0:00

This is The Reason Interview

0:02

with Nick Gillespie. My guest

0:04

today is Harvard economist Jason

0:06

Furman, who is one of

0:08

Barack Obama's chief economic advisors

0:10

and an architect of the

0:12

Affordable Care Act. These days,

0:14

he's in the news for

0:17

his withering critique of Joe

0:19

Biden's dismal economic record and

0:21

his two -fisted attacks on Donald

0:23

Trump's trade policy. We talk

0:25

about Furman's defense of global

0:27

markets, Obamacare's unfulfilled promises, and

0:30

why he now thinks the national

0:32

debt and ai are bigger deals

0:34

than he used to believe here

0:36

is the reason interview with jason

0:38

firman uh

0:42

jason firman thanks for talking to reason

0:44

really excited to do so love love

0:46

reading you So

0:48

you recently wrote a piece

0:50

talking about the failure of

0:52

Bidenomics and what you called

0:54

the post -neoliberal delusion. Let's

0:57

start with that. You know, Irving Kristol

0:59

famously said that a neoconservative is a

1:01

liberal who got mugged by reality. What

1:04

is a neoliberal, the way that you're defining

1:06

that term? Sometimes I think a

1:08

neoliberal is like whatever I happen

1:10

to think that somebody else disagrees with

1:12

and they bundle it all together. It's

1:15

often defined more in the negative

1:17

as an epithet. If I had to

1:19

do it in the positive, it

1:21

is one, understands gains from trade,

1:24

that when you have markets, when

1:26

you have exchange, you can end

1:28

up with more for everyone. Number

1:31

two, respects.

1:33

a corollary that follows from that, that

1:35

decentralized decisions can result in the

1:37

best outcome. Now, you may want regulation

1:39

or taxes to move them in

1:41

a certain direction, and I might want

1:43

more of those than you would

1:45

want. I suspect that's true. Number

1:48

three, that scarcity

1:50

is at the heart of

1:52

all interesting public policy problems,

1:54

and so you have to

1:57

respect trade -offs. budget constraints.

1:59

At a macro level, that means

2:01

you have to worry about the

2:03

budget deficit. At a micro level,

2:05

that means you have to do

2:07

cost benefit analysis. And finally, something

2:09

that follows from all of that,

2:11

that you don't have an automatic,

2:13

easy answer to any question. You

2:15

really have to sort of work

2:17

through it, do the analysis, and

2:19

figure it out. So there's some

2:21

technocratic mindset associated with it as

2:23

well. You said that neoliberal is

2:25

often invoked as a kind of

2:27

term of derision or dismissal. And

2:29

I think I'm right about this, correct me

2:32

if I'm wrong, that it started being

2:34

used in that way around the end

2:36

of the 90s, you know, after NAFTA

2:38

had been passed, started by, I guess,

2:40

Reagan, Bush, then, you

2:42

know, ultimately done by Clinton. And

2:44

then at the end of the

2:47

90s, the beginning of the World

2:49

Trade Organization. And this is when...

2:51

people who were interested in kind

2:53

of increasing global trade and lowering

2:55

tax or tariff barriers and things

2:57

like that started to be called

2:59

pejoratively neoliberals. Does that track? That

3:02

tracks, but it really exploded about

3:04

a decade ago because there wasn't,

3:06

at least on the democratic side,

3:09

a very well articulated alternative

3:11

to it. And in some ways

3:13

there isn't, but there are a

3:15

set of them. Modern monetary theory

3:17

with deficits don't matter. A neo

3:19

-Brandeisian approach to antitrust that says

3:21

big companies are presumptively bad. Right,

3:23

just bigness is a curse. And

3:26

the fact that they're big means

3:28

we should be thinking bad. Breaking

3:30

them up. Yep. During

3:32

the inflation debates, the focus on

3:34

greed as a cause of inflation

3:37

and stopping greed as a solution

3:39

to inflation. And then people writing

3:41

about industrial policy. So

3:43

these are some different

3:45

strands. But they all

3:47

combine together in a

3:49

rejection of a lot

3:52

of those ideas of

3:54

gains from trade, decentralization.

3:58

and the importance of taking trade off seriously.

4:00

So this is, I mean, you were in

4:02

the Obama administration. I mean,

4:04

this is a pretty lonely place

4:06

you're occupying as somebody who identifies with

4:08

the Democratic Party, right? It's

4:11

not totally lonely. I mean, President

4:13

Obama put a lot of effort into

4:15

negotiating the Trans -Pacific Partnership, a free

4:17

trade agreement with our friends in

4:19

the Pacific. He had a proposal that

4:21

I spent a lot of time

4:23

on, on corporate tax reform that would

4:25

have lowered the corporate tax rate.

4:27

He proposed 28%. At the time, it

4:29

was 35%. And he was, I think,

4:32

pretty sympathetic to the markets. But

4:34

Obama left office a decade ago. Sorry.

4:36

At this point, it's

4:38

mixed. You know, it's

4:40

not like people don't listen to economists.

4:43

There's economists in government. There's economists

4:45

in Fed. There's a lot of economists

4:47

around. In the Biden administration, I

4:49

would say the policy took a turn

4:51

against this. And the rhetoric took

4:53

a massive pivot against this. In

4:55

some ways, the language turned even more

4:58

than the policy did. I mean,

5:00

you know, you would have been

5:02

working with Joe Biden, you know,

5:04

in the Obama White House. I

5:06

mean, did you see it coming

5:08

that? his presidency was going to

5:10

be as, you know, as tilted

5:13

towards kind of progressive concerns as

5:15

opposed to kind of, you know,

5:18

market -friendly liberal concerns?

5:20

Or did that kind of

5:22

blindside you? It mostly blindsided

5:24

me compared to Obama. Obama's

5:27

quite cerebral, comes at things from

5:29

the perspective of ideas and policy,

5:31

may end up with a big

5:33

dose of politics. He wouldn't have

5:35

gotten where he was without that.

