0:03

Bloomberg Audio Studios, Podcasts,

0:06

Radio News.

0:20

Hello and welcome to another episode of

0:22

the Odd Thoughts podcast. I'm Tracy Alloway.

0:25

And I'm Joe Wisenthal.

0:26

Joe, I know we did one episode on

0:29

pod shops, Yeah, on multi

0:32

strategy hedge funds, but

0:34

it was primarily focused on their impact

0:37

on the market, and I have to say I still

0:39

came away from that conversation sort

0:41

of wondering if I worked

0:44

at a pod shop, what is it

0:46

exactly that I would be doing all day?

0:49

I would love to know the exact same

0:51

thing. I mean, like, I guess I have this very

0:53

vague sense of sort of they have a bunch

0:55

of people all focused on

0:57

their specific areas in the sort of ever

1:00

out, and they net out a bunch of stuff and it's

1:02

capital efficient and you

1:04

know, it's like market neutral

1:06

and theory and et cetera. But beyond

1:09

that, like, I still don't really

1:11

like understand. The only thing I know is like they've

1:13

done really well and many people are

1:15

launching more of them.

1:16

Yes, they seem to be all the rage. They

1:18

seem to be where everyone kind of

1:20

wants to go in the quantitative

1:22

finance space. At least everyone's

1:25

sort of aiming for these big

1:27

names, you know, places like Citadel

1:30

Millennium. Maybe yeah, But my

1:32

question is like, why was

1:34

it just that they're minting money they're expected

1:36

to continue minting money in the future, or

1:39

is there something that's like fundamentally

1:41

intriguing and attractive about

1:44

working in that space that means lots

1:46

of people want to get in.

1:47

I mean, I think that could be two ways of

1:49

saying the same thing. If they're minting money,

1:51

then that probably is fundamentally attractive

1:53

to people in that space. But I

1:55

do think like backing up the questions, like what

1:58

we know is that many and

2:01

including apparently even like B tier

2:03

C tier funds have done like very well. So

2:05

I'm just like curious like how and why?

2:08

And then yeah to the question of like what does it

2:10

take to succeed in them or who

2:12

is the type of person who can

2:14

succeed in this environment?

2:16

All right, well, I'm glad you put it that way, because today

2:18

we're going to be speaking with someone who has

2:20

done exactly that succeeded in this particular

2:23

environment. We have the perfect guest. We're going

2:25

to be speaking with Giuseppe Palielogo

2:27

aka Gappy. He describes himself

2:30

as a constant gardener, someone

2:32

who's on gardening leave quite a lot. He

2:34

is also the author of Advanced Portfolio

2:37

Management, a quant's guide for fundamental

2:39

investors, and I have to say it is one

2:42

of the funniest books that I've read

2:44

in quant finance. I can't say it's the

2:46

funniest because I did read My Life as

2:48

a Quant from Emmanuel Derman, but it's

2:50

definitely up there. And Joe, I know you

2:53

enjoyed it too.

2:54

I did you know, I like skipped over all

2:56

the numbers.

2:56

And equations and you just looked at the jokes.

2:59

And Greek letters. But it's very breezily

3:01

written for what it is, and I did.

3:03

Actually, I think maybe I learned

3:05

a little bit even in my sort

3:07

of basic reading of it. Extremely well

3:09

written. I'm extremely excited about

3:12

this conversation. You know, you mentioned that our guest

3:14

is the king of gardening leave. If you

3:16

look in his LinkedIn, it really is many

3:19

different roles. Well.

3:20

I also have to say he is the only person I know

3:22

who has both an alpha and a

3:25

beta tattoo on his shoulder.

3:27

Oh wow.

3:28

You know, some people do get the alpha symbol,

3:30

but he has both. So you know, a well

3:33

balanced portfolio of tattoos

3:35

all around. Yeah, so gappy.

3:38

Thank you so much for coming on all thoughts.

3:40

Hi Tracy, Hi Joe.

3:42

So maybe to begin with, I'm going to

3:44

let you explain your

3:46

previous job history because there is quite a

3:48

lot. What is it that you've been doing in this industry.

3:52

I'm not sure, I'm not sure, Okay,

3:54

good question. Well, I got into this industry

3:57

almost accidentally. I was for a

3:59

few years a researcher

4:01

in the math department at IBM Research,

4:04

and then I got a little bit bored.

4:06

So the only place that you can the

4:08

only industry you can work in New York other

4:10

than you know, IBM or tech, is finance.

4:12

So I got into finance almost accidentally.

4:16

And then again, there is no major

4:19

plan to you know, to to my career

4:21

choices. When I was getting

4:23

bored for some reason, somebody called

4:26

me and offered me a more interesting job.

4:28

And so I have been working mostly

4:31

on the so called buyside of the industry,

4:33

so the part of the industry that invests,

4:35

actively invests and takes risks. So I've

4:37

worked for Citadel twice

4:40

for a small edge fund as a portfolio manager,

4:43

and then Millennium and Hudson River Trading,

4:45

and I've kind of taken turns between

4:47

doing quantitative research and risk

4:50

management. So most recently

4:52

I was at Hudson River Trading until the beginning

4:54

of November.

4:55

I think when people think about

4:57

like multi strategy hedge fund or

5:00

shop or whatever, maybe sort of Millennium

5:03

is the first one that would come to mind

5:05

for people. If someone asks you, how

5:07

does Millennium make money? And they seem to have made

5:09

a lot of money over the years, what's the answer?

5:12

Okay, I hope without saying anything

5:14

that is proprietary, but

5:17

I think.

5:18

That like the business model of Millennium, Yeah.

5:20

I think that what Millennium has

5:23

excelled at has been the ability

5:25

to scale up, so to adapt

5:28

its existing platform to accommodate

5:32

new strategies and new portfolio

5:34

managers and so

5:36

sometimes actually in some of

5:38

their marketing material they called it something

5:41

like an investment operating system. So

5:43

it's a system that is a firm that

5:46

is willing to absorb

5:48

some relatively new strategy

5:51

and create an environment for that strategy

5:53

to succeed. And so because of

5:55

that, I think they might be having

5:58

right now the highest number of individual pods

6:00

maybe close to three hundred and

6:03

hovering around sixty billion dollars

6:06

of AUM of assets under management. But

6:08

I would say, what is their superpower

6:11

is really their ability to scale in

6:14

number of pods.

