Investing Billions - E162: Why Most Investors Quit Before Winning w/Cliff Asness
Episode Date: May 9, 2025Cliff Asness is one of the most influential minds in quantitative investing and the Founder, Managing Principal and Chief Investment Officer at AQR Capital Management, which oversees over $100 billion... in assets. In this wide-ranging conversation, we go deep into what makes a successful long-term strategy, how Cliff thinks about building portfolios, and why most investors misjudge both volatility and leverage. He also shares what it was like launching AQR after his early work on momentum strategies at Goldman Sachs, and what he’s learned about investor behavior across cycles. This is one of the most insightful and entertaining conversations we’ve had on the podcast—and Cliff brings both humor and hard-earned wisdom to the table.
Transcript
Discussion (0)
During our toughest period rounds to zero.
My concern about my personal losses in that, because I am a believer.
I've seen this movie several times.
I've stuck with our strategy and seen it work.
Devising great strategies that have very decent, attractive, positive
sharp ratios and are uncorrelated is only step one.
Step two is sticking with and convincing others
to stick with those strategies long-term.
98% of it is when you manage money for others,
you feel a responsibility.
Bringing people bad news is no fun for anyone.
I don't care who you are.
And you're often dealing with very smart people
who wanna do the right thing,
but they face what are called agency problems.
This is a big part of what makes running an institutional portfolio really difficult.
Hedge funds, unlike any other asset class in the world, is literally a black box.
And it's very mysterious to people. I was walking around your office here, there's 600 employees.
Talk to me about how a strategy goes from a thesis to being tested, to being integrated
into.
Today, I'm excited to welcome Cliff Asness, billionaire co-founder and CIO of AQR Capital,
one of the largest quantitative hedge funds on the planet.
Cliff is renowned for his pioneering research on factor investing, his outspoken takes on
markets and his willingness to question even a long-held
industry beliefs.
In this episode, we dive into Cliff's approach to risk, why he believes the market is still
inefficient, and how he balances cold rationality with the realities of client psychology.
We'll also discuss building resilience through tough periods, the evolution of quant strategies,
and what it really takes to stick with convection in the face of market pressure.
You've written about pulling the goalie as a finance metaphor for taking calculated risks.
What did you mean by this metaphor?
In ice hockey, if you're losing with very little time left, doesn't matter if you lose
by two, doesn't matter if you lose by three.
Typically in a normal point of the game, it's really important to have a goalie.
If you took your goalie out and put six skaters on instead of five, your
chance of scoring goes up myself and a colleague, Aaron Brown wrote a paper,
uh, trying to calculate what the optimal time to pull a goalie.
We found they should be radically different than what the coaches do.
They should be pulling the goalie five minutes before if they're down by one.
If they're down by two, they should be pulling the goalie with 10 plus minutes left in the
game, which I don't think I've ever seen.
And then the way we tied it to finance, we ended up writing this
paper in the journal of portfolio management on hockey, which was a little
bit odd and the reason it ended up there was these links we can make to, to.
Investing essentially, you got to step back and ask yourself, why
don't people do what's optimal?
Are they just stupid and not doing the math right?
And we don't think that's it.
We think they face incentives.
Most of the time it doesn't work.
Most of the time you still lose the game.
If you actually follow Aaron Brown and I's strategy and pull down by two with 10 and
a half minutes left, you usually lose by three, four or five.
Because our strategy is cold blooded. It just says, well, you're down by four. Yeah. Your chance
of winning is minuscule, but keep the goalie pulled. So in real life, coaches don't like to be
horribly embarrassed. They don't like to differ from the crowd for real, but small gains in expected wins.
If you look at investing, I'll use one that's dramatically self-serving because we do this,
trend following strategy.
They can go through five, even 10 years of, you usually don't lose a lot of money, but
not making a lot of money, but not making a lot of money,
you know, from the GFC for about a decade afterwards, markets went straight
up and trend following strategies with, they made money, but, but they
were disappointing to people.
And then suddenly there's a year or two that pays for all the other years.
And it's, it's kind of when you really want to get paid.
It's when things are falling apart.
That can be really hard to maintain.
This actually applies to almost any active strategy that goes through a bad period.
I picked one that can have a decade.
But doing something that is reputationally harmful,
that often doesn't work, but you should do anyway,
is a fair amount of good investing.
It's interesting, the analogy, the hockey analogy,
is like playing to two audiences.
There's the, you said cold blooded score,
the probability of winning the game,
but there's also the fan base.
If your fans go to a game and they lose two to one
versus five to one,
it might be the same outcome on paper, but it feels differently. And ultimately hockey is an
entertainment sport. Same with investing. If you look at investing, if you're investing your own
money, you can be hyper rational. But when you have clients, I've heard you talk about this rule of
not wanting to lose someone's money for three years. So talk me about how you build around this, the client psychology. Well, you actually don't want to lose someone's money for three years. So talk me about how you build around the client psychology.
Well, you actually don't want to lose someone's money for three minutes.
That's just not feasible.
I've joked that three years is about the longest,
a lot of institutional investors, like you said,
not necessarily the final investors, but someone who reports to someone else.
And I'm not putting them down.
It's harder when you do that.
That's about as long as people can often stick with an underperforming strategy, which is
understandable, but it's bad news because there are a lot of things in investing that
can go through periods that long or longer.
We haven't had one in a long time.
Actually, it's not even that long ago.
The stock market from the dot com bubble for 10 years out was a net loser.
And people have learned that one and maybe only that one reasonably well.
The whole stocks for the long run has gotten better, but there are a lot of
strategies I can think of that have had much shorter
periods than 10 years that just become very, very hard to stick with.
Another concept that I've thrown out there, everyone in finance tries to seem smarter
by using physics terms and I'm no different.
In physics, in relativity, they talk about time dilation.
Our version is what I call statistical time, call that rational time.
I look at a back test and I go, this strategy makes a lot of money over a hundred years,
but has had three year periods of difficulty four times.
And you ask someone, would you stick with that?
And they go, oh, of course I'd stick with that.
Partly because they see where it ends at the end, real life.
You don't get to see the end.
You're living, you're, you're living through it.
So real life time behavioral time is often quite shorter than statistical time.
I'll tell you, people like me are a little bit hypocritical in complaining about this
because if everyone were coldly rational, I don't think there'd be as many fun things
to do in active management.
Every active manager wants the world to be wildly quote inefficient, except as soon as
they put the position on the world to instantly realize that they were right and to fix that
inefficiency.
And as much as I, like everyone else would like that to be the case, that's, that's just not it. So if, if, if you want a world where there's
value to be added by an active manager, you also want a world where it can be very tough
to be an active manager occasionally.
So you want this emotional market, but you don't want your clients to be emotional. You
want your clients to be emotional.
If tax season felt harder than it should have, it's time for a change.
Juniper Square offers a smarter way to handle fund administration.
With industry-leading solutions and team of in-house experts who understand venture capital,
we simplify everything from LP reporting to audit prep.
No fund too complex, no better time to switch.
Visit junipersquare.comcom slash VC to connect with our team.
Yes. The perfect world is a, a really, really emotional downright silly market.
And your clients are all Vulcans.
Um, it's, it's, it's hard to pull that one off.
So when you evaluate the strategy, you looked at the strategy that
performed over a hundred years, but had three years, three year
periods, multiple times, are these strategies that you don't pursue or do you, do you market them a
certain way?
How do you deal with that kind of trend?
Um, well, no, we certainly pursue quite a few of them.
Um, you try to put them together in a portfolio where hopefully there are some
other parts that are not having that three-year period at the exact same time.
At the end of the day, if you're trying to do something at institutional scale, I do
believe there are strategies, and AQR doesn't really play in this world, but that we could
also do that are much smaller in capacity, that have much smaller bad periods, that are
effectively higher sharp ratio in the industry parlance. But if you're going to do something at institutional scale, I mean, tens, hundreds of billions
of dollars taking decent amount of risk.
I don't think there are, quote, two, three sharp ratio strategies at that scale.
So even after putting a bunch of them together, you're going to have some periods that are
quite difficult.
So we try to deal with that probably in three ways.
One is just a pre-education.
Everyone will talk about this.
It sounds very obvious and it is very obvious, but you got to do it.
If you don't, I mean,
I literally talk to a client and I'll show them that 100 year back test and those
periods and I'll take them through.
Let's just method act and imagine we're going through this.
Now don't just look at the graph, but you're going back to your board after two years and
you're saying, well, it happened again.
One great thing is if you go to someone after a year and have a tough
period, as long as you have a reasonable story and you've historically done well, that usually
goes fine.
You go back after two years and often the story is pretty much the same, especially
if it's, let's say a bubble or an irrationality that just got bigger.