5:37

Biden pretty much starts and ends

5:39

with politics. He

5:43

has certain things he likes. He

5:45

definitely likes labor unions quite a

5:47

lot. So that part of it

5:49

was not a surprise. But the

5:51

degree to which he adopted something

5:53

that was much more towards the

5:55

left within the Democratic Party, yeah,

5:57

that did surprise me after he

6:00

ran in the primaries as the

6:02

centrist. Why do you think

6:04

the neoliberal consensus, and again,

6:06

your article is called The

6:08

Post -Neoliberal Delusion, why did

6:10

that neoliberal coalition

6:12

within the Democratic Party breakdown?

6:16

I don't know. There's one argument

6:19

that it is China entered

6:21

the WTO, lots of suffering resulted,

6:23

and there was a populist

6:25

revolt in the form of Donald

6:27

Trump, and you needed to

6:30

listen to that and learn. I'm

6:32

much less sure that that's the

6:34

case. If you look at inequality

6:36

in the 20 years after China

6:38

entered the WTO, it actually went

6:40

up much less than it did

6:42

in the 20 years before China

6:44

And most economists at this point,

6:46

to the extent they talk about

6:48

a China shock, it had kind

6:50

of been absorbed by around 2010

6:53

anyway. Yes, you can debate the

6:55

original China shop because there were

6:57

definitely negative sides. There were also

6:59

a bunch of positive sides. I

7:01

think the positives that weighed, the

7:03

negatives. Regardless, everyone agreed it was

7:05

basically over. I think it's also

7:07

possible, and this is real speculation

7:09

on my part, that there are

7:11

ways in which the Democratic Party

7:13

got culturally out of touch with

7:15

Americans, that that led to some

7:17

backlash. And that was then misinterpreted

7:19

and used as an argument for

7:21

changing. on economic policy. What do

7:23

you think the role of somebody

7:25

like Bernie Sanders is in this,

7:27

whether it's a rhetorical or a

7:29

policy shift? He obviously, you know,

7:31

the fact that Hillary Clinton won,

7:33

or rather Hillary Clinton lost in

7:35

what should have been an easy

7:37

win. I mean, I think that's

7:39

a legitimate statement, but Sanders, clearly

7:41

his insurgency in 2016 hurt her

7:43

on some level. And is

7:45

it then that, like, if Hillary may

7:47

or may not have been a neoliberal

7:49

the way you're talking about it, but

7:51

Sanders definitely was not. He was anti -WTO.

7:53

He was anti -trade, very pro -union protectionist.

7:55

Is that where some of the energy

7:57

for rejection of neoliberalism in the Democratic

7:59

Party comes from? Look, I'm not a...

8:01

expert. I'm a hobbyist. I pay a

8:03

bit of attention. I have much more

8:05

conviction around what I think is good

8:07

for the world and bad for the

8:09

world. And by the way, if you're

8:11

governing and you do things that are

8:13

good for the world, I think you

8:15

get rewarded. Do things that are bad

8:17

for the world, you get punished. And

8:19

we just saw that in the last

8:21

election. But I

8:24

don't know. Americans

8:26

are a strange mixture of

8:29

contradictory views. I myself have

8:31

a set of contradictory views.

8:34

You can tap into a sort

8:36

of fear and loathing of billionaires.

8:38

And every billionaire is a policy

8:41

failure. But people can also laud

8:43

and be thrilled about the things

8:45

that billionaires invented and did and

8:47

changed their lives and understand that

8:49

and both of those together at

8:52

once. Yeah, on that point, Forbes,

8:54

I guess, came out with its newest

8:56

billionaires list. And Jerry Seinfeld is a

8:58

billionaire now. And it's like, I don't

9:00

think people think about Jerry. Seinfeld

9:02

and Elon Musk in the same category.

9:05

But who has Jerry Seinfeld killed, right?

9:07

Who has he hurt? Yeah. I mean,

9:09

Robert Nozick in his book had an

9:11

example of Wilt Chamberlain. I do that

9:13

in my class. I'm not sure they've

9:15

all heard of Wilt Chamberlain. So I

9:17

put Taylor. Wilt Chamberlain post -MeToo, I

9:19

don't know. I put Taylor Swift up

9:21

on the screen. And, you know, did

9:24

she make you better? I do

9:26

say that I'm not endorsing

9:28

her music. But,

9:30

you know, did she make people

9:32

worse off by being a billionaire

9:34

or did she make people better

9:36

off and in the course of

9:39

that became a billionaire? Ken, you

9:41

know, as you noted, I mean,

9:43

there's a shift on the right

9:45

also. You know, Trump is post

9:47

-neoliberal. He does not like global

9:49

trade. He does not like. you

9:52

know, kind of market forces. To

9:54

some degree, he's more pro -business, I

9:56

think, in a real way than

9:58

Hillary Clinton, Biden, or Kamala Harris

10:00

might have been. But he's antagonistic

10:02

towards the liberal project, if we're

10:05

talking about things like global trade

10:07

and decentralized, you know, open borders,

10:09

open services, and things like that.

10:12

Why, from an economic point of view,

10:14

we're not talking politics here, You

10:17

know, the period of, you know, in

10:19

the post -war generally, but then since

10:21

the collapse of the Soviet Union

10:23

and since, you know, the WTO was

10:25

created at the beginning of this

10:27

century, things have gotten better for the

10:29

vast majority of people in America,

10:31

you know, on a material basis. Why

10:33

then are people, why do

10:36

they not see that and why

10:38

are they captivated by things

10:40

like, you know, stronger tariff policies

10:42

and more protectionism? Yeah. So

10:44

first of all, on the premise

10:46

of your question, I agree. President

10:49

Trump, in some ways, is

10:51

an even bigger rejection of anything

10:53

resembling economic analysis, thinking about

10:55

trade -offs, thinking about gains from

10:57

trade and the like. And for

11:00

the most part, has eschewed

11:02

any intellectual rationalization of it. There's

11:04

some people out there like

11:06

Orrin Cass that provide some. intellectual

11:09

foundations for it or attempt to

11:11

provide some intellectual foundations for it. But

11:13

for the most part, that's actually

11:15

not what Donald Trump is interested in.

11:17

They make a lot of noise.

