 market beta, .31 on size, .26 on value, and importantly .23 on TERM (so much higher there, helping to pull in tails). So clearly with just long only funds you can diversify more toward risk parity. Then see the worst years and the SR and the best years.

And then can add other sources of risk should one desire like quality or reinsurance and carry or whatever to create even more risk parity with long only funds.

Bottom line don't need to add long-short if willing to have very high tilts. Now of course would add international exposure as well.

Best wishes
Larry
Larry, then what the heck do we need AQR Style Premia Fund QSPIX for? Why would we need AQR at all? The answer would be that these type of Alt funds would be useful if you didn't want to do the extreme tilts that you show above. As far as the Alts that are interval funds, hard to access without an advisor, we are in a largely do-it-yourself forum.

I get your point but most folks here won't do the extreme tilts.

The portfolios here at Bogleheads need stocks to propel their returns and most of that return comes from Market Beta. Even a stock portfolio of 100% Small Cap Value stocks will derive most of its returns from Market Beta. It seems that the premiums you get on Small/Value stocks vs. Total Market is 2% to 3% a year. We can stretch it to 4% for discussion's sake. So if you got let's say 12% a year from Small/Value, probably 2/3 of that return or more is from Market Beta. I would venture to say that most of that risk is also from Market Beta. Indeed, Small/Value is also more volatile than just Market Beta.

Your example shows that you need pretty extreme tilts in a long only portfolio to achieve what you described. I for one am not willing to put all my stocks into Small Value even if they are internationally diversified. The "Larry portfolio" is a good one but one that I would not do for myself for pretty deep philosophical reasons. I just could not ignore Large-Caps. It gives you great factor diversification but not the diversification across asset classes that I would desire.

I know that your strategy guards against extreme distribution of investment returns, particularly on the negative side of the bell-shaped curve. It has a very good chance of working this way in the future but obviously no guarantees. What you say makes lots of sense but in my world, I would have a tough time implementing that. One problem is if Small/Value has an extended period of underperformance relative to Total Market. Good thing this hasn't been a problem over the last 10 years. Over the last 5 years, not so good. A "Larry portfolio" would necessitate extreme patience. Also in a low interest rate environment, the bonds won't help much with performance.

What I will say is that it appears my milder tilts seem to have made a difference with my portfolio. Unfortunately, Value has been on an extended vacation so my tilts have been a drag on performance over the last decade. In a Large Growth environment that we have seen since the 2008-2009 financial crisis, that is about what one would expect, no surprises there.
A fool and his money are good for business.

User avatar
3CT_Paddler
Posts: 3260
Joined: Wed Feb 04, 2009 5:28 pm
Location: Marietta, GA

Re: What is wrong with AQR?

Post by 3CT_Paddler » Sat Jun 15, 2019 5:55 am

larryswedroe wrote:
Fri Jun 14, 2019 9:35 pm


What you see is very similar returns and SD (B slightly better in both cases), but loadings for A are .6 on market beta and .16 on term. For B it's like .41 market beta, .31 on size, .26 on value, and importantly .23 on TERM (so much higher there, helping to pull in tails). So clearly with just long only funds you can diversify more toward risk parity. Then see the worst years and the SR and the best years.

And then can add other sources of risk should one desire like quality or reinsurance and carry or whatever to create even more risk parity with long only funds.

Bottom line don't need to add long-short if willing to have very high tilts. Now of course would add international exposure as well.

Best wishes
Larry
Larry, I respect your work, but this approach has gone further and further into complexity beyond my understanding. Bogle's Costs Matter Hypothesis and diversification is a free lunch are the only two things I need to understand to get superior returns to most active investors.

What would Bogle say about AQR? I know you would probably disagree with his simple approach, but the long term results and the solid fundamentals of his argument win the day for me.

bgf
Posts: 922
Joined: Fri Nov 10, 2017 9:35 am

Re: What is wrong with AQR?

