HomeVideos

AI Dominates Economy and Markets with Torsten Slok | The Real Eisman Playbook Ep 68

Now Playing

AI Dominates Economy and Markets with Torsten Slok | The Real Eisman Playbook Ep 68

Transcript

1769 segments

0:05

Hey, this is Steve Eisman. This is

0:06

another episode of The Real Eisman

0:08

Playbook. There are just so many issues

0:10

going on these days. Crossurrence,

0:13

politics, war. But from an economic

0:16

perspective, I think the biggest issues

0:18

are the impact of AI,

0:20

re-industrialization,

0:22

short-term, the so-called big beautiful

0:24

bill, unemployment,

0:27

the deficit, and the overall investing

0:30

environment. And these are major issues,

0:32

and we've actually never had an

0:34

economist on to sort of plow through all

0:37

of them. So today I've invited Torstston

0:40

Sllock who is the chief economist of

0:42

Apollo, an excellent economist and we're

0:45

going to go through a ton of issues and

0:48

afterwards I'm going to come back and

0:50

talk about lessons learned. See you

0:52

soon.

0:56

Hi, this is Steve Eisman and welcome to

0:58

another episode of the real Eisman

1:00

playbook. So there's so much going on

1:03

these days. AI,

1:06

K-shaped economy, oil prices,

1:10

the deficit. I mean, it it's just hard

1:13

to keep up.

1:15

>> So, today we have as our guest chief

1:18

economist at Apollo, Torson Slack, who's

1:20

going to help us, I hope.

1:23

>> I certainly hope my best to go through

1:27

what must be like every day, you can't

1:29

believe how much information is coming.

1:30

So, Torson, welcome first of all.

1:31

>> Thank you, Steve. Thanks for having me.

1:33

So let's start with

1:36

a simple topic. Give us your very broad

1:39

overview about the health of the US

1:41

economy. A simple question.

1:43

>> Well,

1:44

>> and then we'll then we'll dig deep.

1:46

>> Well, thanks all again first for having

1:48

me. But it's really quite simple

1:50

initially. What is the headwinds and

1:52

tailwinds to the economy at the moment?

1:54

Because there are three very important

1:56

tailwinds that are driving growth. First

1:57

of all, we have an AI spending boom

1:59

because of the data centers and the

2:00

energy associated with the data centers.

2:02

We calculate that that contributes at

2:05

the moment about 1%ish point to GDP

2:06

growth. Normally GDP growth at two and

2:08

now 1% is coming from the AI spending

2:10

boom alone.

2:11

>> So in other words, GDP growth this year.

2:14

>> You estimate up about two and 50% of it

2:17

is purely from AI spending.

2:19

>> Absolutely. So that's both the boom

2:20

coming from the data center and energy

2:22

buildout, but also the associated wealth

2:23

effects from the stock market being high

2:25

and the increase in consumption,

2:27

especially for high-end consumers. So

2:28

this is a very important source of

2:30

growth. That's very unusual. We have not

2:32

seen this source of growth for literally

2:34

decades before where we've seen one

2:36

sector in such a significant way

2:38

contributing even with housing. We

2:39

didn't see a contribution that was as

2:41

high as 1 percentage point because this

2:43

sector is just really really

2:44

significant. The second source of growth

2:47

is the industrial renaissance.

2:48

Industrial renaissance means politicians

2:50

that want to do home shoring or

2:52

production of semiconductors. This was

2:54

the chips act under Biden home shoring

2:56

of pharmaceuticals prescription drugs

2:57

and of course importantly also home

2:59

shoring and production of defense. You

3:01

know, I've had industrial economists on

3:04

and when I would ask them, is there

3:07

evidence

3:08

that of of this? In other words, people

3:11

building factories in the United States,

3:13

their response generally be that was

3:16

that literally. So you're saying it's

3:19

stronger than eh. It is a bit stronger

3:20

than a it only adds 0.3% GDP. So not a

3:23

full percentage point like the AI

3:25

spending boom, but we are seeing

3:27

especially after the chips act, we saw

3:29

significant increase in manufacturing

3:30

capacity for semiconductors and we've

3:33

also seen significant increase in

3:34

capacity for manufacturing generally. So

3:36

that means the ISM in the last five six

3:38

months has started to go up. So the

3:39

manufacturing sector is doing a little

3:41

bit better than a but it's not as big a

3:44

source of growth compared to the AI

3:46

spending boom.

3:47

>> So 1% from AI

3:49

>> 0.3

3:50

>> 0.3 from re-industrialization.

3:52

>> Exactly. From the globalization

3:54

reversing.

3:54

>> That's out of two.

3:55

>> That that's out of two. And then the

3:56

last thing we have is 0.9 coming from

3:58

the one big beautiful bill. Remember the

4:00

one big beautiful bill which was signed

4:02

last year. It lowered taxes

4:03

retroactively for consumers so that

4:05

taxes were lowered starting January 1,

4:07

2025. The consequence of that is when

4:10

people are filing their taxes this year,

4:12

both those who did in April and those

4:13

who had extension, last year the average

4:15

tax refund was around $3,000 and this

4:18

year the average tax refund for

4:19

households is about $4,000. That means

4:21

over the next 6 months we will continue

4:23

to see consumption do really well.

4:25

That's a very strong tailwind coming to

4:27

consumers because of the one big

4:29

beautiful bill that was implemented last

4:30

year.

4:31

>> So let me pause you on that one. Okay,

4:32

I'll grant you that. But that that's one

4:36

time.

4:37

>> That is indeed one time.

4:38

So 2027

4:41

90 basis points of growth. If what

4:43

you're saying is 90 basis points of

4:44

growth this year is from the one big

4:46

beautiful bill next year it's not going

4:48

to be there.

4:48

>> That's absolutely correct. So it will be

4:50

smaller and that's why if you add those

4:52

things up AI about 1% the industrial

4:54

renaissance about 0.3 and 0.9 from the

4:57

one big bill you get to a little bit

4:59

more than 2% this year right

5:00

>> but what's most important about these

5:02

three different sources of growth is

5:04

that they are not sensitive to interest

5:05

rates. Right? In other words, this is

5:07

not your traditional economic situation

5:09

where interest rates go up and the

5:11

economy slows down. We are seeing the

5:13

sectors that are sensitive to interest

5:14

rates, namely housing and autos are not

5:16

doing well because they are very

5:18

sensitive to interest rates that have

5:19

gone up in the front end and in the long

5:21

end of the yield curve. But at the

5:22

moment, because these sources are not

5:24

sensitive to interest rates, that's why

5:25

Kevin W is dealing with a strong

5:27

economy, high inflation, that's why long

5:29

rates continue to be high because the

5:31

economy is just not slowing down. So

5:33

that's why you're right when we come to

5:34

2027 that's a different discussion but

5:37

for the next 69 months we still have

5:39

strong tailwinds coming from the AI boom

5:41

strong tailwinds from the industrial

5:42

renaissance and strong tailwinds coming

5:43

from one big beautiful bill

5:45

>> okay so since you brought up wash let me

5:47

press wash a little bit given all what

5:50

you've just said would you agree that

5:52

the probability of the Fed cutting rates

5:54

this is zero is zero

5:55

>> yes

5:55

>> zero

5:56

>> zero it's not going to happen

5:57

>> not going to happen economy is too

5:58

strong inflation is high for a number of

6:00

different reasons partly because the

6:01

economy is also because of tariffs, also

6:04

because of of course oil prices that

6:06

have gone up and we're also seeing now a

6:08

contribution to inflation of 0.3 coming

6:10

from the AI and data center buildout

6:12

because semiconductors are more

6:14

expensive, labor to build data centers

6:16

is more expensive and you also have

6:17

equipment is also more expensive and

6:19

energy also being more expensive is also

6:21

adding to inflation. So there is

6:22

literally zero chance that he will cut

6:24

interest rates this year.

6:25

>> How about raise them? Well, the market

6:27

as we speak today are pricing that the

6:30

Fed will be hiking rates in September

6:31

and in December. So that's two hikes.

6:33

Wow. So that's pretty a shift. It is

6:35

very dramatic shift. As you and I know

6:37

very well, in the beginning of the year,

6:39

the dot plot was clearly saying that the

6:41

Fed is going to cut cut cut and rates

6:42

are going down. And now suddenly we have

6:44

a situation where the market is pricing

6:46

that well maybe especially after his

6:48

latest press conference where he said

6:49

I'm not going to give any forward

6:50

guidance. And when you don't give

6:52

forward guidance, the market has to

6:53

start guessing. And that's why you and I

6:54

are now guessing and the best guess that

6:56

the market has at the moment is that we

6:58

will see hikes coming because the

7:00

economy is really strong and we have

7:02

some upward lift. Inflation at the

7:03

moment is 3 and a half%. That means we

7:05

have some strong tailwinds not only to

7:06

GDP growth also to inflation and that

7:08

just makes it impossible for wars to cut

7:10

rates over the next 6 months.