11:19

On the pro -business, in the wake

11:21

of the tariff announcements, I thought it

11:23

was striking how many people from

11:26

the administration, including the Treasury Secretary, went

11:28

out to talk down big business

11:30

as a globalist project, as something for

11:32

the elite. We don't care about

11:34

the stock market. We don't care about

11:36

the businesses. pretty

11:38

pivoting pretty quickly away

11:40

from business. So what's

11:43

the commonality here? There's

11:46

always the sense that things used

11:48

to be better. And that if

11:50

we could just get back to

11:52

the golden age that was, that's

11:54

been a powerful thing throughout humanity.

11:56

That siren. Song has been more

11:58

powerful in the wake of financial

12:00

crises. It's something you see. That's

12:02

where populists often arise. And when

12:04

Trump first rose, we weren't that

12:06

far past a financial crisis. We've

12:08

just had another type of enormous

12:10

amount of dislocation in our country,

12:12

both from COVID and through a

12:14

set of extremely rapid cultural changes.

12:17

And so I tend to blame

12:19

that. rather than looking at the

12:21

statistics on median family income, which

12:23

has grown quite strongly, or even

12:25

wage inequality, which has actually fallen.

12:27

Yeah, and I mean this, it

12:29

grew under Trump, you know, and

12:31

then there's the COVID moment. And

12:33

then after that, under Biden, real

12:35

median household income, individual income all went

12:38

up, which is a good sign. It's

12:40

a sign that the economy is kind

12:42

of delivering. but people

12:44

aren't interested in that. When

12:46

you said there have been radical

12:48

or rapid social transformations, what

12:50

do you mean? I just meant

12:52

the Black Lives Matter movement,

12:54

how rapidly things change related to

12:56

trans and the like. There

12:59

was a sense that a lot

13:01

of people in the country

13:03

were being caught in a very

13:05

big and very rapid set

13:07

of changes. When

13:10

you're talking about the failure of

13:12

Bidenomics, Biden spent a lot of

13:14

money on a couple of big

13:16

projects or big spending bills. Can

13:18

you talk about them and how

13:20

they failed to deliver what he

13:23

and his supporters claimed that they

13:25

were going to do? Yeah, let's

13:27

just tick through five quickly. There's

13:29

the American Rescue Plan, $1 .9

13:31

trillion. And the main

13:33

thing we got from that

13:35

was inflation. It's possible there was

13:38

a slightly faster recovery, but

13:40

the U .S. recovery wasn't actually

13:42

much faster than any other place,

13:44

at least in 2021 and

13:46

into 2022. Second, student

13:48

loan relief. This to me

13:50

in many ways was the most

13:53

egregious. This was done in August

13:55

2022. Inflation was raging.

13:57

The deficit was high. Interest

13:59

rates were high. And another

14:01

$500 billion was poured on

14:03

top of that without Congress. authorizing

14:06

that money in a way that

14:08

I think is quite abusive of

14:10

presidential authority. Third,

14:12

infrastructure. Here I

14:15

have a mixed view. I think

14:17

it's a noble goal. I think

14:19

it is a worthwhile goal. But

14:21

it was accompanied with so many

14:23

new rules and restraints on how

14:25

the money would go out. Moreover,

14:27

so much money was spent so

14:29

quickly in an economy that was

14:31

already overstretched that it ended up

14:33

driving prices up. rather than getting

14:35

more infrastructure. That's where I think

14:37

things like the abundance agenda are

14:39

quite important that are about building

14:41

more stuff. Fourth is

14:43

climate. Here, I like a

14:45

lot of it. I would love a

14:47

carbon tax. To me, that's the

14:49

simple solution. I think libertarians should love

14:51

that solution too. Well, I think

14:54

in principle they do. And then like

14:56

when you start. looking at the

14:58

politics against dice here, but in any

15:00

case. So that's what I'd love.

15:02

You can't pass one. That's not Biden's

15:04

fault. There was a combination of

15:06

subsidizing the use of wind and electricity.

15:08

I think that's okay. Those use

15:10

less carbon. But there was also a

15:12

lot of remaking American industry. So

15:14

we would make solar panels in the

15:16

United States. That's where I... get

15:18

off the train. I'd rather just buy

15:21

the cheapest ones from around the

15:23

world. And finally, there's the chips program

15:25

where from an economic perspective, I

15:27

think we're going to get worse microchips

15:29

and not create great jobs. National

15:31

security and resilience, I think it will

15:33

be a plus. And that plus

15:35

outweighs for me the economic minus. I'm

15:37

willing to pay a cost in

15:39

terms of efficiency to have microchips made

15:41

here, advanced ones, instead of all

15:43

in Taiwan. They're more

15:45

expensive and they're less good. But they're ours.

15:48

That's not my goal. Yeah, but I would

15:50

say that about very few things. And

15:52

that's where, again, the cost benefit comes

15:54

into play. If you tell me that

15:56

you think your microchip plan is going

15:58

to create more jobs and better jobs

16:00

and increase economic growth, I'm not going

16:02

to believe you. If you come and

16:04

tell me, you know what? It's going

16:06

to cost some money. It's going to

16:08

be a little bit inefficient. but

16:10

here's what the national security benefit is,

16:13

then we can go back and forth.

16:15

How big is the economic cost? How

16:17

big is the national security benefit? Is

16:19

it worth it? Is it not? Some

16:21

of these programs, and there were some

16:23

of them under Trump and under Biden

16:25

when we're talking about COVID. Can you

16:27

explain from an economic point of view?

