AQR fund family's performance relative to its benchmarks

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
User avatar
Topic Author
nisiprius
Advisory Board
Posts: 42244
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

AQR fund family's performance relative to its benchmarks

Post by nisiprius »

See also AQR and Vanguard funds performance vs. stated benchmarks, updated 8/31/2020
________________________________________________________________________________

I was sufficiently intrigued by bjr89's comment, on AQR funds, that "they pretty much all underperform their benchmarks." Eventually it got to me, to the point where I was willing to point and click and copy and paste fifty-six times in order to find out the answer. As is so often the case, it's tricky to figure out how to make a fair summary of the data. The most difficult problem for me is that quite a few of AQR's funds list two benchmarks--for example, QCELX, ASMOX, ATIMX, and ATSMX are compared both to the Russell 2000 and the Russell 2000 Growth index. Also, for quite a few they list several different returns having to do with taxes.

Anywhere, here's what I did. I got data for every AQR fund listed on their Total Returns page (except QVPIX, the Volatility Risk Premium Fund, for which no data is shown). I obtained returns since inception. I ignored all "returns after taxes" lines (thus in every case using a number favorable to AQR). I went to the fund's page and opened the "Standardized Performance" tab. If two benchmarks were shown, I entered two lines in my spreadsheet, listing the fund twice. I copied the benchmark name, in some cases abbreviating it and/or normalizing it when they used slightly different names for the same benchmark index. Finally, I sorted the results by benchmark name (which also groups funds by general category) and charted them.

I tried to be as objective as possible, simply using AQR's data with as little selection or "processing" (other than calculating performance relative to benchmark). Or rather, I tried to do this for the first chart.

In many of the cases where a fund outperformed its benchmark, the benchmark was "Bank of America ML 3 Month T-Bill Index." I find it disturbing that a fund company is allowed to benchmark a fund like QSPIX, with an intentionally targeted volatility of 10%, against a benchmark with a volatility of 3%; I can't regard that as appropriate. So, after doing the chart "straight," I editorialized by adding a second chart from which I arbitrarily removed every fund benchmarked against Treasury bills.

In calculating averages and percentages I just took the numbers and counts as I had tabulated them, not trying to do anything about funds with two benchmarks being counted twice.

I calculated "performance relative to the benchmark" as the "geometric difference." I could have just subtracted the numbers and it would make virtually no difference, but I think "geometric difference" is more strictly correct. The geometric difference of X% "minus" Y% is (1+X%)/(1+Y%)-1. For example, the geometric difference of 12% and 2% is 1.12/1.02 - 1 = 9.8% (instead of 10%).

No, I haven't memorized the AQR tickers, but after fiddling around with Excel's charts I just could not find any good way to get the names onto the charts. You'll just have to look them up, but you should be able to guess the general category from the benchmark.

Here is the result for the full list:

Image

27% beat their benchmarks, 73% failed to beat their benchmarks, and the (raw, arithmetic, double-counting some funds) difference from the benchmark was -1.05%.

If we throw out the funds that benchmark to T-bills, we see this:

Image

17% beat their benchmarks, 83% failed to beat their benchmarks, and the average difference from the benchmark was -1.41%.
Last edited by nisiprius on Sun Sep 20, 2020 10:14 am, edited 4 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
tarheel
Posts: 412
Joined: Fri Sep 17, 2010 8:33 am

Re: AQR fund family's performance versus its benchmarks

Post by tarheel »

Nisi, thanks for taking the time, very interesting. As an AQR investor, my take is that things actually look pretty good.

As far as equities, I have significant allocations to QSMLX (-0.39%), QCELX (+0.07%), QICLX (-0.79%) and QEELX (-1.18%).

This looks about right to me - these multistyle funds all had inception dates in the 2013-2014 range, and without question factor based investing, and in particular the value premium, has performed poorly over the last five years. I would expect to underperform.

Once again, people should not be investing in these funds unless they are in it for the long term and understand what they are investing in.

My small allocations to QSPIX (+4.28%) and QMHIX (-0.85%) look fine too.

I will add that IMHO making any kind of blanket judgement over relative performance in a five year period is tricky at best.
User avatar
nedsaid
Posts: 13955
Joined: Fri Nov 23, 2012 12:33 pm

Re: AQR fund family's performance relative to its benchmarks

Post by nedsaid »

Nisiprius, you have raised good points here. From the threads, we are told to expect equity like returns from Alternatives but they are benchmarked against Treasury Bills. There are arguments that such funds are their own benchmarks. They are too transparent or not transparent enough. The reality is that these funds are unique creatures, hard to define just what these are since each fund is different, much less find a good benchmark. For myself, if I owned such investments, I would expect returns somewhere between the returns of bonds and the returns of stocks. I would be satisfied with 5% to 7% returns with low correlation to stocks but I am not clear that these funds even deliver the returns of bonds. That is my take, 5% to 7% returns with low correlation to stocks or they are not worth owning. Last I looked, QSPIX, AQR Style Premia Fund had a five year return of 5.05%, which would be acceptable to me. But, you would want to look at a basket of such funds and my sense is that the return of the basket would be about the same as bonds.
A fool and his money are good for business.
gtwhitegold
Posts: 441
Joined: Fri Sep 21, 2012 1:55 pm

Re: AQR fund family's performance relative to its benchmarks

Post by gtwhitegold »

I would argue that the alternatives should be benchmarked against a mixed equity and bond portfolio consisting of intermediate term treasuries and a total world index targeted to either 10 or 15% volatility depending on the strategy if you are determined to directly compare it to traditional assets, however the best measurement is probably to compare a similar portfolio targeting a specific average volatility with and without the liquid alternative.
bjr89
Posts: 151
Joined: Sat Apr 23, 2016 8:15 am

