September 5, 2024 • 54 mins
One of the problems in investing or trading is that — to use a common disclaimer — past results are no guarantee of future success. Someone can have a great track record in their stock picks, but maybe they just got lucky. Or maybe they were particularly well-dialed into one market regime that inevitably shifts. Or maybe they're actually just better than other traders. For multi-strategy hedge funds or "pod shops," there's an ongoing battle to hire or train the next great portfolio manager. But how can managers tell who is actually good and who isn't? On this episode of the podcast, we speak with Joe Peta, who was previously the head of performance analytics at Point72 Asset Management and has had a long career in the trading world. He's also an avid fan of sports gambling, and the author of the recent book, Moneyball for the Money Set, which attempts to take some of the talent analytical principles that originated in Major League Baseball and apply them to evaluating portfolio managers. He talks us through the traditional approach funds use to find or create superstars, and how these approaches can be improved upon using more rigorous, quantitative methods.
Mentioned in this episode:
Hedge Fund Talent Schools Are Looking for the Perfect Trader
How to Succeed at Multi-Strategy Hedge Funds
Only Bloomberg.com subscribers can get the Odd Lots newsletter in their inbox each week, plus unlimited access to the site and app. Subscribe at bloomberg.com/subscriptions/oddlots
See omnystudio.com/listener for privacy information.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:03):
Bloomberg Audio Studios, Podcasts, Radio News.
Speaker 2 (00:20):
Hello and welcome to another episode of The Odd Lots Podcast.
Speaker 3 (00:24):
I'm Joe Wisenthal and I'm Tracy Alloway.
Speaker 2 (00:26):
Tracy, do you know this? Sometimes I wonder, like, you know,
one in the morning, if I can't sleep, I think
to myself, in a different life, could I have been
the next Steve Cohen? Yeah? No, for real though, And
I don't, you know, need to talk about it. There's
a lot and I've brought it up before, you know.
I did get an offer at a prop trading shop
right after college to be a stock trader at this
(00:49):
place where they're gonna let use to your capital. And
I think Steve Cohen started off like as a prop
trader at some shop before being one of the great
hedge funders of all time. And I didn't take that
job for reasons I still can't explain to myself twenty
five years later, But I always wonder whether could have
cut it. Maybe I could have been a good trader.
I don't know.
Speaker 3 (01:07):
It's good you have a healthy level of self confidence, Joe. No,
When I lay awake at night, I think like, oh, shoot,
what did I say? Something stickid on the podcast, and
that's what keeps me up. But yes, good for you, Joe.
Speaker 2 (01:19):
No, I don't really think I could have, and I
actually do not think I would have been a good trader.
I don't think like that. I'm not that good at
poker other things. I'm not a natural better I don't
do like sports betting. I don't think that. But I
do sort of, you know, wonder about what my life
had been different if I had said yes to that.
Speaker 3 (01:38):
Yeah, fair enough. I mean we know, we know from
multiple episodes of the podcast this year alone. Yes, like
there are a lot of hedge fund traders out there,
especially in multi strats, who seem to be making a
lot of money, and everyone's sort of talking about them
up until recently. Maybe I should say, we're recording this
on August seventh, so maybe those bonuses a little bit
(02:00):
less this year given the market sell off. But up
until this month, people seem to have been doing relatively well,
and there was all this intrigue and interest in the
world of traders. And I'm sort of curious. This has
come up a couple times now, but what makes a
good trader and how are traders actually evaluated? Because my
(02:22):
impression was always like, Okay, well it depends on how
much money you make, but what's the timeframe for making
that money? And then also what about people who are
working in for instance, these particular pods who are doing
one specific thing, like what is the benchmark against which
they are judged?
Speaker 2 (02:40):
You mentioned that maybe they're not making so much money
this week or this month, But Tracy, I think we're
told all the time they're so neutral on everything. Their
market neutral, they're beta neutral, they're neutral every factor you
can think of. Why should they be losing money right
now They're supposed to like be neutral all.
Speaker 4 (02:55):
Of this time.
Speaker 3 (02:56):
Yeah, I'm sure they're making loads shore. I'm absolutely sure.
Speaker 2 (02:59):
No, but you're right. And look, we've been doing a
lot on hedge fund structure, and we did that episode
with Giuseppe Polyoligo, and we did that episode with Rich
falk Wallace, various aspects of like how hedge funds measure
risk and try to isolate alpha and all this stuff.
But they're just like so many questions in my mind,
Like I feel like we're just scratching the surface because
(03:21):
you know, we haven't even really talked about like idea generation.
So it's one thing to you know, talk about like, okay,
here's how you like factor out all of these exposures
that you don't on have like market beta, et cetera.
It's another thing to talk about like okay, but like
how do you pick the stocks to buy or go short?
Speaker 3 (03:37):
Well, yeah, we have gotten into this a little bit,
but you're right, there's more we could do. There are
all these questions about like how do you size your positions?
And if you're convinced that one thing is going to
be the next big thing, then why don't you just
have like one hundred percent positions?
Speaker 4 (03:50):
Yeah?
Speaker 2 (03:50):
Right, how do you make money if you can't just
go one hundred percent leverage long and video in video.
Speaker 3 (03:55):
Yeah.
Speaker 2 (03:55):
Anyway, so there's a lot more we can do. But
to my original, very egotistical start to this episode, I
do wonder like it's.
Speaker 3 (04:03):
Okay, Joe, it's good to have self confidence. I'm being serious,
thank you.
Speaker 2 (04:08):
I do wonder like this big question of like you know,
and a lot of people are probably interested in this
because these hedge one PM jobs or trader jobs seem
pretty great and as you mentioned, lucrative, and so it
would be interesting to know how a fund or anyone
goes about identifying like the next great trader who gets
to have that seat, so to speak.
Speaker 3 (04:26):
Well, I also think if you can identify what makes
a good trader at a hedge fund, then you can
get more into the business model of what they're actually
doing on a day to day basis. It helps us
understand what.
Speaker 2 (04:37):
They're really good at and what they can do specifically. Well,
I'm very excited today because we really do have the
perfect guest. We're going to be speaking with, Joe Peta.
He is the author of a recent book, Moneyball for
the money set, which is the name sort of implies
tries to, you know, figure out new ways or the
best ways to identify talent. I'm sure there's a lot
(04:57):
of old heuristics like they had in Bay, you know,
and they're like, well, this guy looks like he has
a good hustle, and then Moneyball came along. He's like, no, actually,
you want to really look at his like you know,
on base percentage or whatever it is, and stop looking
at like his like spirit or you know, his hustle
ahead of him. Anyway, and prior to that, in his career,
he's been in this industry for a long time. He
was the head of performance analytics at point seventy two.
(05:21):
So this speaks right to the question of how do
you evaluate traders. We also hit him on years ago
one of our really early episodes where he talked about
sports betting with some of these same ideas, et cetera.
So I'm thrilled to have Joe back to talk about
this basic question of how it's good trader. So thanks
for coming back, Joe.
Speaker 4 (05:40):
Oh, it's great to be here Joe and Tracy, and
nice to do it in person. Seven years ago, Tracy,
I believe you were in Hong Kong and yeah, Joe,
you just had a garage band instead of selling out venues.
Speaker 2 (05:50):
Now that's right, that's right. Still mentioned you were head
of performance analytics at point seventy two. How did you
get that job at a point seventy two? Steve Cohen's
big yes.
Speaker 4 (06:00):
So that goes right back to my appearance seven years ago.
So when I was on in twenty seventeen, I had
written a book called Trading Basis, which really looked at
the critical reasoning overlap between asset management, sports betting, and
the moneyballization of baseball. And you had ask me, Joe,
I think it was you asked me a specific question
of well, I mentioned that somebody from all three of
(06:22):
those constituents could learn something from the other two. And Joe,
you asked me for a specific example of how they
look at things differently, and I said, well, if you
go onto a trading floor, or you go to a
mutual fund and you ask them, hey, who's your best
trader or who's your best PM, Inevitably they will point
(06:42):
to the individual who had the highest return in the
prior year, either the biggest P and L or the
highest return on capitol. But I contrasted that that if
you went into the front office of a baseball team
and asked them who their best player was, they wouldn't
look at you know, which picture necessarily had the lowest
ra or the most wins. They would answer that question
based on skill sets, and so they it's a subtle difference.
(07:05):
Instead of looking at results, they would look at skills
because they know that the skills, there's so much noise
and results that the skills. If you can identify the skills,
you have a better chance of predicting who will do
better going forward. And as it was told to me,
a member of the c suite, at zero point seventy
two listen regular listener heard that episode and played a
(07:29):
portion of it for Steve. In fact, I think it
was the part I just mentioned, And I was told,
as it was relayed to me that Steve said, find him.
I want to talk to him. And I guess that's
not a surprise because in twenty twelve, and this is
all public knowledge. In fact, there's a book by Molly
Knight called The Best Team Money Can Buy that chronicles
(07:50):
the Dodger's ownership through the turbulent court years Frank McCourt's ownership,
and that team was sold in twenty twelve to the
Guggenheim Group. But Steve also bid for that team and
came very close to buying the Dodgers in twenty twelve.
And of course we all know him now as New
Yorker's nome his uncle Steve owner of the New York Mets.
(08:11):
So he has, I believe, always had an interest in
an analytical approach, and I think he always wondered, well
could that work in the hedge Fund. And I came
away from those meetings with the bunch of different people
in the investment committee, and I kind of came away
with three queries that I thought could sort of be
my marching orders and how I could help, And that
(08:33):
was I think at all these pod shops, when somebody
has a good year, they ask for more money, and
in terms of buying power, not cash, but in terms
of buying power. And so the question that management would
have is, well, is what they did repeatable? And at
the same time, as you know, there's turnover at these firms, right,
(08:55):
And I think another question is, well, sometimes we let
people go too early that then thrive elsewhere just because
they had a bad start to their career in terms
of results. Is there a way that we can avoid
that mistake? And then finally, when a team does well,
inevitably there's a bit of way, right because we know
that these four or five huge firms are all very
competitive and they're trying to steal talent. And so the
(09:19):
question is I know what a PM and or her
team may have made me in the past, but what
are they worth going forward? And all of those queries
can be answered by looking at skills, which is a
little different than what the traditional quants do at these firms.
Speaker 3 (09:35):
Okay, so here's my question, who should we bill for
the finder's fee fee? Should we send it directly to Steez?
Speaker 2 (09:42):
He do you right?
Speaker 4 (09:44):
It would be the firm, right, they probably saved a
lot of money as opposed to going through a traditional headhunter.
Speaker 3 (09:51):
Okay, I'm joking. Obviously that's fantastic to hear. I've loved
stories like that. Before we get into the existing model
of compensation. There's one question that I wonder because I
think we've done a number of Moneyball episodes at this point,
but it's been a while since we've talked about that approach,
and all I remember is the movie and Brad Pitt
(10:12):
kind of unconvincingly playing a guy that understands math. Could
you maybe explain, like what it is about the Moneyball
approach that seems to attract people in finance, Like why
is there that analogy that seems to come up again
and again.
Speaker 4 (10:29):
Yeah, I think if you're attracted to critical reasoning, and
that's the big thing and all of this industry is
you know, Joe said, what have I succeeded here? And
I always think the biggest question is do you have
the mentality in the stomach to make decisions and commit
capital based on incomplete information? Whether you have the skills
(10:50):
to you know, build models for you know, and and
understand companies and read documents. It's really can you make
decisions based on income information? And it's true at the
poker table, all right, and it's certainly true when you're
building a sports team, right you're like, how much is
this free agent worth? And before there were a lot
(11:13):
of Joe like you say heuristics, and I even mentioned
that in the book. I feel that still goes on
at the allocator level in this industry. Allocators they do
the interviews and you will hear things like, well he
just got divorced, you know, or there's a Bentley in
the parking lot. He must not be hungry anymore. Oh absolutely.
(11:33):
And one of the reasons is because they don't have
they don't take a different approach that might be more databased.