Post by bgf » Sat Jun 15, 2019 7:13 am

larryswedroe wrote:
Fri Jun 14, 2019 9:35 pm
nedsaid, very much disagree with this statement
but I have seen that you need shorting to isolate the factors from market beta to get a source of return that doesn't correlate to the US Stock Market
And easy to show it's wrong. Here's example with start date the inception of the DFA SV fund in April 93 through year end 2018.
So run a 60/40 (A) with US TSM and say intermediate Treasury using Vanguard for Portfolio A and then because own higher EXPECTED returning DFA SV fund you do 40/60 (B), with same Vanguard intermediate Treasury.

What you see is very similar returns and SD (B slightly better in both cases), but loadings for A are .6 on market beta and .16 on term. For B it's like .41 market beta, .31 on size, .26 on value, and importantly .23 on TERM (so much higher there, helping to pull in tails). So clearly with just long only funds you can diversify more toward risk parity. Then see the worst years and the SR and the best years.

And then can add other sources of risk should one desire like quality or reinsurance and carry or whatever to create even more risk parity with long only funds.

Bottom line don't need to add long-short if willing to have very high tilts. Now of course would add international exposure as well.

Best wishes
Larry
what is wrong with AQR?

why is it that following the strongest criticism of modern financial theory by behavioral economics and, in my uneducated opinion, by mandelbrot, who wrote the clearest explanation of when, how, and why the tools of modern financial theory fail, have AQR and others basically doubled down on all the same assumptions with factors?

are factors not dependent on the same fundamental assumptions and tools? just another patch and complication to save a failing theory before the next paradigm shift?

i apologize, but i see stuff like "0.41 market beta, 0.31 on size, 0.26 on value" and red flags pop up. the method and tools at your disposal are incapable of supporting such a conclusion. 0.41 market beta? not 0.42? not 0.40? i just can't buy it. i can't.

i'd love to read more about this in a fashion not full of math, which i wouldn't be able to follow as a layman.
“TE OCCIDERE POSSUNT SED TE EDERE NON POSSUNT NEFAS EST"

Forester
Posts: 130
Joined: Sat Jan 19, 2019 2:50 pm
Location: UK

Re: What is wrong with AQR?

Post by Forester » Sat Jun 15, 2019 7:21 am

nisiprius wrote:
Fri Jun 14, 2019 10:11 pm
larryswedroe wrote:
Fri Jun 14, 2019 8:49 am
...I would note that trend following and managed futures...tend to perform poorly in most periods and then do well when needed most in EXTENDED crashes, so most people cannot hold on. We have not had an extended crash since 08. So no big surprise it has not done well. Really nice piece by Jack Vogel showing how if look at MF for last 10 years the SR is bad, but go back 1 year and it looks GREAT.
]
Image
The grey line is the Credit Suisse managed futures index. The blue "managed futures" is the Morningstar category average for managed futures funds. I'm not seeing either of them "looking great" or "doing well when needed most" during 2008-2009. They didn't crash by -50%, but neither did my Vanguard Total Bond fund.

I do not see any place at all in the blue line (managed futures category average) that could be described as "doing great."
The Vogel piece was primarily about trend following equity indices, the managed futures index you posted is not a like-for-like comparison.

User avatar
vineviz
Posts: 4036
Joined: Tue May 15, 2018 1:55 pm

Re: What is wrong with AQR?

Post by vineviz » Sat Jun 15, 2019 7:35 am

bgf wrote:
Sat Jun 15, 2019 7:13 am
i apologize, but i see stuff like "0.41 market beta, 0.31 on size, 0.26 on value" and red flags pop up. the method and tools at your disposal are incapable of supporting such a conclusion. 0.41 market beta? not 0.42? not 0.40? i just can't buy it. i can't.
The factor loadings you mention are simply the outputs of a linear regression: they are measurements of a specific sample at a specific time. You never know what the measurement for the entire population is, but you can reasonably estimate it using common and accepted statistical tools.

Even if it's not reported, each of those regression betas has a confidence interval associated with it. That "0.41 market beta" result might have a 95% confidence interval of 0.36 to 0.46.

Could it plausibly be 0.42? Absolutely.

Could it plausibly be 0.00? Absolutely not.