And they're like, that's what you said last year.
And they're not looking happy.
And you're like, I know it happened again.
And if you I've never had to do three full years, I've done decent subset of three, three
years and our longest drawdown is about two and a half years.
But you go back to the end of three years.
I think you're pretty much talking to your mom.
She's the last person in the fund.
And she is she's putting a redemption notice.
She just hasn't fully filed it yet.
So number one, we try to pre-educate.
All these are obvious actually, but I'm going to tell you them anyway.
Number two is we keep trying to do better.
We went through one from part of 18 through part of 20.
So again, it was not quite
three years, but call it a few months into 18 to near the end of COVID, where value strategies
in particular, which are far from all we do, I'm very defensive about this, the last four
or five years have been amazing for us and not for value. So we're not always like that.
But over this period we were, and it varies through time. Sometimes you something's more extreme, other things
make up for it, sometimes they don't. But you, we probably and I'm bragging now,
but I think in the last five years, I think we've had our best five years of
research ever. This will be tested at some point, I don't pretend you know, I
just say it and it happens. But you're always trying to, um, not passively sit back and go, well, it's going to stink
sometimes, even if that's true, you work on, on getting better and minimizing
those periods and trying to do some different things.
Finally, if you unfortunately find yourself going through one of those
tough periods, you got to strike a real balance between having an open mind,
but not caving too quickly.
I have a mother-in-law who's from the Great Plains and she has some wonderful wisdom.
I don't know if she made this up or she heard it and repeated it, but at one point, she
actually doesn't have much of an accent, but somehow I do
this in a, in a thick Midwestern accent, which I won't do for you now.
Cause if I try to do an accident, it's very bad, but I heard her say once an
open mind is good, but not so good that your brains fall out.
Very evocative.
And what, in this case, what it means is you've researched
strategies, say for 25 years. strategies, say, for 25 years.
You've traded them for 25 years.
You've seen them go up and down.
You've seen them go up well more than down.
You've seen tough periods, though.
You know to expect it occasionally.
You can't see a tough period, even a really tough period, and just throw the baby out
with the bath water and cave.
Because it's obvious.
You're not going to see long-term success if you do that.
But if you go too far the other way and say, you know, people always say,
this time is different.
It never is.
You're going too far.
This time is different.
I will venture to say, as I think a lot of people in investing
would, is usually wrong. Obviously there are differences in technology, for instance, but
the general principles of, you know, you shouldn't pay trillions of dollars for sales in a company
already selling a lot. That's probably usually not a good deal, historically.
A lot of other things.
So, if you,
even if you apply this rule and say,
most of the time the world doesn't change,
most of the time the principles you've been using will work,
doesn't mean you're not in the one out of 20
times that it did change. It could be a mundane thing. It could be an accounting change where
you're using some measure that's broken. Now, this came up a lot during values drawdown in 18
through 20 in particular with intangible assets. That Some value measures don't deal very well with intangible assets, the most famous being if
you measure value by price to book, which to be honest is quantitative value circa 1990,
and it gets, I think, too much attention.
But if you do that, there's a strong argument that book misses intangible value.
And so if something's intangible value has made it 10 times more valuable, it's going
to look like it has a price to book 10 times more than it really does.
That argument was thrown at us a lot during this tough period for value.
You've got to keep an open mind and you got to keep an open mind because,
A, you should, it's in your own interest.
If things really did change,
sticking with something that is broken is not a good idea.
That's not a bold statement.
Two, for your investors' confidence.
I think they have to really believe
that you are walking both sides of this line.
That you have a stick-to-it it, you believe in what you do,
but it's not so dogmatic, you won't question it when it's not working,
or even question when it's working.
I'm only dealing with bad periods.
Frankly, if you come up with a hypothesis
for why something might be broken going forward,
that can happen in good times.
You tend to think about it more when things aren't working.
But on that same example
I brought up, we took seriously this intangible idea. We started out saying a very small fraction
of what would make us consider a stock cheap is something like price to book. And many other
value measures are less affected or unaffected by intangibles.
To go to a different end of the value spectrum,
anything that looks like a price to sales.
But price to sales has nothing to do with intangibles.
If it's not showing up in sales,
there's no mismeasurement humanly possible.
And as one thing we looked at, we said,
well, what if we only did price to sales?
Well, that did about as badly as price to book and something I haven't touched on, but we spent a lot of time looking at the disparity in valuations between what we
would call cheap and expensive had hit a record in 75 years of history, back at
the peak of the tech bubble in the late 90s, early 2000s.
According to our measures, it broke that record during COVID.
If it broke that record only on price to book, but on price to sales, say, looked pretty
reasonable, then you'd have to go, maybe there's some mis-measurement going on.
Maybe it is this intangible thing.
Quite simply, we didn't find that.
In fact, it gets to be, and my clients were subjected to this, which I will
endlessly apologize for, a very repetitive exercise if you go through a bad
period, because you take on all comers.
You're doing it right.
Anyone who has a theory that isn't absolutely demonstrably nuts, you know,
it's all about astrology.
You, you, people have the wrong sign.
I probably would dismiss that one out of hand.
Maybe not at the depths of a drawdown.
I would dismiss it out of hand.
But if it's a very plausible thing like intangibles or interest rates matter so
much that, that then you got to take it seriously, but you're often in this, this,
this productive loop because you're proving things, but still a repetitive loop
where the answer is we looked at that really hard.
Here's our evidence.
That's not it either.
And you know, the old Sherlock Holmes, when you've eliminated the impossible,
you have to consider the improbable.
Frustratingly, it's never a proof.
There's always something that would make you wrong for real, not wrong just in current returns,
but wrong in what you're doing that you just haven't thought of.
And then no client has thought of that no one out there has proposed publicly.
Um, so it's a little unsatisfying in that, that you're trying to, uh, you
know, the famous, you can't disprove a negative, something's wrong in your
process, all you could do is take on all comers with an open mind rigorously,
it, to the extent, something is rigorously testable, try to investigate it.
And at the end of the day, you have a strategy that has worked for you for a quarter century live,
has worked for a hundred years of back tests.
You've truly taken on all comers as to why it might be broken and truly gotten to the point
where you are really comfortable that none of them are what's going on.
Then you still don't bet your life.
You know, no one, no client has a hundred percent of their money with us. that none of them are what's going on, then you still don't bet your life.
You know, no one, no client has a hundred percent of their money with us.
Sadly, except maybe me.
I'm fairly confident in it.
But at the end of the day, then you plant your feet and you say, I will not be moved.
The process of gathering feedback could also increase increase your resolve not just not just well if you're honoring
Um, like some people in the room, I don't mean you
Um, yeah, it can uh, but you try not to be emotional in either
in either direction, um
But good feedback and I think this is what you mean that has a reasonable hypothesis
That you disprove.
Yeah.
Makes you each one again, you never get to that satisfied.
I've proven everything because there is no everything, but each one you go through.
There's a little bit of nervousness while doing it to carry your point even further.
Um, because finding out something is broken would be unpleasant.
Um, it, it would be two-sided.
I never even thought about this one.
This is interesting.
It would be a relief because you'd have something to do.
A lot of times when you have a good process going through a tough period,
the main thing to do is to stick with it.
The human impetus to do something is often wrong.
Sometimes your best thing to do is nothing.
Nothing is still trade your portfolio, follow your process.
So in one hand, if you found an answer, it would be a relief in that you'd say, oh, okay,
we fixed this.
But you also did something wrong and that money's not coming back.
If you are right and you did nothing wrong, there's a very good chance that money's
coming back when the world returns. It is kind of a nerve wracking in both directions,
emotional exercise. I think on net, you feel good when you really know you've tested something
with an open mind and your process passed again, but you never get to perfect certainty.
Perfect certainty is the stuff of, of mad men.
Karius, you mentioned you have, you mentioned you have all your money in AQR.
You have hundreds of billions of dollars from clients.
It's very stressful when things go wrong.
Do you ever focus on building a prepared mind and do you do anything like yoga, meditation, and you know, anything that prepares you for the...
I'm never going to write a memoir because it would be very boring.
I don't think anyone wants a QuantGeeks memoir.
But a working title I have in my head is world's worst meditator.
I last about 15 seconds before I'm thinking about something to do with work or family.
So I have tried.
Um, that's not the answer.
I will tell you, I'll, I'll take the question a little different, uh, direction.
Nobody's immune, but even in tough periods, by the way, we've had many more
good periods, cumulating to much more than our tough periods.
I feel that commercial moment is necessary given how often we're, how
much we're talking about tough periods.
But during our toughest period, it rounds to zero my concern about my personal losses
in that because I am a believer.
I've seen this movie several times.