16:30

I am merely, I'm a simple

16:32

English major and a psychology

16:34

major. Why is it

16:36

bad when to... Tell people

16:38

at a time like COVID,

16:41

we're going to double or

16:43

triple your unemployment benefits and

16:45

extend them for long periods

16:47

of time. Why does that

16:49

end up having a bad effect on

16:51

the economy? Right. Well, if your goal

16:53

is to get people not to work,

16:55

and that might have been a goal

16:57

in March and April of 2020, that's

17:00

a perfectly fine way to do

17:02

it. The policy lasted

17:04

through September 2021. At

17:06

that point, there were record

17:08

job openings. People were

17:10

either vaccinated. or had caught

17:12

COVID and were less susceptible to

17:14

it again. And by that point,

17:16

it really started to interfere with

17:19

people getting back into work. I

17:21

mean, why would you possibly want

17:23

to work when you could get

17:25

more money not working? I think

17:27

that almost every day until my

17:29

retirement, I'm sure. And then the

17:31

second issue in the whole response

17:34

is if you give people more

17:36

money, let's say we as a

17:38

country were producing and buying and

17:40

selling $1 ,000 a year. 1 ,000

17:42

units of whatever. You give people

17:44

enough money to buy 1 ,100. At

17:47

the same time that something's happened

17:49

to the economy, so it can only

17:51

make 900, the result is you're

17:53

not going to get more stuff. You're

17:55

just going to get higher prices

17:57

for the existing amount of stuff. is

17:59

more dollars chasing the same number

18:01

of goods. This is textbook inflation. Yeah,

18:03

exactly. So it made sense to

18:05

protect people. you

18:08

don't want your living standards to go

18:10

way, way down because of COVID. But

18:12

what we ended up doing was not

18:14

doing the math and figuring out what

18:16

you needed to do to protect people,

18:18

not doing the targeting to get to

18:20

the people who needed it. And so

18:22

the result was just much, much, much

18:24

more money than we could possibly make

18:26

in terms of stuff. And so it

18:28

drove up the price of that stuff.

18:30

Where does it come from? I mean,

18:32

has it always been this way that

18:35

You know, what you're saying is

18:37

obvious, but then it became

18:39

politically inexpedient to admit that, particularly

18:41

when Biden was in office

18:43

and this was happening or when

18:45

he was starting to run

18:47

for president and then Harris. I

18:50

mean, is it just simply

18:52

political posturing that Democrats are

18:54

not going to admit the

18:56

policies that their guys pursued,

18:58

you know, caused inflation? just

19:01

in the same way that Republicans

19:03

won't admit it when it's their

19:06

guy in the dock. I think

19:08

there's a real problem with the

19:10

way information is aggregated and transmitted.

19:12

There were some people I know

19:14

who, when that first stimulus bill

19:16

were passed, who were emailing

19:18

me, hey, this is really too large,

19:21

but I don't want to undermine Biden

19:23

by saying that. Or,

19:25

you know, this is too large, but

19:27

I'm afraid people are in the student

19:29

loan case. This is a bad policy,

19:31

but I don't want to get destroyed

19:34

the way, Jason, I saw you get

19:36

destroyed on Twitter. Some people,

19:38

though, don't even quite see that. So

19:40

I've had people reach out to me

19:42

who have said, hey, you know, aren't

19:44

the Trump tax cuts going to cause

19:46

inflation, Jason? You should write about that.

19:48

And I'm like, wait, weren't you the

19:50

same person that a couple of years

19:52

ago told me the Biden fiscal expansion?

19:55

wasn't causing inflation? Like, why does the

19:57

fiscal impact on inflation depend on the

19:59

political party of the person? And it's

20:01

hard to think outside your own bubble

20:03

and your own ecosystem. And all of

20:05

this creates a sort of false sense

20:07

of what it is people actually think

20:09

when a lot of people aren't saying

20:11

what they think. And some of those

20:14

people are tenured professors who aren't going

20:16

to be, you know, in no risk

20:18

of being fired. I

20:20

guess in a similar

20:24

You know, we're hearing this now

20:26

because Trump has, you know, Trump

20:28

has levied tariffs. I mean, depending

20:30

on what day of the week

20:32

and when this airs, who knows

20:34

what the case will be. But,

20:36

you know, currently he's suspended his

20:38

tariffs or put them on ice

20:40

for 90 days except for China, etc.

20:43

But you hear a lot of

20:45

Republicans now saying, well, you know, tariffs

20:47

aren't taxes. and tariffs

20:49

are paid for by the country

20:51

that is sending goods, not

20:54

the people who are importing them

20:56

here in America. You recently

20:58

wrote a piece in the New

21:00

York Times saying that Trump's

21:02

tariffs are based on an economic

21:04

theory that makes no sense. Can

21:07

you explain that? What is

21:09

wrong with the tariffs the

21:11

way that Trump is doing?

21:13

Yeah, basically every step of

21:15

Trump's mentality on trade is

21:17

wrong. First of all, imports

21:20

are good. They're a wonderful

21:22

thing. We like to import

21:24

coffee. We like to import

21:26

cars. We like to import

21:28

the inputs we need for

21:30

American manufacturing. Second of all,

21:32

trade deficits don't reflect other

21:34

countries' tariffs and trade policies.

21:36

There's countries with huge tariffs

21:38

that run large trade deficits.

21:40

There's countries with low tariffs

21:42

that run trade surpluses. Third,

21:44

our retaliation against them is

21:47

going to shrink imports. but

21:49

it's also going to shrink

21:51

exports. And that hurts consumers

21:53

on the import side. It

21:55

hurts American workers on the

21:57

export side. And so just

21:59

every step of his reasoning

22:01

and fixation on trade deficits

22:03

is just the wrong, wrong

22:05

way to think about trade.

22:07

Is there a consensus forming

22:09

that either that that's a

22:12

bad policy, a kind of

22:14

bipartisan or transpartisan? consensus

22:16

forming that, okay, these tariffs are not

22:18

working and we're getting strong signals from

22:20

global markets as well as, you know,

22:22

the things like the S &P 500

22:24

and the Dow Jones. But is,

22:26

you know, is there also a

22:28

kind of transpartisan, you know, coalition

22:30

sort of saying like, oh, maybe

22:32

we should be more protectionist? There's,

22:35

President Trump went so far that

22:37

I think he discredited it

22:39

for even many people on the

22:42

protectionist side. So people like

22:44

free trade more than they ever

22:46

did before right now, more

22:48

than they ever would after listening

22:50

to me extol the benefits

22:52

of free trade. So I don't

22:55

think he's building that much

22:57

of a coalition against it. But

22:59

he has shifted the Overton

23:01

window enormously. I mean, he did

23:03

this initial reciprocal tariff announcement

23:05

on Liberation Day. Then

23:07

he took away. You've lost.