Re: AQR fund family's performance relative to its benchmarks

Post by bjr89 »

-----
Last edited by bjr89 on Fri Nov 16, 2018 7:11 pm, edited 1 time in total.
Theoretical
Posts: 1514
Joined: Tue Aug 19, 2014 10:09 pm

Re: AQR fund family's performance relative to its benchmarks

Post by Theoretical »

bjr89 wrote: Mon Nov 12, 2018 4:14 pm As a hedge fund investor, you want to pay performance fees so they actually have an incentive to try. Otherwise you get stuck with an under-performing asset gatherer like AQR.
This is precisely the justification behind one of the oldest CTAs, Dunn Capital Management, which has a 0/25% cost structure and runs an extremely volatile -- as in more volatile than Emerging Markets stocks, but still thoroughly diversified, portfolio. It can and does make gobtons of money along with losing it. Since this is an important question for these funds, this CTA is investable on the retail side through Arrow Funds at Fidelity - MFTNX - since 2015.

I think investors' being inherently averse to volatility really makes these kinds of funds have a temptation to run them essentially as bond alternatives. The problem is you need a lot of it then to actually make a measurable difference on the portfolio, which gets even more expensive. At the end of the day, you end up with extremely expensive bonds with a nasty ER drag and limited upside for the vast majority of these funds.
hdas
Posts: 1342
Joined: Thu Jun 11, 2015 8:24 am

[Deleted]

Post by hdas »

[Deleted]
Last edited by hdas on Wed Jan 29, 2020 11:41 am, edited 1 time in total.
....
JackoC
Posts: 1869
Joined: Sun Aug 12, 2018 11:14 am

Re: AQR fund family's performance relative to its benchmarks

Post by JackoC »

1. ...If the odd institution wishes to protect itself in this way there is no contradiction, but if they all do, the risk of destabilising short term market behaviour will once again be high.

2. Winton is a leader in this industry, a beneficiary of some of the trends described and loser from others. That our industry is now an accepted part of the asset management community whilst we are a leader within it is positive. That there are now many cheap providers who implicitly question the real value offered by our services is obviously not so welcome to us.

1. This isn't much of an argument why 'the less sophisticated' (too *retail*!) trend following strategies wouldn't work. it's an argument on some kind of a macro basis, 'this is bad for the market'. It's another new twist. First we had the switch from 'AQR is too much of a black box' to 'AQR is too transparent'. Now a new argument from a 'macro prudential' POV about momentum/trend following. If everyone is selling into downturns, it would seem a prisoner's dilemma situation where as one investor I should also want to. Portfolio insurance didn't work as advertised in the 1987 crash, but as one investor would I have been better off the afternoon of 10/19/87 selling equity index futures like the PI guys were, or buying them? If that's bad for the market and society, that's a regulation and ultimately political issue, not about 'individual and actionable'.

2. I view this article very much through that prism without them even telling me to. :D

I'll repeat some disclaimers because people don't always remember or pay attention to previous posts. I'm not a partisan of AQR. And I also find it altogether plausible the simple market anomalies could persist for a long time, but not forever, and a bigger industry chasing after them *could* be a reason for that. Although, I still see 'dead alpha marketed as beta' as more of a clever phrase than a proven hypothesis that alternative beta or 'style premia' is just pure market inefficiency that can be (and has been) driven out. If that were entirely so it's puzzling how long some such anomalies have persisted. It's not as if the market has gotten that much more efficient just in very recent years.

It could be combination of various factors that reduces the return on what actually *are* alternative risks. For example some academic work has been done on explaining a risk basis for currency carry (this forum tends to always come around to stocks, but not all these things are stock market, the most discussed AQR fund here, one that hasn't done badly, QSPIX, is mainly not stock). Ie that going long forwards on currencies with higher short term rates against ones with lower short term rates has generally made money, the 'market expectation' as some people too loosely the term the fact that lower rate currencies are more expensive forward, 'doesn't fully come to pass', on average. To the extent that's less true now, it is pretty clearly not because of retail oriented alternatives funds in recent years. There were legions of people trading on currency carry long before. One explanation might be less synchronized inflation cycles in various countries in the past than now. IOW it is/was a real beta/'style premium' to some degree. That still doesn't gtee it will always exist.
Last edited by JackoC on Tue Nov 13, 2018 12:08 am, edited 1 time in total.
fennewaldaj
Posts: 979
Joined: Sun Oct 22, 2017 11:30 pm

Re: AQR fund family's performance relative to its benchmarks

Post by fennewaldaj »

JackoC wrote: Mon Nov 12, 2018 11:55 pm Now a new argument from a 'macro prudential' POV. If everyone is selling into downturns, let's just say, that means I individually should not? No, it doesn't.
It might though right? If a bunch of known selling is happening there may be some edge from finding a signal that sells before all this other selling. On the other side too much systematic selling could lead to overselling every time the market is down-trending which might make it better to always be buying in these spots.
JackoC
Posts: 1869
Joined: Sun Aug 12, 2018 11:14 am

Re: AQR fund family's performance relative to its benchmarks

Post by JackoC »

fennewaldaj wrote: Tue Nov 13, 2018 12:03 am
JackoC wrote: Mon Nov 12, 2018 11:55 pm Now a new argument from a 'macro prudential' POV. If everyone is selling into downturns, let's just say, that means I individually should not? No, it doesn't.
It might though right? If a bunch of known selling is happening there may be some edge from finding a signal that sells before all this other selling. On the other side too much systematic selling could lead to overselling every time the market is down-trending which might make it better to always be buying in these spots.
If you knew in advance what was going to happen you'd always have a huge edge. But the point they are making is that lots of selling into downturns, by the less sophisticated hoi polloi of trend followers from their POV, is bad. And that could right, but would be bad for the whole market and by extension society. I edited too late adding the concrete example. On Oct 19 1987 shops attempting 'portfolio insurance'* were selling in larger volume as the market declined, overwhelming buyers. But was it better to join them or to to buy? It was bad for the market to join them, but not for the individual seller joining them. The article is not saying you can make money off the trend following lower classes when it hits the fan, it's saying all that selling could destabilize the market, make it go down more. Again that's a regulatory perspective, not an individual and actionable one.