6:16

So you mentioned creating an

6:18

environment for success there,

6:20

what does that look like at

6:23

an organization like that? What are the

6:25

sort of like conduits that allow

6:27

trades in that particular organization

6:30

to be successful?

6:33

So I would give a sort of an idiosyncratic

6:35

maybe a story around, please the

6:37

rationale for success of platforms.

6:41

So I see platforms a little bit like managing

6:44

an arbitrage or some kind of gap

6:47

between the single platform, the single

6:49

manager, or the small hedge funds and

6:52

the fund of funds. So if you're a fund

6:54

of funds, you do have the scale,

6:57

but you do not have the

6:59

ability to observe from

7:01

a close distance the performance of your

7:04

vehicles for investment. And

7:07

let's say that they don't perform well, you

7:09

have to wait a year in order to take your

7:11

money back. In the case of a hedge

7:13

fund platform, you could actually

7:15

not only observe the performance

7:18

of pms or volume managers their

7:20

skill from a very close distance, but

7:22

you can also help them perform better. So

7:25

you can centralize some of the functions

7:27

that make them better capital

7:30

access, corporate access, risk

7:33

management. If they perform well,

7:35

to give them more capital. If they don't perform

7:37

well, to take capital away from them

7:40

or let them go. And at the

7:42

same time you also

7:44

solve two for two other problems. So one

7:46

is there is a risk transfer

7:48

happening because a platform

7:50

almost by design otherwise is not really a

7:52

platform, has a pass

7:55

through fee structure that's fundamental

7:58

for the existence of a platform that makes really

8:00

a platform what it is instead of a just

8:02

multi manager hedge fund like the show.

8:05

So this means that a portfolio

8:07

manager is not paid with

8:09

the incentive fee that the hedge fund

8:11

as a whole receives from the limited partners,

8:14

but instead the portfolio managers

8:16

are paid a percentage of their p and L.

8:19

This payment is passed

8:22

through directly to the limited

8:24

partners to the investors,

8:26

and this basically transfers the risk

8:29

directly basically from

8:31

the PM into the limited partner,

8:33

And so this makes the system more robust

8:36

in a sense, right, And combine

8:38

this with the diversification across

8:41

investment styles and the number

8:43

of pms, and now you start having a

8:45

mote around a platform that

8:47

makes it successful.

9:04

If a entity has three hundred pods

9:07

and everyone's doing their own thing, et cetera,

9:10

why doesn't the return just become

9:12

the market return like it seems like because

9:14

there's a right Like, one intuition

9:17

could be that this model wouldn't

9:19

scale. I mean, I know it does, but one intuition could

9:21

be that this model wouldn't scale that

9:23

the more you add, you overdiversify

9:26

and then you just end up with like whatever,

9:28

like you know, like buy the VTI ETF

9:30

or something like that. Why doesn't it work out that.

9:32

A simplest explanation for this is actually

9:35

just to look at what a retail

9:37

investor right would hold in their

9:39

portfolio. So let's say that they

9:41

are, you know, long Apple and IBM.

9:44

Okay, they have a little bit of an imperfect

9:47

version of the market, right, But what

9:49

makes their skill is

9:52

how different are the weights

9:54

of their Apple and IBM holdings compared

9:56

to the market. Okay, So you can

9:58

decompose your performance in

10:00

your personal account into the sum

10:03

of let's say the market and your idiosyncratic

10:06

bets into these stocks. Now,

10:08

what the hedge funds do is

10:11

they do the same, but they completely eliminate

10:14

as much as they can their exposure

10:17

or their investment in the market, So they run

10:19

purely market neutral and factor

10:21

neutral portfolios. So

10:23

there is diversification, but these indiosyncratic

10:26

bets don't get diversified away

10:28

into a big market, but they actually

10:30

become essentially a bunch of independent

10:33

bets that by the law of large

10:35

numbers, they tend to have better

10:37

and better risk adjusted profiles.

10:39

So I still see some platform

10:42

heads describe like the overall

10:44

tilt as market neutral. So

10:47

what do they mean by that? Exactly?

10:49

I mean they typically run a wide

10:51

range of strategies, so let's focus

10:53

because it's more relatable. Let's focus

10:55

on discretion ory long short equities

10:57

and systematic equities because everybody knows star.

11:00

I've love that you think systematic equities is

11:02

relatable.

11:03

Yeah, yeah, I mean relatively to I

11:05

don't know, treasury basis or sell

11:07

involved. So they mean that

11:10

typically they do have a so called factor

11:12

model, and a factor model

11:14

is a little bit like having a market model

11:16

on steroids, So you have a

11:19

market term, so you can see your

11:21

portfolio as having

11:23

exposure to the market, so behaving a little bit

11:26

like a market. And then it's

11:28

also behaving a little bit like a portfolio that

11:30

has momentum okay, and then it

11:33

also has maybe a tilt in

11:35

terms of value. The platforms

11:37

tend to run portfolios that have

11:40

no market exposure whatsoever, and

11:42

then they also tend to have controlled

11:44

exposure in these more exotic factors.

11:47

How do they know that? I mean, so there's

11:49

someone up there at the center, there's

11:52

all that three hundred pods. The data gets

11:54

probably aggregated and sliced in various ways,

11:56

but what is the job or how

11:58

do they actually sure that

12:00

on that their portfolio managers

12:03

don't have that market beta.

12:05

Yeah, they

12:07

typically have at the very minimum.