The whole idea of the moneyball approach is to tease
out skill from or the signal from these very noisy results,
because both athletes and asset managers, their results are filled
(11:55):
with influences over which they have no control. To answer
this question later, you both were talking about like market
neutral PMS, neutral everything pms, Why would they be having
the worst week this week than before. And there's an
actual real answer to that that has nothing to do
with their skills. Yeah, so this can apply to any
(12:19):
time period. We're looking at days or months, a year.
Let's go to sort of an economics one OHO one
like holding all else equal. Let's say we have a
PM that has one long and one short, okay, and
that's their entire portfolio. And of course they never would
this goes to something else you said in the intro
because of career risk. Right, even if it's their best
idea long and best idea short, they'll still fill it.
(12:40):
But let's say this is their portfolio and on any
given day or any period we could measure, but let's
keep it at a day.
Speaker 2 (12:47):
It's a perfect.
Speaker 4 (12:47):
Portfolio in that the long produces alpha and the short
produces alpha. So the long outperforms the market and the
short underperforms the market. Right, So that's a perfect portfolio.
What is the expected re turn for that portfolio for
like I say, any period, but for a day, And
the answer is there's a way to figure out. And Tracey,
you're gonna love this because the answer is dispersion. And
(13:10):
I know you light up when when you have the
But this is a little different dispersion than the quants
and the derivative traders make their life around. This dispersion
is and it's going to be very context specific for
where the PM toils. Right, so we know what these
pod shops they tend to be. They have you know,
(13:31):
subject matter expertise in sectors. Right, So you might have
an energy PM, and so let's say this is a consumer
discretionary PM, and you would say, okay, well, I'm going
to look at his or her universe and maybe that's
the S and P fifteen hundred consumer discretionary Maybe it's
say a portfolio of just consumer discretionary stocks that he
has modeled, so there might only be eighty or so
he and his team, but whatever it is, we'll say
(13:53):
that it's the all the consumer discretionary stocks in the
S and P five hundred or fifteen hundred. Well, the
way to figure out what the expected return is is
to simply look at all those stocks and say, here's
the skill neutral return, which would be the average return
of all those holdings. And then you look at the
ones that outperformed what was their average And you look
at all the stocks that underperformed and what was their average.
(14:15):
And the difference between those two numbers is the dispersion
between outperformers and underperformers, and that varies greatly from day
to day, and it can very greatly from year to year.
Speaker 3 (14:28):
It's not like the maximum that you can produce dispersions.
Speaker 4 (14:32):
Not the maximum, because you could have the very best
outperformer and the very best underperformer. But if you're looking
at a POD, so I'm taking all the pod from
all the shops across the street that are focused on
consumer discretionary, I'm going to be dead on by saying
the average of all those of all those perfect portfolios
(14:52):
is going to be the average of all the outperformers
and the average of all the underperformers. And it's invisible
to investment commits, to CIOs, to the pms themselves. They
can be just as skilled from one day or one
period and one year to the next, but the payoff
is different. And this is sort of the moneyball. Look
at hey, once we get this all context neutral, we
(15:15):
might say that a neutral everything PM that had a
seven percent return one year and a five percent cent
return the next year, he may have even been more
skilled than the five percent year, but the dispersion wasn't
there to pay off that skill.
Speaker 2 (15:30):
Oh, I see what you're saying. So in other words,
it's like, Okay, this person's up five percent. In order
to establish like whether that's good or bad or not,
you have to have some sort of like holistic view
of what dispersion on average looked like. In that that's
exactly well, that makes sense. It also seems kind of obvious,
(16:07):
you know, I know, the divorce and the Bentley is
probably like extreme examples. They're sort of funny. But you know,
thinking about the moneyball thing, and I mentioned in the
old days like oh, that guy looks he's a good
ey or whatever, just like all these sort of unquantified
his hustle, you know, his heart whatever. And then you know,
Brad Pitt or the being came along and actually put
some numbers on this. If they're not doing that, what
(16:30):
are the sort of like old heuristics that aren't the
extreme ones that the investment committees or the hiring committees
of the firing committees would have been using two aviliate.
Speaker 4 (16:40):
Pro So there's no question that it seems obvious, and
it's just the first building block. And this isn't Black
Shawl stuff in terms of complexity.
Speaker 2 (16:48):
Yeah.
Speaker 4 (16:49):
I started this sort of journey and analytics by working
for a company called Novus, and Novas was one of
about fifteen years ago, was at the forefront of portfolio analytics,
and in fact, they had read my book and I'm like, hey,
this is what we try to do. And I worked
for them. So I've seen just about every package out there,
whether it is from a vendor in terms of analytics
(17:11):
or you know inside firms. I've you know, worked with allocators.
I have never seen dispersion quoted Michael Mobison has written
a paper on it. So there are academics who are
aware of it, but I don't think people realize that
is the calculation for the fruit on the tree, the
(17:32):
meat on the bone, for these pod shops, there has
to be dispersion to pay off a non factor you know,
a factor neutral portfolio. So what the quants really do
and this is like what Gappy touched on when he
talked about you know, the day in the life of
a quant and your other guest within the last month
whose name I can't read. Yes, there was lots of
(17:54):
talk about risk management, right because of course it's of
utmost importance when you have a leveraged firm, right, you
have to understand every factor that's bouncing around in there,
and that's really their job, and they will, of course
because draw downs in a leverage firm, drawdowns are to
be avoided as much as possible. So the sharp ratio
(18:14):
really drives the way the quants are looking at pms.
But they're all backwards looking, sort of in my view,
So they do strip out everything. But once they get alpha,
or as I know, one firm calls it idiosyncratic alpha.
What I then do is the next step. I don't
change the definition of alpha, but then I break that
into a skill framework so that once you get different skills,
(18:38):
you can say this one's more repeatable than another skill,
et cetera.
Speaker 3 (18:42):
So like dispersion weighted basically like weighted by the opportunity
that's available.
Speaker 4 (18:47):
To you, Yes, exactly, And that's what allows you Tracy
to compare the energy trader to the consumer discussionary trader
because and I make an analogy in the book, it's
like looking at NFL punt. You know, PM's job is
to make as much money as possible, and essentially a
punter's job is to kick the ball as far as possible.
So before sports analytics came along, punters were judged on
(19:09):
and in fact, I think there was even award for
the punter that had the biggest average at the end
of the year, right the distance of all his punts
divided by total number of punts. But what sports analytics
people quickly figured out is that, well, hey, if the
best punter is averaging forty four yards to punt, and
you've got another punter whose coach is so conservative that
he's constantly punning from the opponent's thirty five yard line
(19:33):
or the opponent's forty yard line, he can't even get
a forty four yard punt off. So the way to
measure that is to say, okay, when a punts from
his own fifteen yard line, I'm going to measure that
against every other punt from the fifteen yard line. And
now you each punt is then evaluated. And I think
(19:54):
what's really important to the work I do too, is
or to note, you don't measure it now by the distance.
You measure it by plus or minus the average punter.
So you can say someone is on average one and
a half yards better per punt, and then you can
put a value on that. And that's the same way
a lot of you know, my framework is, it's not
(20:15):
saying you know it especially sort of like that that
canned package. You will see a canned batting average on
all analytics platform. It's meaningless. In fact, it's worthless. But
if you express it the way I just talked about punters,
sort of this skill neutral and to say, oh, his
batting average is one or two percent above, you know,
(20:36):
over the year he averages one percent a day. Well,
you know, in a fifty percent portfolio, that would be
you know, one more winner than expect it every other day.
Then you can compare that to the dispersion world that
he lives in, and you can put an absolute value
on his skill. Now it might differ from the actual,
but that's because of stuff out of the PM's control.
(20:58):
So that's the approach, and it's sort of marrying the
sports analytics approach. And again you kind of said, like,
why isn't this done? I do have some thoughts on
that because I got dropped into a fish out of
water quant division And they're brilliant, right, they are brilliant,
but they're not very flexible in they're thinking. They tend
(21:18):
to think the same way. And I found that when
I was interviewing for a quant developer, you know, because
I'm sort of building my framework on Excel and then
you need some production around it to make it usable
in a big firm or to clients. And I was,
you know, an interviewing for a quant developer. I couldn't
get them to stop talking about factors because that's sort
(21:39):
of the way they're trained. And I'm like, okay, right,
we're gonna strip out factors. How would you evaluate skill?
And again it get you know, it came down it
was very.
Speaker 3 (21:48):
Good start talking about factors again.
Speaker 4 (21:50):
Yeah, yeah, And like I say, they're brilliant, but I
think sort of an approach outside the industry, Yeah, it
can really help. You can uncover different stuff by sort
of marrying two different industries.
Speaker 2 (22:02):
So a lot of this stuff, so far as you've
described it is intuitive as you describe it, like, yeah,
it makes a lot of sense that you know, you
have to if you're going to compare two different pods
that are trading consumer discretionary, you have to understand that
dispersion and how they compare to a.
Speaker 3 (22:20):
Or comparing someone trading consumer discretionary versus like utilities.
Speaker 2 (22:23):
Totally, and it makes sense to me that there's more
than just volatility adjusted return sharp ratios. And it makes
sense to me that punters shouldn't just be measured on
pure length because you don't know where their coaches have
them punt from. And maybe sometimes you want to punt
shorter for various reasons because you want to have a
chance that you know, if you're catcher something like that. Okay,
I get all of that. Talk to us a little
(22:46):
bit more about the art of measuring skill, specifically outside
of returns, because this is the moneyball thing, which is
like every day they're coming up with new metrics and
vanity metrics that they these conferences where it's like vorp
and all these things. And I know that vorp is
like that was like twenty years ago that someone invented vorp, right,
(23:07):
but you know, there's all of these new things that
I was trying to come up with something that will
unlock this is the guy who produces a lot of
extra wins or something for the baseball team. What are
some of the other techniques or maybe what are the
other skills sure that you can measure a traitor on
other than just looking at xpos factor returns to justify risk.
Speaker 4 (23:27):
Right, Yes, so's that's a great question. And I'm laughing
as you talk about the acronyms because obviously the sport
channel at the community is famous for their acronyms. So
I in creating my framework, I have five skills that
explain alpha, okay, and it doesn't reinvent alpha or in
any way, it just breaks it down, and of course
(23:49):
I use acronyms to describe, and with a nod to
the industry that inspired them, I've named them after five
different baseball players from you know, the night teen seventies
when I was an impressionable baseball fan, And those skills
by name are sever Aaron, carew Rose, and then lumb
(24:09):
Lum is something you probably a name you haven't heard of,
but that is named for.
Speaker 3 (24:13):
A Yeah, that is named for it.
Speaker 2 (24:18):
Sorry, I'm not the other four Tom.
Speaker 4 (24:23):
Aaron, Aaron so all Hall of Fame level players, even
though Pete Rose isn't Lan but so interestingly, and I
won't go in deeply into this.
Speaker 2 (24:31):
But Rose measures the degree to which they're betting on
the side.
Speaker 4 (24:34):
How good are they at yes, at being well, Pete Rose,
it's a good one. So this isn't actually a descriptive acronym.
So Rose stands for return on sector excellence. So why Rose?
And you know why this? Well, Pete Rose made more
All Star teams at different positions than anybody else in baseball.
He made an All Star team at second base, outfield,
third base, and first base. So he was good at
(24:55):
sector rotation, right, So that that's sort of what that
skill is, measuring the lumb for luck uncontrolled by the manager.
Lum and what that really references, Tracy, It's a lot
of what we were talking about in terms of the
dispersion and really sort of the average stock in a
portfolio versus what the benchmark might be, because the average
(25:15):
stock is really what the skill neutral performer. Well, those
differences are sort of luck that is either a tailwind
or head wind uncontrolled by the manager. And Mike Lum
was a journeyman player who happened to play on an
Atlanta Braves team with Hank Aaron and Davy Johnson Daryl
Evans when they all hit forty home runs. They're the
only team that did that, and that inflated all of
(25:36):
Mike Lum's performance too, and obviously it's something he couldn't control.