Linear regressions are not TERRIBLY sophisticated statistical tools, but the field of statistics heavily depends on math so I don't think it's likely to find a completely satisfactory in-depth explanation that doesn't involve math.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

marcopolo
Posts: 2094
Joined: Sat Dec 03, 2016 10:22 am

Re: What is wrong with AQR?

Post by marcopolo » Sat Jun 15, 2019 8:12 am

vineviz wrote:
Sat Jun 15, 2019 7:35 am
bgf wrote:
Sat Jun 15, 2019 7:13 am
i apologize, but i see stuff like "0.41 market beta, 0.31 on size, 0.26 on value" and red flags pop up. the method and tools at your disposal are incapable of supporting such a conclusion. 0.41 market beta? not 0.42? not 0.40? i just can't buy it. i can't.
The factor loadings you mention are simply the outputs of a linear regression: they are measurements of a specific sample at a specific time. You never know what the measurement for the entire population is, but you can reasonably estimate it using common and accepted statistical tools.

Even if it's not reported, each of those regression betas has a confidence interval associated with it. That "0.41 market beta" result might have a 95% confidence interval of 0.36 to 0.46.

Could it plausibly be 0.42? Absolutely.

Could it plausibly be 0.00? Absolutely not.

Linear regressions are not TERRIBLY sophisticated statistical tools, but the field of statistics heavily depends on math so I don't think it's likely to find a completely satisfactory in-depth explanation that doesn't involve math.
Regression is backwards looking. How stationary are these statistics?

How come when we build these models based on PAST results, no one seems concerned about "resulting" as Larry calls it, but when we try to evaluate how well strategies built on those models perform using results after the models are built, all of a sudden, results are meaningless, it is only the strategy that matters?

By the way, another thread about AQR...
Once in a while you get shown the light, in the strangest of places if you look at it right.

Random Walker
Posts: 3902
Joined: Fri Feb 23, 2007 8:21 pm

Re: What is wrong with AQR?

Post by Random Walker » Sat Jun 15, 2019 8:29 am

marcopolo wrote:
Sat Jun 15, 2019 8:12 am
How come when we build these models based on PAST results, no one seems concerned about "resulting" as Larry calls it, but when we try to evaluate how well strategies built on those models perform using results after the models are built, all of a sudden, results are meaningless, it is only the strategy that matters?
Results are very important. Out of sample tests are a very important test for any model as well. We can use a past data set to help generate a model, but then need to look out of sample to see if the model holds up. One can look in different time periods, different markets, and sometimes even different asset classes.

Dave

marcopolo
Posts: 2094
Joined: Sat Dec 03, 2016 10:22 am

Re: What is wrong with AQR?

Post by marcopolo » Sat Jun 15, 2019 8:37 am

Random Walker wrote:
Sat Jun 15, 2019 8:29 am

Results are very important.
Dave
We may have finally found an area where you and Larry appear to disagree.
Look at his recent posts about the dangers of "resulting".
Once in a while you get shown the light, in the strangest of places if you look at it right.

larryswedroe
Posts: 15835
Joined: Thu Feb 22, 2007 8:28 am
Location: St Louis MO

Re: What is wrong with AQR?

Post by larryswedroe » Sat Jun 15, 2019 9:05 am

Nedsaid
Think viniez answered your question. And it doesn't matter if .41 or .42 anyway, it matters that you have more of a risk parity approach. Note pretty smart people like David Swensen and Ray Dalio invest this way!

And no you don't need to use such a high tilt, there are trade offs though, the more you tilt the more you achieve risk parity. Or can add other assets like REINSURANCE and selling volatility insurance (the VRP) to move in that direction. It's all about direction you are moving, and individual decisions.

You need AQR because you cannot do long-short on own, nor trade efficiently nor leverage as cheaply nor do it across multiple asset classes. Now if don't do this then need to tilt more or add other unique sources of risk to move in that direction.

Best
Larry

bgf
Posts: 922
Joined: Fri Nov 10, 2017 9:35 am

Re: What is wrong with AQR?