7:13

>> Sounds like that's pretty bad for

7:14

housing.

7:15

>> Well, because housing is

7:17

>> not that housing is so good right now

7:18

either. You are the world expert in this

7:20

of course but it is absolutely the case

7:22

that housing lives and dies on what's

7:24

going on with mortgage rates and housing

7:26

is already experiencing very little

7:27

supply. The home builders have been very

7:29

reluctant to produce and create more

7:31

housing. And if you're on top of that on

7:32

the demand side also have that rates are

7:35

very high for a very long period because

7:36

we now have a strong economy and upward

7:38

pressure on inflation. It is indeed the

7:40

case that the most sensitive parts of

7:41

the economy, housing and autos are just

7:43

not doing very well at the moment. And

7:44

we expect that to continue because the

7:47

growth is not coming from traditional

7:49

sources of growth that are sensitive to

7:50

interest rates. It's coming from these

7:52

really unique three areas of the AI

7:54

boom, the one big beautiful bill, and

7:55

the industrial renaissance. All right,

7:57

so let's dig down into AI a little bit.

8:00

So tell me if you agree with this or

8:03

disagree with this. seems to me, I mean,

8:06

not a day goes by that something

8:08

dramatic doesn't happen with AI. It's

8:10

it's kind of it's pretty hard to keep

8:11

up. I get the impression

8:14

that at least part of the AI story has

8:17

really dramatically changed in the last

8:20

I would say not even more than a month.

8:23

And I would say it's along two vectors.

8:26

One, and this is what I'm curious what

8:28

you think about. Number one, this is now

8:31

a very capital intensive business. You

8:34

know, last year, for example, Google

8:35

spent 80 billion on AI and basically

8:40

funded it from its own cash flow. And

8:41

this year, they're spending 190 billion

8:44

and they just raised 85 billion in

8:45

equity. And so, you're starting to see

8:47

more and more companies raise capital

8:48

because the the demands on their balance

8:51

sheets are just so huge. So, that's new.

8:53

And the second thing which maybe is even

8:56

more important is I get the impression

8:58

that there are no moes in this business

9:01

that people flip from Gemini to Claude

9:03

to chat GPT. So you're talking about

9:06

massive companies spending trillions of

9:08

dollars for something that may have no

9:10

moes and

9:13

that's not a recipe for longevity. So I

9:17

I'd be curious what you think about

9:19

that.

9:19

>> Yeah. So absolutely on the first point

9:21

if you look at the free cash flow for

9:22

the hyperscalers has absolutely gone

9:24

from being very very high and literally

9:27

is dropping down over the next 6 12

9:29

months towards zero and it might even

9:31

begin to go negative because the capex

9:33

requirements which is so massive

9:35

>> it's very very very substantial and

9:37

these are and continue to be very

9:39

profitable businesses especially the

9:41

magnificent 7 of course which we have

9:42

most information about have had

9:44

significant cash flows have continued to

9:46

do so well and they have now decided to

9:48

spend an enormous into the trillions as

9:50

you're saying in terms of spending on

9:52

data centers and the energy buildout

9:53

because they really view this clearly as

9:56

existential that they got to have the

9:58

capacity the computing power that is

10:00

needed in this case of course to deliver

10:02

all the demand for compute that's going

10:04

to come along and to your second point I

10:06

think actually my second point is more

10:08

important because if there were moes

10:11

let's assume that there were very high

10:12

moes as an investor I would say okay so

10:15

you're going to spend a lot of money but

10:16

you're going to spend a lot of money and

10:17

I'll give you

10:18

because at the end of the day, you're

10:20

going to have a business that's a

10:22

duopoly or or or very well protected.

10:25

But if you're asking me to give you

10:27

money for a business that has no moes, I

10:30

don't want to give it to you. I' I'd

10:32

rather I' I'd rather buy Cisco that's

10:36

going to supply you is the analogy that

10:38

that that I've drawn is it's kind of

10:40

like comparing airlines to transdime.

10:43

Airlines is a terrible business because

10:45

it's very capital intensive and you have

10:47

and you have no pricing power and

10:49

Transdime which supplies parts to

10:51

airlines is a great business. So I'm

10:53

just curious as an economist if if if

10:56

I'm right what does that mean? What's

10:58

exactly most important about this

11:00

discussion is exactly are there no modes

11:02

for everyone or is it just modes for

11:04

someone

11:05

>> for the hyperscalers?

11:06

>> There could be some of the hyperscalers

11:08

that will end up being the winners and

11:09

others who will end up not being the

11:11

winners. In other words, there are

11:13

clearly modes in the sense that there

11:14

are some including of the private

11:16

hyperscalers that have clear pricing

11:18

power and clear products that they are

11:20

rolling out in a very substantial way.

11:22

But the question becomes of all the

11:24

capacity that's being rolled out. Is

11:26

that all going to have modes or in other

11:28

words, are they going to have pricing

11:29

power? Are they going to have special

11:31

products? Or is there a scenario as

11:33

you're saying where you can begin to

11:34

worry about that some of them may not be

11:36

able to survive in this situation even

11:39

though compute demand continues to go up

11:41

which is absolutely indisputable that

11:43

there will be almost unlimited compute

11:45

demand. The question is what is the

11:47

price that they're going to generate? In

11:49

other words, what's the revenue they're

11:50

going to generate on that compute

11:51

demand? Because if the price of compute

11:53

the price of tokens keeps going down

11:55

towards zero, then it may absolutely be

11:57

the case that there are some modes that

12:00

might be a lot more shallow or be much

12:02

smaller. I mean, let's imagine that one

12:05

of the companies that has no modes is

12:06

chat GPT, Open AI, just hypothetically.

12:10

I I've got no skin in that game, but and

12:13

that one day Open AI is in huge trouble.

12:16

The ramifications of that because so

12:19

much of of what's being spent is related

12:22

one way or another to open AI and and

12:24

and and anthropic are massive. I mean

12:26

Oracle, for example, has a a 600 billion

12:29

backlog, but half of the backlog is open

12:31

AI. I mean, it's a little scary what's

12:33

going on. But the added issue here is

12:35

also because from a pure competitive

12:38

perspective, the competitive landscape

12:39

is also dominated not only by the names

12:41

we're talking about here in the

12:42

hyperscalers. But remember also that a

12:44

lot of this also happens to then turn

12:46

into more open-source models including

12:48

Chinese models, right? So that means

12:49

that if you are a business and you say I

12:52

need some compute to do some things for

12:54

AI, well are you willing to instead say

12:57

it may be that the price of tokens say

13:00

from a Chinese model is only 1% of what

13:03

is the price of a token from a US model?

13:06

It still raises some important

13:07

questions. Are you still willing to go

13:09

after the cheap model because it runs

13:11

the risk that you have to upload your

13:13

data into say Chinese models and

13:14

therefore into something that could

13:16

become a much bigger issue rather than

13:18

just thinking about the cost. So the

13:19

mode is also and should also be in my

13:22

view thought of as there is also this

13:24

proprietary discussion about you're

13:26

right the data is transferable the

13:28

models are replaceable and they can

13:30

replace each other very easily but it

13:32

still ends up being a discussion that

13:33

those that have the cheapest tokens at

13:34

the moment at least they are certainly

13:36

the Chinese models and that becomes

13:37

important because a lot of businesses

13:39

might be able to say and willing to say

13:40

you know what I'm willing to pay for the

13:42

mode over here and for the fact that

13:44

this is a good service because this is a

13:46

US service rather than running the risk

13:47

of doing this is an open source model or

13:49

in a Chinese model. So from that

13:51

perspective, there is some unique um

13:54

characteristics by the US hyperscalers

13:56

relative to the hyperscalers especially

13:58

again from China.

13:59

>> Okay, let's switch gears. Let's talk

14:01

about the K-shaped economy, the K-shaped

14:04

consumer. Why don't you first define it?

14:05

I mean, people throw this term out all

14:08

the time, and half the time I I think

14:10

that when they when they throw the term

14:12

out, they don't even know what they're

14:13

talking about. So I say to you person

14:17

K-shaped economy K-shaped consumer

14:19

define this for me like what is this all

14:21

about?

14:22

>> This is all about three things. Number

14:23

one is about a K-shaped situation in

14:25

wealth that high income households today

14:27

relative to 2019 have literally savings

14:30

that are trillions of dollars higher

14:32

than where they were in 2019. Lowillions

14:35

trillions about one and a half trillion

14:36

dollar higher than where it was in 2019.

14:39

That's why the airlines have been saying

14:42

quite simply basically that they have no

14:44

problem selling business class tickets

14:45

to high-income households, but they're

14:47

having some challenges selling business

14:48

class tickets, sorry, economy class

14:50

tickets to low-inccome households. Okay?