But so these skills, I think the So what they're
really measuring is one is luck, two is sector excellence.
Third is a consistency measure, and that's the Rod carew
and in great batting average. And then there's power. Like
I talk about what the expected return is of that
perfect portfolio, Well, if someone's return is above or below that,
(25:59):
what that's really measuring is their ability to identify the
best of the outperformers and crucially avoid the worst of
the outperformers. And I can quantify that. And then the
final one, the siver is a sizing thing, and you
put all five of those together and you might have someone, Well,
here's a great example of how it's useful on a
multi manager platform. And I should say that all my
(26:22):
work only deals with public equities. Yeah, public equity APMs
evaluating them. So on a POD platform, you might have
four dozen, five dozen different teams, right, and you generally
do not need a model to tell you who the
best two or three are, And to a little lesser extent,
(26:43):
you don't need a model tell you who the worst
two or three are. They're outliers, and they the ones
that are really good are out there every year. But
in the middle you might have three dozen pms that
are tightly bunched around sort of the average production of
all the pms. What the model is really good at
is looking at these very similar returns at the end
(27:04):
of the year, looking at the skills that make them up,
and say, well, I know sizing tends to have a
correlation of zero from year to year. It reverts back
to the mean. So if you have two people with
the same return, but one of them was adding alpha
via their sizing decisions versus someone who was more consistently
picking out performers, and this is what you don't see
(27:27):
if you're just looking at idiosyncratic alpha, even though you've
stripped out all the factors. That's how the framework comes about,
and that's how it is both backward looking in terms
of explaining alpha by skill, but then also it becomes
a forward predictor by knowing what the correlation is between
past and future periods.
Speaker 3 (27:45):
I have so many questions for my next one, and
let me just add a caveat before I ask it,
which is everything I know about baseball I learned from
that one episode of The Simpsons. So that is to say,
I don't know very much at all other than don't
mean to Daryl Strawberry. But my impression, and again I
don't remember a lot about moneyball, but my impression was
(28:07):
like part of that strategy was finding players that are
underpriced by the market and capable maybe of doing one
specific thing very well, and then kind of putting them
together into a team that can work very well, like holistically,
rather than just going after the expensive players that hit
(28:27):
home runs a lot. Yeah, exactly. I guess my question
is I get the approach to evaluating individual traders, but
is part of your approach also looking at how they
like holistically work together and impact each other at all,
or because of the nature of multi strats and the
pod shops, doesn' not matter so much on that.
Speaker 4 (28:48):
That's an insightful question, And I will pick up a
topic that Gappy talked about a couple months ago. He
talked about the different cultures and how like how these
pod shops and the and the multi manager platforms can
be different and a lot of times there's a big
culture difference. And I would say that that is absolutely true,
and I have a great sort of answer to your
(29:10):
question for that. So at some shops, the philosophy is,
we are going to strip out everything a PM does
and cynically they have so many factors, and we'll pay
them on what the idiosyncratic alpha that's left is. And
they have so many factors they're stripping out that you know,
(29:30):
they're trying to get that alpha number down as small
as possible so they don't have to pay off bonuses.
And I remember joking with a PM one time at
one of those shops and he's like, yeah, I feel
like every time I go in there, they tell me like, yeah,
you had a good year, but look year out performance
is due to investing in dividend paying companies where the
CEO went to an IVY League school and we can get.
Speaker 2 (29:51):
That for free, right.
Speaker 4 (29:52):
So that so at those shops, their philosophy is it
doesn't matter because we're taking out everything. I prefer a
little different approach, and there are shops that do it
this way, which is to say, my job as a
CIO is to build a multi manager platform where some
(30:14):
of these offset so that there are different skills and
then instead of stripping out factors at each portfolio level,
more stripping out the factors. Once you put them all together,
you've got this bully of base stew and then you
take the factors out. And that is a different approach
because I think the pms feel a little bit more freedom.
(30:35):
They still have their buffers they have to stay in,
but they don't see the ETFs or the factor anti
factor things getting shoved right into their portfolio. The approach
is more higher. So you can take either approach. I
do prefer the sort of roster construction idea that you
have that you have in sports, but that really is
(30:56):
a difference in you know, I think in from culture.
Speaker 2 (31:00):
Yeah, that's super interesting. So in baseball, a general manager
(31:21):
looking for players can look at other teams, they can
look in the minor leagues, they can look at college sports.
They can start scouting at high school. Probably there's a
farm system and they call it a farm system. What
you've described so far makes sense for evaluating people in
existing seats, either at your shop or perhaps at another shop.
(31:42):
Is there a way to transfer it or to apply
some of these same ideas to people who don't have
the same because there I don't think there's the same
equivalent unless trading, you know, an Mari trader Schwab, which
actually I do think maybe is kind of a thing.
But is there a way to sort of think about,
like how you evaluate someone who just does not have
(32:03):
that much of a track record yet.
Speaker 4 (32:05):
Yes, because of the way these multi manager platforms are
formed now, they don't hire from the street anymore. I
think twenty years ago, thirty years ago, I know when
I was on the street, the researchers that were covering
the companies, they'd get plucked away, sometimes by the shops.
Sometimes traders would get plucked away. Right, that doesn't happen
(32:25):
as much anymore because what these huge firms have done,
and this also goes to their competitive advantage and their
ability to scale, is they are now training these people
right out of school, right. They have you know, universities
or academies or you know these schools essentially where they're
teaching people to be analysts or pms and again sort
(32:49):
of to a culture thing. My favorite ones are the
ones where the firms realize it used to just be
an upper out thing, right, Like you became an analyst
and then you became a PM, and if you weren't
a good analyst, you never became a good PM. And
I think that there are firms now that recognize and
analysts can be a career. You may be a great analyst,
(33:10):
but not necessarily united a capitol committee. You know, there's
a different skill set to being the PM. And they
find out some of these things in the academies and
in the universities. They're in house training schools. This is
the farm system that is coming up. Quite literally, this
is the bench and we see that and they don't
(33:31):
just get thrown in. They do tend to run paper
portfolios or portfolios that feel like they're real because they
are entering trading.
Speaker 2 (33:40):
So in their careers depend on them doing well. So
they're taking risk even if it's paper money.
Speaker 4 (33:44):
Yes, exactly, And you can run the same analytics on
these portfolios. And what I definitely have seen is some
of the newly graduated pms. These firms are good at
who they're training and those are the best pms to
find alpha signals from. Because they're portfolios, they tend to
(34:06):
be small so they can be replicated. It's it's and
this is another job of the quants too. If you
have a very senior PM, who's you know, who has
a contract that allows he or she to run a
two billion dollar biotech portfolio. There's not much left for
the quants to you know, because they're you know, they're
probably a little more thinly cat capitalized. There's not much
(34:27):
room to replicate that portfolio at another quant level in
the firm. But the new people that are coming up,
they're cheap, they're running small portfolios. But if they're skilled,
they're knowing what they're in is just as important as
a more senior PM.
Speaker 2 (34:43):
Yeah, Tracy and listeners. There's a great piece on the
Bloomberg from June nineteenth by our colleagues Nishan Kumar and E.
Liza Tetley about exactly this hedge fund talent schools are
looking for the perfect trader, and it talks about zero
point seventy two and it talks about citadel building these
sort of in how training things. So all all these
pieces are coming together, building the own farm system in
(35:05):
house to see who's going to be good one day.
Speaker 3 (35:07):
We should go to talent school. It's fun talent school,
that's to be clear. Okay, that was the joke. Yeah, Okay, Joe,
you've talked about sizing and you talked about the general
skill set that you're looking for one thing I'm still
unclear on. You alluded to it earlier, but I would
love for you to talk more about it in detail.
(35:29):
Time frame. What is the time frame by which you
are evaluating traders? And I guess how much runway do
you give people to either prove themselves correct or prove
themselves to be disastrously wrong because you know the correlation
they were betting on suddenly breaks down.
Speaker 4 (35:48):
Yeah. So again, great question, and it became a point
of frustration for me from when I first started at
Novus and building this stuff because I was very used
to sports analytics, and specifically baseball, but some some other
sports as well. I'll touch on golf. When you're evaluating
the skill of a picture, and there's three skills that
(36:08):
a picture has that are not dependent on anything else,
not dependent on it's teammates, who's batting it, et cetera.
It's right, not dependent on fielding. Right would be the
strikeout rate of a picture, the walk rate of a picture,
and the ground ball rate of a picture. These are
things that the picture controls, and what happens is after
about fifty plate appearances, you get the strikeout rate for
(36:30):
a picture. That is predictive of you know, it's you know,
from a maths standpoint, the correlation is above zero point seven,
so squared it's above point five. Right, the past explains
more of the future than factors that we haven't identified.
But with PMS, there's much more noise in their result
and it takes a lot longer to find a meaningful correlation.
(36:52):
So although I can do work for like, I can
and I do this for a single day, right, So
every day I generate a report, and I do this
for clients now showing their PMS and exactly what they're
readings of all these skills were each day. And of
course for one day it's just trivia. It's no more
than trivia. But what it is doing is building a
data set. And at the point that you get to
(37:15):
six months, which is about one hundred and twenty five
days trading days. Bigger picture, the full model takes five
hundred the past five hundred results, and that's when you
start getting very different and but more persistent correlations between
all these skills. Right. But what I have found is
(37:36):
that even after one hundred and fifty days, if you
take for the other year and a half, a mean
reversion assumption, and then just every time a new day
comes in, you drop off an assumption you'd get you
have a pretty robust skill reading that starts to mean
something after six months, and after two years, that's when
(37:56):
it really has, you know, has some great predictive power
for the next quarter, and so you're constantly dropping off. Now,
why only two years? I talk about this in the book.
I don't have a great answer for that.
Speaker 3 (38:07):
I suppose you have to start somewhere around.
Speaker 4 (38:09):
Yeah, well, here's what I knew. Two years was better
than three years, which in one sense, why would that be?
And I have talked to different quants about that, and
they have approached this from a much different perspective, and
they also have come to somewhat of a two year conclusion.
The reason seems to be regimes within the stock market,
(38:29):
just something about where you are skilled. You know, I
haven't been able to identify it. And I also know
that we could do a you know, we could we
could run the numbers and find out that, oh, you know,
it's not two years. It's the most predictive thing for
the last three months would have been two years and
(38:49):
forty three days. When you try to get that precise
that's not going to be what the perfect So two
years does seem to work because you're constantly rolling off
whatever happened two years to go, and so there's some
regime change that seems to work. But that is an
unanswered to a question I have too.
Speaker 2 (39:07):
So one thing about baseball is that every GM in
baseball has basically perfect visibility into the performance of every
player on every other team because it's all out on
the field and it's all measured, and we all have
the same information. You know, one of the.
Speaker 3 (39:21):
Most measure heart Jack.
Speaker 2 (39:22):
Yeah, right right, you can't measure Harvey get we all
could see players on base percentage and ops and slop
and all of this stuff, right. You know some of
the most popular alerts that always read spike on the
terminal or it's like consumer Discretionary Manager, Palacity goes to
citadel whatever. People love, people eat that stuff up. Just
from an industry perspective, sitting aside, whether you want to
(39:46):
use a traditional sharp ratio perspective or rose or lum
or whatever, how much visibility does one shop have into
the performance of a pod at another shop that can
then be ported over, or how much insight can you
we have if maybe there's an undervalued player somewhere else,
if you want to bring them over and give them
more capital.
Speaker 4 (40:04):
Than they're getting, extremely limited in terms of it. So well,
and I'll tell you who can solve that problem. What
you will have is, of course, if a team is
marketing itself or being recruited by another firm, they bring
over their returns, right, But they don't bring over They
might talk about portfolio construction, but I'm pretty sure they
shouldn't and probably don't bring over their two years right,
(40:26):
what the portfolio, what their holdings have been for the
last two years. So you don't get that type of visibility.