*in simplest terms this meant an idea of replicating the delta (market position) of owning put options on the stock market, without actually having to buy them. As the market went down, the delta of the hypothetical put went up, meaning selling more and more (futures, most easily). The market's speed of drop got way ahead of what they were supposed to be doing, but anyway they were selling, with the market going down.
User avatar
Topic Author
nisiprius
Advisory Board
Posts: 42244
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: AQR fund family's performance relative to its benchmarks

Post by nisiprius »

I corrected a small error in the original charts. QCELX had been incorrectly compared to the Russell 2000 and Russell 2000 Growth, instead of the Russell 1000.
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.
afan
Posts: 5282
Joined: Sun Jul 25, 2010 4:01 pm

Re: AQR fund family's performance relative to its benchmarks

Post by afan »

The best way to identify a benchmark is start by looking to R2. Then compare the return distributions. Do they have similar standard deviation, skewness and kurtosis?

If so, then you have a good benchmark. If you pick an asset with drastically different SD then that is obviously not an appropriate benchmark.

If the closest one can come is a benchmark that is otherwise suitable but has much lower volatility, then one can correct for that. Leverage up the low volatility benchmark to get the same variance. Then compare returns.

Of course, even that fails to capture what the funds are touted to provide. They are supposed to increase the risk adjusted returns of a stock/bond portfolio. So add 10% magical fund to a 55/35 stock/bond portfolio and compare returns.

Then you have the problem of predicting how long the fund characteristics will remain stable. For an index fund one expects the investment to remain the same. For an actively managed fund there is every reason to expect it to change. The factors chosen can change. The factor weights can change. You start with the idea of choosing factors with low or negative correlation with stocks. After that everything is up for grabs. Whatever the performance may have been over one period, there is no reason to expect it to persist.

It takes huge datasets to explore performance persistence among funds. The results for active funds, which is still most of them, suggest very low persistence. With the miniscule data available for these AQR funds there is no basis for making any predictions.

Although entertaining, I suppose, there is no reason to expect an analysis of past performance of these funds to tell you anything about what they will do in the future.
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
betablocker
Posts: 505
Joined: Mon Jan 11, 2016 1:26 pm

Re: AQR fund family's performance relative to its benchmarks

Post by betablocker »

I'm curious about the hostility that AQR inspires on the board. Bogleheads can be an emotional bunch! AQR has performed well over the years delivering exactly what they promised. Yes there has been a long term underperformance in value and recently the other factors have also underperformed. It happens sometimes just like all things underperform from time to time. In my mind AQR is a model of transparency and low fees for what they offer (long/short, hedge fund strategies, etc.). You might not want to invest in these strategies and all the power to you. I only invest in one myself. They've been accused of being asset gatherers. Funny from a board named after the guy who founded a firm with $5 trillion in....assets! Benchmarks are elusive so I agree that is frustrating but blending all their funds together and saying they should be some benchmark makes no sense. QSPIX for example is doing long short on bonds, currencies, commodities, and equities. Why would you benchmark it to equities? Their other funds are employing all sorts disparate strategies. I don't know where they said you should equal weight your investments across their entire fund family. It's true that buy and hold, low cost, market cap weighted indexing is a great strategy. At times the board treats it more like a religion though.
bjr89
Posts: 151
Joined: Sat Apr 23, 2016 8:15 am

Re: AQR fund family's performance relative to its benchmarks

Post by bjr89 »

-----
Last edited by bjr89 on Fri Nov 16, 2018 7:11 pm, edited 2 times in total.
afan
Posts: 5282
Joined: Sun Jul 25, 2010 4:01 pm

Re: AQR fund family's performance relative to its benchmarks

Post by afan »

This thread is about whether it is true that they have delivered what they promised. The way to test that would be to compare performance of each fund to an appropriate benchmark.

As the above indicates, identifying the appropriate benchmark is not a trivial undertaking. Until you have done that you don't know whether they are producing what they promised.

If someone promises small cap value exposure you have to compare it to small can values indexes to see whether that is what they gave you.

The appropriate benchmark for these AQR funds might be hybrids of multiple factors. If the factors weights were stable then at least it would be possible to create such benchmarks. If the factor weights change over time then it may be impossible. So you will likely never know whether they gave you what they promised.

On the other hand, you can count on the fees, turnover and varying investment selection. Each of those rule them out for me. Still makes for nteresting questions about analyzing their performance.
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
marcopolo
Posts: 3650
Joined: Sat Dec 03, 2016 10:22 am

Re: AQR fund family's performance relative to its benchmarks

Post by marcopolo »

betablocker wrote: Tue Nov 13, 2018 9:43 am I'm curious about the hostility that AQR inspires on the board. Bogleheads can be an emotional bunch! AQR has performed well over the years delivering exactly what they promised. Yes there has been a long term underperformance in value and recently the other factors have also underperformed. It happens sometimes just like all things underperform from time to time. In my mind AQR is a model of transparency and low fees for what they offer (long/short, hedge fund strategies, etc.). You might not want to invest in these strategies and all the power to you. I only invest in one myself. They've been accused of being asset gatherers. Funny from a board named after the guy who founded a firm with $5 trillion in....assets! Benchmarks are elusive so I agree that is frustrating but blending all their funds together and saying they should be some benchmark makes no sense. QSPIX for example is doing long short on bonds, currencies, commodities, and equities. Why would you benchmark it to equities? Their other funds are employing all sorts disparate strategies. I don't know where they said you should equal weight your investments across their entire fund family. It's true that buy and hold, low cost, market cap weighted indexing is a great strategy. At times the board treats it more like a religion though.
That is interesting.