12:09

They will buy some commercial

12:12

factor model, which is a model of the

12:14

market, like of your investment

12:16

universe, how the how

12:18

a stock behaves. How can you decompose

12:21

the performance of the stock in this various

12:23

systematic or let's call them pervasive

12:26

market wide factors and instead

12:28

idiosyncratics. So you buy them off the

12:30

shelves. I mean, they're really expensive and

12:32

they do a job. And so once

12:35

you've bought them, you create some

12:37

kind of user friendly interface so

12:40

that a portfolio manager can always see

12:43

how the portfolio

12:45

looks like at any point in time. It's a little bit like

12:47

having an X ray of you know,

12:49

your body in real time. You know, you

12:51

can see, oh, well, my portfolio is is

12:53

a little bit short, the market is a little bit long momentum,

12:56

maybe there is some crowding, exposure, whatever,

12:58

And so this is in the hands of the portfolio

13:01

manager. And then there is another

13:03

layer on top of that, which is very

13:05

important risk management, which ensures

13:08

that pms are behaving well, that they're

13:11

not going out of scope. You

13:13

know, they're not buying microstocks

13:15

or you know, investing in crazy stuff just.

13:17

Going or just going along in video.

13:19

Or long in Vidia. Yeah, if their idea

13:21

is going long in Vidia, probably that's not an

13:24

ideal portfolio manager. Yeah.

13:27

So the other thing I've been wondering is how

13:29

much visibility are there between

13:32

the different pods within

13:35

one shop. Yeah, And I

13:37

mean that like, I assume there's a centralized

13:41

risk management system of some sort

13:43

that is like netting out positions

13:45

and trying to make use of capital most efficient,

13:47

and that's where a lot of the edge comes from.

13:50

But also, if you're just a trader

13:53

pursuing your own strategy, do you

13:55

know what the guy next to you is

13:58

doing? Do you have that kind of vision ability

14:00

or is the idea to keep everyone sort of

14:02

intellectually separated so that

14:04

they're not influenced by each other.

14:06

Right, that's a good question. So there

14:08

is no really black and

14:11

white answer to this, because historically

14:14

there was a time when platforms

14:16

had more visibility and more collaboration

14:19

among pods or at least pods in the same

14:21

sector, for example. But I would say that

14:23

the historical trend has been

14:26

more and more to give them the tools

14:28

to succeed, but not give them the

14:31

ability to see into each

14:33

other's portfolios for example. And

14:35

the rationale for this is you probably

14:37

prefer having independent bets to having

14:40

maybe corredated bets that could be

14:43

like maybe a little bit more informed. So that's

14:45

the trade off. Let's if we talk,

14:47

maybe we can come up with slightly better ideas.

14:49

Sure, but yeah, I think that the trend

14:51

is more and more towards you are not seeing

14:54

what I am, what I'm having, what I'm holding.

14:57

Talk to us more about the risk management

14:59

component, and again I don't know very much.

15:01

I understand that you know, stop

15:04

losses are very tight and you don't get

15:06

a long leash to lose money, and if you're not doing well,

15:08

your capitals reduced. If you're doing well,

15:11

I guess you get more, and if you do more, you

15:13

get more, et cetera. But from how would

15:15

you describe the sort of the

15:17

essence of risk management at

15:19

the hedge fund level.

15:21

So there are maybe two or

15:23

three core functions that can be described

15:25

in a qualitative way, but you know,

15:27

I think pretty comprehensively. And

15:29

then there there is something that is a little bit more esoteric

15:32

or like domain specific. So let's

15:34

talk about the general principles. Okay, So

15:36

you mentioned stop losses, so this

15:38

is very important. You know, there

15:41

are always stop losses, the ones that you know

15:43

you have and the ones you don't know you have, but

15:46

everybody has stop losses in life. Okay,

15:48

So those are very important because

15:50

you could imagine that a PM is a little bit like somebody

15:53

who's holding a coll option and you

15:56

you know, the PM who's losing

15:59

money has kind of a incentive to go for

16:01

broke maybe sometimes. But the stop laws

16:03

is effectively at

16:06

a sort of a primitive tail insurance

16:08

TAE risk management tool on the left

16:11

tail of a PM. So that's very important. The

16:13

second principle is sort of

16:15

self enforcing, is true diversification.

16:18

So this is where you want to have some kind of risk

16:20

model that tells you what are the

16:22

hidden bets that kind of overlap

16:24

and maybe compound at

16:27

the aggregate level, so that if everybody

16:29

takes a little bit of a factor exposure

16:31

in the same direction and then you sum this

16:33

across three hundred pms, it

16:36

becomes a big factor exposure. So

16:38

a risk management organization needs

16:40

to get that right. The

16:42

third thing is making sure that people

16:44

stay in scope. Okay, so seems

16:47

trivial, but actually that requires a lot

16:49

of domain expertise. So understanding

16:51

the trades, what can go wrong

16:53

from an operational standpoint macrostructure

16:56

standpoint.

16:56

Is this factor drift risk

16:59

as well?

16:59

Or said that scope is more like factor

17:02

drift or in general strategy drift,

17:04

not only factor but whereas being

17:06

in scope is more of a

17:09

pure strategy drift or just taking

17:11

risks that a portfolio

17:13

manager would be possibly

17:15

aware of, but that maybe the head

17:17

of the hedge fund, because it's not an expert in

17:19

that area, is not so aware of. So

17:21

the risk manager has to know

17:25

very well what's going on and

17:28

an alert. Talk to the PM, talk

17:30

to the business head

17:32

and.

17:33

Can you give us concrete examples

17:36

from your experience of the

17:38

kind of things that would set off alarm

17:41

belts. So is there like, I

17:43

guess you don't have to give us specific examples,

17:46

but you know the kind of thing, the types

17:48

of examples, Yeah, the types of examples

17:51

that would catch your eye in a risk

17:53

management position.

17:54

So we covered a little bit the easy stuff,

17:57

right, So the easy stuff is people taking

17:59

too much risk. First of all, it simple,

18:01

but you know, we think in terms of dollar volatility.

18:04

Dollar volatility is a little bit like how much you can

18:06

make or lose in one

18:08

year for it.

18:09

So like value at risk, those kind of kind

18:11

of value risk.

18:11

Yes, I mean most people think

18:13

in terms of all value risk too.

18:16

Okay, yeah, choose your risk metric

18:18

you want to stay within that. Then factor exposures.

18:21

Okay, that's also easy concentration.

18:23

So if you take a mega bet in Nvidia, it

18:25

has to surface. Okay.