23:10

You're using his language, right?

23:12

You call it the Gulf of

23:14

America as well. I don't care

23:16

much about words. Maybe that's a

23:18

mistake. Then he

23:21

took away some of them,

23:23

the so -called reciprocal part.

23:25

So that left the

23:27

United States with tariff rates

23:29

that are the rates

23:31

that prevail in Iran, Venezuela.

23:33

and a bunch of other countries

23:35

that aren't nearly as economically

23:37

successful as Iran and Venezuela. And

23:40

that's now become the reasonable,

23:43

moderate, decent position. And so the

23:45

Overton window has shifted quite

23:47

a lot, and I'm not sure

23:49

how we can shift it

23:51

back. A place

23:53

where your thinking has evolved

23:55

over the years has

23:57

to do with the role

23:59

of debt and deficits

24:01

in larger macroeconomic concerns. I

24:04

think it was in 2019,

24:06

you wrote a piece with

24:09

Larry Summers, another Harvard grad

24:11

and administration official back in

24:13

the Clinton years. where

24:15

you were kind of poo -pooing, you know,

24:18

debt, you know, it's not that big a

24:20

deal and it should never really get in

24:22

the way of doing what's right, you know,

24:24

in terms of helping the poor, you know,

24:26

et cetera. You have kind

24:28

of revised some of that. Can you talk

24:30

a little bit about where you are now

24:32

and how you got to that position? Yeah.

24:34

So to say I've evolved on the debt

24:37

is unfair. I've zigzagged on the debt. So

24:39

I was concerned about it

24:41

in the 90s. Then I

24:43

got increasingly less concerned about

24:46

it. Then I got more

24:48

concerned about it. I'd like

24:50

to think that I'm following

24:52

the data. And when interest

24:54

rates are low, we really

24:56

could afford more debt. When

24:58

interest rates are high, we

25:00

can't afford more debt. It's

25:02

possible my thinking has evolved

25:04

and changed even subject to

25:06

the data. But a world

25:08

with 70 % debt to

25:11

GDP and a 2 % interest

25:13

rate is just very, very

25:15

different from 100 % debt

25:17

to GDP. 4 .5 % interest

25:19

rate. Why does large and

25:21

persistent and growing national debt,

25:23

and whether we're using debt

25:25

held by the public or

25:27

gross debt, which includes intergovernmental

25:29

holdings, why when that grows

25:32

and particularly crosses the 90

25:34

% or 100 % of GDP

25:36

threshold, why is that a

25:38

problem? So I don't know that we

25:40

know what the threshold is. Argentina

25:42

had a massive financial crisis, including

25:44

a fiscal crisis when it

25:46

had debt of 45 % of

25:48

GDP. The UK during the,

25:50

I think it was the Napoleonic

25:52

Wars, got as high as 250%.

25:55

So it does depend on. why

25:57

you're borrowing, do people trust your

25:59

country, and the like. Right, and

26:01

it helps, I mean, the U .S.,

26:03

beyond everything else, because the dollar

26:05

is the global reserve currency, we're

26:07

working with different rules than, you

26:09

know, smaller countries.

26:11

Yeah, even than Germany or Canada. They

26:14

can borrow much less than we

26:16

can. I mean, the first thing I'd

26:18

say is, if you were czar,

26:20

or I was czar, or anyone was

26:22

czar, there is just no way

26:24

that your economic plan would be, we're

26:26

going to spend at the current

26:28

rate, tax at the current rate, and

26:30

then make some larger abrupt change

26:32

20 years in the future. You would

26:35

rather make a smaller change now

26:37

than a larger change in the future.

26:39

In economics, that's called tax smoothing.

26:41

The consequence of not doing that is

26:43

for sure we get less economic

26:45

growth. And we raise the chance of

26:47

a fiscal crisis. And I don't

26:49

know how high that chance is. But

26:51

look, right now, I pay homeowners

26:53

insurance against the risk that my house

26:55

will burn down. The fact that

26:58

my house didn't burn down last year,

27:00

the year before, the year before

27:02

that, doesn't mean I'm like a loser

27:04

and a sucker for having bought

27:06

this. I think of deficit reduction in

27:08

some ways just like that. Why

27:12

does large and growing debt

27:14

reduce economic growth? It means the

27:16

government is borrowing more money,

27:18

which drives up interest rates. Driving

27:20

up interest rates drives down

27:22

business investment, drives down home building.

27:24

The other way in which

27:26

it matters is you might get

27:28

out of that if you

27:31

borrow from abroad to support your

27:33

investment. Your

27:35

economic growth, you have to give

27:37

it to those foreign countries to repay

27:39

them for everything you borrowed. So

27:41

either way, you're either doing less investment

27:43

or you're having to repay foreigners

27:45

if you want to maintain your investment.

27:48

This was during the fiscal crisis

27:50

or the financial crisis or

27:52

shortly after it. Carmen

27:54

Reinhardt, Vincent Reinhardt, and Ken

27:56

Rogoff wrote a paper, one

27:58

version of which then they

28:00

corrected. It really

28:03

said, you know, as you approach

28:05

a 90 percent debt to

28:07

GDP ratio, that these episodes, they

28:09

looked at a series of

28:11

them over the past couple of

28:13

hundred years, you know,

28:15

that it tended to reduce long

28:18

term economic growth, you know, which

28:20

average for them, it was about

28:22

an average of 23 years of

28:24

periods they call debt overhang. And

28:26

it reduced. growth in general from

28:28

something like 3 % annually to 2

28:30

% annually or like 1%. point,

28:33

does that sound about right to

28:35

you? And does that help explain

28:37

the sluggish economic growth that we've

28:39

really been witnessing for a good

28:41

part of the 21st century? I'll

28:43

take your second question first. I

28:45

think the causation might go more

28:48

the other direction, that there was

28:50

a set of sluggishness either from

28:52

slower productivity growth, from demography or

28:54

whatever it was. And in order

28:56

to keep the economy going, We

28:58

needed the extra debt. And what's

29:00

the evidence for that? We didn't

29:02

until 2021 get a lot of

29:05

inflation. So through 2019, the problem

29:07

was low inflation, not high inflation.