I've lived through the dot com bubble in the late 90s, the GFC, I've stuck with, the strategy evolves and
can be very different 20 years later than it was before, but I've stuck with our strategy
and seen it work.
And I think 2% of my angst over those periods are my own investment.
In fact, I've rather aggressively, and clients know this this move to do more when we've suffered,
which you often feel a little stupid about in the short term.
But I know you're really worried about me, but it has worked out for me.
But 98% of it is when you manage money for others, you feel a responsibility.
Bringing people bad news is no fun for anyone.
I don't care who you are.
And you're often dealing with very smart people who want to do the right thing, but they face
what are called agency problems.
They report to somebody else.
And whatever, this is a big part of what makes running an institutional portfolio really
difficult because when they're not seeing the results they want on part of the portfolio, somebody
is asking them about it.
And then they, who probably understand what you do fairly well, are trying to articulate
it.
Now you're playing a little telephone because they might understand it well, but probably
not as well as you do.
And they're not as practiced as explaining it.
And they're explaining it one level up.
And God forbid you get two levels up on the organization.
You know, you start out talking to the head
of the pension fund, you end up,
there's a CFO involved and suddenly
it's the CEO is in the room, you're done.
That's over, there's no communication necessary.
And to loop back in that bias to do something works
dead set against sticking with a strategy.
Especially when the market is tough.
And people have learned lessons from places where the lesson
went does go the other way.
There are plenty of things in life where if it's not working for two years,
you should do something else.
Right?
Yeah.
The Einstein, exactly.
The definition of insanity is, is, is, is doing the same thing over and over again
when it doesn't work or some version of that.
Expecting different results.
Exactly.
You took this from me.
I don't know if I would have done it this time, but I referenced this even though I
mangled it just now because sometimes in good investing, you were exactly fighting Einstein on this.
Now, again, the problem is time scale.
Because on a longer time scale, you're really not, because
you've not seen it not work.
You've seen it work on a longer time scale.
But I should say more accurately, it feels like you're fighting Einstein.
And when you're going through an agency issue where you're talking to someone
who's not the final decision maker and they talk to someone, you're saying,
yeah, it hasn't worked for a while.
And in regular business, if my conveyor belt in my factory didn't work for two
years, that would be one year and 364 days more than I would tolerate.
Investing doesn't work that way.
The stock market has a risk adjusted return of sharp ratio.
Again, it's not a perfect way.
We all use, um, 0.35, 0.4.
It's up about two out of every three years.
It can have a flat to bed decade.
If I could invent something for you as good as the stock market, but
absolutely uncorrelated with it.
And you really believe that.
So the, forget the fact that you should always be a little cynical.
Well, math says you should put as much money in that as the stock market.
That's what I just said.
It's the stock market, just an uncorrelated version.
Personally, that is where the stock market people have, and this is a good
thing in the world.
In my career, I think people have net gotten better at this, the stocks
for the long run, sticking with it.
If you have a strategy, there are strategies we believe in pretty much
every much every, every bit as much as we believe in the stock markets risk.
Risk premium.
Maybe we're crazy, but we do.
It is harder when you're doing something more off the run.
Um, when you're doing something more different than the crowd, it is harder to
stick with the paradox, everything I bring up is seemingly a paradox, but the
paradox there is pretty much the only way to add value is to do something
different than the crowd.
And as long as you're long-term good, there is a big caveat.
Nothing I'm saying absolves you from making money over the long-term, but as
long as you're long-term good, the less it looks like everything else, the
better, but that can paradoxically make it very hard to stick with, particularly
in this case, when everything else is working.
Our hardest times have been we've made and lost money in every market environment.
A lot of what we do strives to be quote market neutral.
In, in a simple framework, if you're doing a long short portfolio, you'd
be short about as much as you're long.
That doesn't mean it's a hedge.
We do run some things that we think are hedges.
Some of those trend following strategies I mentioned, but a hedge does
particularly well when the market falls.
If you're independent in the market, that just means if you make money six out of every
10 months, you have the same chance if the market's up or the market's down.
So you'll see if you imagine two by two quadrants, up down market, you up down, you'll see all
four over time.
By far the hardest one I've ever found is market up, you down.
Because people think you're an
idiot. It's easy to make money now. Why are you not? And of course, a lot of
people intellectually know, well, you're short as much as you're long. It doesn't
really help, but it still feels that way to a lot of people. I've actually in the
rare periods and thankfully they've been rare when markets have had tough periods
and so have we. I've had to tell people, don't let us off the hook so easy.
I've had very nice, very smart clients say, oh, nobody can make money right now.
I'm like, no, that shouldn't be an excuse for me.
I tell you that I'm being market neutral.
I think we're still good.
I think we're going to make you money.
I'm not saying, but they're actually too upset when markets are up and you're down, and even
a little too nice to you when markets are down and you're down.
When you're up, they're happy no matter what.
You're benchmarked against S&P whether you like it or not.
Yeah.
Psychologically.
Well, that's definitely true.
I'll be even more cynical.
If the world were beaten in the S&P, unlike the S&P that has beaten the world for a long
time, you'd be benchmarked against that.
There is certainly some hindsight bias.
Again, there are a lot of really smart clients out there who won't act on that, but some
do, and everyone feels the tendency.
If there was easy money, it'd be made just by doing something simple.
Even if your whole purpose in a portfolio is to zig when that zags, you're still benchmarked
to it in some sense.
I'm curious, hedge funds, unlike any other asset class in the world, is literally a black
box.
And it's very mysterious to people.
I was walking around your office here, there's 600 employees.
Talk to me about how a strategy goes from a thesis to being tested, to being integrated
into.
Well, backing up with what you said about hedge funds, first of all, hedge fund doesn't
really mean anything, right?
If you say active value stock picker, people may have very different opinions.
I'm not even close to saying they're all the same, but you have an idea of what's going on. If you say hedge fund, well, there are hedge funds
that are short biased. There are hedge funds that are very long biased. There are hedge
funds doing esoteric fixed income arbitrage. So hedge fund, first of all, encompasses a
whole set of strategies. The cynical quote for many years has been,
hedge fund is not an investment strategy,
it's a compensation structure.
Warren Buffett.
Yeah, I think that may have been him, you're right.
You know more of the stuff than I do.
Thank you for listening.
To join our community and to make sure you do not miss
any future episodes, please click the follow button above
to subscribe.
I read a lot of names.
It's rude to show up your guests like that though.
So hedge funds are very heterogeneous.
We pissed off a lot of the hedge fund industry.
We wrote a paper in 2001, it's a 24 year old paper now, with the title Do Hedge Funds Hedge?
Where we took the indices of hedge fund returns.
They're very, call call them imperfect is overdoing
it.
There are a lot of funds don't report to the indices.
They're a very incomplete snapshot.
Some argue they're biased to look worse than hedge funds because the really good ones don't
report.
Some argue that they're biased to look better because there's what's called survivorship bias that I don't exactly know how they would do this, but some can game
the system and only report like, I'll start reporting to you, but fill in my prior three
year returns when I report to you. And of course you only start reporting when the prior
three years are good. So the bias can go either way, but just taking it as given that the
indices are the best we got. We found hedge funds weren't fully hedged.
They were about half hedged.
Um, in geek speak 0.4 or 0.5 beta, not a one beta like the market, but not a zero
like a fully hedge beta portfolio would have, uh, but what's worse is all the
active stuff they were doing was real and it was there, but it was much smaller
in terms of contribution to their ups and downs than that 0.5 beta.
That means their correlation with the market was, I think it's like over 0.8.
Again, I don't know, I have various geek levels on your podcast, but 0.8 is frigging high.
Yeah, I'll use a technical term.8 is frigging high. Yeah, I'll use a technical term.
It's frigging high.
And we found, again, using this dodgy set of indices, admittedly, that there was approximately
zero alpha at the end of the day.
I don't know why it upset me.
I'm in the hedge fund industry and it didn't upset me that, that, that were coming up with these results.
Um, active management of any kind and hedge funds are kind of maybe the
ultimate active managers is an inherently arrogant act to argue that on
average they all make money.
You know, Bill Sharp, many years ago, Nobel laureate, Bill Sharp, creator
of the capital asset pricing model.
The simplest law he ever came up with is the average can't beat the average.
Um, so if the passive market is inherently the average of what all of us think.
Everyone who's overweight against the passive market to someone who's
underweight and he just pointed out.
And then if you take out fees and trading costs, the average
underperforms the average.