18:28

So these are relatively simple. There

18:30

are things that are a little bit more complicated,

18:32

like, for example, you take some

18:35

true arbitrage positions

18:38

where you think that something is running

18:40

cheap versus rich in say

18:42

bond versus futures, or you do

18:44

some kind of funding arbitrage trade

18:47

where different agents in the investing

18:49

world have different funding rates

18:51

for their assets, and those

18:54

can break, like in a dislocation, that can

18:56

break. And so the way that typically you manage

18:59

these things a little it's a little bit like in merger arm.

19:01

You give it a max size and

19:04

you want to make sure that this is correct, that

19:06

this size is correct, and it's monitored. So

19:09

this is stuff that can go wrong.

19:11

Two managers like,

19:14

how much do they I mean, I'm sure

19:16

there's sort of I don't know if it's accidental style drift

19:18

or you know, drift is sort of a neutral

19:20

term. How much

19:23

does the risk manager have to watch

19:25

out for I guess intentional drift or

19:27

this is a working I know this is not quite my mandate,

19:30

this is not quite what I was made to trade. But I

19:32

could sort of justify it this way, or I

19:34

just see all these lines up over here going up,

19:36

I need to how much of a risk management

19:38

concern is that?

19:39

Okay, I think that in general the

19:41

principle should be trust, but verify. I

19:43

would say that the vast

19:45

majority of portfolio

19:47

managers are very responsible, and

19:50

because they're in that role, they have been educated

19:53

to control their risks, to understand

19:55

them with occasional screw

19:58

ups, and so that's why you need very fie

20:00

got it?

20:02

Okay.

20:02

On the opposite side of screw

20:05

ups, I'm curious how

20:07

capital gets kind of doled

20:10

out. And if I'm running

20:12

a massively profitable, successful

20:14

training strategy, do I automatically

20:17

start giving start being given

20:20

more money to you know, play

20:22

around with or is there some amount of

20:24

discipline here where you don't want

20:26

people to be bumping up against

20:29

you know, sizing positions or additional

20:32

trading costs and things like that. Imagine

20:34

I am the most popular trader, the

20:36

most successful popular also

20:39

popular, I'm both

20:41

the most popular trader and most

20:43

successful trader at Citadel.

20:45

What is the process for traces getting more money

20:47

to trade?

20:47

How do I get more popular and successful?

20:50

Probably not popular?

20:51

Okay, assume that you're popular and successful,

20:53

okay, So do you get more capital? You

20:55

do get more capital up to a point. So

20:58

there are a couple of factors. The first one is there

21:01

is like a natural limit where

21:03

somebody can be too successful. And without

21:06

giving examples, but there are large

21:08

funds whose daily p and l sometimes

21:11

at points are driven is driven by

21:14

a single strategy Okay, and

21:16

maybe that's justified, right, But there is a

21:18

point where there could be just too much because

21:20

the concentration across strategies.

21:23

Or think of pods as stocks, right,

21:25

you don't want to have ninety percent of your

21:28

savings in Nvidia, So okay, so that's number

21:30

one. So there is some kind of basic heuristics.

21:33

Then there is just a natural limit to

21:35

growth for strategies, like there

21:38

is a trade off

21:40

because your market impact

21:43

is very high and so, or

21:45

there is just a hard size for your

21:47

strategy, so you cannot scale high

21:49

frequency, you cannot scale to

21:51

infinity even index rebalancing. Or

21:53

if you're a consumer PM,

21:56

your costs increase faster than

21:58

the size of your portfolio,

22:01

so your P and L in the absence

22:03

of costs goes more or less

22:05

lilenar linearly, but your

22:07

costs grow faster than linearly. So there is

22:09

a point where you just don't

22:12

want to grow all Right.

22:13

On the flip side, Let's say Tracy

22:16

comes in and she is a PM

22:18

and she has her pod. How

22:22

long is she likely to last and

22:25

what would cause her what would be the

22:27

threshold at which she gets fired.

22:31

I don't have the statistics on the average

22:34

tenure of a PM, Okay, if

22:36

I had them, probably I shouldn't say. Well,

22:39

and also depends a lot on the place.

22:41

Okay, so how long I

22:45

would say that it's like everything in

22:47

life, right, So like ninety percent of everything

22:49

is of poor quality, I'm sorry to say, but the same

22:52

applies to pms. But this is another beautiful

22:54

aspect of platforms, by the way. Okay, so

22:56

let me take a quick ditchure about this again, because

22:58

like a beautiful and under appreciated aspect

23:01

of platforms is that

23:04

they act like sieves. So you

23:07

go through basically every

23:10

possible PM on the market, and

23:14

there is a turnover, let's say, of twenty percent,

23:16

So twenty percent of pms more or less

23:18

are let go every or leave every year,

23:21

but you keep the good ones right, and so

23:23

eventually you have a

23:25

sufficient number of pms who

23:27

really can carry make the business

23:29

sustainable. And a platform

23:32

is an instrument for exploration. Okay. So

23:35

I'm not saying how long they last or whatever, right,

23:37

but okay, how good do

23:39

you need to be? I think that if you have

23:41

a market neutral sharp

23:43

RAI show which for those who are not

23:46

used to this number, this basically

23:48

is a risk adjusted measure of profits.

23:50

So you take your P and L and you divide

23:52

by some measure of risk, and you get the sharp

23:54

pray show. If you don't

23:56

have these kind of market exposures,

23:58

you call it information race show. If you have

24:00

an information ratio of one, and

24:03

you are managing your left tail sufficiently

24:05

wisely, you can survive.

24:08

Okay, So you know, start practicing.

24:11

Okay, okay.

24:13

But on this note, the other thing I wanted to ask you

24:16

was, you know, we tend to talk about

24:18

these things platforms, pod shops,

24:21

multi strat as like this

24:23

one big blob basically

24:26

doing a similar thing. But my

24:28

impression is that the culture varies quite

24:31

substantially across firms. And

24:33

again there aren't that many that

24:35

are doing this, although as Joe said and the

24:37

intro, the number is growing. But when

24:40

we talk about that kind of cultural variation,

24:42

what do we mean exactly?

24:44

To an amazing extent, I

24:46

think that platforms are shaped

24:49

by the personalities of their founders.