Post by bgf » Sat Jun 15, 2019 9:23 am

vineviz wrote:
Sat Jun 15, 2019 7:35 am
bgf wrote:
Sat Jun 15, 2019 7:13 am
i apologize, but i see stuff like "0.41 market beta, 0.31 on size, 0.26 on value" and red flags pop up. the method and tools at your disposal are incapable of supporting such a conclusion. 0.41 market beta? not 0.42? not 0.40? i just can't buy it. i can't.
The factor loadings you mention are simply the outputs of a linear regression: they are measurements of a specific sample at a specific time. You never know what the measurement for the entire population is, but you can reasonably estimate it using common and accepted statistical tools.

Even if it's not reported, each of those regression betas has a confidence interval associated with it. That "0.41 market beta" result might have a 95% confidence interval of 0.36 to 0.46.

Could it plausibly be 0.42? Absolutely.

Could it plausibly be 0.00? Absolutely not.

Linear regressions are not TERRIBLY sophisticated statistical tools, but the field of statistics heavily depends on math so I don't think it's likely to find a completely satisfactory in-depth explanation that doesn't involve math.
the method of least sqaures only works if your data is normally distributed. unfortunately, we know that stock price changes are not normally distributed. so we know that the model is "wrong" with respect to the frequency and amplitude of price change fluctuations. ok fine, we know all models are "wrong." but the model assumes that the largest price changes are nowhere near as large as they actually are and it also assumes that they are far less frequent than they are.

we also know that these outliers cluster and that when the outliers pop up it is most important that the models work. in other words, the models are likeliest to fail just when they are most needed. that's a big problem.

you are correct that linear regression is not complicated, but it also isn't very useful in the context of financial markets.

this is what im hoping would be addressed.
“TE OCCIDERE POSSUNT SED TE EDERE NON POSSUNT NEFAS EST"

bgf
Posts: 922
Joined: Fri Nov 10, 2017 9:35 am

Re: What is wrong with AQR?

Post by bgf » Sat Jun 15, 2019 10:05 am

larryswedroe wrote:
Sat Jun 15, 2019 9:05 am
Nedsaid
Think viniez answered your question. And it doesn't matter if .41 or .42 anyway, it matters that you have more of a risk parity approach. Note pretty smart people like David Swensen and Ray Dalio invest this way!

And no you don't need to use such a high tilt, there are trade offs though, the more you tilt the more you achieve risk parity. Or can add other assets like REINSURANCE and selling volatility insurance (the VRP) to move in that direction. It's all about direction you are moving, and individual decisions.

You need AQR because you cannot do long-short on own, nor trade efficiently nor leverage as cheaply nor do it across multiple asset classes. Now if don't do this then need to tilt more or add other unique sources of risk to move in that direction.

Best
Larry
indeed, a lot of very smart people invest in a variety of ways, to varying degrees of "success."

i don't reject the risk parity approach in broad strokes. for example, i can buy into it on the level of stocks, catastrophe bonds, bitcoin, and antique furniture are likely to be uncorrelated. the problem is, i have neither the inclination nor the ability to invest in all of those things.

i'm OK buying into that correlation argument because they are fundamentally, different THINGS.

i have a far more difficult time buying into the approach by AQR or other factor models that slice and dice within categories of the same THINGS.
“TE OCCIDERE POSSUNT SED TE EDERE NON POSSUNT NEFAS EST"

User avatar
vineviz
Posts: 4036
Joined: Tue May 15, 2018 1:55 pm

Re: What is wrong with AQR?

Post by vineviz » Sat Jun 15, 2019 10:19 am

marcopolo wrote:
Sat Jun 15, 2019 8:12 am
How come when we build these models based on PAST results, no one seems concerned about "resulting" as Larry calls it, but when we try to evaluate how well strategies built on those models perform using results after the models are built, all of a sudden, results are meaningless, it is only the strategy that matters?
Give me some data from the future and I'll use that data instead.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

User avatar
vineviz
Posts: 4036
Joined: Tue May 15, 2018 1:55 pm

Re: What is wrong with AQR?