I had the exact opposite reaction. For a board largely bases on simple, low-cost, passive investing approach, I am quite surprised at the almost cult-of personality-like support expressed by some for these expensive, complex, and to date, poor performing (I know, we just need to be patient) products.

If you took the name off these products, and just presented their costs, strategy, and long term under performance, I wonder if there would still be so much support.

At least the cult-like support found here for simple, low-cost, passive investing has been supported by successful long term outcomes.
Once in a while you get shown the light, in the strangest of places if you look at it right.
afan
Posts: 5282
Joined: Sun Jul 25, 2010 4:01 pm

Re: AQR fund family's performance relative to its benchmarks

Post by afan »

Wishful thinking.

It would make me very happy if someone would sell me a fund that I reasonably could expect to increase my risk adjusted return over the market.

If you look at all the marketing of active funds, they all offer some version of this promise. A huge number of people want desperately to believe.

So they buy the claims and buy the funds. The near impossibility of testing them against a meaningful benchmark and hence ever knowing whether one got what one paid for does not stop those who want to believe.

Nor does the breathtaking cost.

Reminds me of Bill Miller who got rich by underperforming the S&P 500 for a career. Great work if you can get it.
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
User avatar
Random Musings
Posts: 5840
Joined: Thu Feb 22, 2007 4:24 pm
Location: Pennsylvania

Re: AQR fund family's performance relative to its benchmarks

Post by Random Musings »

Big picture, if you look at the aggregate, AQR's avg expense ratio is around 0.8%. For that, they capture the ER and provide you with some slight losses versus the benchmark.

Kinda reminds me of active funds, all you have to do is just pick the winnahs in advance.

RM
I figure the odds be fifty-fifty I just might have something to say. FZ
User avatar
Topic Author
nisiprius
Advisory Board
Posts: 42244
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: AQR fund family's performance relative to its benchmarks

Post by nisiprius »

For laughs, I did the same analysis on Vanguard's stock funds, using those described by Vanguard as either "US stocks" or "international stocks," "active management," and "investor shares." For the time period, I used fund inception if Vanguard shows a benchmark value from fund inception; otherwise, I used the last ten years.

58% of these funds beat their benchmarks, 42% failed to beat them. On the average, these funds beat their benchmarks by 0.44%.

Image
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
Random Musings
Posts: 5840
Joined: Thu Feb 22, 2007 4:24 pm
Location: Pennsylvania

Re: AQR fund family's performance relative to its benchmarks

Post by Random Musings »

nisiprius wrote: Tue Nov 13, 2018 9:31 pm For laughs, I did the same analysis on Vanguard's stock funds, using those described by Vanguard as either "US stocks" or "international stocks," "active management," and "investor shares." For the time period, I used fund inception if Vanguard shows a benchmark value from fund inception; otherwise, I used the last ten years.

58% of these funds beat their benchmarks, 42% failed to beat them. On the average, these funds beat their benchmarks by 0.44%.

Image
Thanks. Of course, some of this comes down to how relative the benchmark chosen actually is. I guess the real question is -are the funds being analyzed generating alpha that is statistically significant or are they not?

However, these charts appear to point to the mantra that costs matter if you want a better chance with active funds.

RM
Last edited by Random Musings on Tue Nov 13, 2018 10:38 pm, edited 1 time in total.
I figure the odds be fifty-fifty I just might have something to say. FZ
User avatar
whodidntante
Posts: 9296
Joined: Thu Jan 21, 2016 11:11 pm
Location: outside the echo chamber

Re: AQR fund family's performance relative to its benchmarks

Post by whodidntante »

Cliff Asness is a known hothead. I wonder if he is reading the recent threads here and throwing furniture/yelling, or if he dismisses these threads at hand. Either way, he's rich while he's doing it.
betablocker
Posts: 505
Joined: Mon Jan 11, 2016 1:26 pm

Re: AQR fund family's performance relative to its benchmarks

Post by betablocker »

marcopolo wrote: Tue Nov 13, 2018 10:49 am
betablocker wrote: Tue Nov 13, 2018 9:43 am I'm curious about the hostility that AQR inspires on the board. Bogleheads can be an emotional bunch! AQR has performed well over the years delivering exactly what they promised. Yes there has been a long term underperformance in value and recently the other factors have also underperformed. It happens sometimes just like all things underperform from time to time. In my mind AQR is a model of transparency and low fees for what they offer (long/short, hedge fund strategies, etc.). You might not want to invest in these strategies and all the power to you. I only invest in one myself. They've been accused of being asset gatherers. Funny from a board named after the guy who founded a firm with $5 trillion in....assets! Benchmarks are elusive so I agree that is frustrating but blending all their funds together and saying they should be some benchmark makes no sense. QSPIX for example is doing long short on bonds, currencies, commodities, and equities. Why would you benchmark it to equities? Their other funds are employing all sorts disparate strategies. I don't know where they said you should equal weight your investments across their entire fund family. It's true that buy and hold, low cost, market cap weighted indexing is a great strategy. At times the board treats it more like a religion though.
That is interesting.

I had the exact opposite reaction. For a board largely bases on simple, low-cost, passive investing approach, I am quite surprised at the almost cult-of personality-like support expressed by some for these expensive, complex, and to date, poor performing (I know, we just need to be patient) products.