24:52

So is Englander

24:54

as a personality, and a personal

24:57

history can grief in as a different

24:59

one, So of wads and river trading

25:01

not a platform, you know, strict

25:03

to censu but you know, to some extent, multi

25:05

strategy, and so and

25:07

so the cultures are very affected by

25:10

this. So if you are a trader

25:14

like Ken Griffin, it's

25:16

more likely that the fund that you work

25:18

in it's as more of a trading as opposed

25:21

to maybe a pure technology culture.

25:23

Millennium is very decentralized. Citadel

25:26

tends to run more like a

25:29

centralized and efficient organization. So

25:31

in the words of a of a Hedgehund manager,

25:33

you know, Citadel is like Singapore and

25:36

Millennium is like the United States. Right,

25:38

Singapore very efficient, efficiently

25:41

run technocratic to some extent, and

25:43

the US is messy

25:46

and inefficient, but it's very

25:48

robust. And in a sense, you know, Millennium

25:50

has these features of robustness of

25:52

it's like an organic creature. It

25:54

does change a lot. So other some

25:57

firms are more collaborative. I think Ballyasni

25:59

for example, tends to be more collaborative than

26:01

these other two firms. But by the way, and your

26:03

marriage may vary between different teams,

26:06

like depending on where you work, you

26:08

know, it can be heaven or it can be hell.

26:11

All Right, someone hears this podcast, maybe

26:13

they're in college studying finance or

26:15

maybe something in tech or something engineering

26:18

or whatever. They're like oh, this sounds really cool. I want

26:20

to work, for one, what is sort of the basic

26:22

path that one winds up maybe

26:25

first in a pod and then running a pod.

26:27

Okay, So first of all, I would like to dissuade

26:29

everybody who's listening from studying

26:31

a career in finance.

26:32

Okay, okay, So everyone's going

26:34

to take that as a challenge, but keep going.

26:36

Of course. And so

26:39

I wrote a small document because I got a

26:41

lot of questions like this from

26:44

students, and the brutal answer is

26:46

that it's very difficult and there is some

26:48

luck involved. So it does

26:50

help to go to schools

26:53

with a brand name, for sure. It

26:55

definitely does help if you want to do quantitative

26:58

stuff to be a very good programmer, and

27:00

you know, you need to have the ability to think quantitatively.

27:03

So that's that's for sure. There are couting

27:05

tests that make the admission a little bit

27:07

more democratic nowadays, but still still

27:11

it's very selective. I am not particularly

27:13

qualified to give advice

27:16

on how to get food

27:18

in the industry. I think I have a better

27:20

view of how to succeed in how

27:22

to be happy, not succeed how to be happy in the industry.

27:25

So that's probably more important, let's hear this.

27:28

Yeah, yeah, So I mean how to

27:30

be happy in the industry. I think that I

27:34

ask a lot the question of what

27:37

makes a good analyst or a good

27:39

quantitative researcher to people, and

27:41

I get very often the same

27:43

answer, which is people who are curious

27:46

do well and seem to be happy. So as

27:49

usual, you need to have passion, you need to

27:51

go, you know, to get into

27:53

the weekend, and not being able not

27:55

to think about a problem. So I think obsession

27:58

helps. Okay, So I think the belongs

28:00

to the obsessed, for good or worse in

28:03

the future. Like you can see this, it's

28:05

a heavy tailed world. So if

28:08

you want to have a more

28:11

stable job and less

28:13

absorbing, I think being a dentisty is a better

28:15

career path. But having some

28:18

level of obsessions into this stuff

28:20

it's good. Otherwise at some point, you know, you leave the

28:22

industry. It's perfectly fine.

28:23

By the way, So

28:26

this actually reminds me of something else I wanted to

28:28

ask you. So you said the world

28:30

belongs to the obsessed, which great line

28:32

is a very good line. But when

28:35

I read books on quantitative

28:38

finance, so much of it seems to be

28:40

about Greek letters for a start,

28:42

but basically sizing and managing

28:45

risk and how to look at your positions

28:48

and all of that, how do you actually generate

28:50

trade ideas? Like where does the

28:52

strategy come from? Am

28:54

I just looking for you know, mathematical

28:57

dislocations in the market and arbitrary

29:00

opportunities? Or am I thinking like I

29:03

want to go big on something like AI

29:05

or clean energy or whatever.

29:08

So I think that there are two dimensions to your question.

29:10

So the first one is how

29:12

objectively do you create alpha? Okay?

29:15

And so there are only a certain finite

29:17

number of ways to go

29:19

about alpha okay. So there are structural,

29:22

structural imbalances that are

29:25

not adaptively filled because

29:28

the market is poorly designed, because we don't

29:30

live in a neoclassical world, okay, and

29:33

so these imbalances persist. And

29:35

how do you exploit this physical

29:38

alpha? Is two

29:40

ways. The first one is you're a freaking genius

29:43

and you face a wall

29:45

for two years, do research, and you come up with

29:47

an originally okay, there are people like this, very

29:49

few. The other is simpler.

29:52

It's like a Renaissance style. You are an apprentice

29:55

in a famous painter's shop and you

29:57

learn the trade, and then you strike

29:59

it on your own and you make it a little bit better,

30:02

and even making it a little bit better

30:04

can make a huge difference. So I

30:06

would say imitation plays

30:08

a big role. And then maybe

30:10

there is another characteristic, which is you

30:13

just have to have the right makeup

30:15

in terms of you know, drive tolerance,

30:19

risk tolerance, so you

30:21

know when you I was actually having lunch

30:23

with a former zero point seventy two pm

30:25

now and his biggest

30:28

jordan was ninety million dollars, which is, by the

30:30

way, not crazy crazy high. If

30:33

you're down half a billion dollars, you're

30:36

literally losing your marbles. Okay,

30:38

your you know, your face looks different.

30:41

So have you seen that?

30:42

Oh sure, yeah, yeah,

30:45

yeah.

30:45

I remember in a flowed by randomness

30:48

to Lev talks about watching

30:50

all of like the hormones of someone who just

30:52

lost a lot of money, like pour out, and how pale

30:55

they look.

30:55

Right.