Post by vineviz » Sat Jun 15, 2019 10:26 am

bgf wrote:
Sat Jun 15, 2019 9:23 am
the method of least sqaures only works if your data is normally distributed.
This isn't actually true. A least squares regression has no requirements, explicit or implicit, that require the data have a Gaussian or any other type of distribution.
bgf wrote:
Sat Jun 15, 2019 9:23 am
unfortunately, we know that stock price changes are not normally distributed.
Yes, and we also have tools to evaluate how closely those data are approximated by a normal distribution and whether the approximation is close enough or not to be sufficient for our needs.

Asset return data generally aren't perfectly Gaussian, it's true, but they are close enough that we can safely use an assumption of normality for virtually every application of consequence. When it's not "close enough", we use some other assumption.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

larryswedroe
Posts: 15835
Joined: Thu Feb 22, 2007 8:28 am
Location: St Louis MO

Re: What is wrong with AQR?

Post by larryswedroe » Sat Jun 15, 2019 10:33 am

BGF
The problem with your argument is that they are in fact different things, with unique sources of risk, which is why they are included in asset pricing models that explain the variation of returns of diversified portfolios. And they are even different across asset classes, so value and momentum and carry and defensive work across asset classes (demonstrating pervasiveness and that they are not result of data mining exercises). Now out of the 600 factors in the zoo, very few meet that criteria.

Marco, fwiw you are completely misunderstanding what I said, of course results matter, but you have to consider what alternative universes might have shown up (perfect example is you don't buy life insurance despite having three young kids and stay at home wife and don't die and think you were brilliant because you didn't die and saved the premium, or you put 100% of assets into say Google). And of course you have to consider time horizon and here is where most investors fail.They don't understand that 10 years in risk assets is noise, or at least can be.
Larry
Last edited by larryswedroe on Sat Jun 15, 2019 10:36 am, edited 1 time in total.

bgf
Posts: 922
Joined: Fri Nov 10, 2017 9:35 am

Re: What is wrong with AQR?

Post by bgf » Sat Jun 15, 2019 10:36 am

vineviz wrote:
Sat Jun 15, 2019 10:26 am
bgf wrote:
Sat Jun 15, 2019 9:23 am
the method of least sqaures only works if your data is normally distributed.
This isn't actually true. A least squares regression has no requirements, explicit or implicit, that require the data have a Gaussian or any other type of distribution.

now we're getting somewhere. what distribution is assumed? power law? cauchy? do they just use one or do they somehow incorporate multiple?
bgf wrote:
Sat Jun 15, 2019 9:23 am
unfortunately, we know that stock price changes are not normally distributed.
Yes, and we also have tools to evaluate how closely those data are approximated by a normal distribution and whether the approximation is close enough or not to be sufficient for our needs.

Asset return data generally aren't perfectly Gaussian, it's true, but they are close enough that we can safely use an assumption of normality for virtually every application of consequence. When it's not "close enough", we use some other assumption.

they weren't so safe for LTCM. but also that brings me back to the original point. are we safely using the assumption of normality or not?
“TE OCCIDERE POSSUNT SED TE EDERE NON POSSUNT NEFAS EST"

larryswedroe
Posts: 15835
Joined: Thu Feb 22, 2007 8:28 am
Location: St Louis MO

Re: What is wrong with AQR?

Post by larryswedroe » Sat Jun 15, 2019 10:46 am

BGF, LTCM failed because they overused leverage at more than 100X at one point. In fact all their positions would have been profitable if they could have held to the end. So not at all a relevant example except to show how dumb excessive leverage can be, and in addition they decided they were so smart that they took positions which had nothing to do with their models, just directional bets.

BTW, here's a thought. The argument goes that should should hold market portfolio because it's efficient. Well consider that we have two very different "market portfolios" depending on the type of investor. You have retail which overweights the right hand side of the factors (the short side) and institutional (like hedge funds) that overweight the left said. And the institutions on gross basis outperform the retail side by pretty wide margins. So that raises the question of why you should not own the market as does the institutional crowd? Just saying.

Larry

Post Reply