14:51

Because low-income households, the

14:53

bottom 20% of the population, people who

14:54

make less than $25,000 a year, their

14:57

savings cumulatively as a group today is

14:59

literally in dollar terms exactly the

15:00

same as where it was in 2019.

15:02

>> So what you're saying is people 25,000

15:05

below have no more savings they had in

15:06

2019. in nominal terms.

15:08

>> And people at the upper end have a

15:10

trillion and a half more savings. That's

15:12

an enormous disparity.

15:13

>> And that's because people at the upper

15:14

end have been benefiting from three

15:16

things. Benefiting from stock prices

15:17

going up, home prices going up, and

15:19

people at the other end also own fixed

15:21

income. So when the Fed still has

15:23

interest rates high and still talk about

15:24

raising interest rates, that means that

15:26

the cash flow you get as a high-income

15:28

household is at the highest level in

15:29

fixed income that has been in decades.

15:31

That means that high- income households

15:32

are not only making money on their

15:34

stocks and on their home prices, but

15:36

they're actually also making money on

15:37

the cash flow that they get because some

15:39

Apollo funds pay like 8 n 10%. And those

15:41

returns you can get in private credit,

15:43

public credit, and fixed income is

15:44

basically at the highest level we have

15:46

seen literally in 20 25 years. So for

15:49

that reason, high income households

15:50

continue to benefit both from asset

15:52

price inflation and also from cash flows

15:54

being very very strong. So this is the

15:55

first answer to your question namely

15:57

when it comes to wealth there is a

15:58

K-shaped situation and that continues to

16:01

the legs are just getting longer in the

16:02

K if you will because the stock market

16:04

obviously continues to do well and the

16:06

cash flows continues to also do well

16:07

there is now also a K-shaped situation

16:09

secondly in wage growth the Atlanta Fed

16:12

has wage growth measures across income

16:13

distribution and people at the bottom

16:15

are seeing lower wage growth relative to

16:17

people that are in the middle and higher

16:18

think income distribution so that means

16:19

that the Kship situation is not only in

16:21

wealth it's also in income growth and

16:24

finally There's also a case shift

16:25

situation when it comes to inflation.

16:27

The New York Fed has measures for

16:29

inflation across income distribution and

16:31

people at the bottom of income

16:32

distribution then spade a bigger share

16:33

of their consumption on food, energy and

16:35

housing and these have seen a much

16:37

bigger increase in inflation. So people

16:39

at the bottom are also facing a higher

16:40

inflation rate related to people at the

16:42

top and in the middle. So from that

16:44

perspective the answer is there's a

16:46

K-shaped situation for wealth, there's a

16:48

cave-shaped situation for income growth

16:49

and there's a K-shaped situation for

16:51

inflation. And lastly, if you look at

16:53

stock prices for baskets of luxury names

16:56

in consumer spending have outperformed

16:58

over the last several years a basket of

17:00

retailers that of course cater to

17:03

discount or value names. So that's why

17:05

this discrepancy you can look up on your

17:06

screen every single day. What is the

17:08

difference? And it just continues to be

17:09

the case that high-income names and

17:11

those who cater to high- income names

17:13

continue to outperform discount

17:14

retailers. So that's why the K-shap

17:16

situation continues to be a major theme

17:18

in the outlook at the moment.

17:20

>> Okay. So grant that what are the

17:23

implications for the economy?

17:24

>> So the long term for this because this

17:26

is not a trend that's going to flippity

17:29

flip in one day. This is long term.

17:31

>> Absolutely. The net effect of the K is

17:33

that in aggregate the top 20% of

17:36

consumers they account for 40% of

17:37

consumer spending. The bottom 20% only

17:39

account for 8% of consumer spending. So

17:42

in aggregate total consumption is

17:44

actually still okay. If you look at the

17:45

weekly data from Redbook for same store

17:47

retail sales. So that mean red book goes

17:49

out a week once a week and ask retailers

17:50

what were your sales this week relative

17:52

to the same week a year ago and that's

17:54

still holding up very nicely. So that

17:55

means that in aggregate despite the K

17:57

getting wider and wider you're still

17:59

seeing in aggregate because the bigger

18:00

part of part of the case still has such

18:02

a big weight that in aggregate the

18:04

consumer is actually still doing well.

18:05

So that's why the answer to your

18:06

question is if the K continues it almost

18:09

instead becomes a political discussion

18:11

what does it mean when you have a bigger

18:13

and bigger share of the lower leg of the

18:15

K that continue to face headwinds not

18:17

only because of the three dimensions I

18:18

mentioned with wealth and income and

18:20

inflation but there's also the added

18:21

issue that when you look at the language

18:23

rates on auto loans have been going up

18:25

the language rates on credit cards have

18:26

been going up and the linguist rates on

18:29

student loans have also been going up

18:31

because there are a lot of households in

18:33

the bottom and the middle of income

18:34

distribution that also are facing higher

18:36

interest rates because they have this

18:37

problem that they have now also not only

18:40

a K-shaped situation for wealth and

18:42

income and inflation but also because of

18:44

this issue that delinquency rates are

18:46

going up especially for people in the

18:48

middle and the bottom of the K.

18:50

>> Sounds pretty grim.

18:51

>> So that means to your question before

18:53

that it means that the share of

18:54

households that are getting impacted on

18:56

the lower leg of the K is unfortunately

18:59

just growing and getting bigger and

19:00

bigger and that's of course why this

19:02

becomes a political discussion. Well,

19:04

what do they do?

19:04

>> Well, the worse it gets, the bigger the

19:06

political discussion

19:07

>> because then they become a bigger part

19:08

of the population and you can then ask

19:09

who do they vote for and what are they

19:11

doing and this is becomes ultimately the

19:12

risk namely meaning risk from a upside

19:15

down side. But what is exactly the

19:16

outcome when you have that decap

19:18

situation is unfortunately continuing.

19:20

>> Let's switch to private credit.

19:23

>> Hi, Steve Eisman here. When my kids were

19:25

little, I woke up one morning in a

19:26

panic. I realized that if something

19:28

happened to me, my family would be

19:29

completely unprotected. I immediately

19:31

went out and got life insurance so that

19:33

my family would be financially secure

19:35

and I have never regretted that

19:36

decision. Stop procrastinating. Fabric

19:39

by Gerber Life makes it so quick and

19:41

easy to make sure your family has more

19:43

financial protection this year. Fabric

19:46

by Gerber Life is term life insurance

19:48

you can get done today. Made for busy

19:50

parents like you. All online on your

19:53

schedule right from your couch. You

19:54

could be covered in under 10 minutes

19:56

with no health exam required. If you've

19:58

got kids, and especially if you're young

20:00

and healthy, the time to lock in low

20:02

rates is now. Fabric by Gerber has

20:05

partnered with Gerber Life. Trusted by

20:07

millions of families like yours for over

20:09

50 years. There's no risk. There's a

20:11

30-day money back guarantee, and you can

20:13

cancel at any time. They have over 1,900

20:16

five-star reviews on Trust Pilot with a

20:18

rating of excellent. Join the thousands

20:20

of parents who trust Fabric to help

20:22

protect their family. Apply today in

20:25

just minutes at meetfabric.comisman.

20:28

That's

20:28

mefabric.comisman.mefabric.com/isman.

20:34

Policies issued by Western Southern Life

20:37

Insurance Company not available in

20:39

certain states. Prices subject to

20:41

underwriting and health questions.

20:44

Starting something new isn't just hard,

20:46

it's terrifying. So much work goes into

20:48

this thing that you're not entirely sure

20:50

will work out. And it can be hard to

20:52

make that leap of faith. Trust me, I

20:54

know. But now I know that I was right in

20:56

believing in myself and launching my

20:58

podcast business despite all the fears

21:00

and hesitations. It also helps when you

21:03

have a partner like Shopify on your side

21:05

to help. Shopify is the commerce

21:08

platform behind millions of businesses

21:11

around the world and 10% of all

21:13

e-commerce in the US, from household

21:15

names to brands just getting started.

21:17

Accelerate your efficiency. Whether

21:19

you're uploading new products or trying

21:21

to improve existing ones, Shopify is

21:23

packed with helpful AI tools that write

21:26

product descriptions, page headlines,

21:28

and even enhance your product

21:29

photography. And did I mention that

21:31

iconic purple Shopay button that's used

21:34

by millions of businesses around the

21:36

world? It's why Shopify has the best

21:38

converting checkout on the planet. It

21:40

also helps boost conversions, meaning

21:42

less carts going abandoned and more

21:44

sales for you. It's time to turn those

21:47

whatifs into with Shopify today. Sign up

21:51

for your $1 per month trial today at

21:53

shopify.com/isman.