30:56

He had a specific comment about that if

30:58

there are only so many geniuses, if

31:00

there isn't an infinite supply of alpha,

31:03

if the structural forces, the physical

31:05

forces as you describe them, you

31:07

know, there's only so many sort of these dislocations

31:10

or reasons why reality is separate

31:12

from the neoclassical world. Does

31:14

it imply that as we see more

31:16

of these launches, and as

31:18

these hedgephones get bigger, that the opportunity

31:20

diminishes. Yes? Cool?

31:26

Wait, why, well,

31:28

because everything has a finite capacity,

31:31

that's it. I mean, And you know, as

31:33

you say, Joe, right, there is there are only that

31:35

many opportunities, and each

31:37

opportunity has a finite capacity, and

31:39

so at some point everybody is

31:42

doing the same thing and you get to some kind of

31:44

equilibrium which is not necessary that everybody

31:46

makes the minimum rate of return. Right, But you

31:49

know.

32:05

You mentioned earlier the systematic

32:07

equities are more relatable than other

32:09

things like the treasury basis trade and I kind

32:11

of my personal experience, I would beg

32:13

to differ because I come from a sort of credit

32:16

background. But it reminded

32:18

me a lot of these firms

32:20

are becoming bigger presences in

32:23

the bond market, bigger market

32:25

making roles and that sort of thing. Does

32:28

the day to day of being

32:30

in equities versus fixed income

32:33

in this kind of world? Is it very

32:35

different or do similar principles?

32:37

Supply?

32:38

I think it's very different, actually, you know, And why

32:41

first, in fundamental equities,

32:44

your edge is mostly informational,

32:46

so you do have a model of the world

32:48

that differs from consensus, and

32:51

you monetize that. It's really informational.

32:54

In the case of a lot of fixed income,

32:56

is truly structural. You know, there are

32:59

predictable flows, there are well

33:01

known imbalances, there are different demands

33:03

for liquidity. So it's

33:05

more of a

33:08

strategy or a class of strategies

33:10

that has skew. So

33:13

you could lose a lot of money, but to collect

33:15

pennies on a regular basis, so

33:18

you need to manage risk for that. You

33:20

need to have more capital for that, and scenarios

33:23

for that. So the risk management. The

33:25

way you think about investment

33:27

is different, is more scenario based, it's

33:30

less diversified. Fundamentally, you have relatively

33:32

correlated bets.

33:33

Why isn't the world actually

33:36

mapped to the neoclassical view of the world

33:38

because there's so much money and there's so much

33:41

investment and effort being put into

33:43

spotting any price dislocation anywhere,

33:46

So why is it with all the money

33:48

and all of the professionals and the geniuses

33:50

and the supercomputers and the AI that are

33:53

like essentially attacking the question

33:55

of finding mispriced securities.

33:57

Why are there still mispriced securities?

34:00

Theory everything should get arbed out.

34:02

Yeah yeah, but

34:04

not in practice.

34:05

Well yeah, but that's why. Why not? Why does

34:07

it even with all the professionals and money

34:09

trying to do this, did there still persist

34:12

in these anomalies or dislocations,

34:14

whatever you want to call it.

34:16

I don't I'm not really qualified to answer,

34:18

but I just see, there is only a finite number

34:20

of professionals, you know, and there

34:23

is only a finite number of professionals with a certain

34:25

risk tolerance. So and

34:28

there are constraints all around their constraints

34:30

on your balance sheet, there are constraints

34:32

on how much money can you lose. So

34:35

there are all sorts of limits to arbitrage

34:37

that go beyond the toy model of you

34:40

know, slife with AMBITIONI but

34:42

they go So that's kind of a

34:44

funding arbitrage. And the mechanism, by

34:47

the way, it's wrong for that paper.

34:49

I mean, it's not realistic, not wrong, it's

34:51

like artificial. But wherever there is a

34:53

constraint, independently of how many

34:56

players you have, you have

34:58

a potential inefficiency and

35:01

it's not going to go away.

35:03

I have a practical question, And I always

35:05

wanted to ask this of someone, and I

35:07

think you're the perfect person to perhaps

35:10

answer this. But if you are a

35:12

risk manager at this

35:14

kind of firm, and I don't

35:16

know you're you come into the office and

35:19

it's let's say it's like the day

35:21

of a FED meeting and Jerome

35:24

Powell comes out and says something completely

35:26

unexpected, or let's say it's

35:28

twenty fifteen and China suddenly announces

35:31

they're devaluing the un And

35:33

you're looking at your computer screen and you're

35:35

looking at the various risk metrics. How

35:38

fast do those move and how

35:40

much of it is calculated in real time

35:42

versus all the numbers having to be run

35:44

at the end of the day when you net out

35:47

trading positions.

35:52

If you have the right model, you

35:55

should be able to either capture

35:58

those risks directly in a sense,

36:00

imagine you have a sensitivity

36:03

to the various

36:05

points in the Yell curve, either in

36:07

your fixed income portfolio

36:09

or in your equities portfolio. If you capture

36:12

those well, so it's a risk that you

36:14

you know you're taking and you can hedge.

36:16

You should see the factor moving, but not your

36:18

porfolio moving. Okay, And by

36:20

the way. You can also not have these factors,

36:23

but you may have factors that are proxying

36:26

these microeconomic drivers, like

36:28

say, for example, momentum is one, crowding

36:30

is another. And so even

36:32

if a portfolio manager doesn't

36:35

think directly in terms of points on the ill curve,

36:37

but they have other related

36:40

ways of thinking, so they can still control for that.

36:42

And then there is, unfortunately the case

36:44

where well we never model this, we

36:46

do not have a proxy for this, and then you're screwed,

36:49

And yeah, you don't want to be in that situation.

36:51

Typically, you know, you can see these effects

36:54

like I mean, there was a big surprise when when

36:56

rates went up a lot of equity portfolios

36:58

moved and they didn't know why,

37:01

and there was no interest rates sensitivity

37:03

in commercial factor models.

37:05

So there you go in theory, on

37:08

a day of some sort of unexpected

37:10

event. Tracy mentioned the China U

37:12

end evaluation. If

37:15

everything is working perfectly and

37:17

you truly do have like completely

37:20

eliminated your market exposure,

37:22

does that show up at that level,

37:25

like does it still show up somehow?