This doesn't dissuade me. I think
our strategies are great and I love them and it's what I do for investing. But you're starting
out knowing you're in a world where the average is not going to work. So we published this
and I got yelled at. We were a little Pisha firm, that's Yiddish for tiny at the at the time. And I had maybe seven famous hedge fund managers call me mad
that we wrote this paper. And then I'll give a shout out by the way, when you I won't I won't do
that. But when nowadays that wouldn't bother me at all. I might even frankly enjoy it a little bit
which I which is not becoming of me but is true. of like the So I get this call after being yelled at by a bunch of people for writing this paper. And my assistant goes, Richard Perry of Perry Capital, or I think that was the name, is
on the line for you.
And my shoulders slump because I'm like, I'm going to get taken to the woodshed again.
I get on the phone with him and he just goes, Cliff, I read your paper.
Good job.
That just sums up the industry really well.
And we ended up, and he probably doesn't even know this,
we ended up a client would occasionally send us,
usually anonymized, but sometimes we could figure it out,
like historical hedge fund data and asked us to analyze it.
And he looked really good even on our analysis.
I felt good about that.
But it was very uplifting. Again, it's hard to imagine when you're in your early 30s and
your firm is having a tough time in the beginning during the dot com bubble and you wrote a nasty
paper about the industry and everyone's yelling at you to have someone famous and well regarded
call up and go, good job. Thank you, Richard. We haven't spoken since then, but thank you.
I think there's this meme in the market and even institutional investors.
So many of them have gone out of the hedge fund class, which is a
catchall for a type of structure.
There's this meme in the market that hedge funds just are basically gambling
with your money and that they're going to be as right as they're going to be
wrong or your paper.
And yet you guys have had great performance breakdown, you know, at whatever level, some
of the strategies that you do without giving away all of your secrets.
But why are your strategies working?
Sure.
I'm trying to think at what level to do this.
Let's start out with 1990. What it would have looked like in 1990. We actually
started trading anything live in 94 when some core part of AQR met and formed a group at
Goldman Sachs. But the data in my dissertation ended in 1990. And my dissertation was one
of the very, it wasn't the paper, but it was one of the very
early papers on adding a version of what today we would call a systematic momentum strategy
to, and it wasn't just Fama and French, but they were the most well known and they were
my advisors to their work on value investing and saying, you know, momentum, the very silly sounding strategy of buying
what's been going up over the last, say, six to 12 months and selling what's been going
down was called about as good as their value strategy, which was buying low multiples and
selling high multiples.
We can get in a huge long discussion of that not really being value.
Graham and Dodd people get very upset when the quants call that value.
Cause value should be contextual.
Some things are worth more short answer to my parenthetical here is the
Graham and Dodd people are right, but the quants aren't doing it wrong.
They're thinking about the same things.
They're just calling them something else.
So the quants named low multiples value long time ago and we can't, we can't fix it.
But when Goldman Sachs asked me to form a group and see if we could trade
Goldman partners, initially Goldman partner money, and then client money
with this, the first thing we did was use simple circa 1990s measures of value
and momentum to go long a basket of global stocks and short a basket of global
stocks where the longs were better than the shorts on these measures.
Some were much better on value and a little worse on momentum.
Some were a little bit better on both.
The portfolio was far better on both.
We also very early on applied this to the macro world saying, can you do this at the
country level?
If you lined up say 25 tradable countries and form valuation and momentum measures,
would it work?
I'm cheating here.
It's a little like a cooking show where I already know the cake worked out in the back.
So I'm taking you through the steps.
I'm going to show you.
It did hold up.
And none of these things work all the time.
Uh, but it.
So don't unpack the intuition there.
People are have made money in the U S in the last few years. So more people will pile into that.
Right now the U S it's starting to fade because the ex U S is coming back, but
for a long time, the U S, um, on these two measures, and again, I haven't even gotten to how much
different things are today than 1994.
But if you were just doing these two measures, the US is probably a little bit of a push
because it's looked expensive for a long time, but had really good momentum for a long time.
And if you put those things together, you get some intuition out of it.
A lot of times the smart thing to do in active trading is nothing.
We don't have a strong signal here on one of the things we like, you know, but,
but if it's really expensive and it starts to really look vulnerable, again,
hopefully better than just six month price momentum over time, you can
improve those things.
Then, then, then might be the time to take a position.
So for a long time, these things were kind of canceling out for the US.
So imagine you're doing that around the world for individual stocks.
Those individual stock strategies are balanced in each country.
So you're trying not to take a country bet, but then you do it for countries.
And you say, if France is half the multiple of Germany, on average, France will win.
Not all the time.
These are risky bets, but on average, if France has started to better than Germany
over the last six to 12 months, on average, France will win.
And if you have both of these things going for you, it's slightly better bet.
You can do that for country stock indices, bond indices.
You have to think about how to measure value.
For country stock indices, you could just use the same exact measures.
If you like price to sales, aggregate that up for the country.
For bonds, it's not that complicated.
You could use real bond yield.
What's the yield minus economist consensus inflation for currencies?
You can use things like purchasing power parity for valuation.
And so we did this macro and individual stock.
Those are still two cornerstones of what we do.
Two other ones over the years that we've added, and we've added these a long time ago now,
are directional macro.
What we did back at Goldman Sachs was
always balanced.
France versus Germany.
It could be 20 countries versus another 20 countries.
Directional macro looks a lot like trend following and managed futures.
And we think we have some innovations there that are not just price trend following, but
we've applied the same principles.
And then we do arbitrage strategies.
I've made fun of the word arbitrage.
I have an article on the things I can't stand in finance. I think I call them my top 10 peeves in finance, which is grossly underestimating
how many peeves I have, but I had to cut it down to 10 in the financial
analyst journal in 2012.
And one of them were a list of phrases I don't like.
And one was arbitrage.
Um, because most of these things, arbitrage literally means riskless profit.
And these are strategies that make money over the long-term, but
occasionally blow up and kill you.
Um, it doesn't sound riskless to me.
Famous ones are convertible arbitrage, merger arbitrage, some off the run, on
the run, treasury arbitrage, which off the run, on the run treasury arbitrage,
which is one we've really never done because the leverage is a limit even how much we'll
lever a strategy.
But many years ago, we teamed up with two professors in academia, Mark Mitchell and
Todd Pulvino, who had done, I'm always fond of saying the best work on testing merger
arb strategies.
I'm always cheating a little bit because they pretty were close to doing the only work on it. They built the database.
This was sweat equity. They had to build the database of all the mergers because one of the
merger arbitrage is, and I'm sure you know this, but a deal is announced. A is buying B for stock
swap. If the deal is announced, the stocks are going to converge to a certain number. I'm making gestures of convergence with my hand.
If anyone's listening to the audio, it doesn't go all the way the first day because there's
some probability the merger won't go through the merger.
Our person, if they believe in the merger, will generally buy the target, sell the acquirer,
earn that last little bit when the deal goes through.
But when the deal doesn't, they lose a lot, much more than they make.
And then it's a probability game.
If they lose much more than they make, but that only happens quite rarely and not enough
to wipe out the profits, it can be a pretty good strategy.
It's like selling insurance. Mark and Todd effectively found that the universe of really smart, and they are really smart,
merger ARB managers, after their rather exorbitant fees, don't beat a much lower fee merger ARB
strategy that essentially, it's not quite this simple.
There's some risk constraints, but essentially does all the mergers.
That does no underwriting. says, I'll ensure anybody.
And so we've been working on those.
We have our, it's not quite everyone anymore.
We have some opinions as to which ones, but we've applied that to mergers, converts
to different capital structure trades.
And probably the furthest to field from pure quant in that you have to
model a specific deal and there can be some judgment involved and that's not a great word
in the quant world. But that's the fourth layer. So it's individual stocks, macro that's hedged,
macro that can be directional and arbitrage. We have portfolios that do all these things.
directional and arbitrage.
We have portfolios that do all these things.
It's kind of a best of what of AQR.
And then a lot of people hire us for subsets or to run them against benchmarks, not just as hedge funds.
You know, I keep talking about hedge funds and we keep talking about hedge funds.
I know it varies through time.
I never keep exact track, but call it half our assets are run against typical
long only benchmarks.
Often we like to be able to short a little bit against them, but often
they don't let us in there just highly traditional looking using the same models.
Um, what have we done?
I could have told you the same story at least 15 years ago and most of it since inception.
What do you do better over time?
Breath and depth.
Add more ways to measure what's good and bad.
So even in the public world and not everything is public, there are things
that people like us discover that we won't, you know, people often ask why we
write so much.
I'm like, you don't see what we don't write.
There are things.
But even in the public world of so called equity factor investing value and momentum was most of
the ball game in 1990. It was also the size effect, which we actually never
really believed in. But I should give that point of pride. It was there. Over
time.
publicly known things buying low beta stocks.
They don't really outperform their high beta versions, but they keep up, which they're
not supposed to.