21:55

Go to shopify.com/isman.

21:58

That's shopify.com/isman.

22:03

Talk to me about from an economist

22:06

perspective, what does the growth in

22:08

private credit mean or not mean? And

22:12

what do you think of the the

22:15

overindexing of private credit to

22:17

software?

22:19

>> So let's back up and think about exactly

22:21

why private credit and why the financial

22:23

system has changed so much since 200

22:25

because dod frank as you know better

22:27

than anyone was implemented and that

22:29

meant that the banks were essentially

22:30

asked to do less and the market was

22:33

asked to do more. So let's just agree

22:35

that this was the rules changes that

22:36

came after the

22:38

>> there's no question that's what

22:38

happened.

22:39

>> This is what happened. This is where we

22:40

are today. Right?

22:41

>> What are the consequence of this? Well,

22:42

where we sit right now, if I just back

22:44

up and you ask me as a macroeconomist,

22:46

what's the situation in credit? Well, if

22:48

you look at default rates in loans and

22:50

in high yield, they have actually been

22:52

going down for the last 12 months. If

22:54

you look at distressed exchanges,

22:55

meaning that I borrow $100 from you, I

22:58

come back two years later and say,

22:59

"Sorry, I can't pay you back." You can

23:01

then decide to say, "It looks like

23:02

you're in distress. Why don't you

23:04

instead give me some equity in your

23:05

business? Why don't we instead extend

23:06

the maturity of the loan? Why don't we

23:08

lower the interest rate of the loan?

23:09

let's make a deal and we will change

23:11

your capital structure because you were

23:13

not able to survive higher interest

23:14

rates. Let's try to find a solution out

23:16

of this. But distress exchanges have

23:18

also been going down. So not only are

23:20

default rates going down, distress

23:21

exchanges are going down and finally

23:22

liability management exercises are also

23:24

going down. So let's agree at the

23:26

highest level credit is actually getting

23:29

better because default rates are going

23:30

down, distress exchanges are going down,

23:32

LME are going down. So what is the

23:33

problem in credit? And this gets exactly

23:35

to your question. The problem is in

23:37

credit that there are some sectors in

23:39

credit that have much higher leverage

23:41

and much lower coverage ratio. And one

23:44

sector that absolutely stands out is

23:46

>> just for for viewers because not

23:47

everybody knows what is coverage ratio

23:49

mean.

23:49

>> The leverage means of course how much

23:51

debt you have in your business. And

23:52

coverage ratio means what are my

23:54

earnings divided by my debt servicing

23:56

cost. In other words, what is my the

23:57

coverage ratio?

23:58

>> What's my ability to pay my interest?

23:59

>> Exactly. Bottom line, am I able to pay

24:00

my debt? Am I not able to pay my debt?

24:02

And if you do a simple scatter diagram

24:04

of all the sectors in credit or really

24:06

all sectors in the economy and ask which

24:08

sectors have a lot of debt, which

24:09

sectors have little debt. So that's a

24:11

very important exercise in equities and

24:12

in credit to understand are these

24:14

sectors highly levered, are they able to

24:16

pay their debts, where are their rate

24:18

mean where is their ability to pay their

24:19

debts relative to other sectors and

24:21

software stands out in a very

24:23

significant way by having significant

24:25

amounts of debt and at the same time

24:27

having actually very little ability to

24:29

service that debt. So even before

24:31

>> and it has very little ability to

24:32

service that debt because

24:34

>> because interest rates are now higher

24:35

for longer because Kevin Walsh is about

24:37

to raise interest rates later.

24:38

>> So the coverage ratio is deteriorating.

24:39

>> So the coverage ratio is deteriorating

24:41

because these companies are

24:42

unfortunately not being helped by the

24:44

Fed hiking rates. Got it.

24:46

>> So that's why if you now even before you

24:48

and I begin to talk about AI disruption,

24:49

we can talk about AI disruption. Some

24:51

software companies in cyber security may

24:53

be better. Some software of companies in

24:55

education may be worse. So from that

24:57

perspective, there are some nuances. But

24:59

the big picture is all software

25:00

companies across the spectrum they have

25:02

very high levels of debt and very little

25:05

ability to service that debt especially

25:07

now that Kevin Walsh is about to raise

25:08

interest rates. So therefore it becomes

25:10

important to look at the maturity wall.

25:12

A lot of these vintages in software were

25:14

originated in 2021 and 22 and a lot of

25:17

these vintages have a 7-year maturity on

25:20

their debt. That means that exactly in

25:21

2028 and 29 we are running into the

25:24

maturity wall for software. That means

25:26

that

25:26

>> so there's some in two 20 2027

25:29

>> a

25:29

>> little bit but most of it is 2 and 29 28

25:32

and 29. So that means if Kevin wars is

25:34

not lowering interest rates before we

25:36

get to 2028 these companies will have

25:39

significant problems rolling over their

25:40

debt. So even before we debate whether

25:42

AI disruption is going to create a

25:44

terminal value of software companies

25:46

that's very low. We already have the

25:48

macroeconomic problem that when interest

25:49

rates are higher for longer because

25:50

inflation is higher for longer, the

25:52

sectors that have a lot of debt and the

25:54

sectors that have little ability to

25:55

service that debt will continue to

25:57

struggle and software unfortunately

25:58

stands out as the number one sector

26:00

that's vulnerable in that environment.

26:01

So that's why yields on loans and

26:03

software continue to trade higher and

26:05

higher. At the moment, you and I can

26:07

take $100 and put into software loans

26:09

and get 12%. I

26:10

>> mean, if we buy it in the open market,

26:12

we're selling it less than par.

26:14

>> Absolutely. So if we buy in the open

26:15

market, you and I could basically say we

26:17

get 12% return in software loans. I mean

26:19

12% that's pretty juicy in any

26:20

investment. But the reason why the

26:22

market still trades that wider and wider

26:23

is moving up towards 12 a.5% is that if

26:26

these companies have a terminal value

26:27

that's zero and at the same time they

26:29

also are not able to roll over their

26:30

debt then they will be facing

26:32

significant headwinds. It's a double

26:33

whammy to software coming from the

26:35

terminal value being questioned and at

26:37

the same time rates higher for longer

26:38

meaning that they're not able to service

26:40

their debt. So the bottom line is there

26:42

is one sector and in particular that

26:44

sector alone there's also some parts of

26:46

healthcare small parts of consumer

26:47

services but the software sector really

26:49

stands out as the number one problem in

26:51

credit and this gets back to what you

26:52

asked about namely that in direct

26:54

lending or in private credit which is a

26:56

$2 trillion market $500 billion of

26:59

private credit is software that was

27:01

originated in the last six seven years.

27:03

So for that reason, software is a

27:05

significant part of private credit and

27:07

is a significant part of public credit

27:09

and it is those parts of the credit

27:10

market that are wrestling with these

27:12

problems of rates higher for longer and

27:13

the terminal value whereas the rest of

27:15

the credit market very broadly speaking

27:17

is actually in good shape exactly

27:19

exemplified by the fact that default

27:20

rates are going down, distress exchanges

27:22

are going down and LME are going down.

27:24

You know, 500 billion sounds like a big

27:26

number, but would you agree that in the

27:28

context of the US economy, which is a $

27:31

31 trillion economy?

27:34

Eh, it's not. In other words, for the

27:36

people who made these loans, God help

27:39

you.

27:39

>> Exactly.

27:40

>> But would you agree that for the the

27:42

implications for the US economy overall

27:44

are not so bad?

27:45

>> Yeah. Because what's also important back

27:46

to this upper crisis, and again, you

27:48

know, much better than anyone, is of

27:50

course that it all becomes a question of

27:51

where are these loans located,

27:53

>> right? are they on very levered balance

27:54

sheets and the generally speaking the

27:56

banking sector of course as you know

27:57

better than anyone was like levered 20

27:59

30 times at the time in the GFC

28:01

>> 40 in some cases 40 and of course now

28:04

you have the BDC's by law are only lever

28:07

twice two to one

28:08

>> yeah so that means that of course if it

28:09

is even in BDCS and if this is in

28:11

pension funds around the world if this

28:12

is insurance companies around the world

28:14

then of course this is indeed a smaller

28:16

number and therefore not a magnifier the

28:18

way that we saw during the GFC when

28:20

subprime was located in balance sheets

28:22

that had to deal very very quickly. So

28:24

in other words, and this might be the

28:26

way that the Fed is thinking about it,

28:28

some people make investments and lose

28:29

money. Some people make investments and

28:30

make money. And to your point,

28:32

>> you're a big boy.

28:33

>> Exactly. You have some losses, you have

28:35

some gains. And 500 billion, yes, it's

28:37

not, of course, it's a huge number in

28:39

some dimensions, but from a

28:40

macroeconomic perspective, the US

28:42

economy is like 33 trillion GDP. Yes,

28:44

it's makes some importance, but it's not

28:47

anywhere near uh those systemic levels

28:49

that we had with subprime in 2006 and 7.