But let me tell you who can. Okay, And this
is I think one of the most important constituents in
our industry because I think they have the purest motive,
and that is the allocators, right allocators. I'm talking about
(40:48):
the huge, multi billion dollar entities which provide the blood
that keeps the heart pumping right at all these hegge file. Sure,
we know that Ken Griffin has a tremendous amount of
the aum is his money, and and you know we
hear that about some other people too, But in general,
(41:09):
these firms it's outside money which which keep these firms afloat.
But the allocators, many of them though, and what I'm
talking about here are foundations, university endowments, sovereign wealth funds, right,
pension plans, and they have a very pure motive. Right,
they are trying to get returns for the retirees or
you know, reduced tuition for future students, et cetera, or
(41:31):
you know, in the case of Norway, the citizens of
the country.
Speaker 2 (41:34):
Right.
Speaker 4 (41:35):
So they are a treasured investor, right if you run
a hedge fund. So when they are doing manager selection,
they have the ability to go to hedge funds. Now,
maybe not Citadel and Millennium, but to all these non
multi manager platforms. They have the ability to go to
them and say, hey, if you want us to really
(41:57):
evaluate you, we need to see podil. Yeah, we need
to see we need position level transparency for the last
two years. Hey if you want, if you don't want
to give us yesterday, start a quarterback so that it's
on a lag. But now they have the leverage to
get those returns, especially if you're talking about emerging managers, right,
young managers are trying to that, you know, to build
a hege Fund, and I don't feel they use that leverage.
(42:21):
And this is to me is like, well, Joe, you're
you know you you talk about this framework and it's
it's applicable to multi manager platforms, and you know an
endowment isn't a leveraged portfolio, so how could they use it? Well,
this is how they could use it, because they can
get that, Joe, and they do ask those questions about
the Bentley and the.
Speaker 2 (42:40):
In fact, can I give you an example? Can I?
Speaker 4 (42:42):
This is I really like this. I've never worked with them.
I should say that I have worked with their brethren,
and I've worked with the endowments that they would measure
themselves against. But the MIT Endowment Matimco is the name
of the uh, the name of the entity. They have
something between twenty and thirty billion dollars under management. So
(43:04):
we know that a portion of that is dedicated to
public equities. And we know because and I won't mention
his name, so I'm not trying to call him out,
but we know that one of those gentlemen that looks
for equity managers is a presence on fin Twitt. He's
actually a great follow, very earnest, and so he'll talk
(43:26):
about things, and sometimes he'll post job postings.
Speaker 2 (43:29):
Right, and what you will find is.
Speaker 4 (43:32):
Everybody who works in that division or and in fact,
you can even see this publicly. I know this. Yale's
Management company has the resumes of every person who's in
that division, and they're all the same. Here's what they
will say. They will say things like, you know, was
president of the investment club at the University of Virginia, right,
and they'd been investing in stock since I had a
(43:55):
paper route.
Speaker 1 (43:55):
Right.
Speaker 4 (43:55):
They're always have this, right. So when they go to
do managers selection, and I've been on that side too
as a as a marketer, they will sit down with
the PM and they'll ask they'll go over each position
in the portfolio and make no mistake about it, they're
passing judgment, right, because if they're not frustrated or want
to be pms, this is how they think about the market.
All right, So I'm gonna put full stop there. Now,
(44:18):
let's go to a The general manager of the Philadelphia
seventy six ers is a gentleman named Daryl Mory.
Speaker 2 (44:23):
Oh yeah, I like Daryl. Darryl.
Speaker 4 (44:26):
Right, he was with Houston and in fact, while he
was at Houston. He really brought moneyball to the NBA.
Mark Cuban was probably maybe the second, but Darryl Moury
right down to the fact that Michael Lewis did a
piece on him in the Sunday New York Times maybe
twenty years ago. So Daryl Morey is the GM of
the seventy six ers, and he has juniors too, right,
(44:46):
And when they're doing their equivalent of manager selection, whether
it'll be drafting a player or looking at free agents,
can you imagine how absurd it would be for Darryl
and the analysts to go down and shoot free throws
with perspective player right, and to judge the player based
on that. But I guarantee you at the endowments they
(45:07):
go back and say, can you believe you know that
managed your short netflix right? Like, so, now, why did
I pick those two? And the example is this, before
Daryl got into basketball, he is a proud graduate of
MIT Sloan. He got his MBA at Sloan School of
Management and he started along with a woman named Jessica Keelman,
(45:29):
he started the Sloan Sports Conference, which started as Bill
Simmons when he was at Grantline described at as Dorcapalooza. Right,
it was just people, kids, guys, and it was almost
all guys back then talking about sports analytics. And it
has morphed into a massive event and it's a job
(45:50):
fare where all these sports teams from all different leagues
are looking for talent, right, and they're essentially looking for
performance analytics people. Right.
Speaker 2 (46:00):
Voros McCracken was the one from what's.
Speaker 4 (46:02):
A exactly exactly, So this is think now, now, look
across the campus at the MIT Sloan Endowment. What they're
actually trying to find is performance analytics. Do you think
there might be anybody right across the campus who may
have never invested in stocks but gets the profit motive?
(46:23):
They would take my work and probably take it three
steps more. But I don't think there's an endowment out
there that things like that. Like, I'm sure they've never
walked across the campus, and even I'm sure Darryl has
never thought to invite them over to the you know,
to hey whyt to interview some of some of our people.
So that's again sort of how I look at, like, Hey,
this is how some of this work. How you somebody
(46:45):
who doesn't have the data can get it at the
allocator level.
Speaker 3 (46:49):
We've been talking very much about, you know, performance evaluation
and metrics from a sort of managerial level. If I
am a trader or quant, you know, a sort of
junior medium level quant, I guess at one of these
multi strat hedge funds, how am I viewing the performance
of others and competition? Is it the case that I'm
(47:11):
trying to move into a particular sector that maybe has
more of an opportunity set in terms of dispersion, where
maybe there's more volatility or more relative value opportunities or
something like that. How am I like viewing my competition?
Speaker 4 (47:29):
It's a good question. Even the work that I do well,
I think there's definitely comparison, right. You douce again to
culture some of the pods, and I think I think
Yapy mentioned that some of the pods there is a
there's a sharing of information, right, and at some shops
there's not. This is a case where I actually prefer
(47:51):
the not sharing of information right, because I would rather
I think the quants would rather know that maybe two
different pms came to the same conclusion independently, as opposed
to they both went to the same idea. Dinner and
then both decided to buy the stock. There's more of
a signal in somebody coming to it independently, so I
(48:13):
believe they're aware of what the returns are of their
other pms. In addition, and I don't know if this
was in that article Joe you just referenced, but these
firms all have coaching teams too, and I certainly found
that the older pms that you know had been in
the business since the nineties, they're setting their ways right.
(48:34):
They don't want a quant to come in with a
laptop and start telling them that's spin rate, right, the
spin rate of their pictures. But the younger people, I
think there's more of a hey, if there's data that
you can give me to help me get better, I
think maybe in some ways they might be looking for that.
Speaker 2 (48:52):
Joe Peter, this is super fun. Thank you so much
for coming on the podcast again.
Speaker 4 (48:57):
Oh it's always great to be here. And I'll see
you in seven years.
Speaker 2 (49:00):
Yeah, exactly when whatever your next job is from.
Speaker 3 (49:02):
The Sun, Thanks so much.
Speaker 4 (49:04):
Joke.
Speaker 3 (49:04):
Yeah, even though there were baseball references, I enjoyed it.
Speaker 2 (49:07):
Yeah, Tracy, that was a really fun conversation.
Speaker 3 (49:23):
I love hearing stories when people get jobs off the
back of Authoughts appearances.
Speaker 2 (49:26):
That's nothing sort of flatters are egos and sense of self.
Speaker 3 (49:31):
Well know, I also like it when people say they're
listening to A Thoughts episodes while going to the gym.
Oh yeah, because I hate going to the gym. I
hate running and things like that. But it makes me
feel nice that like people are listening to us or
to offset something that's kind of like a chore.
Speaker 2 (49:48):
No, But beyond all that, it was very fun. I
feel like we could just talk about these businesses forever.
It seems so rich, you know, like we still have
to do something on like compensation struggle. Yeah, but also
like just the sort of like fundamental point that everyone knows,
which is like manager identification is really difficult because and
(50:10):
you know, first of all, there's all these questions about, well,
it's beating the market really possible because of efficient markets
and stuff. And then you could identify someone to all
this person beat the market seven years in a row,
but have a pool of one thousand managers. There's gonna
be a lot of people who beat the market seven
years in a row, and so it seems like a
very interesting problem to solve.
Speaker 3 (50:29):
It's kind of funny that you're trying to like select
traders on a factor neutral basis, who are themselves able
to be factor neutral in some respects, Like you're kind
of you're trying to separate them from like these circumstances
that they are operating in, or trying to wait them
against the value of the opportunity that they are currently facing.
(50:51):
Right that dispersion that Joe was mentioning, that's kind of funny.
I've thought about it. Not to go all media and
naval games, so sure, but you know, it's sort of
similar to journalist beats in some respects, where you can
get really lucky and be on a really interesting beat
where there's tons happening, and suddenly you know, all your
(51:12):
stuff is getting read and you're getting all these major scoops,
and then maybe two years later, to go back to
that timeframe point, it's sort of faded into the distance
and there's not as much to write about, And how
do you judge the talent of a particular journalist or
a trader from their particular set of circumstances.
Speaker 2 (51:30):
That's a great example. I remember, you know, like when
I was in a business insider years ago. It's like the
reporters who covered Apple on days of like iPhone announcement,
they got we were like measured on traffic back then
they got all the traffic.
Speaker 4 (51:43):
You know.
Speaker 2 (51:43):
It's like, oh, this isn't fair, Like I'm talking about
like the Bank of England decision. This is nonsense. Okay.
Speaker 3 (51:48):
I just read a really good analysis of like US payrolls,
and people only want to read about the next iPhone.
Speaker 2 (51:53):
I explained, I explained Mario Draggy's new OMT thing really
well and like ten people ready, But no, like this
is like it's all like version of the same problem.
By the way, my hedge fund media metaphor that I
use in my head is like alpha decay. So it's
like the first person who ever came up with like
here's what you need to know or the answer will
(52:14):
shock you, like probably like did crazy. Well, but by
the time that was you, wasn't it Yeah, And then
by the million person who did like the answer will
shock you, it stopped working. So it's like there's the
same thing of like alpha decay, where it's like you
can be the first on a strategy and then everyone
discovers it and then the excess returns from.
Speaker 3 (52:30):
That move on the crowding in effect. Yeah, no, I
did think actually that timeframe point was really interesting and
the fact that Joe kind of I guess gravitated towards
two years or five hundred trading days. But then he
was talking about how others seem to have sort of
alighted on that same time period. Yeah, I wonder why that.
I mean, I get that you have to at some point,
(52:51):
you just have to choose, like age horizon, but it is. Yeah,
it's an interesting one.
Speaker 2 (52:56):
Very interesting stuff. Plenty more to come on this topic.
Speaker 3 (52:59):
All right, shall we leave there.
Speaker 2 (53:00):
Let's leave it there.
Speaker 3 (53:01):
This has been another episode of the Oudlots podcast. I'm
Tracy Alloway. You can follow me at Tracy Alloway.
Speaker 2 (53:07):
And I'm Joe Wisenthal. You can follow me at the Stalwart.
Follow our guest Joe Peta, He's at Magic Rat SF
and check out his book Moneyball for the money Set.
Follow our producers Carmen Rodriguez at Carman Erman dash Ol
Bennett at Dashbot, and Kilbrooks at cal Brooks. Thank you
to our producer Moses One. More odd Lags content go
to Bloomberg dot com slash odd Lots, where you have transcripts,
(53:29):
a blog, and a newsletter, and you can chat about
all of these topics twenty four to seven in our discord,
discord dot gg, slash odlogs.
Speaker 3 (53:38):
And if you enjoy odd Lots, then please leave us
a positive review on your favorite podcast platform. We will
accept positive reviews in lieu of finders' fees. And remember,
if you are a Bloomberg subscriber, you can listen to
all of our episodes absolutely ad free. All you need
to do is connect your Bloomberg account with Apple Podcasts.