If the famous capital asset pricing model were true, they'd make less money.
So it's a little free lunchy that they keep up.
More profitable companies.
Market seems to underestimate how sticky that is.
That's a wonderful offset to value too, because it's often negatively correlated
to value, the expensive looking high multiple stock, if it's more profitable,
maybe you lay off some of that, if you have both of that in your model.
If it's not more profitable and it's expensive, you start going,
now I'm really on something.
Firms that are buying back shares and not issuing shares. There's a slew of things that have been added to that universe
that have made it better.
Um, ironically, and this is a little humbling, I think for quant.
Not all of them, some are very specifically quantitative and those are maybe more
the ones I won't be chatting with you about, but a lot of the advancements in quant
factor investing are making it look more and more like what a Graham and Dodd investor.
If you go read Graham and Dodd, they didn't say buy low multiples.
They said buy low multiples that have a margin of safety that are not too risky, that have
large current profits.
And I do feel sometimes like, I think we're done with this part of the field,
but for about 20 years from like 1995 to 2015,
that was the quant world rediscovering
that these G&D managers were actually onto something.
Quants do it differently, they do it systematically
and in a very diversified way.
But it's just, I think good investing is good investing,
whether you're a gram and Dodd manager or a quant.
Finally, you can get deep into the weeds.
And this is something that just every year, I think it's over that we can't
come up with ways to measure these things better, and I've been perpetually surprised.
The last maybe five, even five to 10 years, it's been the, I feel bad talking
about machine learning because it's like,, I feel bad talking about machine learning
because it's like, how many quants do you get?
How many quants are not gonna drop machine learning
at some point?
And yet I'm going to do it anyway.
Because that's an example where you can have a philosophy
and you can be measuring it rigorously and trading it
and even making money from it,
but it doesn't mean it can't be made better.
Let me give you an example, natural language processing.
Being able to take text and get an inference from that text.
This is very imprecise and probably drives ML people crazy.
But I kind of think of this as somewhat the opposite of what a chat GPT does.
Chat GPT, you put a query in and you get a bunch of texts.
This looks at texts and it's not jeopardy, it's not in the form of a question, but it
kind of gives you a, what is this very summed up?
What does this text mean?
And one classic example was looking at earnings calls with management.
And quants for years had tables of good words and bad words.
And you just mechanically go through.
And if the word was increasing plus two, if the word was ebbing minus one, maybe ebbing is not as bad as decree, you know what I'm saying?
And at the end you get a score and you're probably immediately thinking that's
dumb, what if the sentence is massive embezzlement has been increasing?
I don't think you need to be quantitative.
Yeah.
Yeah.
Embezzlement should have a minus 10.
Yes.
But the increasing word should not have a plus one.
In that case, hopefully embezzlement is in your lexicon.
It turns out modern NLP, natural language processing techniques,
which are pure ML at its finest.
We have a ton of data.
Pure ML at its finest.
We have a ton of data. ML usually needs a lot more data
than traditional statistical techniques
because it's looking for less structured,
less linear patterns that are harder to find.
But NLP is just way better at saying,
is this good or bad news for the stock?
Language is inherently non-linear thing.
Even the examples you and I have been doing
are too simplistic. They're within one sentence. It might've been five
sentences ago that tells you whether this was good news. And it's not made
these strategies perfect arbitrages, but it's made them much better.
So I'm curious, when you look at these factors like earning calls, like your
example, embezzlement is increasing, that might be negative, but it might be
maybe 52% of the time increasing is positive. What kind of signal and delete genius seeing a factor could it be 50.2% in
order for you to be like, yeah, put it in the model.
Do you have to have a much more sizable signal and talk to me about, you know,
how you look for signal and how strong it has.
The only way 50.2% that's going to be a very low sharp ratio signal. A would get into our process would be as if it was also highly
negatively correlated with positive sharp ratio signals. You
can have a role for a signal that's flat that on average
doesn't make any money long term. If it greatly reduces the risk
of what you're doing already. It's
effectively increasing the Sharpe ratio of the portfolio, even if it doesn't have one.
Those are few and far between. 50.02% would probably not cross our barrier. For one thing,
as a quant and even non-quants, they might face it more informally, but they're prisoners of what they've seen.
But in quant we call it data mining and we generally don't use it as a compliment. In some fields
it's used as a compliment.
The reason we don't use it as a compliment is if you only look for what's worked in the past,
you'll always find something that looks good. If you test a thousand independent things, one of them will look one out of a thousand
amazing, even if they were all random numbers.
You fight hard not to be over influenced by this.
In the old school kind of pre ML, pre what we call adaptive, which is, which is
a something we've been doing in recent years with Bayesian techniques and ML techniques choose our weights more for us
But in the old world we chose the weights at the end of the day on the factors and you might try really hard to avoid
data mining
But you're not perfect. It squeezes in so the barrier 50.02
Would make me very cynical. It's even positive
Because you have a bias to keep looking
and to find things that are good.
So how do you come up with theories or test strategies
that are not historic, that don't have a historical context?
How do you only build strategies based on the future?
You don't.
We need data of some kind.
Some things, there's a world called alternative data, which is also a probably last five to
10 year phenomenon where a lot of databases are being built that just didn't exist anymore.
They're legal publicly available information, just no one collated and put into machine
readable form that somebody put the sweat equity into doing.
And that's maybe one exception where sometimes you only have a few years of data.
But in general, we are always prided ourselves on looking for two things.
Has it worked long term?
With again, long term can be longer for some things than others, but over a decent enough
period.
And does it make economic sense?
That economic sense can be just common sense.
Gee, buying more profitable firms is better than buying less profitable firms.
Or it can be a formal economic theory.
But we used to say we're about 50-50, the two of those, and there was some judgment.
You know, I'm saying it's 50-50, but the second part is judgmental. So with the advent of ML, I think we're probably more two-thirds, one-third data and maybe
going up.
This was a hard transition.
I think I slowed our firm down by a year or two at least in being uncomfortable to turning
more things over to the machines because we've always prided ourselves on saying, no, it has to make sense and have worked.
Some sector on a specific day of the week, if they were showing that that was...
Yeah. No, exactly. There are seasonal effects where you look at the same day of the week over
the last five years and similar things keep happening. And I got no story for that.
Sorry.
Basically the worst, the story, the better the data has to, has to be.
It's always a mix, but that mix ML by its nature relies more on data and less on story.
We actually pride ourselves compared to real ML people, serious, you know, only ML people.
I think we're real ML people.
We pride ourselves on still requiring more of a story than they do, but less than we
used to.
We've moved on the spectrum.
And if the techniques and the methodologies get better at doing something, we should change. I'm curious about these high data or higher data, lower story situations.
What percentage of the time do you find a story later on?
And at what point do you have any strategies that you've been running for a decade that you
still don't have an intuition? You know, it's very funny.
One of them, I called common sense a second ago. I'm going to come clean and say,
still makes me a little queasy little queasy buying more profitable companies.
It does feel like common sense to the average investor.
I was cheating there a little bit when I said, sounds reasonable.
Why you should get paid extra for owning a portfolio that everyone would prefer all
SQL.
I don't think academia or practitioners have come up with a great story.
Remember earlier I said you could have a zero sharp thing or even a minor negative sharp thing if it was very diversifying?
Part of say profitability could get into your models because it is so diversifying to like a value
strategy.
So in that sense, even if it didn't make money on its own, it cleans up value.
It says there are 100 super cheap stocks, but 20 of them in fact do look like disasters.
They make no money.
Stay away from those.
They're value traps.
So it can get in there, but it is a healthy, positive sharp and one of the most robust
results in finance.
And here we, like I think almost every other quant in the world, do say, all right, it
has some common sense to it, I know, but the theory is not good.
But the data and the empirics and the hedging aspects of it are so good that we will put
it in our process.
So there are things like that.
Um, I think one of the great papers to be written, I don't know if it ever will be
written, cause I don't know if it's possible, uh, would be a really cogent story.
Why this is either using behavioral finance, irrationality, um, or efficient
markets, which is coming up with a risk-based rational explanation that would
just nail the profitability factor.
Um, I've tried myself.
I've not done a great job.
Sounds like a challenge for the listeners.
Hey, it's good to have something left in your career to do.
In terms of products and different strategies, you have momentum, you have value, you have
arbitrage.
Are these essentially modules that you could put together into specific products?
And how do you marry the strategy with the product?
Yeah, they are modules.
We do have in various forms from LPs to mutual funds to usage accounts,
ones that do rounding to everything.
We do in the best risk adjusted form we can put.
So we will do the unconstrained.
You have all these modules.
How would you combine them?