28:52

So I actually think the software story

28:54

is even worse than what you're saying.

28:56

Not not from a macro perspective, but

28:58

just from from a micro perspective in

29:00

the sense that if you look at any all

29:01

the public software companies, you know,

29:04

like Salesforce, Service Now, they're

29:06

down 50 60% from from their peaks. So,

29:10

if you're the lender

29:13

and it's 2028

29:15

and the loan is now due, the discussion

29:19

is not just, well, you're a riskier

29:21

company. I want to charge you more

29:23

interest. You're you're going to the

29:24

private equity owner of the company and

29:27

you're saying, dude, the value of your

29:30

equity is is basically gone. you got to

29:33

pony up more money. Otherwise, we ain't

29:34

going to make we not we're not going to

29:36

we're not even going to have a

29:36

discussion about lending you, you know,

29:39

rolling over your loan until you pony up

29:41

more equity. And then the question is if

29:43

the private equity has to decide whether

29:44

they want to do that or not.

29:45

>> Absolutely. And we've seen some examples

29:47

of this more recently, of course, but

29:48

you're right. It's absolutely the case.

29:50

This is page one in your finance

29:51

textbook, right?

29:52

>> This is private credit. You are senior

29:55

to, of course, private equity. And if

29:57

you are the equity in these businesses

29:59

and if private credit in software is in

30:00

trouble then of course private equity in

30:02

software is almost in even more trouble.

30:04

>> Let's move on to um just general

30:07

employment. How's the what's the

30:08

employment situation like in the United

30:10

States?

30:10

>> What's remarkable about the discussion

30:12

we had earlier about AI? There's so many

30:14

stories being told about mass

30:16

unemployment going up 10 20%. People

30:18

freaking all losing our jobs. But there

30:20

are two important dimensions of this in

30:22

the data at the moment. Number one is

30:24

non-farm payrolls continues to be

30:26

incredible. Why is that incredible?

30:28

Because we have the tailwinds from the

30:30

AI spending, the one big bill for bill

30:31

and the industrial renaissance. But it's

30:33

also the case that it's incredible

30:35

because clearly AI is not resulting in

30:38

mass unemployment. So the first

30:40

conclusion is we are still creating a

30:42

lot of jobs in this economy and this is

30:43

despite that immigration has been

30:45

slowing down. Remember it was the case

30:46

in 2022, 23 and 24 that net immigration

30:49

legal and illegal into the US was 3

30:51

million people came in every single year

30:53

to the US. Now

30:54

>> 3 million

30:55

>> 3 million people every single year.

30:56

Today basically net immigration is zero.

31:00

>> That has resulted in our friends at the

31:02

Federal Reserve putting out working

31:03

papers, blog posts saying, "Well, hold

31:04

on. If we're not having 3 million coming

31:06

in every year, non-farm payrolls has

31:08

dropped from when it was 3 million, it

31:10

was 200,000 every month. Now they break

31:12

even for non-farm payrolls according to

31:13

the Dallas Fed is 30,000. In other

31:16

words, a dramatic drop in the number of

31:18

jobs because we simply have much fewer

31:19

people coming into the country. So from

31:20

that perspective, 172,000 in job growth,

31:24

100,000 job growth is a phenomenal

31:26

number, way, way higher than the numbers

31:28

that you would be getting if you just

31:29

looked at the demographics alone. So

31:31

that's why the labor market is actually

31:32

in really really good shape which is

31:34

likely also again back to the reason why

31:36

Kevin Moss and the FOMC is worried about

31:38

maybe they have to hike rates is because

31:40

it's not only inflation is three three

31:41

and a half but it's also the fact that

31:43

we have a labor market they're actually

31:44

quite strong. Even if you look finally

31:46

at another indicator of the labor market

31:48

let's look at the unemployment rate for

31:49

people that are between 20 and 24. It's

31:51

been getting a lot of attention that

31:53

young people can't find a job. Young

31:54

people

31:55

>> it's hopeless.

31:55

>> It's hopeless. This is the anecdote in

31:57

the urban myth including here in the

31:59

streets of Manhattan when you hear this

32:00

story at the moment. But if you actually

32:02

look at the BLS data for the

32:03

unemployment rate for people between 20

32:05

and 24 years old, it has actually gone

32:07

down in the last 6 months and it's gone

32:09

down more than the aggregate

32:10

unemployment for everyone else. So maybe

32:13

we have many more dorm room

32:14

entrepreneurs that are sitting at home

32:15

inventing new businesses and they're

32:17

much more solo entrepreneurs, individual

32:19

people who basically now have access to

32:21

tools in AI, access to loops, access to

32:24

agents, access to chat GBT to basically

32:27

start a new business. And I am of the

32:29

strong view that because of that the

32:32

labor market is actually benefiting from

32:34

if a fraction of all the new businesses

32:35

that are creating at the moment which by

32:37

the way is at the highest level ever in

32:38

US history in the weekly data from the

32:40

census we have never seen so many

32:42

businesses being created as we're seeing

32:43

at the moment if a fraction of them are

32:45

successful they will also create

32:46

employment. So if I didn't get a job at

32:48

a bank or in consulting or legal

32:50

services why don't you and I coming out

32:52

of college open a new business together

32:54

100%. And that's become easier than ever

32:56

before. So that's why I think that yes

32:58

there is a net displacement effect in

33:00

particular in t marketers and others

33:02

where people might be losing their jobs

33:03

because of AI at current rates about

33:05

100,000 people are losing their jobs in

33:07

t marketing but at the same time the net

33:09

effect of that is relatively small

33:10

compared to the hundreds of thousands of

33:12

people who are basically out there

33:14

inventing new things and coming up with

33:15

new businesses. We get a much more

33:16

dynamic capitalist economy as a result

33:19

of AI and we should all be very excited

33:21

about this.

33:22

>> So let me go on a little bit of a

33:24

tangent off of what you just said. So

33:26

what you're you're you're saying is that

33:28

the US economy is incredibly dynamic.

33:31

>> Exactly.

33:32

>> I I I would say the US I mean hopefully

33:34

AI lasts lasts forever and we're all

33:37

great but I mean we'll be a year from

33:39

now you'll be back and we'll have

33:40

another discussion about it but as of

33:42

now the US economy I I would say is more

33:44

dynamic than it's ever been in it maybe

33:46

in its history or certainly for a very

33:48

very long time.

33:48

>> I agree. So I worked at the OECD in

33:50

Paris which looks at structural issues

33:52

in economies and they sometimes point

33:54

out that the health care system has some

33:55

challenges in the US. There's some other

33:57

challenges with pensions and other

33:58

things but broadly speaking the number

34:01

one indicator in all OECD work that

34:03

looks at what are dynamic economies is

34:05

that is it easy to fire and hire workers

34:08

in France and Germany it's incredibly

34:11

complex to fire and hire workers and the

34:13

US has the most dynamic labor market.

34:15

That's good. Of course, if you're an

34:17

employer and if you are a good worker in

34:19

your job, it's actually also good for

34:20

you and me. So, in that sense, a very

34:22

dynamic labor market is a critical part.

34:24

A very competitive product market is a

34:26

very critical part. And perhaps most

34:27

importantly, a financial system that's

34:29

willing to finance risk is also not what

34:32

we have in Europe. Unfortunately, we

34:33

don't have in Japan.

34:34

>> That's actually my my my my tangent,

34:37

which is you're starting to answer. So

34:39

let me my get my question in which is

34:42

>> why is Europe so incredibly sclerotic in

34:46

terms of its I mean it's it's almost an

34:48

embarrassment. I mean I was looking at

34:50

statist I I had a um a guest on um last

34:54

year who who wrote I would recommend

34:56

this book to you. It's called Kaput the

34:58

end of the German economic miracle by

35:00

Wolf Gang Munch. He's an excellent book

35:02

and and we're going to have him back on

35:04

soon again. But I was looking at um so

35:07

because of I I interviewed him and I

35:09

read the book. I always try and keep up

35:10

like what's going on in Germany.

35:12

>> Yeah.

35:12

>> German GDP hasn't grown a dollar in like

35:15

the last three years. This like what is

35:18

going on in Europe?

35:19

>> Six months are going unfortunately

35:20

further down. Negative.

35:21

>> So the answer to that question is

35:22

exactly the things we just talked about

35:24

namely the three areas where Germany

35:26

unfortunately still needs to do a lot of

35:27

homework. Number one it is still very

35:30

difficult to hire and fire workers in

35:32

Germany. It's hard to hire.

35:34

>> Also difficult to hire.

35:35

>> Why? Explain that to me. How just

35:38

mechanically why would it be hard hard

35:40

to hire someone?