37:27

It still can show up in weird ways, right,

37:29

So for example, you can be market neutral.

37:32

Yeah, the market has a big drowdown

37:34

and you still lose money. Why because

37:36

the market the drawdown

37:40

starts weird processes of

37:42

the risking that affect your

37:45

portfolio. So even if I'm market neutral,

37:47

somebody is selling my stock

37:50

to reduce their risk and it's affecting even

37:53

though I'm perfectly market neutral. So weird things

37:55

can happen. Unfortunately, you know, so

37:57

there is no perfect model, that's the short answer.

38:00

Unfortunately, you mentioned crowding

38:02

in multi strap and the idea that

38:05

maybe you know, eventually you would reach

38:07

a limit for the efficacy of some of

38:09

this type of trading. What's

38:12

next for hedge funds? So we

38:14

went from fund of funds to pod shops.

38:17

They became the hot new thing. What comes after

38:19

pod shops?

38:20

What's exciting?

38:21

I'd love to know. It's

38:23

for the next guest to answer, I don't know.

38:25

This is where you reveal where you're

38:28

gardening, your current gardening leave ends,

38:30

and where you're gonna wind up next.

38:31

Oh yeah, my best job is always

38:34

the next I don't know. But so

38:38

what's next in terms of

38:40

business model? Would be very interesting to know what's

38:42

next. So there

38:45

are some interesting ideas. So there is the idea

38:48

of alpha capture, which is kind of a big

38:50

umbrella. And

38:52

you know, alpha capture has

38:55

has an interesting story. So there was external

38:58

external sale set alpha capture. That's

39:00

historically like kind of a creation of

39:02

Martial Ways, an English hatch fund

39:05

that in two thousand and three or four study

39:07

at program called tops where they gathered

39:10

ideas from the cell side, and that

39:13

for a while was very profitable and also

39:15

has lots of other byproducts that are

39:17

great. Now I think it's kind of arbitraged

39:20

out now there is a similar concept

39:22

of byside external alpha capture.

39:24

So there are firms that are trying

39:27

to get ideas from hedge funds,

39:29

small edge funds. They don't have scale, they

39:31

can aggregate them and then they make into

39:33

a portfolio. That's a new business model.

39:36

I don't know how scalable it is, how sustainable

39:39

it is, but that's an idea. There

39:41

is definitely an expansion

39:43

into privates. I have like zero

39:45

skill or zero divisibility to this stuff, so that's

39:48

really another question for somebody else.

39:50

And then there is always product innovation. Every

39:52

strategy is continuously innovating,

39:54

has to change. So just look

39:56

at where fundamental Equities

39:59

was one hundred years ago. Go right, the recommendation

40:01

was invest in a railway single

40:03

stock and you know, be happy. And

40:06

now we have, you know, and now we spend hundreds

40:08

of millions of dollars in alternative data

40:10

and there are tools and stuff. So what is it

40:12

in ten years I don't know, but it will

40:14

be very different than it is today.

40:16

I remember, you know, when I was over twenty

40:18

years ago and I first got interested in markets,

40:20

picking up the Intelligent Investor because

40:22

of course, you know, Buffett and Munger were into it

40:24

and like reading is like and so if you buy the

40:26

Brooklyn rail bond

40:29

yielding eight percent, I was like, what is this?

40:32

Yeah?

40:32

I just thought it seems so disconnected from me. I

40:34

mean, I'm sure there's a lot of deep wisdom and I probably would

40:36

have like internalized it. Yeah, but just in terms of

40:38

like what they were talking about, it seems so funny because of how

40:41

antique it all seemed.

40:42

Totally. Yeah. And so now pms

40:45

are quantitative. Fundamental pms tend to

40:47

be quantitatively quite literate. In

40:49

the future they will be even

40:52

different. Maybe they will be prompt experts.

40:54

I don't know.

40:54

Can you be a fundamental PM by

40:57

just being a domain expert

41:00

in a certain area, say like you're really

41:02

understand biotech, or say you really

41:04

understand the semiconductor industry and you want

41:06

to trade chip stocks versus and not

41:08

really have that sort of quant background but some other

41:11

expertise.

41:11

So being a domain expert is definitely

41:14

a necessary condition. You absolutely

41:16

need to be a domain expert. And since

41:19

you make the example of healthcare super

41:21

domain experts, so a lot of good healthcare

41:24

pms have either worked in

41:26

healthcare companies they have never practiced,

41:28

but they are domain expert. Is is

41:30

it sufficient to be just a domain expert.

41:32

No.

41:33

I think that you need to be able also

41:35

to monetize and to risk manage your

41:37

portfolio, and that's very difficult.

41:40

So that's not sufficient, but it's definitely

41:42

necessary.

41:43

How important are the data sets? Like what

41:45

if I'm just really good at finding original

41:47

and alternative data someone's analysts.

41:50

Yeah, it varies

41:52

a lot, so some pms,

41:54

well, okay, first of all, for systematic it

41:56

matters a lot period

41:59

Unconditionally, for discretionary

42:01

pms, it varies a lot. So some pms

42:04

will use alternative data,

42:07

some will do deep research and

42:09

think three months to

42:11

a year ahead. And the

42:13

reality is that there are not that many data

42:15

that really help you think at that horizon.

42:17

So we don't live in the world of really really

42:20

big data for fundamental

42:22

thinking. So I think that's interesting.

42:25

I have just one more question, which is what

42:27

do you find most satisfying

42:30

about your job? What gives you

42:33

the needs yeah or jobs? Yeah?

42:35

What gives you the most pleasure on a data

42:38

day basis? Do you feel fantastic

42:40

if China devalues the un and you

42:42

look at positioning

42:44

across the firm and you're not going

42:46

under, Or do you feel great if you

42:49

identify a particular strategy or something

42:51

like that.

42:52

Now, the thing that gives me most pleasure when

42:54

I work is when I do something

42:57

that is useful and it works for others.