29:09

That is generally indicative of demand

29:11

being too low. So I don't

29:13

think the debt was overheating our

29:15

economy. I think it was making

29:17

up for these problems. And that

29:20

was part of why Larry and

29:22

I thought we needed more debt.

29:24

We overshot what we might possibly

29:26

have imagined. In

29:28

terms of Ken Rogoff

29:30

and Carmen Reinhart's research, those

29:32

papers were suggestive. There

29:34

was a lot of correlation,

29:36

not causation. So directionally,

29:38

I think they're right. But

29:40

I'm not convinced about

29:43

some nonlinear threshold. In terms

29:45

of dealing with debt

29:47

and national debt and deficits,

29:49

I mean, you now

29:51

kind of propose We should

29:53

be trying to balance

29:55

the budget, which we've done

29:57

in our lifetime. Balance

30:00

the budget excluding interest payments.

30:02

Balance what's called the primary The

30:04

primary deficit. And also invoking

30:06

a kind of strict PAYGO policy

30:08

for when we spend more

30:10

or spend less. Can you describe

30:12

how PAYGO would work and

30:15

why that's important? I think PAYGO

30:17

is important for two reasons.

30:19

It says that anything you do

30:21

you have to pay for.

30:23

So you want to give money

30:25

for student loan debt relief. Great.

30:28

You can pass a law that

30:30

costs $500 billion. You then need to

30:32

raise taxes by $500 billion or

30:34

cut spending by $500 billion. That

30:37

has a macroeconomic rationale that when you're

30:39

in a hole, you don't want to

30:41

keep digging. And this stops you from

30:43

digging. I think perhaps even more important,

30:45

it lets you figure out whether the

30:47

things you're doing are a good idea

30:49

or not. Almost every idea

30:51

that involves giving money to

30:53

some group is a good idea

30:56

until you ask where the

30:58

money is coming from. If

31:00

you can say, here's where the

31:02

money is coming from, here's where the

31:04

money is going, and on net

31:06

this is still a good idea, it's

31:09

much more likely that this is

31:11

a choice that's worth society making. But

31:13

if you obfuscate that choice and

31:15

you give the benefit to the people

31:17

today for sure, and you... given

31:19

uncertain costs that might even be even

31:21

larger to people in the future,

31:23

then I don't trust that we did

31:25

the right social cost benefit to

31:28

decide whether or not that idea was

31:30

worth it. Do you think there's

31:32

any chance in hell that we're going

31:34

to get back to a budgeting

31:36

process? And this is politics, but it

31:38

stems from economics that we'll get

31:40

to a budgeting process where any of

31:42

this is actually taken seriously. I

31:44

think there's three broad ways that we

31:46

could get back to that, but

31:49

I agree. Maybe we don't. One is

31:51

higher interest rates could really start

31:53

to put pressure on politicians to take

31:55

the debt more seriously. Even with

31:57

the interest rate increases we've seen over

31:59

the last couple of years, we're

32:01

nowhere near the mortgage rates we had

32:03

30 years ago, 40 years ago.

32:05

So higher interest rates, I think, would

32:07

get people's attention. A second is

32:10

there's what is really a political forcing

32:12

event, which is the Social Security

32:14

and Medicare trust funds get exhausted. And

32:16

there's no real economics there. The

32:18

government could just pass a law saying

32:20

ignore the trust fund. But it

32:22

has a sort of... significance.

32:24

And that's a good thing. And we

32:26

should want to hold on to

32:29

that. So that could force action. The

32:31

third thing is just for whatever

32:33

reason, voters could start to care more

32:35

about it. In 1992, Ross Perot

32:37

centered his third party presidential bid around

32:39

debt and deficit in the last

32:42

cycle. And being anti -NAFTA. And being

32:44

anti -NASTA. In the last cycle, RFK

32:46

Jr. centered around all sorts of things,

32:48

but I don't remember the deficit

32:50

being on his list. Yeah. Do you

32:52

think, I mean, is, you know,

32:54

maybe this is one of the limits

32:57

of economics, but where does that,

32:59

you know, okay, Ross Perot is kind

33:01

of an alien who has, you

33:03

know, dropped onto the planet and starts

33:05

talking about deficit. But

33:07

other people talked about that. I

33:10

mean, it was rhetorically, but where, when...

33:13

do voters get motivated to care

33:15

about this rather than that?

33:17

Because the past couple of years,

33:19

we've spent a lot of

33:21

time talking about stuff that seems

33:23

pretty marginal to anybody's real

33:25

life, but it happens. I'm

33:27

never sure how much of

33:29

these issues are it's already out

33:31

there in the electorate and a

33:34

skilled politician discovers it versus a

33:36

skilled politician. helps change people's

33:38

mind and convince them that something

33:40

is an issue. So I don't

33:42

know which way the causation went.

33:44

Did the people cause Ross Perot

33:46

or did Ross Perot cause the

33:49

people? Either way, Ross Perot

33:51

did cause Bill Clinton to take

33:53

deficit reduction and campaign on it

33:55

more seriously and resulted in him

33:57

doing more deficit reduction than he

33:59

probably otherwise would have done. And

34:02

obviously, it's also he elected or

34:04

he had two years. where he

34:06

pretty much got to do what

34:08

he wanted, what his priorities were.

34:10

Republicans were elected to majorities in

34:13

the House and the Senate for

34:15

the first time in decades. And

34:18

that, you know, by the end of

34:20

the 90s, things had seemed to work out

34:22

pretty well in terms of things like budget

34:24

deficit. And the end of the Cold War,

34:26

so you could reduce defense spending and whatnot.

34:29

What is, I mean, can you explain

34:31

what's going on with the federal

34:33

budget when we talk about things like

34:35

in 2019? These are

34:38

nominal dollars. We spent

34:40

$4 .4 trillion. 2020,

34:43

COVID happens. It's $6 .7

34:45

trillion. It

34:47

reduced in 2023

34:49

to $6 .1

34:51

trillion. We're currently on

34:53

track to spend $7 .2

34:56

trillion. We're years past COVID. What

34:59

are we spending so much

35:01

more on? Well, you clearly... large

35:04

numbers more than I do because I just

35:06

divide everything by GDP and do it all as

35:08

a share of GDP. Spending

35:10

has increased some as a share of

35:12

GDP. One of the biggest things that

35:14

increased as a share of GDP is

35:16

interest on the debt. And that, by

35:18

the way, is, you know, in some

35:21

sense, it's technically spending, but it's the

35:23

consequence of taxes being lower than spending.