35:41

>> Because if you once you hire someone, if

35:43

you turn out that you hired me for a job

35:45

and you send well this guy is not really

35:48

working, it's really difficult for you

35:49

to get rid of me again

35:50

>> because then you need to go through e

35:52

metal, the trade unions, their organized

35:54

systems. And France is the extreme of

35:56

this case. Namely, you can't even get on

35:57

permanent contracts. You have to be on

35:59

temporary contracts. That creates all

36:00

these dual labor markets. Unfortunately,

36:03

Europe and Germany and France are at the

36:05

peak of this on meaning in a bad way

36:07

that it's just become still very

36:08

difficult despite that we're sitting

36:10

here in 2026 to hire and fire workers.

36:12

That means that if you and I have a good

36:13

idea and we want to open a business and

36:15

we say let's go out and hire some people

36:16

to help us open this business.

36:17

>> We don't want to do it in Germany.

36:19

>> Reluctant to do that because you say if

36:20

I hire this person, I can't get rid of

36:22

them again. If we do see a slowdown in

36:23

demand, right? That's very different

36:25

from the US. If we go out and hire

36:26

someone here in New York City, well, if

36:28

our business does great, we can go and

36:29

hire a lot more people. We may have to

36:31

pay for them. But if we have some

36:32

problems of course then we have to fire

36:33

these people quite quickly and we can do

36:35

that in the US. So the labor market is

36:37

just very rigid and the product market

36:39

is also very rigid in Europe including

36:41

in Germany. There are issues of course

36:42

also when it comes to product market

36:44

competition there we see as indicators

36:46

comparing competition in the US relative

36:48

to Germany and Europe and it's also the

36:50

case that it's not very competitive

36:51

there all kinds of monopolies there all

36:53

kinds of problems with pricing there's

36:55

also all kinds of problems with tariffs.

36:56

So a lot of things are also making

36:58

product markets less competitive. And

36:59

finally financial markets. Unfortunately

37:01

to your point if you think about the

37:04

sclerotic situation in the US sorry

37:06

European financial system there are

37:08

basically traditionally people talk

37:10

about the European financial system as

37:11

bank-based and the US system is market

37:14

based. Correct.

37:15

>> And we want the European system to also

37:17

be market based because think about it

37:19

you and I a company in Germany. We would

37:20

like to borrow some money. We can go to

37:22

a bank in Germany or in France and we

37:24

would like to borrow some money. And if

37:25

they say yes, it's great. If they say

37:27

no, we really have other places to go.

37:30

But if you and I go to a bank here in

37:32

Manhattan and say we like to borrow some

37:33

money and they say no, you and I will

37:34

say great. We have some good friends in

37:35

venture capital. We have some good

37:37

friends in private equity. We may have

37:38

some good friends in private credit. We

37:40

could also do IPO. We could also do

37:42

various things when it comes to

37:43

borrowing in even secondary markets. So

37:46

the financial system is just not very

37:48

diverse in Europe, unfortunately. And

37:51

that's a problem for Europe that they're

37:52

still working on the capital market

37:53

union on the financial system generally

37:56

being able to provide more riskwing

37:58

capital the way that we have. Go to

38:00

Silicon Valley and you can get money for

38:02

just a piece of paper on a a very and

38:04

simple idea. So that means that in the

38:06

European situation, we just have

38:08

unfortunately much more red tape, much

38:10

more regulatory complex environments and

38:12

the financial system is just not very

38:14

good at allocating money to a lot of

38:16

good ideas. And that's why unfortunately

38:18

for the Europeans, a lot of Europeans go

38:20

to Silicon Valley, come to New York City

38:21

to basically say I would like to borrow

38:23

some money here rather than borrow and

38:25

do my little business in the Euro area.

38:26

And then fortunately the consequence is

38:28

that a lot of growth is literally all

38:29

good ideas are coming to the US and

38:31

that's what is the main problem there.

38:33

some ideas and some corners of Europe is

38:34

moving a little bit in the right

38:35

direction. But the big answer to your

38:37

question is that it's difficult to hire

38:38

and fire. The product markets are not as

38:40

competitive as in the US and the

38:41

financial system unfortunately is not as

38:43

diversified. It doesn't provide the same

38:46

type of resources available to people

38:48

who have a good idea like we have in the

38:50

US.

38:51

>> Do you think there's a growing

38:52

recognition in Europe that this is a

38:54

problem or not really?

38:56

>> So the drugy recommendations

38:58

>> Okay. So I I'm going to challenge you on

39:00

that. Okay. So, so before I I found Wolf

39:03

Gang last year,

39:05

>> I when I was starting my podcast, I took

39:07

out a piece of paper and I and I wrote

39:10

down

39:11

all the topics I want wanted to do on my

39:14

podcast. And one of them was why is

39:16

Europe so bad? Yeah.

39:17

>> And so then I started looking around for

39:20

for something to read and um friend of

39:23

mine put me on to Mario Draggy's white

39:26

paper. So I it's 100 pages long.

39:29

>> Yeah.

39:30

and I started to read it and by the time

39:32

I got to page 10 I was asleep because he

39:36

his his paper basically said we have a

39:39

problem but I don't want to upset

39:41

anybody about and talking about the

39:43

problem and and so I said this is

39:45

ridiculous and eventually I found Wolf

39:47

Gang's book which which I much more

39:49

helpful so if if that's what everybody

39:51

points to is the draggy white paper it's

39:53

hopeless

39:54

>> I know he did so he was commissioned to

39:57

write a white paper or a report and say

39:59

what do we need to see can you come with

40:01

some specific policy proposals and he

40:03

came with basically 200 different things

40:05

that he wanted to see changed so that's

40:07

why there are now institutions including

40:09

bugal in Brussels which is a think tank

40:10

basically similar to Brook kings in DC

40:13

and they basically tried to track of all

40:15

the things that he suggested now they're

40:17

almost two years ago how many of these

40:18

things have been implemented and the

40:20

answer is this is now two years ago and

40:22

of all his proposals only 10% in round

40:25

numbers have been implemented so yes it

40:28

is it's not quite falling asleep, but it

40:31

really the speed with which the

40:32

Europeans are moving. So, I both have a

40:33

European and US passport to be clear,

40:35

but the speed with which the Europeans

40:36

are moving is just not very impressive

40:38

and it's not helping themselves.

40:40

>> They're not panicked.

40:41

>> They're not helping themselves that

40:42

they're not doing their own homework and

40:44

it's very unfortunate because they

40:45

absolutely need especially with this new

40:47

situation that China is also leading on

40:49

AI and that's beginning to become an

40:51

issue also of course for the sector.

40:53

Absolutely. And that's why if you now

40:54

have that anyone who has an AI and D

40:56

idea in Europe actually goes to the US

40:59

then again they're not helping

41:00

themselves. I think they are waking up a

41:02

little bit. Of course they woke up a lot

41:03

on defense for a number of different

41:05

reasons but I think they're also

41:06

beginning to wake up more on AI. But

41:08

that's why from an Apollo perspective,

41:09

we need financing, a lot of strategic

41:12

financing for the industrial

41:13

renaissance, not only the US, but also

41:14

in the European case for defense, for

41:17

infrastructure. Exactly. For data

41:18

centers, for things that require

41:20

financing to make sure that Europeans

41:22

also can catch up and continue to be

41:23

competitive in the global economy.

41:25

>> Let's switch gears one more time. Let's

41:27

talk about the US deficit. So I have my

41:30

own views about this but I'd be curious

41:32

as to yours which and let me just intro

41:34

in introduce the concept in this

41:36

wonderful deck you point out that uh

41:39

federal US debt to GDP is around like

41:43

100% or so when it's going to 175%.

41:46

You know when you watch CNBC not a week

41:49

goes by that somebody doesn't come on

41:51

and and does what I like to call virtue

41:53

signaling when it comes to the deficit.

41:55

Meaning I am so against the deficit.

41:58

your guest last week, he said he was

42:00

against the deficit, but I'm much more

42:02

against the deficit than him. And and

42:04

and each each guest strings out this um

42:08

disaster scenario, which by the way,

42:10

Pete Peterson strung out 40 years ago.

42:12

>> Yeah.

42:13

>> So,

42:14

>> what's fact, what's fiction? What do you

42:17

think? What's really interesting about

42:19

that discussion is absolutely we have an

42:21

enormous budget deficit and of course we

42:23

have significant deficits every year.

42:26

The government deficit at the moment is

42:27

about 5%. And we have debt levels that

42:30

of course continue to just go up

42:31

literally since 1776. We are entering a

42:34

period where we'll have the highest

42:36

level of debt for the government ever. I

42:38

mean in US history. So let's just start

42:39

out by concluding that the trend in this

42:41

is not our friend. This is a major

42:43

challenge. So now this becomes important

42:45

because the question is of course well

42:47

why are interest rates then still so

42:49

relatively low?