If you took the name off these products, and just presented their costs, strategy, and long term under performance, I wonder if there would still be so much support.

At least the cult-like support found here for simple, low-cost, passive investing has been supported by successful long term outcomes.
Isn’t vanguard the largest active manager in existence? They also run multi factor and long short strategies. If you took the name off the products, described the evidence supporting those products and then placed them in the universe of products by cost and transparency of course I would buy them.
betablocker
Posts: 505
Joined: Mon Jan 11, 2016 1:26 pm

Re: AQR fund family's performance relative to its benchmarks

Post by betablocker »

nisiprius wrote: Tue Nov 13, 2018 9:31 pm For laughs, I did the same analysis on Vanguard's stock funds, using those described by Vanguard as either "US stocks" or "international stocks," "active management," and "investor shares." For the time period, I used fund inception if Vanguard shows a benchmark value from fund inception; otherwise, I used the last ten years.

58% of these funds beat their benchmarks, 42% failed to beat them. On the average, these funds beat their benchmarks by 0.44%.

Image
I mean come on. Are we supposed to take this seriously?
lack_ey
Posts: 6701
Joined: Wed Nov 19, 2014 11:55 pm

Re: AQR fund family's performance relative to its benchmarks

Post by lack_ey »

Thanks for running the numbers here.

I think everybody is just taking the numbers they want to fit their existing narrative, so if you'll allow me as well... :wink:


The main divide I see is actually between the long funds and the market neutral / alt funds.

Basically all the long equity funds lost (those benchmarked against MSCI Emerging, Russell 1000, Russell 2000, and so on). For the stock funds this is basically just DFA-style factor tilts, with the moderate turnover needed to keep to that strategy. Also more or less the same thing as smart beta funds, what Vanguard is doing with their active factor funds. Except AQR usually charges more for these than most competitors, so I've never liked any of them. They all lost, whether targeting value, momentum, defensive, or multiple factors, usually doing worse than the ER might suggest. Maybe some bad luck.

Then also the risk parity funds lost. In hindsight we know that commodities were bad during this period, and as such the result is not much a surprise. The risk parity funds relative to the benchmarks have more term risk exposure and certainly a lot more (not zero) commodity exposure, and I think less equity beta as well, so... combined with a high ER does not sound great over this period. In general, I'm not a huge fan but it's more of a shrug for me.

The actual alternative funds have had less of a bad record, some pretty good. For many of these, T-bills are a rather unsatisfying benchmark, but for those with very little equity or bond betas, what would you suggest we use instead? These are harder to really analyze and it's in any case a little more subjective. It really depends on the goals and perspective.

Now, the curious part to me is that many of the alt funds, including those which have done well, are trading on many of the same signals as the long-only funds. Now, a lot of their funds have done really poorly for the same reasons this year. But in prior years, there was a divergence where a fund targeting some factors in long-only stocks would lose to the benchmark while a market-neutral fund targeting those factors would manage to gain money, quite a lot of it. (I think the difference has to do with some of the returns happening to show up on the short side with the long side being meh, and maybe something to do with a couple of extra signals maybe used in the alt funds).

Of course you can't just shift the goalposts and make up benchmarks as you go along. And there's no excuse for diversification as deworsification, in the short term, or if you hold for the long term. You never want that.



As to bjr89's point, I thought everybody knew they're just selling beta, not targeting alpha? At least in the sense where we consider factors and related things to generally be betas in a multi-factor model.

Just repeatable, quantifiable, well-known strategies, charging mid-to-high fees (which is lower than what some others running these strats charge, sometimes). I think they make a distinction of "true alpha" that they do not attempt. Though they would claim to have some implantation (trading / execution) edge over some competitors. Supposedly.

I guess it's a bit arguable to say event-driven strategies (long/short the usual mergers/acquisitions stuff) and so on would count as beta, but I would have a hard time calling that alpha either.
afan
Posts: 5282
Joined: Sun Jul 25, 2010 4:01 pm

Re: AQR fund family's performance relative to its benchmarks

Post by afan »

The actual alternative funds have had less of a bad record, some pretty good. For many of these, T-bills are a rather unsatisfying benchmark, but for those with very little equity or bond betas, what would you suggest we use instead? These are harder to really analyze and it's in any case a little more subjective. It really depends on the goals and perspective.
The appropriate benchmarks would be a combination of the factor bets the vendor promises to make. Long only factor bets for a long only fund. Long/short factor bets for a long/short fund. Either AQR is delivering those factor bets, in the promised proportions, or it is not.

This of course assumes they tell you upfront what the factors and weights are supposed to be. It also requires that they either maintain those bets throughout the term under study or they announce changes in the bets as they occur. That would permit one to adjust the benchmark to reflect the new goals as they change.

One could try to figure out ex post what those bets were with a principal components analysis, again assuming the bets were constant. If they changed without telling you, one MIGHT be able to decipher when that occurred. But we are already dealing with tiny datasets. Splitting them up would make this underpowered study far worse.

Treasury bills as a benchmark is easy to analyze.

If you wanted a fund that behaved like T bills would you be happy if you got something that had the much higher variance of these funds? I think, rather than satisfied, one would be outraged.

So, no, T bills are not an appropriate benchmark.

If you cannot identify a benchmark because you cannot simulate the performance of a fund that had the promised factor bets then you cannot know whether the funds gave you what they promised.