(53:59):
In order to do that, just find the Bloomberg channel
on Apple Podcasts and follow the instructions there. Thanks for listening.
Bloomberg Audio Studios, Podcasts, Radio News.
Speaker 2 (00:20):
Hello and welcome to another episode of The Odd Lots Podcast.
Speaker 3 (00:24):
I'm Joe Wisenthal and I'm Tracy Alloway.
Speaker 2 (00:26):
Tracy, do you know this? Sometimes I wonder, like, you know,
one in the morning, if I can't sleep, I think
to myself, in a different life, could I have been
the next Steve Cohen? Yeah? No, for real though, And
I don't, you know, need to talk about it. There's
a lot and I've brought it up before, you know.
I did get an offer at a prop trading shop
right after college to be a stock trader at this
(00:49):
place where they're gonna let use to your capital. And
I think Steve Cohen started off like as a prop
trader at some shop before being one of the great
hedge funders of all time. And I didn't take that
job for reasons I still can't explain to myself twenty
five years later, But I always wonder whether could have
cut it. Maybe I could have been a good trader.
I don't know.
Speaker 3 (01:07):
It's good you have a healthy level of self confidence, Joe. No,
When I lay awake at night, I think like, oh, shoot,
what did I say? Something stickid on the podcast, and
that's what keeps me up. But yes, good for you, Joe.
Speaker 2 (01:19):
No, I don't really think I could have, and I
actually do not think I would have been a good trader.
I don't think like that. I'm not that good at
poker other things. I'm not a natural better I don't
do like sports betting. I don't think that. But I
do sort of, you know, wonder about what my life
had been different if I had said yes to that.
Speaker 3 (01:38):
Yeah, fair enough. I mean we know, we know from
multiple episodes of the podcast this year alone. Yes, like
there are a lot of hedge fund traders out there,
especially in multi strats, who seem to be making a
lot of money, and everyone's sort of talking about them
up until recently. Maybe I should say, we're recording this
on August seventh, so maybe those bonuses a little bit
(02:00):
less this year given the market sell off. But up
until this month, people seem to have been doing relatively well,
and there was all this intrigue and interest in the
world of traders. And I'm sort of curious. This has
come up a couple times now, but what makes a
good trader and how are traders actually evaluated? Because my
(02:22):
impression was always like, Okay, well it depends on how
much money you make, but what's the timeframe for making
that money? And then also what about people who are
working in for instance, these particular pods who are doing
one specific thing, like what is the benchmark against which
they are judged?
Speaker 2 (02:40):
You mentioned that maybe they're not making so much money
this week or this month, But Tracy, I think we're
told all the time they're so neutral on everything. Their
market neutral, they're beta neutral, they're neutral every factor you
can think of. Why should they be losing money right
now They're supposed to like be neutral all.
Speaker 4 (02:55):
Of this time.
Speaker 3 (02:56):
Yeah, I'm sure they're making loads shore. I'm absolutely sure.
Speaker 2 (02:59):
No, but you're right. And look, we've been doing a
lot on hedge fund structure, and we did that episode
with Giuseppe Polyoligo, and we did that episode with Rich
falk Wallace, various aspects of like how hedge funds measure
risk and try to isolate alpha and all this stuff.
But they're just like so many questions in my mind,
Like I feel like we're just scratching the surface because
(03:21):
you know, we haven't even really talked about like idea generation.
So it's one thing to you know, talk about like, okay,
here's how you like factor out all of these exposures
that you don't on have like market beta, et cetera.
It's another thing to talk about like okay, but like
how do you pick the stocks to buy or go short?
Speaker 3 (03:37):
Well, yeah, we have gotten into this a little bit,
but you're right, there's more we could do. There are
all these questions about like how do you size your positions?
And if you're convinced that one thing is going to
be the next big thing, then why don't you just
have like one hundred percent positions?
Speaker 4 (03:50):
Yeah?
Speaker 2 (03:50):
Right, how do you make money if you can't just
go one hundred percent leverage long and video in video.
Speaker 3 (03:55):
Yeah.
Speaker 2 (03:55):
Anyway, so there's a lot more we can do. But
to my original, very egotistical start to this episode, I
do wonder like it's.
Speaker 3 (04:03):
Okay, Joe, it's good to have self confidence. I'm being serious,
thank you.
Speaker 2 (04:08):
I do wonder like this big question of like you know,
and a lot of people are probably interested in this
because these hedge one PM jobs or trader jobs seem
pretty great and as you mentioned, lucrative, and so it
would be interesting to know how a fund or anyone
goes about identifying like the next great trader who gets
to have that seat, so to speak.
Speaker 3 (04:26):
Well, I also think if you can identify what makes
a good trader at a hedge fund, then you can
get more into the business model of what they're actually
doing on a day to day basis. It helps us
understand what.
Speaker 2 (04:37):
They're really good at and what they can do specifically. Well,
I'm very excited today because we really do have the
perfect guest. We're going to be speaking with, Joe Peta.
He is the author of a recent book, Moneyball for
the money set, which is the name sort of implies
tries to, you know, figure out new ways or the
best ways to identify talent. I'm sure there's a lot
(04:57):
of old heuristics like they had in Bay, you know,
and they're like, well, this guy looks like he has
a good hustle, and then Moneyball came along. He's like, no, actually,
you want to really look at his like you know,
on base percentage or whatever it is, and stop looking
at like his like spirit or you know, his hustle
ahead of him. Anyway, and prior to that, in his career,
he's been in this industry for a long time. He
was the head of performance analytics at point seventy two.
(05:21):
So this speaks right to the question of how do
you evaluate traders. We also hit him on years ago
one of our really early episodes where he talked about
sports betting with some of these same ideas, et cetera.
So I'm thrilled to have Joe back to talk about
this basic question of how it's good trader. So thanks
for coming back, Joe.
Speaker 4 (05:40):
Oh, it's great to be here Joe and Tracy, and
nice to do it in person. Seven years ago, Tracy,
I believe you were in Hong Kong and yeah, Joe,
you just had a garage band instead of selling out venues.
Speaker 2 (05:50):
Now that's right, that's right. Still mentioned you were head
of performance analytics at point seventy two. How did you
get that job at a point seventy two? Steve Cohen's
big yes.
Speaker 4 (06:00):
So that goes right back to my appearance seven years ago.
So when I was on in twenty seventeen, I had
written a book called Trading Basis, which really looked at
the critical reasoning overlap between asset management, sports betting, and
the moneyballization of baseball. And you had ask me, Joe,
I think it was you asked me a specific question
of well, I mentioned that somebody from all three of
(06:22):
those constituents could learn something from the other two. And Joe,
you asked me for a specific example of how they
look at things differently, and I said, well, if you
go onto a trading floor, or you go to a
mutual fund and you ask them, hey, who's your best
trader or who's your best PM, Inevitably they will point
(06:42):
to the individual who had the highest return in the
prior year, either the biggest P and L or the
highest return on capitol. But I contrasted that that if
you went into the front office of a baseball team
and asked them who their best player was, they wouldn't
look at you know, which picture necessarily had the lowest
ra or the most wins. They would answer that question
based on skill sets, and so they it's a subtle difference.
(07:05):
Instead of looking at results, they would look at skills
because they know that the skills, there's so much noise
and results that the skills. If you can identify the skills,
you have a better chance of predicting who will do
better going forward. And as it was told to me,
a member of the c suite, at zero point seventy
two listen regular listener heard that episode and played a
(07:29):
portion of it for Steve. In fact, I think it
was the part I just mentioned, And I was told,
as it was relayed to me that Steve said, find him.
I want to talk to him. And I guess that's
not a surprise because in twenty twelve, and this is
all public knowledge. In fact, there's a book by Molly
Knight called The Best Team Money Can Buy that chronicles
(07:50):
the Dodger's ownership through the turbulent court years Frank McCourt's ownership,
and that team was sold in twenty twelve to the
Guggenheim Group. But Steve also bid for that team and
came very close to buying the Dodgers in twenty twelve.
And of course we all know him now as New
Yorker's nome his uncle Steve owner of the New York Mets.
(08:11):
So he has, I believe, always had an interest in
an analytical approach, and I think he always wondered, well
could that work in the hedge Fund. And I came
away from those meetings with the bunch of different people
in the investment committee, and I kind of came away
with three queries that I thought could sort of be
my marching orders and how I could help, And that
(08:33):
was I think at all these pod shops, when somebody
has a good year, they ask for more money, and
in terms of buying power, not cash, but in terms
of buying power. And so the question that management would
have is, well, is what they did repeatable? And at
the same time, as you know, there's turnover at these firms, right,
(08:55):
And I think another question is, well, sometimes we let
people go too early that then thrive elsewhere just because
they had a bad start to their career in terms
of results. Is there a way that we can avoid
that mistake? And then finally, when a team does well,
inevitably there's a bit of way, right because we know
that these four or five huge firms are all very
competitive and they're trying to steal talent. And so the
(09:19):
question is I know what a PM and or her
team may have made me in the past, but what
are they worth going forward? And all of those queries
can be answered by looking at skills, which is a
little different than what the traditional quants do at these firms.
Speaker 3 (09:35):
Okay, so here's my question, who should we bill for
the finder's fee fee? Should we send it directly to Steez?
Speaker 2 (09:42):
He do you right?
Speaker 4 (09:44):
It would be the firm, right, they probably saved a
lot of money as opposed to going through a traditional headhunter.
Speaker 3 (09:51):
Okay, I'm joking. Obviously that's fantastic to hear. I've loved
stories like that. Before we get into the existing model
of compensation. There's one question that I wonder because I
think we've done a number of Moneyball episodes at this point,
but it's been a while since we've talked about that approach,
and all I remember is the movie and Brad Pitt
(10:12):
kind of unconvincingly playing a guy that understands math. Could
you maybe explain, like what it is about the Moneyball
approach that seems to attract people in finance, Like why
is there that analogy that seems to come up again
and again.
Speaker 4 (10:29):
Yeah, I think if you're attracted to critical reasoning, and
that's the big thing and all of this industry is
you know, Joe said, what have I succeeded here? And
I always think the biggest question is do you have
the mentality in the stomach to make decisions and commit
capital based on incomplete information? Whether you have the skills
(10:50):
to you know, build models for you know, and and
understand companies and read documents. It's really can you make
decisions based on income information? And it's true at the
poker table, all right, and it's certainly true when you're
building a sports team, right you're like, how much is
this free agent worth? And before there were a lot
(11:13):
of Joe like you say heuristics, and I even mentioned
that in the book. I feel that still goes on
at the allocator level in this industry. Allocators they do
the interviews and you will hear things like, well he
just got divorced, you know, or there's a Bentley in
the parking lot. He must not be hungry anymore. Oh absolutely.
(11:33):
And one of the reasons is because they don't have
they don't take a different approach that might be more databased.
The whole idea of the moneyball approach is to tease
out skill from or the signal from these very noisy results,
because both athletes and asset managers, their results are filled
(11:55):
with influences over which they have no control. To answer
this question later, you both were talking about like market
neutral PMS, neutral everything pms, Why would they be having
the worst week this week than before. And there's an
actual real answer to that that has nothing to do
with their skills. Yeah, so this can apply to any
(12:19):
time period. We're looking at days or months, a year.
Let's go to sort of an economics one OHO one
like holding all else equal. Let's say we have a
PM that has one long and one short, okay, and
that's their entire portfolio. And of course they never would
this goes to something else you said in the intro
because of career risk. Right, even if it's their best
idea long and best idea short, they'll still fill it.
(12:40):
But let's say this is their portfolio and on any
given day or any period we could measure, but let's
keep it at a day.
Speaker 2 (12:47):
It's a perfect.
Speaker 4 (12:47):
Portfolio in that the long produces alpha and the short
produces alpha. So the long outperforms the market and the
short underperforms the market. Right, So that's a perfect portfolio.