35:25

So tax cuts increase interest and spending

35:27

increases increase interest. If you

35:29

look at the non -interest spending, entitlement

35:31

growth is the things that have

35:34

been growing, Social Security, Medicare, and Medicaid.

35:36

With Medicare, the growth has actually

35:38

been much slower than people were counting

35:40

on and hasn't been that rapid. But

35:43

Medicaid is covering a lot more people than

35:45

it was a few years ago. I probably actually

35:47

think that's a good thing. As

35:49

one of the architects of Obamacare,

35:51

you would like that because Medicaid is

35:53

one of the primary ways that

35:55

more people got covered. Exactly. Then

35:58

the goal there was to expand marketplaces,

36:00

but it ended up the marketplaces went

36:02

up a little bit and Medicaid went

36:04

up quite a lot. And is that

36:06

the perfect health system to have in

36:08

our country? Probably not. Do

36:10

I think this is a different

36:12

category of service? Yeah. I

36:14

mean, unless you as a society

36:16

are willing to let the

36:18

person lie outside the hospital. you

36:20

know, dying of whatever it

36:23

is and not treat them. We

36:25

want to have some explicit

36:27

form of care rather than rely

36:29

on a patchwork of, you

36:31

know, bad debt that hospitals are

36:33

forgiving. How do, you know,

36:35

how do we get any kind

36:37

of handle on things like

36:40

Social Security? And Medicare will at

36:42

various points, you know, start

36:44

to grow as well. you

36:46

know, faster than it has

36:48

been. You know, what is the

36:50

preferred, your preferred solution to

36:52

deal with entitlement problems? So if

36:54

I had to write it

36:56

on a specific social security plan,

36:58

I would, for example, extend

37:02

the payroll tax to cover health insurance. That

37:04

would bring in a lot of additional revenue. It

37:06

would do it in a reasonably efficient and

37:08

non manner. By that, you mean just - The

37:10

employer contribution to health insurance. You would lose the

37:13

exclusion for it. So you'd get more revenue

37:15

in. Right. And so it would be, I mean,

37:17

John McCain actually talked about this in 2008.

37:19

He talked about it for income taxes. So I

37:21

just said it for payroll taxes on Social

37:23

Security, but same type of idea. I had written

37:25

about it at the time. I think that's

37:27

a good idea. You have to worry about, does

37:29

it undermine the - system, but as long as

37:32

you have other systems, I don't worry about

37:34

that as much. I would then

37:36

take steps that would cut

37:38

Social Security benefits from their current

37:40

growth path for, say, the

37:42

top one -third of beneficiaries. Doing

37:44

more of that with the retirement

37:46

age, both delaying when you can start

37:48

getting benefits, which has been at

37:50

62 for a very long time. and

37:53

the normal benefit age. And then there's a

37:55

bunch of other changes you could make. And

37:57

I would even strengthen what they do to

37:59

reduce poverty. But these are all sort of

38:01

different tweaks. So you're

38:03

not in for a radical

38:05

rethinking of, say, of

38:07

Social Security, which is kind

38:09

of a bad deal

38:11

for most people. I mean,

38:14

for lower income people, it's

38:16

a good deal. But for

38:18

other people, it's taken out

38:20

in taxes. And then

38:22

you get a bad return on it. Most

38:26

of that bad return comes from

38:28

the fact that we already gave benefits

38:30

to a lot of people in

38:32

the past. And so if you switch

38:34

to a new system now, unless

38:36

you're willing to say for 25 years,

38:38

we're just going to not pay

38:40

any Social Security benefits, then you get

38:43

a good deal. If the new

38:45

people are putting money in private accounts,

38:47

plus they're paying. For the first

38:49

generation of retirees, once you take that

38:51

adjustment into account, you're not actually

38:53

getting much of a better deal out

38:55

of it. So largely, I would

38:57

tweak the existing structure. If accounts are

38:59

important to you, making them an

39:01

add -on account that... adds on to

39:03

the existing system. And if you do

39:06

that, you could do more benefit

39:08

cuts and maintain people's retirement security. That's

39:10

broadly what I would do. From

39:12

a process perspective, though, I really do

39:14

think this is a sort of

39:16

old saying that people say. It needs

39:18

to be bipartisan. If you look

39:20

at the options, the tax -only options

39:22

are really unappealing. The spending -only options

39:24

are really unappealing. Neither party is going

39:27

to want to do this on

39:29

their own. There are a set of,

39:31

by the way, gimmick

39:33

options. We'll transfer

39:35

general revenue. There's proposals to let

39:37

the government borrow money and

39:39

invest that money in stocks and

39:41

leverage it all up. So

39:43

there's a bunch of gimmicks. But

39:45

anything that's real requires pain.

39:48

Anything that requires pain, both parties

39:50

need to be involved with.

39:52

I want to talk about your

39:54

kind of evolution, if we

39:56

can use that term, on AI

39:58

in a second. But has

40:00

any Any of the

40:02

kind of developments since you

40:04

left the Obama administration,

40:06

has any of that shaken

40:08

your faith in a

40:11

kind of Keynesian economics? Maybe

40:13

not a pure Keynesian

40:15

economics, but the idea that

40:17

the government should be

40:19

managing overall supply or demand

40:21

in terms of as

40:23

a kind of counter -cyclical

40:25

intervention to what's going on

40:28

in the economy. So

40:30

as an economic matter, nothing

40:33

has shaken my faith in

40:35

a Keynesian approach. In fact, if

40:37

anything, I think we've seen

40:39

that it really can make quite

40:41

a big difference, both monetary

40:43

and fiscal policy. At the level

40:46

of politics and human decision

40:48

making, I'm now less enamored. by

40:50

the expertise and genius of

40:52

our Congress. And also the Fed,

40:54

I think, has done on

40:56

balance a good job. But there

40:58

have been times I think

41:00

it has gotten things quite wrong.