But you can know how much you paid an advisor to put you in the mystery funds and what expense ratios you paid.
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
marcopolo
Posts: 3650
Joined: Sat Dec 03, 2016 10:22 am

Re: AQR fund family's performance relative to its benchmarks

Post by marcopolo »

afan wrote: Wed Nov 14, 2018 11:12 am
But you can know how much you paid an advisor to put you in the mystery funds and what expense ratios you paid.
You would think that would be the case, but evidently we can't really trust what the funds themselves say are the expense ratios:
viewtopic.php?f=10&t=167241&p=4108066&h ... e#p4107955
Once in a while you get shown the light, in the strangest of places if you look at it right.
afan
Posts: 5282
Joined: Sun Jul 25, 2010 4:01 pm

Re: AQR fund family's performance relative to its benchmarks

Post by afan »

Of course I could say "I don't know or care what factors, leverage, or weights are represented in the fund. I just know I want it."

In which case it would reasonable for one to ask WHY I want it. WHY do I want a fund whose investment properties are unknown to me?
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
KnowNth
Posts: 98
Joined: Mon Apr 18, 2016 12:54 pm

Re: AQR fund family's performance relative to its benchmarks

Post by KnowNth »

Frankly I am shocked that so many posters, including some I greatly respect, on bogleheads, the site that preaches passive low cost index investment, invest not only in these active managed funds, but funds whose fees are well over 1%.

I read the other long thread about AQR funds, I get that they are for diversify, but I still don't understand why they are better than the simple 3 fund portfolio. If the performances are not stellar, that why do you invest in them? Lower risks? Better future returns?
User avatar
oneleaf
Posts: 2510
Joined: Mon Feb 19, 2007 5:48 pm

Re: AQR fund family's performance relative to its benchmarks

Post by oneleaf »

KnowNth wrote: Wed Nov 14, 2018 5:04 pm Frankly I am shocked that so many posters, including some I greatly respect, on bogleheads, the site that preaches passive low cost index investment, invest not only in these active managed funds, but funds whose fees are well over 1%.

I read the other long thread about AQR funds, I get that they are for diversify, but I still don't understand why they are better than the simple 3 fund portfolio. If the performances are not stellar, that why do you invest in them? Lower risks? Better future returns?
I feel the same way. I think, in the end of the day, the desire to outperform the market or get a better risk/adjusted return is just too tempting. The entire foundation of Bogle's principles is to not get suckered into other people's claims of financial expertise.

And this applies just as much to the Edward Jones snake-oil salesmen as it does the Ivy-League graduating hotshot quants. Maybe it just takes a special kind of snake-oil to get a Boglehead... a message highly connected to academics and professional peer-reviewed studies and all that.

When one of these strategies proves disappointing, the reaction is always the same: "the fund is doing exactly what it is supposed to do" and "you simply don't understand the strategy." Not to mention how many "experts" all agree.

Bogle's message made it super easy... just ignore all of it and focus on costs. This way you don't need to worry about which expert advice is expert enough. And now that we have such a plethora of good passive options (ETF's and index funds alike), I cannot see why anyone is still trying to access an advisor to get access to AQR, Stone Ridge, or who else is the "not cheap but for some reason is worth it" flavor of the year.
afan
Posts: 5282
Joined: Sun Jul 25, 2010 4:01 pm

Re: AQR fund family's performance relative to its benchmarks

Post by afan »

By way of comparison, Vanguard's multifactor fund. If you run it through factor analysis on portfolio visualizer, even with its very short lifetime, you get a near perfect R2. That means the factors included in the model explain the performance of the fund.
For the AQR funds, I don't think the results would be so clear. Unlike Vanguard, which promises to deliver minimum volatility, momentum, or whatever, and do so consistently, AQR promises the factors will change. So there is no way to check whether AQR is doing what it promises.

Since these funds are sold on hope, faith and wishful thinking, one may as well extend that hope, faith and wishful thinking to the question of what they are actually doing.
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
lack_ey
Posts: 6701
Joined: Wed Nov 19, 2014 11:55 pm

Re: AQR fund family's performance relative to its benchmarks

Post by lack_ey »

afan wrote: Thu Nov 15, 2018 12:17 pm By way of comparison, Vanguard's multifactor fund. If you run it through factor analysis on portfolio visualizer, even with its very short lifetime, you get a near perfect R2. That means the factors included in the model explain the performance of the fund.
For the AQR funds, I don't think the results would be so clear. Unlike Vanguard, which promises to deliver minimum volatility, momentum, or whatever, and do so consistently, AQR promises the factors will change. So there is no way to check whether AQR is doing what it promises.

Since these funds are sold on hope, faith and wishful thinking, one may as well extend that hope, faith and wishful thinking to the question of what they are actually doing.
Which regression or analysis are you running on which factors?

https://www.portfoliovisualizer.com/fac ... sion=false

I see the multifactor long-only US equity funds having all high R^2 above 90%, lower for Vanguard's ETF than the AQR mutual funds.

Maybe makes sense to include quality, so again:
https://www.portfoliovisualizer.com/fac ... sion=false

The equity market neutral funds, including Vanguard's, have similarly low values. With AQR Equity Market Neutral this is not quite right as that also uses non-US stocks and I'm benchmarking against US factors (though bets are currency hedged).

So I'm not sure if I see the premise as being right in the first place.


Backing up from the example, I don't generally think it necessarily makes sense to look for higher R^2 on multifactor regression. Deviation from consistent factor loads, and some other differences may come from implementation in real world with cost-aware trading, among other things. Factor agreement (which stocks are on the long sides and short sides of different factors) also shifts over time.
afan
Posts: 5282
Joined: Sun Jul 25, 2010 4:01 pm

Re: AQR fund family's performance relative to its benchmarks

Post by afan »

lack_ey wrote: Thu Nov 15, 2018 9:20 pm
afan wrote: Thu Nov 15, 2018 12:17 pm By way of comparison, Vanguard's multifactor fund. If you run it through factor analysis on portfolio visualizer, even with its very short lifetime, you get a near perfect R2. That means the factors included in the model explain the performance of the fund.
For the AQR funds, I don't think the results would be so clear. Unlike Vanguard, which promises to deliver minimum volatility, momentum, or whatever, and do so consistently, AQR promises the factors will change. So there is no way to check whether AQR is doing what it promises.