If you were doing one investment for all your money going forward, you'd also have to choose
how much risk to take in that investment.
That's the problem.
I end up investing in almost everything we do because whenever we start a new product,
they often want to see co-investment. I end up investing in almost everything we do because whenever we start a new product,
they often want to see co-investment.
So I'm very pan AQR.
But in anything I have personal discretion over and I haven't promised to be in this
fund, yeah, that's what I would tend to.
It'd be a little weird if I didn't tend to.
Right.
You should put out all your money.
Yeah.
And I didn't do it.
It doesn't mean I'm...
Look, even if it's not quant, I'm
a human. I might have an opinion at some point that, oh, the next three years are a good
time for trend following. I actually kind of have that opinion now, given I think uncertainty
is very, very high in the world. So I'm not above having an opinion, but call that the
thing I would invest in. Almost everything else is because clients differ in their needs and beliefs.
For instance, there are a fair amount of clients who can't do a long, short, levered hedge
fund in every asset class around the world.
There are a fair amount of clients whose central problem is beat an equity index. We have models that if it's the U S only, we think can beat a U S equity
index, and if it's global, we think we can beat it by both picking stocks
and using those country and currency models I mentioned before to maybe
shift the portfolio.
So there's one example, but even on the hedge fund side, I mentioned
trend following multiple times.
Trend following has this interesting property.
It is not the single highest standalone risk adjusted return we can create.
It's actually not close to that, but it's a good risk adjusted return.
And it has a property that the geeks will call positive convexity.
In English, when the world falls to crap, it tends to have done fairly well.
And that's not a proof.
It doesn't mean it will every time, but it has nine out of 10 big drawdowns for the equities.
It's made a lot of money.
And I think in the 10th, it broke even.
Partly because most of those big drawdowns didn't just happen in a day.
Something developed and the trend caught it.
So there are plenty of clients out there who say for the bucket of my
portfolio, I'm looking to get cheap protection.
Protection is this convexity property and cheap is the fact
that it makes money on average.
Um, you can buy puts all day and you will get protection in a crash,
not necessarily a bear market.
Bear market that goes on for two years.
You can keep rolling your puts and overpay from an, and it work or, or not work.
But in a bolt from the blue crash, yeah, put will work.
But the drain of having a put on all the time makes it a very big
negative expected return.
So trend following is attractive to those.
Then people will differ because they'll look at the cliff portfolio.
They do one of everything and they'll say, well, frankly, we think AQR is great at two
of these, maybe the best at two of these.
I'm bragging now.
Now I'm not going to brag.
But we think you're not the best and we can find better at these other two.
Of course, these people are dramatically wrong.
They are fools.
We are the best at everything.
I'm kidding.
Um, they're more than allowed to have this opinion.
The best portfolio I can create using my, my own stuff at AQR is not necessarily
the best portfolio they can create if parts of what we do, they think someone
else is better at.
So a decent amount of what we do is to fit into niches for different people who
have different beliefs of what else is out there.
And finally, the one I like least is forgetting even the competition.
Some people just might not believe in one aspect.
They might just say, I don't think people can make money doing macro
trading or have absolute restrictions.
We can't run using leverage more than one and a half to one.
There are some head strategies that are amenable to that.
There are some, if you go into a fixed income strategy, you got to lever that a
lot more than an equity strategy because, and you can quote me on this fixed income
is fricking boring with leverage.
It becomes non boring and occasionally terrifying.
But you need to do that.
And there's some who are just restricted.
So I do think of what we do.
And mentally, when I look at it, I get a P and L of like all of our funds, which,
and they're literally, they're green numbers and there are red numbers and they
blink and I tell clients, I spend my life telling clients, you should look at your portfolio
once every three years.
And of course I look at this every three years.
It's funny to show that returns are correlated in behavioral finance, especially among masses.
Well that is true.
And I've gotten really good at not touching the portfolio.
So I hope I'm not the counter example to this because when I'm traveling,
I'm actually pretty good.
I'll check, you know, twice a day, beginning of the day, end of the day.
When I'm in the office, it's just up on my screen and you're not human.
If you're not in a good mood, when most of the numbers are green and a bad
number and bad mood, when most of the numbers are red, even if you know,
every day your edge should be positive, but very slight.
You're not making money 70% of the days. Yeah, Jim Simons, who's probably the goat of what we do,
lost money on many a day. If you make money six out of 10 days, you have an insanely good
process. You have to learn to live with that. So the portfolio that I would create for a one-stop shop, this is what I want to
invest in, it's the best of AQR balanced across our stuff in the best way I think
possible run at the risk I think a long-term investor should run at is for a
lot of people that is a great portfolio, but to a lot of people doesn't fit them
at all because of all these differences and heterogeneity.
We will not do anything where we think we don't have a decent edge.
We're not brokers.
We won't run something where we just go, there's demand for it and we can put on that exposure,
but it's a coin toss.
We have to think that if you stick with us for that three plus year period,
you will make money a lot more often, at least than you don't, hopefully all the time.
You've said many times that investors are under levered. How should investors think about levering?
Well, those two questions go great together because you can easily be over levered if your version of investing is to buy the three times
ETF on Kathy Woods arc fund. I
Think and you put all your money in that I think you're over levered
But I will not share my opinion on even the one-time version, but at three times, I think we you're over levered
leveraged used to blow up a
concentrated bet.
Look, if you buy a whole bunch of lottery tickets, some are going to win,
but is ex ante a pretty bad strategy?
Um, leverage though can be necessary if your goal is to smooth your
risk across different strategies.
Let's say you find 10 strategies and you really think their correlation is low.
I emphasize really think.
Some of the great problems in investing have been thinking correlations are low and finding
they're not so low when the blind kids or whatever.
We've actually been run levered portfolios now for more than a quarter of a century and
we've had good and bad times, but that's never been a really our issue.
So we've been pretty good at that, but you need to be reasonably convinced
these are diversifying.
This is a problem the hedge fund to fund industry has historically had.
They get a whole bunch of managers they think are good.
They, in fact, they very well might be they, maybe they can pick the better
ones in this, in this universe.
That's kind of not bad, but not great either.
So they pick up a whole bunch of good funds, but they end up with a portfolio
with in geek speak, what's called, you know, 2% volatility and a 4% expected return.
That's a sharp ratio.
I'm ignoring cash.
That's a sharp ratio too.
That's phenomenal.
It's still a 4% expected return.
And because they're invested in all these separate managers, it's
not easy just to gear it up.
But if you were doing all of that internally, and you truly believe
that these were diversifying, you'd say, let me lever that a few times to one.
Still not that crazy.
And now I'm making pretty decent money with a sharp ratio of, of, of two.
You know, you can take it'll give you three asset class examples.
Imagine you have an equally good hedged equity strategy,
fixed income strategy trading 10 year government bond futures
around the world and commodity strategy,
trading various commodities against each other.
If you invest in equal amount of dollars across the three,
you're not a bond manager. Your, your bond stuff is rounding error.
You are a little bit of an equity manager and you're mainly a commodity manager.
Cause I will again get very technical on you.
Commodities viewed alone are pretty fricking scary.
And are more volatile. There, there are, there are no, to my knowledge, there are no hundred percent, maybe some
really esoteric emerging market, but there aren't a hundred percent steady
state vol equity markets.
There are commodities.
Now imagine in a risk adjusted sense, you think you're about equally
skilled at all three of these.
Forgetting how much you lever the whole portfolio, you use
intra the three leverage to say I'm going to take, put most of my
dollars in fixed income, the middle amount of dollars in
equities and the smallest amount of dollars in bonds, excuse me,
backwards commodities.
Thank you.
I'm doing that again.
These are models.
These are imperfect.
You want to test your assumptions, you want to stress test,
you never want to overdo it.
Over relying on diversification could be dangerous too.
But if you do this well,
you've now created a considerably better
risk adjusted return,
because you have three things that are equally good
that aren't very correlated,
pulling on the rope at a similar strength.
If you put equal dollars,
commodities is pulling on that rope like me at 800
pounds and bonds are just hanging out.
So leverage can be a very useful tool to equilibrate bets, to spread your
bets better than you can in a world where you can't lever up some and
lever down some.
The argument about leverage, and I always stress this, it's not
just an argument to leverage up.
With commodities, I'm saying you need to lever them down, at least vis-a-vis
the other asset classes or they dominate.
So I think leverage is an extremely useful tool.
There are strategies.
I touched on this earlier.
I said the off the run, on the run strategy.
This is where the treasury that they issued three years ago or five years ago that now
is the same maturity as the new one issued five years later or maybe 10 years later.
There's a 20 that's rolled down to a 10, whatever.
The old one sells a slightly higher yield than the new one.