What is the expected re turn for that portfolio for
like I say, any period, but for a day, And
the answer is there's a way to figure out. And Tracey,
you're gonna love this because the answer is dispersion. And
(13:10):
I know you light up when when you have the
But this is a little different dispersion than the quants
and the derivative traders make their life around. This dispersion
is and it's going to be very context specific for
where the PM toils. Right, so we know what these
pod shops they tend to be. They have you know,
(13:31):
subject matter expertise in sectors. Right, So you might have
an energy PM, and so let's say this is a consumer
discretionary PM, and you would say, okay, well, I'm going
to look at his or her universe and maybe that's
the S and P fifteen hundred consumer discretionary Maybe it's
say a portfolio of just consumer discretionary stocks that he
has modeled, so there might only be eighty or so
he and his team, but whatever it is, we'll say
(13:53):
that it's the all the consumer discretionary stocks in the
S and P five hundred or fifteen hundred. Well, the
way to figure out what the expected return is is
to simply look at all those stocks and say, here's
the skill neutral return, which would be the average return
of all those holdings. And then you look at the
ones that outperformed what was their average And you look
at all the stocks that underperformed and what was their average.
(14:15):
And the difference between those two numbers is the dispersion
between outperformers and underperformers, and that varies greatly from day
to day, and it can very greatly from year to year.
Speaker 3 (14:28):
It's not like the maximum that you can produce dispersions.
Speaker 4 (14:32):
Not the maximum, because you could have the very best
outperformer and the very best underperformer. But if you're looking
at a POD, so I'm taking all the pod from
all the shops across the street that are focused on
consumer discretionary, I'm going to be dead on by saying
the average of all those of all those perfect portfolios
(14:52):
is going to be the average of all the outperformers
and the average of all the underperformers. And it's invisible
to investment commits, to CIOs, to the pms themselves. They
can be just as skilled from one day or one
period and one year to the next, but the payoff
is different. And this is sort of the moneyball. Look
at hey, once we get this all context neutral, we
(15:15):
might say that a neutral everything PM that had a
seven percent return one year and a five percent cent
return the next year, he may have even been more
skilled than the five percent year, but the dispersion wasn't
there to pay off that skill.
Speaker 2 (15:30):
Oh, I see what you're saying. So in other words,
it's like, Okay, this person's up five percent. In order
to establish like whether that's good or bad or not,
you have to have some sort of like holistic view
of what dispersion on average looked like. In that that's
exactly well, that makes sense. It also seems kind of obvious,
(16:07):
you know, I know, the divorce and the Bentley is
probably like extreme examples. They're sort of funny. But you know,
thinking about the moneyball thing, and I mentioned in the
old days like oh, that guy looks he's a good
ey or whatever, just like all these sort of unquantified
his hustle, you know, his heart whatever. And then you know,
Brad Pitt or the being came along and actually put
some numbers on this. If they're not doing that, what
(16:30):
are the sort of like old heuristics that aren't the
extreme ones that the investment committees or the hiring committees
of the firing committees would have been using two aviliate.
Speaker 4 (16:40):
Pro So there's no question that it seems obvious, and
it's just the first building block. And this isn't Black
Shawl stuff in terms of complexity.
Speaker 2 (16:48):
Yeah.
Speaker 4 (16:49):
I started this sort of journey and analytics by working
for a company called Novus, and Novas was one of
about fifteen years ago, was at the forefront of portfolio analytics,
and in fact, they had read my book and I'm like, hey,
this is what we try to do. And I worked
for them. So I've seen just about every package out there,
whether it is from a vendor in terms of analytics
(17:11):
or you know inside firms. I've you know, worked with allocators.
I have never seen dispersion quoted Michael Mobison has written
a paper on it. So there are academics who are
aware of it, but I don't think people realize that
is the calculation for the fruit on the tree, the
(17:32):
meat on the bone, for these pod shops, there has
to be dispersion to pay off a non factor you know,
a factor neutral portfolio. So what the quants really do
and this is like what Gappy touched on when he
talked about you know, the day in the life of
a quant and your other guest within the last month
whose name I can't read. Yes, there was lots of
(17:54):
talk about risk management, right because of course it's of
utmost importance when you have a leveraged firm, right, you
have to understand every factor that's bouncing around in there,
and that's really their job, and they will, of course
because draw downs in a leverage firm, drawdowns are to
be avoided as much as possible. So the sharp ratio
(18:14):
really drives the way the quants are looking at pms.
But they're all backwards looking, sort of in my view,
So they do strip out everything. But once they get alpha,
or as I know, one firm calls it idiosyncratic alpha.
What I then do is the next step. I don't
change the definition of alpha, but then I break that
into a skill framework so that once you get different skills,
(18:38):
you can say this one's more repeatable than another skill,
et cetera.
Speaker 3 (18:42):
So like dispersion weighted basically like weighted by the opportunity
that's available.
Speaker 4 (18:47):
To you, Yes, exactly, And that's what allows you Tracy
to compare the energy trader to the consumer discussionary trader
because and I make an analogy in the book, it's
like looking at NFL punt. You know, PM's job is
to make as much money as possible, and essentially a
punter's job is to kick the ball as far as possible.
So before sports analytics came along, punters were judged on
(19:09):
and in fact, I think there was even award for
the punter that had the biggest average at the end
of the year, right the distance of all his punts
divided by total number of punts. But what sports analytics
people quickly figured out is that, well, hey, if the
best punter is averaging forty four yards to punt, and
you've got another punter whose coach is so conservative that
he's constantly punning from the opponent's thirty five yard line
(19:33):
or the opponent's forty yard line, he can't even get
a forty four yard punt off. So the way to
measure that is to say, okay, when a punts from
his own fifteen yard line, I'm going to measure that
against every other punt from the fifteen yard line. And
now you each punt is then evaluated. And I think
(19:54):
what's really important to the work I do too, is
or to note, you don't measure it now by the distance.
You measure it by plus or minus the average punter.
So you can say someone is on average one and
a half yards better per punt, and then you can
put a value on that. And that's the same way
a lot of you know, my framework is, it's not
(20:15):
saying you know it especially sort of like that that
canned package. You will see a canned batting average on
all analytics platform. It's meaningless. In fact, it's worthless. But
if you express it the way I just talked about punters,
sort of this skill neutral and to say, oh, his
batting average is one or two percent above, you know,
(20:36):
over the year he averages one percent a day. Well,
you know, in a fifty percent portfolio, that would be
you know, one more winner than expect it every other day.
Then you can compare that to the dispersion world that
he lives in, and you can put an absolute value
on his skill. Now it might differ from the actual,
but that's because of stuff out of the PM's control.
(20:58):
So that's the approach, and it's sort of marrying the
sports analytics approach. And again you kind of said, like,
why isn't this done? I do have some thoughts on
that because I got dropped into a fish out of
water quant division And they're brilliant, right, they are brilliant,
but they're not very flexible in they're thinking. They tend
(21:18):
to think the same way. And I found that when
I was interviewing for a quant developer, you know, because
I'm sort of building my framework on Excel and then
you need some production around it to make it usable
in a big firm or to clients. And I was,
you know, an interviewing for a quant developer. I couldn't
get them to stop talking about factors because that's sort
(21:39):
of the way they're trained. And I'm like, okay, right,
we're gonna strip out factors. How would you evaluate skill?
And again it get you know, it came down it
was very.
Speaker 3 (21:48):
Good start talking about factors again.
Speaker 4 (21:50):
Yeah, yeah, And like I say, they're brilliant, but I
think sort of an approach outside the industry, Yeah, it
can really help. You can uncover different stuff by sort
of marrying two different industries.
Speaker 2 (22:02):
So a lot of this stuff, so far as you've
described it is intuitive as you describe it, like, yeah,
it makes a lot of sense that you know, you
have to if you're going to compare two different pods
that are trading consumer discretionary, you have to understand that
dispersion and how they compare to a.
Speaker 3 (22:20):
Or comparing someone trading consumer discretionary versus like utilities.
Speaker 2 (22:23):
Totally, and it makes sense to me that there's more
than just volatility adjusted return sharp ratios. And it makes
sense to me that punters shouldn't just be measured on
pure length because you don't know where their coaches have
them punt from. And maybe sometimes you want to punt
shorter for various reasons because you want to have a
chance that you know, if you're catcher something like that. Okay,
I get all of that. Talk to us a little
(22:46):
bit more about the art of measuring skill, specifically outside
of returns, because this is the moneyball thing, which is
like every day they're coming up with new metrics and
vanity metrics that they these conferences where it's like vorp
and all these things. And I know that vorp is
like that was like twenty years ago that someone invented vorp, right,
(23:07):
but you know, there's all of these new things that
I was trying to come up with something that will
unlock this is the guy who produces a lot of
extra wins or something for the baseball team. What are
some of the other techniques or maybe what are the
other skills sure that you can measure a traitor on
other than just looking at xpos factor returns to justify risk.
Speaker 4 (23:27):
Right, Yes, so's that's a great question. And I'm laughing
as you talk about the acronyms because obviously the sport
channel at the community is famous for their acronyms. So
I in creating my framework, I have five skills that
explain alpha, okay, and it doesn't reinvent alpha or in
any way, it just breaks it down, and of course
(23:49):
I use acronyms to describe, and with a nod to
the industry that inspired them, I've named them after five
different baseball players from you know, the night teen seventies
when I was an impressionable baseball fan, And those skills
by name are sever Aaron, carew Rose, and then lumb
(24:09):
Lum is something you probably a name you haven't heard of,
but that is named for.
Speaker 3 (24:13):
A Yeah, that is named for it.
Speaker 2 (24:18):
Sorry, I'm not the other four Tom.
Speaker 4 (24:23):
Aaron, Aaron so all Hall of Fame level players, even
though Pete Rose isn't Lan but so interestingly, and I
won't go in deeply into this.
Speaker 2 (24:31):
But Rose measures the degree to which they're betting on
the side.
Speaker 4 (24:34):
How good are they at yes, at being well, Pete Rose,
it's a good one. So this isn't actually a descriptive acronym.
So Rose stands for return on sector excellence. So why Rose?
And you know why this? Well, Pete Rose made more
All Star teams at different positions than anybody else in baseball.
He made an All Star team at second base, outfield,
third base, and first base. So he was good at
(24:55):
sector rotation, right, So that that's sort of what that
skill is, measuring the lumb for luck uncontrolled by the manager.
Lum and what that really references, Tracy, It's a lot
of what we were talking about in terms of the
dispersion and really sort of the average stock in a
portfolio versus what the benchmark might be, because the average
(25:15):
stock is really what the skill neutral performer. Well, those
differences are sort of luck that is either a tailwind
or head wind uncontrolled by the manager. And Mike Lum
was a journeyman player who happened to play on an
Atlanta Braves team with Hank Aaron and Davy Johnson Daryl
Evans when they all hit forty home runs. They're the
only team that did that, and that inflated all of
(25:36):
Mike Lum's performance too, and obviously it's something he couldn't control.
But so these skills, I think the So what they're
really measuring is one is luck, two is sector excellence.
Third is a consistency measure, and that's the Rod carew
and in great batting average. And then there's power. Like
I talk about what the expected return is of that
perfect portfolio, Well, if someone's return is above or below that,
(25:59):
what that's really measuring is their ability to identify the
best of the outperformers and crucially avoid the worst of
the outperformers. And I can quantify that. And then the
final one, the siver is a sizing thing, and you
put all five of those together and you might have someone, Well,
here's a great example of how it's useful on a
multi manager platform. And I should say that all my
(26:22):
work only deals with public equities. Yeah, public equity APMs
evaluating them. So on a POD platform, you might have
four dozen, five dozen different teams, right, and you generally
do not need a model to tell you who the
best two or three are, And to a little lesser extent,
(26:43):
you don't need a model tell you who the worst
two or three are. They're outliers, and they the ones
that are really good are out there every year. But
in the middle you might have three dozen pms that
are tightly bunched around sort of the average production of
all the pms. What the model is really good at
is looking at these very similar returns at the end
(27:04):
of the year, looking at the skills that make them up,
and say, well, I know sizing tends to have a
correlation of zero from year to year. It reverts back
to the mean. So if you have two people with
the same return, but one of them was adding alpha
via their sizing decisions versus someone who was more consistently
picking out performers, and this is what you don't see
(27:27):
if you're just looking at idiosyncratic alpha, even though you've
stripped out all the factors. That's how the framework comes about,
and that's how it is both backward looking in terms
of explaining alpha by skill, but then also it becomes
a forward predictor by knowing what the correlation is between
past and future periods.