42:50

>> Yes.

42:50

>> And the answer is that the rest of the

42:52

world is still buying a lot of US

42:55

assets. Importantly, they're still

42:56

buying a lot of US treasuries. The rest

42:58

of the world, by the way, is also still

43:00

buying a lot of US credit. And the rest

43:01

of the world is also still buying a lot

43:03

of US equities. And why is that? That's

43:05

because, back to what we spoke about

43:06

before, if you are a pension fund in

43:08

Europe, you have to be invested in AI,

43:12

you must be invested in the US. So,

43:14

pension funds in Europe have significant

43:15

allocations in dollars to US AI. If you

43:18

are pension fund in Europe, you see your

43:20

own interest rates at relatively low

43:21

level. You see higher returns in the US.

43:23

you again say I got to also allocate

43:26

more to the US that also helps finance

43:28

US deficits because the level of

43:30

interest rates is simply higher in the

43:31

US than it is in all European countries.

43:34

So the reason why this is still able the

43:37

government is still able to finance the

43:38

deficit is that there is an incredible

43:40

willingness especially among foreigners

43:42

to still buy US government debt and buy

43:44

US credit and also buy US AI meaning

43:47

stocks and other products of course that

43:49

gives you AI exposure. The first answer

43:51

to your question is there's a remarkable

43:53

willingness especially among foreign

43:55

investors are buying US government debt

43:57

because the low interest interest rates

43:59

is higher and that's of course helping

44:00

when you want to cut coupons and you are

44:02

a pension fund or insurance company in

44:04

Japan in Europe in Taiwan and of course

44:06

also in Canada. So now here's the other

44:08

side of the problem. If you look at the

44:09

domestic investors

44:10

>> I haven't heard a problem yet. I've I've

44:12

only heard that it's not a problem.

44:15

>> So far we have the foreigners are happy

44:16

to come to the US with money. But the

44:18

problem now is in the US that there are

44:20

two problems. Both when it comes to

44:22

institutional demand for treasuries and

44:24

also when it comes to retail, meaning

44:26

household demand for treasuries.

44:28

Remember, normally if you are a pension

44:30

fund and insurance company in the US,

44:32

you would have some 30-year liabilities.

44:33

You need a 30-year asset. Historically,

44:35

you would say, "I'm buying US treasuries

44:37

because that matches my 30-year

44:39

liabilities in my insurance company."

44:40

But today, insurance companies and

44:42

pension funds are not buying US

44:43

treasuries. They are buying privately

44:45

issued longduration assets. They're

44:47

buying privately issued longduration as

44:48

in data centers, infrastructure,

44:50

climate, energy transition, you name it,

44:52

longduration assets that have a better

44:54

risk return profile. That means that

44:56

from an asset allocation perspective,

44:58

pension insurance in the US has been

45:00

moving towards privately issued

45:01

longduration assets instead of buying

45:03

longduration US treasuries. That's a

45:05

challenge. That's a headwind. That's why

45:07

the of course market is worried about

45:09

that the Treasury and the Tback, the

45:11

Treasury Bing Advisory Committee is at

45:13

risk that if they issue more

45:14

longduration assets, then there will not

45:16

be enough demand. So that's the

45:17

institutional side has been switching

45:18

towards privately issued longduration

45:20

assets and the retail the household side

45:22

has been also switching in the last

45:24

several years away from instead of

45:27

buying longduration US treasuries in

45:28

ETFs their flows have continued to go

45:30

down instead households are now buying

45:33

money market funds and short duration

45:35

government bonds so that's another way

45:37

of saying why do you think that is

45:38

>> because the yield curve is a lot flatter

45:40

now and you suddenly get a very high

45:42

return when the Fed keeps rates higher

45:43

for longer

45:44

>> in other words you're getting enough and

45:46

the short end of the curve if you're a

45:47

household. Why do I need to buy

45:49

something 30 years out?

45:50

>> And and in response, the Treasury both

45:52

under Janet Jillen and under Scott

45:54

Besson have been issuing much more T

45:55

bills because hey, now there's all this

45:57

demand from households to buy short

45:59

duration assets. So that's why now

46:00

households are willing to cut coupons in

46:02

the very front end. So there's a

46:03

different way of saying in summary that

46:05

for a number of different reasons,

46:07

there's less appetite for the long end

46:08

from institutions because they're now

46:10

buying other privately issued

46:11

longduration assets. And there's also

46:13

less demand from households because I

46:14

get less out of buying long duration US

46:16

treasuries. If I can cut coupons in T

46:18

bills, that basically gives me a return

46:20

that's also quite decent. So that's why

46:22

the challenge at the moment is that when

46:24

the debt level continues to move higher

46:25

and higher and higher, we run into the

46:27

risk of course that at some point then

46:29

the Treasury needs to think about where

46:31

on the curve are we issuing and there's

46:33

just less and less institutional demand

46:34

in the long end, less demand from

46:36

households in the front end. It's only

46:38

really the foreigners that have been

46:39

holding up demand in a very substantial

46:41

way. Especially private investors,

46:43

foreigners have been kick cutting

46:44

coupons and putting money into the front

46:46

end. So that's why if you segment who

46:48

the different players are in the

46:49

treasury market, it used to be that it

46:51

was China which was not interest rate

46:52

sensitive but all these entities namely

46:55

foreigners and institutions and

46:56

households they are very interest rate

46:58

sensitive. So we have a situation where

47:00

you could worry about a spring coil

47:01

effect where everyone is saying great

47:03

rates are high, rates are high, so now

47:04

I'm plowing money into treasuries. But

47:06

if the Fed succeeds with cutting rates a

47:07

lot, then foreigners might not be buying

47:09

so much, households might not be buying

47:10

so much. And suddenly the interest rate

47:12

sensitivity will become a very important

47:14

part of why there is a risk that the US

47:16

government deficit cannot continue and

47:17

the US government debt level can

47:19

continue to be at these very very high

47:21

levels.

47:22

>> How worried are you about this

47:23

>> at this point? Because the AI boom

47:25

continues and at this point because

47:27

rates are higher for longer and the Fed

47:28

is about to hike rates. I'm not worried

47:30

about this. Definitely not this year.

47:32

But I am worried about the dynamics that

47:34

we have shifted from Chinese being not

47:36

an interest rate sensitive buyer to now

47:38

having these different groups of much

47:39

much more interest rate sensitive

47:41

buyers. And by the way, the basis trade

47:42

and hedge funds have also been

47:44

benefiting a lot of course from some of

47:45

these developments. That also means that

47:47

if these new entities or buyers suddenly

47:49

are much more interest rate sensitive.

47:51

If we do get a situation where the Fed

47:52

will have to cut rates dramatically down

47:54

to zero, then suddenly there might be

47:56

much more risk involved with treasuries

47:58

because now we suddenly have a much

47:59

bigger group of investors who are much

48:01

more interested in what is actually the

48:03

yield that I get on this investment that

48:04

I that I'm doing relative to when it was

48:06

China where it was purely done for FX

48:08

reasons to protect the exports and not

48:10

so much with consideration to what the

48:12

level of interest rates were at. So the

48:13

answer is I'm not worried about that

48:15

over that the next several years I still

48:17

think we'll be okay but it's very clear

48:19

that the trajectory that we are on as J

48:21

Palway always was saying and Jenna and

48:23

Benanken that is an unsustainable

48:24

trajectory and at some point this will

48:27

come home to roast and be something

48:28

that's important for financial markets

48:29

but we're just not quite there yet.

48:32

Let's just quickly about China and then

48:34

about big risks. Um is there any risk

48:37

that China ever dumps treasuries?

48:39

>> Well the issue of course is that this

48:41

has been getting a lot of attention.

48:42

China used to have at the peak $1.3

48:44

trillion in US treasuries. Now they're

48:46

down closer to around 700 billion. So

48:48

China has already been offloading

48:50

treasuries over the last 5 years. So

48:53

there's already a development where

48:54

China is more gradually lowering their

48:56

holdings of treasuries for a number of

48:57

different reasons. Now they have less

48:59

trade directly with the US. They trade

49:01

more with others which is not in

49:02

dollars. So there's a number of

49:03

different dimensions to why that's been

49:05

happening. But in short, if they were to

49:08

do that, the risk of course would be

49:09

that the US economy, if you really saw a

49:11

significant spike in long-term interest

49:13

rates would begin to slow down very very

49:15

hard. And if this slowdown would be very

49:17

hard in the US economy, that will also

49:18

begin to hurt therefore Chinese exports

49:20

to the US. So that's why they are

49:22

probably having a strategic

49:24

consideration. Yeah. Because they don't

49:26

want to slow their own economy. They

49:27

still depend importantly on exports to

49:28

the US. although they have been

49:29

diversifying away to Europe and other

49:31

emerging markets then they are generally

49:33

not interested in slowing and crashing

49:35

the US economy because that would also

49:37

result in much less demand from you and

49:39

me and others in the US buying Chinese

49:41

goods. So I take that they're probably

49:43

trading very carefully when they think

49:44

about how they want to think about that

49:46

topic.