That's nonsensical in a theoretical sense.
They're the government cash flows.
The world where the government goes bankrupt but only pays off the on-the-run treasuries,
I don't think anyone seriously thinks that's in there.
But they serve a liquidity purpose.
On-the-runs usually trade a little bit more expensive because they're the currency of
that market.
So there is a trade out there.
It's still done.
It varies how much it's done based on how recently it's blown up.
But where you go long off the runs and short a similarly close to cashflow matched as you
can set of on the runs.
In theory, you've locked in an arbitrage.
You have positive government cash flows in all the scenarios, but that's only as they
pay off.
It's not about tomorrow.
When there's a liquidity event, on the runs, the liquidity benefit of on the runs goes
through the roof.
And off the runs, that yield difference where they yield more gets to be much more.
And it's a very easy strategy to blow up on.
If you wanted to put Arv on, you would put it into a portfolio that has a lot of liquidity.
So you would not need that.
You better.
The death combination in the world is tremendous leverage and illiquidity.
How my friends in private equity pull it off a little bit, I don't know.
I don't think they use leverage.
Well, they don't use leverage anything like off the run, on the run, but a very
leveraged strategy in a strategy that's maybe not illiquid like PE where you
never look, but maybe it's just hard to trade a ton of it.
That can be a real problem.
We don't do any of the on-the-off-the-run strategy because the left tail seems to be
too big relative to even if it makes money on average.
We have limits.
We use judgment.
We've been pretty good at this for, again, more than a quarter of a century.
But once you go into the leverage world, there's a set of skills you want to probe on.
Um, someone starts out tomorrow.
I'm all for young people.
I'm the parent of four young people.
Um, but you know, someone, uh, coming out of school saying I got a great new
strategy, but you only have to lever at 14 to one I'm an old fogey on that.
I'll be, I'm going to get you after you blow up and relaunch your fund four years later.
You're also an advocate of having very high vol strategies and small amounts in portfolios.
Talk to me about that and what's the rational amount to hold my portfolio?
Well I'll take the second question first.
At reasonable assumptions for good alternative investments, no one's going to invest the
rational amount.
I think a lot of them could be run in at least the sharp ratio of global equities and bonds,
which case you should have at least half your money in them.
I shouldn't say zero.
I've met very few investors who'll do that.
How much you should put becomes an exercise into how much you can tolerate.
And again, the method acting thing I talked about before, how much you think you can
honestly stick with if it goes through its bad relative or absolute period.
Now, why I think they should be higher of all, first of all, it has to be done
with dramatically open eyes.
I don't know what dramatically open eyes, eyes wide open.
The worst thing on earth is to not realize something is very high volatility.
In plain English, bigger ups and downs than the S&P 500.
Bigger ups and downs over a week, bigger ups and downs over a few years.
Over time, the goal was to make more and to be unrelated to it.
So you make your money.
But be that as it may, it was a high ball alternative.
Probably 25 of them, somebody said, love what you're doing.
I'm not mentioning the many meetings where they didn't love what we were doing.
But in 25 meetings, someone said, love what you're doing.
Want to invest, but I don't need 22 and a half percent.
Well, um, I need a quarter of that.
I responded like a flip obnoxious 30 year old would respond.
There's a reason for that.
I was a flip obnoxious 30 year old.
So it all kind of kind of worked.
Um, but I said, Oh, you want a quarter of the ball?
Give us a quarter of the money.
Works like a charm.
In fact, it's dramatically fair.
The fee goes down, the fixed fee at least goes down the performance.
He, excuse me, goes down by a quarter and the fixed fee goes down by a quarter.
You're doing a quarter of what you do.
If you, you can scale up or down the size, but I was flipped and told them to do it that way
and didn't accommodate them and give them the low volatility bucket.
It's just been a new product.
New product, even simple in that it can usually just be a feeder fund
into a high volatility.
That's essentially holding the cash for them.
It's just presented.
Exactly what they wanted, but with a higher fee for you.
We were not going to cheat.
We were going to cut the fee proportional, but actually one of the
things I've called out in the hedge fund industry is there is too much of having
a few good years at high vol, lowering the vol and not lowering your fee.
Because then you are effectively raising your fee because you're giving them less of what you do at the same price.
So if they want as much of what you do as before, they got to give you more money, higher fee.
I promise you we don't do that. We don't do that, right, Kevin?
All right. There's a guy over there who's confirmed for me that we don't do that. We don't do that. Right, Kevin? All right. There's a guy over there who's confirmed for me that we don't do that. After many years of learning that lesson,
I'm talking about both now. There are some investors who know themselves, know thyself
is pretty important. And if I'm going to have to report up and downs on 22.5% vol, I don't
care how many times Cliffs told me,
I only put a quarter of the money in,
so my dollar losses are the same as a much smaller.
I'm gonna get stopped out.
Someone's gonna make me do that.
That person should invest in the low vol version.
But there is a tremendous amount of cash efficiency
to investing in the high vol version.
And it's as simple as this.
I wrote a blog on this with some examples that are just stocks, bonds, and one not too
aggressive, plausible, sharp ratio, simple alternative, made up Excel examples.
I'm old.
I use Excel still for these things.
Essentially a high vol alternative, same exact thing, just run at a higher risk level,
means you have to put much less money in it and means you can invest in other things.
If you're truly uncorrelated, it's a form of leverage again.
It's not literal leverage of the whole fund, but you can, and you get these interesting things.
Imagine you had a 60-40 stock bond portfolio and a low vol all was your only option.
And the low vol all was uncorrelated to stocks and uncorrelated to bonds, but
pretty decent risk adjusted return.
Most opt, most optimizations at reasonable risk levels end up looking like mostly
at reasonable risk levels end up looking like mostly equities and the alt. Because the alt's a little better than bonds, and it makes that substitution.
If you have a high vol alt, it puts less in the alt and suddenly brings back bonds.
Because bonds are still good, and they're diversifying. They're just too much of a drag at their low risk level to waste too many dollars on them
if the other stuff is low vol 2.
So I think great cash efficiency and we're trying an experiment here of offering both.
So I'm flying in the face of my original failure of saying no 25 years ago to to people who wanted Lowval, but I'm not totally, because we're still offering that.
And I have a theory, we've done this for years, but we're really amping it up, that having
people consciously choose between the two, presenting them both, saying you do both,
will lead to better outcomes for everyone.
When you tell someone, and this is just psychology,
this isn't great quant finance,
you learn at the University of Chicago,
but when you tell people the only way we do it
is 22 and a half percent vol,
they might understand the math,
but that decision's on you.
If you tell people we can do 22 and a half percent
or we can do five and change percent
and we'll adjust the fees fairly few, you decide.
They have made a volitional conscious choice, which they've hopefully vetted internally to go, no, we get it. It's going to swing like crazy. And we think that's going to be better.
That's probably a behavioral aspect there as they do low while their
investment committee gets comfortable. They may move out the spectrum.
how many gets control? They may move out the spectrum.
Absolutely.
We secretly hope so.
Maybe now that I've said it on our podcast,
not that secretly.
Talk to me about some virtue of complexity
that your colleague Brian Kelly wrote about.
It's a wonderful title, first of all. Obviously, everyone always talks about the virtue of simplicity.
I tried to get him to change the title. Occam's razor, the simplest decision, is usually best.
I wanted him to title the paper, Occam was Wrong, which has a very similar flavor. I shouldn't have
even said that because we may use that title one day. Everyone forget I said that. Virtue of complexity. I'm not going to sum it up as well as Brian. Brian Kelly has machine
learning for AQR as a Yale professor and an AQR partner. I'll brag for him in academic finance.
I think he's the leading light of applying ML to finance. I'm not saying that someone in the dungeon
of Renaissance who's not ahead of
Brian on some fronts. I think Brian would accept that also, but he's a superstar. But
a lot of this paper with examples are to take us out of our intuition that I mentioned held
me back for a few years. That you need straightforward, explainable models. That's very correlated to simple, right?
To have any faith in them.
Machine learning changes the game somewhat.
It is better at processing data.
A lot of what good machine learning does
is internally exactly fit the data
with a pretty heavy loss and penalty functions
for overfitting the data.
And there are lots of ways, model averaging, resampling to mitigate.
You can never fully get rid of the data mining problem.
So again, I always get scared explaining his stuff that he's going to, you know, as soon
as this comes out, he's going to text me, you know, that's it.
Just a little, but I think I'm getting, I'm yeah, I think I'm getting it right.
The general idea of the virtue of complexity is to, is to shock people in the title with
the counterintuitive notion that in the new world of ML complex, maybe less
disadvantageous and perhaps even advantageous to simple, at least in the
places where it's applicable.
And I just think it's fun.