Since these funds are sold on hope, faith and wishful thinking, one may as well extend that hope, faith and wishful thinking to the question of what they are actually doing.
Which regression or analysis are you running on which factors?

https://www.portfoliovisualizer.com/fac ... sion=false

I see the multifactor long-only US equity funds having all high R^2 above 90%, lower for Vanguard's ETF than the AQR mutual funds.

Maybe makes sense to include quality, so again:
https://www.portfoliovisualizer.com/fac ... sion=false

The equity market neutral funds, including Vanguard's, have similarly low values. With AQR Equity Market Neutral this is not quite right as that also uses non-US stocks and I'm benchmarking against US factors (though bets are currency hedged).

So I'm not sure if I see the premise as being right in the first place.

I tried QSPIX. Not interested enough to try others since AQR promises the factors and loads will change.
.
Backing up from the example, I don't generally think it necessarily makes sense to look for higher R^2 on multifactor regression. Deviation from consistent factor loads, and some other differences may come from implementation in real world with cost-aware trading, among other things. Factor agreement (which stocks are on the long sides and short sides of different factors) also shifts over time.

I agree that one expects deviation from consistent factor loads with these funds. That being the case, one cannot determine whether they are delivering what they promise.

If they were claiming to deliver consistent factor loads then one could try a multifactorial regression, assuming one knew what those factors were supposed to be. If you don't know the factors (there are hundreds of candidates with more constantly proposed) then it would be pointless to run a regression with an arbitrary small subset of possible ones.

For those who think these funds are delivering on what they promise, how are they coming to that conclusion? Apparently it is not by comparing fund performance to that of the factors it is supposed to deliver.

If there are no appropriate benchmarks, how does one analyze the faithfulness to the promise?
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
User avatar
HomerJ
Posts: 15349
Joined: Fri Jun 06, 2008 12:50 pm

Re: AQR fund family's performance relative to its benchmarks

Post by HomerJ »

betablocker wrote: Tue Nov 13, 2018 9:43 amThey've (AQR) been accused of being asset gatherers. Funny from a board named after the guy who founded a firm with $5 trillion in....assets!
https://about.vanguard.com/what-sets-va ... p-matters/
You find out the importance of ownership when you follow the money.

Typical investment management companies are owned by outside stockholders. These companies have to charge fees to pay their owners, which can reduce investors' returns.

At Vanguard, there are no outside owners, and therefore, no conflicting loyalties. The company is owned by its funds, which in turn are owned by their shareholders—including you, if you're a Vanguard fund investor. Our unique client-owned structure allows us to return profits to our fund shareholders in the form of lower expenses.
Jack Bogle collected a salary, not a percentage of all the money invested in Vanguard. Sure, a very GOOD salary, but "gathering assets" did not pay fees directly into his pocket. He founded Vanguard 44 years ago, and at the age of 89, he's worth around $80 million.

Cliff Asness is worth $3.6 billion. And two of his partners are worth more than $1 billion. AQR has $226 billion assets under management. 1%-2% fees a year, every year, on that pile is a LOT of money for them.
A Goldman Sachs associate provided a variety of detailed explanations, but then offered a caveat, “If I’m being dead-### honest, though, nobody knows what’s really going on.”
User avatar
Topic Author
nisiprius
Advisory Board
Posts: 42244
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: AQR fund family's performance relative to its benchmarks

Post by nisiprius »

Following my own curiosity, I re-did the charts to show performance before expenses. That is, I added (geometric sum) the stated expense ratio to the stated total return. Notice that according to Larry Swedroe, the SEC definition of the expense ratios for the Style Premia Alternatives fund and probably the other funds benchmarked to Treasury bills is (if I understand correctly) simply wrong, and ought to be about half the official number, but in any case I just took it as stated and added the full amount.

Before expenses--and again using my weird way of counting, in which every fund with two benchmarks is simply counted twice, once against each benchmark--the funds in the AQR family exceeded their AQR-chosen benchmarks 44% of the time, with an average difference of -0.03% relative to the benchmark.

Before expenses, those funds in the Vanguard family that were categorized by Vanguard as "stock funds" or "international [stock] funds," and as "actively managed," exceeded their benchmarks 65% of the time, with an average difference from the benchmark of +0.63%. The purpose of including the Vanguard funds is not to prove that Vanguard is better; it's not an apples-to-apples comparison for many reasons, and limiting it to long-only stock funds means we are looking at less ambitious funds; but I did want to see what happens in a different fund family.

Image

Image
Last edited by nisiprius on Fri Nov 16, 2018 2:55 pm, 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.
betablocker
Posts: 505
Joined: Mon Jan 11, 2016 1:26 pm

Re: AQR fund family's performance relative to its benchmarks

Post by betablocker »

HomerJ wrote: Fri Nov 16, 2018 11:14 am
betablocker wrote: Tue Nov 13, 2018 9:43 amThey've (AQR) been accused of being asset gatherers. Funny from a board named after the guy who founded a firm with $5 trillion in....assets!
https://about.vanguard.com/what-sets-va ... p-matters/
You find out the importance of ownership when you follow the money.

Typical investment management companies are owned by outside stockholders. These companies have to charge fees to pay their owners, which can reduce investors' returns.

At Vanguard, there are no outside owners, and therefore, no conflicting loyalties. The company is owned by its funds, which in turn are owned by their shareholders—including you, if you're a Vanguard fund investor. Our unique client-owned structure allows us to return profits to our fund shareholders in the form of lower expenses.
Jack Bogle collected a salary, not a percentage of all the money invested in Vanguard. Sure, a very GOOD salary, but "gathering assets" did not pay fees directly into his pocket. He founded Vanguard 44 years ago, and at the age of 89, he's worth around $80 million.