Speaker 3 (27:45):
I have so many questions for my next one, and
let me just add a caveat before I ask it,
which is everything I know about baseball I learned from
that one episode of The Simpsons. So that is to say,
I don't know very much at all other than don't
mean to Daryl Strawberry. But my impression, and again I
don't remember a lot about moneyball, but my impression was
(28:07):
like part of that strategy was finding players that are
underpriced by the market and capable maybe of doing one
specific thing very well, and then kind of putting them
together into a team that can work very well, like holistically,
rather than just going after the expensive players that hit
(28:27):
home runs a lot. Yeah, exactly. I guess my question
is I get the approach to evaluating individual traders, but
is part of your approach also looking at how they
like holistically work together and impact each other at all,
or because of the nature of multi strats and the
pod shops, doesn' not matter so much on that.
Speaker 4 (28:48):
That's an insightful question, And I will pick up a
topic that Gappy talked about a couple months ago. He
talked about the different cultures and how like how these
pod shops and the and the multi manager platforms can
be different and a lot of times there's a big
culture difference. And I would say that that is absolutely true,
and I have a great sort of answer to your
(29:10):
question for that. So at some shops, the philosophy is,
we are going to strip out everything a PM does
and cynically they have so many factors, and we'll pay
them on what the idiosyncratic alpha that's left is. And
they have so many factors they're stripping out that you know,
(29:30):
they're trying to get that alpha number down as small
as possible so they don't have to pay off bonuses.
And I remember joking with a PM one time at
one of those shops and he's like, yeah, I feel
like every time I go in there, they tell me like, yeah,
you had a good year, but look year out performance
is due to investing in dividend paying companies where the
CEO went to an IVY League school and we can get.
Speaker 2 (29:51):
That for free, right.
Speaker 4 (29:52):
So that so at those shops, their philosophy is it
doesn't matter because we're taking out everything. I prefer a
little different approach, and there are shops that do it
this way, which is to say, my job as a
CIO is to build a multi manager platform where some
(30:14):
of these offset so that there are different skills and
then instead of stripping out factors at each portfolio level,
more stripping out the factors. Once you put them all together,
you've got this bully of base stew and then you
take the factors out. And that is a different approach
because I think the pms feel a little bit more freedom.
(30:35):
They still have their buffers they have to stay in,
but they don't see the ETFs or the factor anti
factor things getting shoved right into their portfolio. The approach
is more higher. So you can take either approach. I
do prefer the sort of roster construction idea that you
have that you have in sports, but that really is
(30:56):
a difference in you know, I think in from culture.
Speaker 2 (31:00):
Yeah, that's super interesting. So in baseball, a general manager
(31:21):
looking for players can look at other teams, they can
look in the minor leagues, they can look at college sports.
They can start scouting at high school. Probably there's a
farm system and they call it a farm system. What
you've described so far makes sense for evaluating people in
existing seats, either at your shop or perhaps at another shop.
(31:42):
Is there a way to transfer it or to apply
some of these same ideas to people who don't have
the same because there I don't think there's the same
equivalent unless trading, you know, an Mari trader Schwab, which
actually I do think maybe is kind of a thing.
But is there a way to sort of think about,
like how you evaluate someone who just does not have
(32:03):
that much of a track record yet.
Speaker 4 (32:05):
Yes, because of the way these multi manager platforms are
formed now, they don't hire from the street anymore. I
think twenty years ago, thirty years ago, I know when
I was on the street, the researchers that were covering
the companies, they'd get plucked away, sometimes by the shops.
Sometimes traders would get plucked away. Right, that doesn't happen
(32:25):
as much anymore because what these huge firms have done,
and this also goes to their competitive advantage and their
ability to scale, is they are now training these people
right out of school, right. They have you know, universities
or academies or you know these schools essentially where they're
teaching people to be analysts or pms and again sort
(32:49):
of to a culture thing. My favorite ones are the
ones where the firms realize it used to just be
an upper out thing, right, Like you became an analyst
and then you became a PM, and if you weren't
a good analyst, you never became a good PM. And
I think that there are firms now that recognize and
analysts can be a career. You may be a great analyst,
(33:10):
but not necessarily united a capitol committee. You know, there's
a different skill set to being the PM. And they
find out some of these things in the academies and
in the universities. They're in house training schools. This is
the farm system that is coming up. Quite literally, this
is the bench and we see that and they don't
(33:31):
just get thrown in. They do tend to run paper
portfolios or portfolios that feel like they're real because they
are entering trading.
Speaker 2 (33:40):
So in their careers depend on them doing well. So
they're taking risk even if it's paper money.
Speaker 4 (33:44):
Yes, exactly, And you can run the same analytics on
these portfolios. And what I definitely have seen is some
of the newly graduated pms. These firms are good at
who they're training and those are the best pms to
find alpha signals from. Because they're portfolios, they tend to
(34:06):
be small so they can be replicated. It's it's and
this is another job of the quants too. If you
have a very senior PM, who's you know, who has
a contract that allows he or she to run a
two billion dollar biotech portfolio. There's not much left for
the quants to you know, because they're you know, they're
probably a little more thinly cat capitalized. There's not much
(34:27):
room to replicate that portfolio at another quant level in
the firm. But the new people that are coming up,
they're cheap, they're running small portfolios. But if they're skilled,
they're knowing what they're in is just as important as
a more senior PM.
Speaker 2 (34:43):
Yeah, Tracy and listeners. There's a great piece on the
Bloomberg from June nineteenth by our colleagues Nishan Kumar and E.
Liza Tetley about exactly this hedge fund talent schools are
looking for the perfect trader, and it talks about zero
point seventy two and it talks about citadel building these
sort of in how training things. So all all these
pieces are coming together, building the own farm system in
(35:05):
house to see who's going to be good one day.
Speaker 3 (35:07):
We should go to talent school. It's fun talent school,
that's to be clear. Okay, that was the joke. Yeah, Okay, Joe,
you've talked about sizing and you talked about the general
skill set that you're looking for one thing I'm still
unclear on. You alluded to it earlier, but I would
love for you to talk more about it in detail.
(35:29):
Time frame. What is the time frame by which you
are evaluating traders? And I guess how much runway do
you give people to either prove themselves correct or prove
themselves to be disastrously wrong because you know the correlation
they were betting on suddenly breaks down.
Speaker 4 (35:48):
Yeah. So again, great question, and it became a point
of frustration for me from when I first started at
Novus and building this stuff because I was very used
to sports analytics, and specifically baseball, but some some other
sports as well. I'll touch on golf. When you're evaluating
the skill of a picture, and there's three skills that
(36:08):
a picture has that are not dependent on anything else,
not dependent on it's teammates, who's batting it, et cetera.
It's right, not dependent on fielding. Right would be the
strikeout rate of a picture, the walk rate of a picture,
and the ground ball rate of a picture. These are
things that the picture controls, and what happens is after
about fifty plate appearances, you get the strikeout rate for
(36:30):
a picture. That is predictive of you know, it's you know,
from a maths standpoint, the correlation is above zero point seven,
so squared it's above point five. Right, the past explains
more of the future than factors that we haven't identified.
But with PMS, there's much more noise in their result
and it takes a lot longer to find a meaningful correlation.
(36:52):
So although I can do work for like, I can
and I do this for a single day, right, So
every day I generate a report, and I do this
for clients now showing their PMS and exactly what they're
readings of all these skills were each day. And of
course for one day it's just trivia. It's no more
than trivia. But what it is doing is building a
data set. And at the point that you get to
(37:15):
six months, which is about one hundred and twenty five
days trading days. Bigger picture, the full model takes five
hundred the past five hundred results, and that's when you
start getting very different and but more persistent correlations between
all these skills. Right. But what I have found is
(37:36):
that even after one hundred and fifty days, if you
take for the other year and a half, a mean
reversion assumption, and then just every time a new day
comes in, you drop off an assumption you'd get you
have a pretty robust skill reading that starts to mean
something after six months, and after two years, that's when
(37:56):
it really has, you know, has some great predictive power
for the next quarter, and so you're constantly dropping off. Now,
why only two years? I talk about this in the book.
I don't have a great answer for that.
Speaker 3 (38:07):
I suppose you have to start somewhere around.
Speaker 4 (38:09):
Yeah, well, here's what I knew. Two years was better
than three years, which in one sense, why would that be?
And I have talked to different quants about that, and
they have approached this from a much different perspective, and
they also have come to somewhat of a two year conclusion.
The reason seems to be regimes within the stock market,
(38:29):
just something about where you are skilled. You know, I
haven't been able to identify it. And I also know
that we could do a you know, we could we
could run the numbers and find out that, oh, you know,
it's not two years. It's the most predictive thing for
the last three months would have been two years and
(38:49):
forty three days. When you try to get that precise
that's not going to be what the perfect So two
years does seem to work because you're constantly rolling off
whatever happened two years to go, and so there's some
regime change that seems to work. But that is an
unanswered to a question I have too.
Speaker 2 (39:07):
So one thing about baseball is that every GM in
baseball has basically perfect visibility into the performance of every
player on every other team because it's all out on
the field and it's all measured, and we all have
the same information. You know, one of the.
Speaker 3 (39:21):
Most measure heart Jack.
Speaker 2 (39:22):
Yeah, right right, you can't measure Harvey get we all
could see players on base percentage and ops and slop
and all of this stuff, right. You know some of
the most popular alerts that always read spike on the
terminal or it's like consumer Discretionary Manager, Palacity goes to
citadel whatever. People love, people eat that stuff up. Just
from an industry perspective, sitting aside, whether you want to
(39:46):
use a traditional sharp ratio perspective or rose or lum
or whatever, how much visibility does one shop have into
the performance of a pod at another shop that can
then be ported over, or how much insight can you
we have if maybe there's an undervalued player somewhere else,
if you want to bring them over and give them
more capital.
Speaker 4 (40:04):
Than they're getting, extremely limited in terms of it. So well,
and I'll tell you who can solve that problem. What
you will have is, of course, if a team is
marketing itself or being recruited by another firm, they bring
over their returns, right, But they don't bring over They
might talk about portfolio construction, but I'm pretty sure they
shouldn't and probably don't bring over their two years right,
(40:26):
what the portfolio, what their holdings have been for the
last two years. So you don't get that type of visibility.
But let me tell you who can. Okay, And this
is I think one of the most important constituents in
our industry because I think they have the purest motive,
and that is the allocators, right allocators. I'm talking about
(40:48):
the huge, multi billion dollar entities which provide the blood
that keeps the heart pumping right at all these hegge file. Sure,
we know that Ken Griffin has a tremendous amount of
the aum is his money, and and you know we
hear that about some other people too, But in general,
(41:09):
these firms it's outside money which which keep these firms afloat.
But the allocators, many of them though, and what I'm
talking about here are foundations, university endowments, sovereign wealth funds, right,
pension plans, and they have a very pure motive. Right,
they are trying to get returns for the retirees or
you know, reduced tuition for future students, et cetera, or
(41:31):
you know, in the case of Norway, the citizens of
the country.
Speaker 2 (41:34):
Right.
Speaker 4 (41:35):
So they are a treasured investor, right if you run
a hedge fund. So when they are doing manager selection,
they have the ability to go to hedge funds. Now,
maybe not Citadel and Millennium, but to all these non
multi manager platforms. They have the ability to go to
them and say, hey, if you want us to really
(41:57):
evaluate you, we need to see podil. Yeah, we need
to see we need position level transparency for the last
two years. Hey if you want, if you don't want
to give us yesterday, start a quarterback so that it's
on a lag. But now they have the leverage to
get those returns, especially if you're talking about emerging managers, right,
young managers are trying to that, you know, to build
a hege Fund, and I don't feel they use that leverage.