I just think my title is better.
But that's...
We don't have enough time to talk about the less efficient market hypothesis in totality,
but maybe for another podcast.
But the question is, how do markets... What are the catalysts for markets to efficiently price?
So you have this bias in the market and you mentioned it takes up to three years to reprice.
What is that catalyst and what's the function?
I don't want that three year to become like the Asinist rule.
It can be very different.
Some things, you know, that was-
When you make three years the rule and you turn around a year, you look like a-
You look like a genius, yeah.
That was really for what I consider decent active strategies.
There are some things, Japan can underperform world equities for 30 years.
They blew away three years, 27 years into a drawdown.
I don't think I'd be very successful encouraging someone to stick with us
after 27 bad years.
As much as I may believe it, I can test every hypothesis known to man at 27 years.
It's, it's, it's not going to be, it's, it's not going to be happening.
Um, the catalyst is the hardest question.
And I'm, I'm main, I'm going to talk, but I'm mainly going to be
avoiding your question because nobody knows.
In the micro sense, I think of a lot of what we do on momentum, both fundamental
and price momentum, as trying to embed some catalyst work into a rational
valuation framework.
And by that, again, I don't mean simple quant value strategies.
It may have all the Graham and Dodd flavor of profitability, low risk,
efficient use of capital, whatever.
But sometimes those things can get worse before they get better.
And something like price momentum does act as saying, well, let's wait
to see it start to work.
So you're not even identifying the catalyst, but you're assuming it's
occurred for some of the most celebrated crazy things in the markets. So you're not even identifying the catalyst, but you're assuming it's occurred.
For some of the most celebrated crazy things in the markets, we don't even know the catalyst
for what popped the bubble after the fact.
A lot of your readers are too young for this, but AQR cut its teeth on the dot com bubble
again in 99 and 2000.
It's going on for a few years before that, but that's really when it crescendoed.
It peaked in March of 2000.
Um, and NASDAQ 100 peaked about a week and a half away from the NASDAQ composite. They were both in that, in that month.
Looking back, I can't tell you why it didn't peak six months earlier, six months later.
I can't tell you why it didn't peak six months earlier or six months later.
There was one theory, Barron's, the financial magazine, had a cover story
with flames, very dramatic cover, with flames.
And it was about the quote burn rate at the dot com firms that they were just burning through cash and they only had X months of cash left.
And it was great article and it was prescient, but a lot of people
literally pointed to that as the catalyst.
And even at the time, a bunch of us were saying, yeah, they've been
burning through cash for seven years.
Now they keep refinancing.
People keep giving them more cash.
The catalyst has to be why they're not going to give them more cash this time.
So the old, uh, is it Herb Stein?
Who, who said the remarkably obvious, but still insightful.
If something can't go on forever, it won't.
It's kind of also completely.
It does.
It does.
Um, though I think it was some brilliant guy who said it, not
the Yogi wasn't brilliant.
I use Yogi quotes all the time.
My favorite quote for international diversification
is yogis, you have to go to other people's funerals, otherwise they won't come to yours.
Obviously with funerals, you can't each go to each other, but for markets you can. You can be in the
bad market if you're also going to be out of your market to some degree when it's the bad market.
So catalysts for major events, sometimes they're obvious.
I don't think anyone's confused about the March of 2020
catalysts that sent the market way, way down,
you know, global pandemics.
I don't think, I don't think,
yeah, I don't think there's literally anyone on earth
who disputes that that was the catalyst for the market fall.
But when you have long, slow buildups and things get mispriced, which again, we don't have time for it,
but the less efficient market hypothesis is me not running away from my ex-professor, Gene Fama.
He's still one of my heroes.
But I do think markets are somewhat less efficient and more bubble prone
than I did when I was his student 30 years ago.
And I think they've gotten a little more so over my career.
I told you at the beginning, at the end of the day, I'd be punting on your question
because if you're sure you're right and sure, put sure in quotes, as close to sure as you can get
after doing all that work we talked about before, you want to be there for when the catalyst
occurs. But waiting for the
catalyst, aside from maybe a little bit of price momentum in your model, which we
do believe in, even after the fact for most things that aren't COVID
level obvious, we don't usually have very direct catalysts. And the best you could do is have the prepared mind of knowing that's my take away.
Yeah.
And sit around and wait.
Hope you don't.
When you said meditation and other things, I should be better at it because devising
great strategies that have very decent, attractive positive Sharpe ratios and are uncorrelated
is only step one. Step two is sticking with and convincing others to stick with those strategies
long-term and neither step works without the other.
Sticking with a crap strategy, if I may say so, it's really not going to help anyone.
All the discipline in the world is not going to, it's not going to make that good.
And having a great strategy that you just can't stick with is, is,
is not going to do it either.
So that magic combination in any form, some strategies mean heavier
on one than the other.
Um, but that's, that's a lot of investing in a nutshell.
Having the client sell back you.
Yeah.
Yeah.
No, it's a partnership.
It, and you know, you always remember the world.
I think we're all biased to remember the bad meetings, the client who didn't buy it, who
left, but we've had tremendous gratification from clients.
Again, we've had so many more good times than bad times.
You got me only talking about the bad.
Last five years, wonderful.
I'm not even know if I'm allowed to say it, but let me, with value not doing work, I got
to do a little commercial given how much I've talked about to say it, but let me, with value, not doing work. I got to do a little commercial, given how much I've talked about bad times.
But if I'm rational about it and not getting emotional and remembering the disappointments,
the far more happy experiences of clients who really got what we were doing, stuck with
it, or even in many cases added to it when it was going through a tough time.
When you yourself are adding to it, you get some credibility with clients.
Absolutely.
What do you see as the future of AQR and what will it look like in five, 10, 20 years?
Oh, you asked the hardest question for the end.
Um, well, let me say the obvious.
If I really had a solid answer, we'd be there already.
Right.
And it's always like the, what's the next great innovation? Like, innovation? That's what they call it, innovation. I don't currently know. Again,
buzzwords, but I'll do it anyway. The move to ML is real. There are cynics out there
say it's exaggerated. I'm not talking about ML stocks. We don't do individual stocks.
They may be worth it. They may be wildly overpriced. I'm not making a comment on that. But ML changing the world and I don't know more or less, maybe
more because we're a quantitative field, but the lives of quant is very real. I think the reason
these things are so hard to forecast is it's a constant battle to innovate,
to get better with the knowledge that getting better might mean staying the same.
And what I mean by that is the world is always trying to take away your edge.
If you found, let's say you're the first person to find a good new strategy, it could be a
factor, it could be something more esoteric that maybe you wouldn't call it factor.
It's not going to be yours alone forever.
Probably not.
So you got to get better just to stay as good as you used to be.
Because if you don't get better, you're getting worse.
That's I sound like Ricky Bobbie.
If you're not first, you're last from from from Talladega nights.
And Griffin believes that his main...
Not as peer, I'm as P on, but go on.
Um, your, your peer Ken Griffin believes that his main competitive advantage is
recruiting.
That's his only sustainable edge.
Do you fall into that camp and how do you, how do you stay competitive
in the hyper competitive?
Well, it's recruiting for us in a different sense.
I would agree.
Recruiting is, is incredibly important.
in a different sense, I would agree recruiting is incredibly important.
More than Ken because he runs a multi-strat.
And we run, it's always confusing. We run multi-strat portfolios, but they're all our strats.
Multi-strat in Ken's sense is the famous pod shop where you're putting money out.
So there is not, and Ken could, if Ken disagreed with me, listen to Ken, but I
don't think there's one overarching
investment philosophy.
In fact, he's probably looking for the opposite, even more diversification from, we have a
lot of themes that run through everything we do.
I'm comfortable having it be all my portfolio, but a lot of other people would say, no, you
have some themes.
We're going to give you a 10% allocation.
And I'm fine with that.
So there's one difference where the philosophy, it's not more important than
the people, but it's up there being right about the basics of what you believe in.
If you're not a firm that writes down ever the basics of what you believe in,
you just farm that out to pods.
That's obviously which pods you choose is a hundred percent of it.
For us, a cogent theory we believe in and will stick to is right up there, but I could
not be happier with the AQR team.
Even through some tough times, we've largely kept it together.
It is probably tied with philosophy and having a generally good investment plan for success.
And I'll say the cliche thing old men like me always say at this point,
I couldn't get a job here today. They'd look and they'd go, that's pretty good. But have you
considered going to law school? Cliff, I really appreciate you taking the time and look forward to catching up.
This was fun.
Thank you.
Thanks for listening to my conversation.
If you enjoyed this episode, please share with a friend.
This helps us grow, also provides the very best feedback when we review the episode's
analytics.
Thank you for your support.