Cliff Asness is worth $3.6 billion. And two of his partners are worth more than $1 billion. AQR has $226 billion assets under management. 1%-2% fees a year, every year, on that pile is a LOT of money for them.
As a believer in capitalism I wouldn’t begrudge a founder who offers an innovative product his or her reward. I also think Bogle is a hero. Asness could have made billions no matter what he did. The fact that he helped democratize, make transparent and lower the fees for hedge fund strategies was a great service to the investing community. Btw, Bogle and Asness are friends and highly respect each others’ firms.
User avatar
packer16
Posts: 1414
Joined: Sat Jan 04, 2014 2:28 pm

Re: AQR fund family's performance relative to its benchmarks

Post by packer16 »

betablocker wrote: Sat Nov 17, 2018 10:27 am
HomerJ wrote: Fri Nov 16, 2018 11:14 am
betablocker wrote: Tue Nov 13, 2018 9:43 amThey've (AQR) been accused of being asset gatherers. Funny from a board named after the guy who founded a firm with $5 trillion in....assets!
https://about.vanguard.com/what-sets-va ... p-matters/
You find out the importance of ownership when you follow the money.

Typical investment management companies are owned by outside stockholders. These companies have to charge fees to pay their owners, which can reduce investors' returns.

At Vanguard, there are no outside owners, and therefore, no conflicting loyalties. The company is owned by its funds, which in turn are owned by their shareholders—including you, if you're a Vanguard fund investor. Our unique client-owned structure allows us to return profits to our fund shareholders in the form of lower expenses.
Jack Bogle collected a salary, not a percentage of all the money invested in Vanguard. Sure, a very GOOD salary, but "gathering assets" did not pay fees directly into his pocket. He founded Vanguard 44 years ago, and at the age of 89, he's worth around $80 million.

Cliff Asness is worth $3.6 billion. And two of his partners are worth more than $1 billion. AQR has $226 billion assets under management. 1%-2% fees a year, every year, on that pile is a LOT of money for them.
As a believer in capitalism I wouldn’t begrudge a founder who offers an innovative product his or her reward. I also think Bogle is a hero. Asness could have made billions no matter what he did. The fact that he helped democratize, make transparent and lower the fees for hedge fund strategies was a great service to the investing community. Btw, Bogle and Asness are friends and highly respect each others’ firms.
IMO the real question here is not complexity but how much each man thought they were entitled too. Bogle thought shareholders should get the lion share of the gains, Asness did not. Asness is more comparable to your traditional mutual fund manager versus a hedge fund manager but that still does not change the fact that he thought he was worth billions versus a more modest amount for Bogle. Asness' products are more complicated & dependent upon many more events going right versus wrong for them to be winners and the human expertise require to make these many decisions. This why I think AQR is active but quantitative vs. passive index finds that do not require judgement beyond the selection of the firms in the index. OTOH, Bogle's funds are based upon simplicity & only a few things going right for them to be winners.

Packer
Buy cheap and something good might happen
betablocker
Posts: 505
Joined: Mon Jan 11, 2016 1:26 pm

Re: AQR fund family's performance relative to its benchmarks

Post by betablocker »

packer16 wrote: Sat Nov 17, 2018 10:47 am
betablocker wrote: Sat Nov 17, 2018 10:27 am
HomerJ wrote: Fri Nov 16, 2018 11:14 am
betablocker wrote: Tue Nov 13, 2018 9:43 amThey've (AQR) been accused of being asset gatherers. Funny from a board named after the guy who founded a firm with $5 trillion in....assets!
https://about.vanguard.com/what-sets-va ... p-matters/
You find out the importance of ownership when you follow the money.

Typical investment management companies are owned by outside stockholders. These companies have to charge fees to pay their owners, which can reduce investors' returns.

At Vanguard, there are no outside owners, and therefore, no conflicting loyalties. The company is owned by its funds, which in turn are owned by their shareholders—including you, if you're a Vanguard fund investor. Our unique client-owned structure allows us to return profits to our fund shareholders in the form of lower expenses.
Jack Bogle collected a salary, not a percentage of all the money invested in Vanguard. Sure, a very GOOD salary, but "gathering assets" did not pay fees directly into his pocket. He founded Vanguard 44 years ago, and at the age of 89, he's worth around $80 million.

Cliff Asness is worth $3.6 billion. And two of his partners are worth more than $1 billion. AQR has $226 billion assets under management. 1%-2% fees a year, every year, on that pile is a LOT of money for them.
As a believer in capitalism I wouldn’t begrudge a founder who offers an innovative product his or her reward. I also think Bogle is a hero. Asness could have made billions no matter what he did. The fact that he helped democratize, make transparent and lower the fees for hedge fund strategies was a great service to the investing community. Btw, Bogle and Asness are friends and highly respect each others’ firms.
IMO the real question here is not complexity but how much each man thought they were entitled too. Bogle thought shareholders should get the lion share of the gains, Asness did not. Asness is more comparable to your traditional mutual fund manager versus a hedge fund manager but that still does not change the fact that he thought he was worth billions versus a more modest amount for Bogle. Asness' products are more complicated & dependent upon many more events going right versus wrong for them to be winners and the human expertise require to make these many decisions. This why I think AQR is active but quantitative vs. passive index finds that do not require judgement beyond the selection of the firms in the index. OTOH, Bogle's funds are based upon simplicity & only a few things going right for them to be winners.

Packer
I do like that Bogle set up the firm as a coop. Still don’t forget that vanguard is the largest active manager in the business as well.
Post Reply