(42:21):
And this is to me is like, well, Joe, you're
you know you you talk about this framework and it's
it's applicable to multi manager platforms, and you know an
endowment isn't a leveraged portfolio, so how could they use it? Well,
this is how they could use it, because they can
get that, Joe, and they do ask those questions about
the Bentley and the.
Speaker 2 (42:40):
In fact, can I give you an example? Can I?
Speaker 4 (42:42):
This is I really like this. I've never worked with them.
I should say that I have worked with their brethren,
and I've worked with the endowments that they would measure
themselves against. But the MIT Endowment Matimco is the name
of the uh, the name of the entity. They have
something between twenty and thirty billion dollars under management. So
(43:04):
we know that a portion of that is dedicated to
public equities. And we know because and I won't mention
his name, so I'm not trying to call him out,
but we know that one of those gentlemen that looks
for equity managers is a presence on fin Twitt. He's
actually a great follow, very earnest, and so he'll talk
(43:26):
about things, and sometimes he'll post job postings.
Speaker 2 (43:29):
Right, and what you will find is.
Speaker 4 (43:32):
Everybody who works in that division or and in fact,
you can even see this publicly. I know this. Yale's
Management company has the resumes of every person who's in
that division, and they're all the same. Here's what they
will say. They will say things like, you know, was
president of the investment club at the University of Virginia, right,
and they'd been investing in stock since I had a
(43:55):
paper route.
Speaker 1 (43:55):
Right.
Speaker 4 (43:55):
They're always have this, right. So when they go to
do managers selection, and I've been on that side too
as a as a marketer, they will sit down with
the PM and they'll ask they'll go over each position
in the portfolio and make no mistake about it, they're
passing judgment, right, because if they're not frustrated or want
to be pms, this is how they think about the market.
All right, So I'm gonna put full stop there. Now,
(44:18):
let's go to a The general manager of the Philadelphia
seventy six ers is a gentleman named Daryl Mory.
Speaker 2 (44:23):
Oh yeah, I like Daryl. Darryl.
Speaker 4 (44:26):
Right, he was with Houston and in fact, while he
was at Houston. He really brought moneyball to the NBA.
Mark Cuban was probably maybe the second, but Darryl Moury
right down to the fact that Michael Lewis did a
piece on him in the Sunday New York Times maybe
twenty years ago. So Daryl Morey is the GM of
the seventy six ers, and he has juniors too, right,
(44:46):
And when they're doing their equivalent of manager selection, whether
it'll be drafting a player or looking at free agents,
can you imagine how absurd it would be for Darryl
and the analysts to go down and shoot free throws
with perspective player right, and to judge the player based
on that. But I guarantee you at the endowments they
(45:07):
go back and say, can you believe you know that
managed your short netflix right? Like, so, now, why did
I pick those two? And the example is this, before
Daryl got into basketball, he is a proud graduate of
MIT Sloan. He got his MBA at Sloan School of
Management and he started along with a woman named Jessica Keelman,
(45:29):
he started the Sloan Sports Conference, which started as Bill
Simmons when he was at Grantline described at as Dorcapalooza. Right,
it was just people, kids, guys, and it was almost
all guys back then talking about sports analytics. And it
has morphed into a massive event and it's a job
(45:50):
fare where all these sports teams from all different leagues
are looking for talent, right, and they're essentially looking for
performance analytics people. Right.
Speaker 2 (46:00):
Voros McCracken was the one from what's.
Speaker 4 (46:02):
A exactly exactly, So this is think now, now, look
across the campus at the MIT Sloan Endowment. What they're
actually trying to find is performance analytics. Do you think
there might be anybody right across the campus who may
have never invested in stocks but gets the profit motive?
(46:23):
They would take my work and probably take it three
steps more. But I don't think there's an endowment out
there that things like that. Like, I'm sure they've never
walked across the campus, and even I'm sure Darryl has
never thought to invite them over to the you know,
to hey whyt to interview some of some of our people.
So that's again sort of how I look at, like, Hey,
this is how some of this work. How you somebody
(46:45):
who doesn't have the data can get it at the
allocator level.
Speaker 3 (46:49):
We've been talking very much about, you know, performance evaluation
and metrics from a sort of managerial level. If I
am a trader or quant, you know, a sort of
junior medium level quant, I guess at one of these
multi strat hedge funds, how am I viewing the performance
of others and competition? Is it the case that I'm
(47:11):
trying to move into a particular sector that maybe has
more of an opportunity set in terms of dispersion, where
maybe there's more volatility or more relative value opportunities or
something like that. How am I like viewing my competition?
Speaker 4 (47:29):
It's a good question. Even the work that I do well,
I think there's definitely comparison, right. You douce again to
culture some of the pods, and I think I think
Yapy mentioned that some of the pods there is a
there's a sharing of information, right, and at some shops
there's not. This is a case where I actually prefer
(47:51):
the not sharing of information right, because I would rather
I think the quants would rather know that maybe two
different pms came to the same conclusion independently, as opposed
to they both went to the same idea. Dinner and
then both decided to buy the stock. There's more of
a signal in somebody coming to it independently, so I
(48:13):
believe they're aware of what the returns are of their
other pms. In addition, and I don't know if this
was in that article Joe you just referenced, but these
firms all have coaching teams too, and I certainly found
that the older pms that you know had been in
the business since the nineties, they're setting their ways right.
(48:34):
They don't want a quant to come in with a
laptop and start telling them that's spin rate, right, the
spin rate of their pictures. But the younger people, I
think there's more of a hey, if there's data that
you can give me to help me get better, I
think maybe in some ways they might be looking for that.
Speaker 2 (48:52):
Joe Peter, this is super fun. Thank you so much
for coming on the podcast again.
Speaker 4 (48:57):
Oh it's always great to be here. And I'll see
you in seven years.
Speaker 2 (49:00):
Yeah, exactly when whatever your next job is from.
Speaker 3 (49:02):
The Sun, Thanks so much.
Speaker 4 (49:04):
Joke.
Speaker 3 (49:04):
Yeah, even though there were baseball references, I enjoyed it.
Speaker 2 (49:07):
Yeah, Tracy, that was a really fun conversation.
Speaker 3 (49:23):
I love hearing stories when people get jobs off the
back of Authoughts appearances.
Speaker 2 (49:26):
That's nothing sort of flatters are egos and sense of self.
Speaker 3 (49:31):
Well know, I also like it when people say they're
listening to A Thoughts episodes while going to the gym.
Oh yeah, because I hate going to the gym. I
hate running and things like that. But it makes me
feel nice that like people are listening to us or
to offset something that's kind of like a chore.
Speaker 2 (49:48):
No, But beyond all that, it was very fun. I
feel like we could just talk about these businesses forever.
It seems so rich, you know, like we still have
to do something on like compensation struggle. Yeah, but also
like just the sort of like fundamental point that everyone knows,
which is like manager identification is really difficult because and
(50:10):
you know, first of all, there's all these questions about, well,
it's beating the market really possible because of efficient markets
and stuff. And then you could identify someone to all
this person beat the market seven years in a row,
but have a pool of one thousand managers. There's gonna
be a lot of people who beat the market seven
years in a row, and so it seems like a
very interesting problem to solve.
Speaker 3 (50:29):
It's kind of funny that you're trying to like select
traders on a factor neutral basis, who are themselves able
to be factor neutral in some respects, Like you're kind
of you're trying to separate them from like these circumstances
that they are operating in, or trying to wait them
against the value of the opportunity that they are currently facing.
(50:51):
Right that dispersion that Joe was mentioning, that's kind of funny.
I've thought about it. Not to go all media and
naval games, so sure, but you know, it's sort of
similar to journalist beats in some respects, where you can
get really lucky and be on a really interesting beat
where there's tons happening, and suddenly you know, all your
(51:12):
stuff is getting read and you're getting all these major scoops,
and then maybe two years later, to go back to
that timeframe point, it's sort of faded into the distance
and there's not as much to write about, And how
do you judge the talent of a particular journalist or
a trader from their particular set of circumstances.
Speaker 2 (51:30):
That's a great example. I remember, you know, like when
I was in a business insider years ago. It's like the
reporters who covered Apple on days of like iPhone announcement,
they got we were like measured on traffic back then
they got all the traffic.
Speaker 4 (51:43):
You know.
Speaker 2 (51:43):
It's like, oh, this isn't fair, Like I'm talking about
like the Bank of England decision. This is nonsense. Okay.
Speaker 3 (51:48):
I just read a really good analysis of like US payrolls,
and people only want to read about the next iPhone.
Speaker 2 (51:53):
I explained, I explained Mario Draggy's new OMT thing really
well and like ten people ready, But no, like this
is like it's all like version of the same problem.
By the way, my hedge fund media metaphor that I
use in my head is like alpha decay. So it's
like the first person who ever came up with like
here's what you need to know or the answer will
(52:14):
shock you, like probably like did crazy. Well, but by
the time that was you, wasn't it Yeah, And then
by the million person who did like the answer will
shock you, it stopped working. So it's like there's the
same thing of like alpha decay, where it's like you
can be the first on a strategy and then everyone
discovers it and then the excess returns from.
Speaker 3 (52:30):
That move on the crowding in effect. Yeah, no, I
did think actually that timeframe point was really interesting and
the fact that Joe kind of I guess gravitated towards
two years or five hundred trading days. But then he
was talking about how others seem to have sort of
alighted on that same time period. Yeah, I wonder why that.
I mean, I get that you have to at some point,
(52:51):
you just have to choose, like age horizon, but it is. Yeah,
it's an interesting one.
Speaker 2 (52:56):
Very interesting stuff. Plenty more to come on this topic.
Speaker 3 (52:59):
All right, shall we leave there.
Speaker 2 (53:00):
Let's leave it there.
Speaker 3 (53:01):
This has been another episode of the Oudlots podcast. I'm
Tracy Alloway. You can follow me at Tracy Alloway.
Speaker 2 (53:07):
And I'm Joe Wisenthal. You can follow me at the Stalwart.
Follow our guest Joe Peta, He's at Magic Rat SF
and check out his book Moneyball for the money Set.
Follow our producers Carmen Rodriguez at Carman Erman dash Ol
Bennett at Dashbot, and Kilbrooks at cal Brooks. Thank you
to our producer Moses One. More odd Lags content go
to Bloomberg dot com slash odd Lots, where you have transcripts,
(53:29):
a blog, and a newsletter, and you can chat about
all of these topics twenty four to seven in our discord,
discord dot gg, slash odlogs.
Speaker 3 (53:38):
And if you enjoy odd Lots, then please leave us
a positive review on your favorite podcast platform. We will
accept positive reviews in lieu of finders' fees. And remember,
if you are a Bloomberg subscriber, you can listen to
all of our episodes absolutely ad free. All you need
to do is connect your Bloomberg account with Apple Podcasts.
(53:59):
In order to do that, just find the Bloomberg channel
on Apple Podcasts and follow the instructions there. Thanks for listening.
Hosts And Creators
Joe Weisenthal
Tracy Alloway
Popular Podcasts
On Purpose with Jay Shetty
I’m Jay Shetty host of On Purpose the worlds #1 Mental Health podcast and I’m so grateful you found us. I started this podcast 5 years ago to invite you into conversations and workshops that are designed to help make you happier, healthier and more healed. I believe that when you (yes you) feel seen, heard and understood you’re able to deal with relationship struggles, work challenges and life’s ups and downs with more ease and grace. I interview experts, celebrities, thought leaders and athletes so that we can grow our mindset, build better habits and uncover a side of them we’ve never seen before. New episodes every Monday and Friday. Your support means the world to me and I don’t take it for granted — click the follow button and leave a review to help us spread the love with On Purpose. I can’t wait for you to listen to your first or 500th episode!
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com