49:46

>> Okay, let's just finish up with from an

49:49

economist perspective what do you think

49:50

the biggest risks in the market are?

49:52

Well, I think one thing that is very

49:54

important in markets at the moment is

49:56

that AI has absolutely turned out to be

49:58

almost everywhere. If you and I think

50:00

about the 6040 portfolio and I think

50:02

about 60% equities, 40% fixed income.

50:04

Let's talk about what is in my equity

50:06

first. Well, the S&P 500, the 10 biggest

50:09

stocks now make up 42% of the index. So,

50:11

let's just agree that returns for the

50:13

last 5 years, basically half of it has

50:15

been coming because of AI. So AI plays a

50:17

very important role in my returns in

50:19

equities have played for the last

50:20

several years and at the moment have

50:22

such a big weight that it continues to

50:23

be a huge bet if I put money into the

50:25

S&P 500. So the first conclusion is

50:27

let's just agree there's one factor

50:28

playing out in AI is the key factor in

50:31

equities. But even now in fixed income

50:33

in credit because the hyperscalers are

50:36

issuing so much debt that means that the

50:38

IG index is changing. It used to be that

50:40

IG was government bonds. IG is

50:42

investment grade

50:43

>> investment grade credit and it also used

50:45

to be banks. Those were the two main

50:47

components. There's also a little bit

50:48

industrials but mainly banks and also

50:50

government bonds but now there's a new

50:52

player in investment grade credit and

50:53

that is hyperscalers that are issuing

50:56

$700 billion in debt this year. That

50:58

means that AI is suddenly also becoming

51:00

a very important part. So that means

51:01

that in my 40, not only do I have a lot

51:03

of AI in my 60 in my 6040 portfolio, but

51:05

I also have a lot of AI in my 40. And

51:07

finally, if you and I also put money in

51:09

venture capital, venture capital used to

51:11

be pharma, biotech, prescription drugs,

51:13

new medical products, but now 87% of

51:16

venture capital is also AI. So now I

51:19

wake up in 2026 and I look at my 6040

51:21

portfolio or 60

51:23

>> 6040.

51:24

>> It's it's I mean it looks 60/40,

51:26

>> but it's basically all AI in my equity

51:28

portfolio. It's a lot of AI in my fixed

51:30

income portfolio. It's also AI in my

51:31

venture capital portfolio.

51:32

>> So AI better work.

51:34

>> This AI, I think, better work out. it

51:35

better work out

51:36

>> because if that doesn't work out then

51:38

your portfolio will be in trouble.

51:39

That's why ironically the best

51:41

investment recommendation today is the

51:42

new 6040 is really to do 60 maybe AI and

51:46

40 non AI. So in other words, the best

51:48

recommendation for investors is to

51:50

invest in nonAI things that are not

51:51

correlated with this one factor. Because

51:53

if there's one thing we have learned in

51:55

finance since the financial crisis is

51:56

factor investing, you don't want to be

51:58

exposed just to one factor. And at the

52:00

moment there's one factor staring all of

52:01

us right in our eyes and that is AI is

52:03

literally everywhere. And that's of

52:05

course means that value investing which

52:06

you will appreciate more than anyone

52:08

else is actually superior because I'm

52:10

already exposed to AI everywhere. But

52:12

the problem with that thesis, which is

52:14

wonderful, is that all the stuff that

52:18

you would want to that you like if we

52:20

drew up a list like what can I invest in

52:23

that's not correlated and then I look at

52:25

the chart of those things, the chart

52:28

looks terrible of every single one of

52:30

those things. It's like hasn't moved in

52:32

years like consumer staples for example

52:34

>> 100%. But that's exactly why those

52:35

things haven't moved for years. But if

52:36

you now are going to see back to our

52:38

token discussion and demand for comput

52:40

and data centers and if you if there

52:42

truly is no mode as you were saying of

52:44

course then we will have some problems

52:45

in the AI world and if that's the case

52:48

then of course these things are about to

52:49

take off like a rocket because then

52:51

investors will be saying I got to buy

52:53

something

52:53

>> I got to buy something else which is not

52:54

this thing that is the one factor that

52:56

is now the biggest risk. this I to be

52:58

sure large language models I have seven

53:01

on my phone they are incredibly helpful

53:02

they will change your life my life is

53:04

changing all of our lives but that's not

53:06

the same thing as saying that the

53:07

revenues that are coming in for the AI

53:10

firms is going to come at the speed that

53:12

is priced in markets today

53:13

>> right

53:14

>> okay Torson thank you very much

53:17

we'll have you back

53:18

>> thank you and we're back so I thought

53:20

one of the first interesting things that

53:22

Torston said is that this year GDP will

53:26

grow a little bit more 2% and if you

53:28

divide it up 1% of that 2% comes from AI

53:33

spending 3/10en of 1% comes from the

53:37

re-industrialization and onshoring

53:40

and 90 basis points comes from the

53:44

consumer getting a lot of money back

53:46

from tax refunds from the big beautiful

53:48

bill actually raises an interesting

53:49

issue in that those tax refunds won't

53:52

exist next year so you know the base of

53:54

GDP growth um will be sub should be sub

53:58

2% in 2027 unless something else

54:01

happens. You then we started talking a

54:03

lot about AI and the dramatic impact

54:07

it's had on the US economy. We then move

54:11

to how dynamic the US economy really is

54:14

that interestingly enough despite all

54:16

the you know news stories that you hear

54:19

about people losing their jobs the

54:21

unemployment rate is actually still

54:23

excellent. Job creation is very very

54:26

strong and job creation is actually very

54:28

very strong statistically amongst young

54:30

people which belies the stories that

54:32

that you hear about. So the US economy

54:34

is is very dynamic. It's still growing

54:37

but it's unbelievably AI dependent and

54:40

you know we talked about some of the

54:42

bare case stories of AI which are that

54:44

it's become more capital intensive.

54:46

There are potentially no moes and these

54:48

are things that everybody should keep in

54:50

mind about future risks. Then we moved

54:53

on to Europe where Torson basically

54:55

agreed that Europe is sclerotic and

54:57

nothing's going to change anytime soon.

55:00

And we ended up with an interesting

55:02

comment from Torson about investment

55:05

risks that people think that they're

55:07

diversified because they have 60% of

55:09

their money in equities and 40% of their

55:11

money in debt. And what they're missing

55:14

is that of the 60 because the large

55:17

companies now make up 40% of the S&P and

55:20

so much is AI related, if you own the

55:23

equity markets or own the general

55:25

indexes, you are very heavily AI

55:28

indexed. And then on the debt side, you

55:31

would normally think that would be

55:32

diversification, but because of all the

55:34

debt being issued by AI data centers,

55:37

debt is now becoming overindexed to AI.

55:40

So people are incredibly overindexed to

55:43

AI. This AI story better work because if

55:46

it doesn't work, the losses that people

55:48

are going to experience are going to be

55:49

mammoth. And I think that was the

55:53

concluding message that I wanted to

55:54

bring home. Thanks for watching. See you

55:56

soon.

56:02

This podcast is forformational purposes

56:04

only and does not constitute investment

56:06

advice. The hosts and guests may hold

56:08

positions in stocks discussed. Opinions

56:10

expressed are their own and not

56:12

recommendations. Please do your own due

56:14

diligence and consult a licensed

56:15

financial adviser before making any

56:17

investment decisions.

Interactive Summary

Steve Eisman interviews Torsten Sllock, Chief Economist at Apollo, on the US and global economic landscape. Sllock details the US economy's current strength, attributing it to robust AI spending, re-industrialization efforts, and significant tax refunds, noting these drivers are largely insensitive to interest rates, making Fed rate cuts unlikely this year. They delve into the capital-intensive nature and potential lack of 'moats' in the AI sector, alongside the widening 'K-shaped' economy where high-income households thrive while lower-income ones face increasing headwinds. The discussion also covers the specific vulnerability of the software sector within private credit due to high leverage and deteriorating coverage ratios, though the systemic risk to the broader US economy is deemed low. Sllock highlights the dynamism of the US labor market, contrasting it with Europe's sclerotic economy, characterized by rigid labor/product markets and a less diverse financial system. Finally, they examine the US deficit, currently sustained by strong foreign demand for US assets, and conclude with a major market risk: the pervasive over-indexing to AI across all asset classes, warning of potential significant losses if the AI narrative falters.

Suggested questions

10 ready-made prompts