42:59

So I just love the

43:01

social aspect of working, Like it's

43:05

actually a job where you can be of some

43:07

use to other people, and I just

43:10

enjoy that. So when things work out,

43:12

like you come up with an idea after multiple

43:14

failures and it works, you implemented,

43:17

and somebody else uses it or

43:19

finds a value to this, and everybody

43:21

is happier and like and we get drunk together.

43:24

That's great, all

43:26

right.

43:27

Giuseppe Palia logo aka

43:29

Gappy, Thank you, so much for coming on all blots.

43:31

Really appreciate it.

43:32

Thank you, thank you.

43:33

That was fantastic, Joe.

43:50

I feel like that's good life advice. If it all

43:52

ends in people getting drunk, it's usually

43:54

no, wait, that doesn't make sense. Sometimes it's really bad,

43:57

yeah, say never mind. But sometimes

43:59

it's great.

44:00

Sometimes it's good. I love that

44:02

line. I feel like the world belongs

44:05

to the obsessed. It's just like a really good

44:07

line. That's sort of ominous to me because I

44:09

don't really get obsessed with anything besides country

44:11

music. And then the rest of my time, I'm just like

44:13

I want to talk about hedge funds one day,

44:16

and then the next day I want to talk about like how energy.

44:18

Yeah, I was going to say, it's not really like get

44:20

obsessed. It's just you flip. Then I'm an obsession

44:22

to obsession, so.

44:23

It's not real obsession. It's kind of delet Wait,

44:26

Tracy, have I told

44:28

you about when I got a job offer

44:31

at a prop trading shop?

44:33

This vaguely rings a bell.

44:35

So can I tell a quick story?

44:36

Go for it?

44:37

So I had traded stocks in college

44:39

just because it was like the dog Come era. It was fun, it was

44:41

very easy. Everything was going up. I

44:43

managed to sell for excellent reasons

44:45

a good time, and I didn't lose all my money anyway. I

44:47

was always I got interested in markets. Then I

44:49

graduated with my useless liberal

44:52

arts degree and I had

44:54

a job. I was making minimum wage working at a

44:56

Delhi and I saw this help wanted ad

44:58

at a prop trading shop and all Austin, Texas,

45:01

and it didn't seem like they had many requirements,

45:03

so I went. They asked me about

45:05

my personal trading. I played

45:08

ping pong against the CEO. I

45:10

played this video game that involved me using

45:13

two joysticks. One was to control

45:15

the tilt of a triangle and the other one was to

45:17

control the space and I kept it in the square already,

45:19

its weird. And I did this other

45:22

thing where I like typed without like too many typos

45:24

and stuff like that. And there were like two hundred people applied

45:27

and second round, I got

45:29

one of the four spots that they offered,

45:33

and for reasons that still

45:35

allude me to this day, I

45:38

didn't take the job. I was enjoying making

45:40

minimum wage at the deli. All

45:43

my friends worked there. It was like the cool place to

45:45

work in Austin. I didn't feel like giving that up,

45:47

and I didn't, and I just like, I always think

45:49

about what if, what does my life look

45:51

like if I took that job? The strangest

45:53

most inexplicable career decision

45:56

I could ever imagine, not taking a trading job

45:58

from a five dollar minimum way job or

46:00

whatever it is at the time. Anyway, I'll never

46:02

know.

46:03

Okay, Well, I once got offered a specialty

46:05

sales position in bank

46:07

equities at a Swiss bank, and

46:10

I never question what my

46:12

future would have been had I taken

46:14

that job. I'm very satisfied, but I

46:16

actually have a question. Do you think you were put off by

46:18

the weirdness no interview process?

46:21

Like did you think that you were going to be playing ping pong

46:23

and like moving joysticks as

46:25

part of the job.

46:26

That was fun? And I didn't even beat the CEO

46:28

in ping pong. She beat me, but she still

46:31

hired me. I don't, no,

46:33

I don't know why I can't. The only thing I could

46:36

explain is that in my post college life,

46:38

I had a cool job where I got to hang out

46:40

with my friends in the back of this deli at a grocery

46:42

store. I didn't really feel like giving it.

46:44

Up, just yet all right, Well, I

46:46

do feel like coming

46:49

out of that conversation with Giesseppe, I feel

46:51

like I have a much better conception of how

46:53

multistrat actually works and what people are sort

46:55

of doing on a day to day basis, and

46:57

also just maybe a better understanding of some

46:59

of the terminology around the industry totally.

47:02

So now we'll probably do more episodes, but

47:04

I feel like I'm now like roughly grounded

47:06

in at least some core ideas here.

47:09

Yeah, and everyone should definitely check

47:11

out Gappy's Buyside quant

47:13

Job Advice. It's nine pages

47:16

and it actually it goes into some detail

47:18

on the structure of the industry itself of

47:20

how you know quantitative hedge funds actually

47:23

work, and like who are the big names and things

47:25

like that. So anyone's interested

47:28

in the space, definitely check it out. Shall

47:30

we leave it there?

47:30

Let's leave it there.

47:31

This has been another episode of the Odd Thoughts

47:33

podcast. I'm Tracy Alloway. You can follow

47:36

me at Tracy Alloway and.

47:37

I'm Jill Wisenthal. You can follow me at

47:39

The Stalwart. Follow our guest Joseppi Polyioligo

47:42

aka Gappy. He's Double

47:44

Underscore Palyioligo on

47:46

Twitter Follow our producers Carmen Rodriguez

47:49

at Kerman Ermann Dashil Bennett a Dashbot,

47:51

Killed Brooks at kill Brooks. Thank

47:54

you to our producer Moses on Them. For

47:56

more odd Lots content, go to Bloomberg dot

47:58

com slash odd Lots, where we have transfer a

48:00

blog and a newsletter. And check out the

48:02

discord where you could chat with fellow listeners

48:04

twenty four to seven discord dot

48:06

gg slash odd Lots.

48:08

And if you enjoy all Lots, if

48:10

you like it when we talk to practitioners

48:12

in a particular space such as multi strat,

48:15

then please leave us a positive review

48:17

on your favorite podcast platform. And

48:19

remember, if you are a Bloomberg subscriber,

48:22

you can listen to all of our episodes absolutely

48:24

ad free. All you need to do is connect

48:26

your Bloomberg account with Apple Podcasts.

48:29

Thanks for listening.