41:02

Does it generally get things

41:04

wrong by keeping interest rates low

41:07

or? Letting them go too

41:09

high. The error of recent years has been

41:11

keeping them, has been on interest rates

41:13

on the low side. When we're talking like

41:15

recent years, we're really talking, what, about

41:17

like 30 or 35 years? No, I mean,

41:19

we didn't have high inflation prior to

41:21

2020. So really since 2020 is where they

41:23

kept interest rates too low. So what

41:26

does this lead me to? So I think

41:28

monitoring fiscal policy really matter, just the

41:30

way John Maynard Keynes was sketching out. But

41:32

I think people often misapply them. So

41:34

that means I would like more rules. Automatic

41:37

fiscal rules. The unemployment rate goes up,

41:39

money goes out. The unemployment rate comes

41:41

down, money comes back. Same thing with

41:43

interest rates, basically a formula to set

41:45

interest rates. Now with interest rates, I

41:47

would not hand monetary policy over to

41:49

ChatGPT. I don't think it's quite good

41:51

enough to do it yet. But what

41:53

I would do is have the Fed

41:55

chair, every time they do something different

41:57

from what the formula said, get up

41:59

there and say, here's what the formula

42:01

said to do. Here is why I

42:03

went out and chose to do something

42:05

different than that formula. So you'd give

42:07

more, you'd make it more

42:10

of the default. So I have now

42:12

more sympathy for policy rules. And

42:14

is that coming kind of out of

42:16

a Friedmanist monitorist, or maybe not

42:18

monitorism per se, but the idea of

42:20

having predictable default rules then? And

42:22

if you deviate from them, you've got

42:24

to explain why. Yeah. And Milton

42:26

Friedman had one for the growth of

42:28

the money supply. That's not one

42:30

that I think really anyone supports anymore

42:32

because the relationship between the money

42:34

supply and both inflation and output really

42:36

broke down. But John

42:39

Taylor, who's in some ways an heir

42:41

to Milton Friedman, he came up with the

42:43

Taylor rule. And for a long time,

42:45

Republicans were like, hey, this is the thing

42:47

you should do. Democrats were saying, they're

42:49

so wise. Why would they want to listen

42:51

to this rule? It's much better to

42:53

listen to their wisdom. I now

42:55

have a little bit of

42:57

the conservative understanding and skepticism of

42:59

the wisdom even of the

43:01

very best experts. And that says

43:03

a dumb formula isn't perfect,

43:05

but it probably is at least

43:07

less biased and systematic errors

43:10

than you might get from people.

43:12

Why did you become an

43:14

economist? I really

43:16

liked... and a

43:18

logical way of thinking about the world.

43:20

But I really cared about all the

43:22

issues that we're talking about now. I

43:24

mean, I wrote an article in my

43:27

high school newspaper about the budget deficit.

43:29

And I haven't found a copy of

43:31

that. I'd love to find what I

43:33

wrote. Were you for it or against

43:35

it? Whatever it is, I've changed my

43:37

view since that article many times since

43:39

then. So I really like the set

43:41

of issues, but I liked being able

43:43

to think about them in a logical

43:46

and rigorous way. And economics let me

43:48

do that. Do you feel like economics?

43:50

as an academic discipline, is it becoming

43:52

more and more tainted by politics and

43:54

ideology, or is that just a wrong

43:56

way to think about things? If you

43:58

go to an economics there

44:01

is a huge premium on

44:03

coming up with novel and

44:05

counterintuitive ideas. Somebody comes

44:07

up with a conclusion that you

44:09

like, you don't want to... applaud

44:11

them. You want to point out

44:13

some really stupid mistake that they

44:15

made and have everyone think you're

44:17

smart. So I think the process

44:19

of refereeing, of debate, et cetera,

44:21

is healthier than it is in

44:23

most of the social sciences or

44:25

in all of the social sciences.

44:27

That being said, if you look

44:29

under the age of 50 in

44:31

the economics profession, most people are

44:33

on the left. They're not on

44:35

the right. And that does

44:37

affect the questions that people ask,

44:40

the moral approach that underpins their

44:42

work, and maybe the amount of

44:44

effort they put into trying to

44:46

find errors in things. So I

44:48

would love an intellectual ecosystem in

44:50

economics that was more ideologically balanced

44:52

than what we have now. I

44:55

don't think it's terrible. There's certainly

44:57

not discrimination. A lot of the

44:59

key institutions are run by people

45:01

that are older and that actually

45:03

are Republican. I

45:06

think we'll get better answers to

45:08

the questions we care about with

45:10

people coming at it from different

45:12

perspectives. And the main different perspective that

45:15

we're missing is one that's skeptical

45:17

of government, skeptical of the wisdom

45:19

of the people turning the knobs

45:21

and doors. Where does that kind

45:23

of change come from? There's groups like

45:25

Heterodox Academy, which is trying to

45:27

create ideological diversity among faculty. How

45:31

does that happen? Essentially

45:34

through hiring practices or?

45:37

Yeah, look, I mean, people talk about

45:39

disparate treatment where you have discrimination against

45:41

a group. You don't want to hire

45:43

a conservative. I think that might be

45:45

true in some departments. I don't think

45:47

it's true in economics. Then

45:50

there's disparate impact where you just end

45:52

up with different ratios. When

45:54

you have disparate impact, the first step is

45:56

to just be aware of it. And

45:58

then maybe at every stage might

46:00

find something. In history, for example, if

46:02

you went out and rather than

46:04

hiring someone to do sort of ethnic

46:06

studies, you hired someone to do

46:08

military history and you made your military

46:10

history search completely non -ideological, you're more

46:12

likely to pick up someone from

46:14

the conservative side advertising for that job

46:17

than a different type of job.

46:19

So it's the types of questions you

46:21

ask is part of it. You

46:23

know, in the economics of crime. How

46:25

much are you studying the problems

46:27

that police cause with the way they

46:29

police? How much are you studying

46:32

the problems that crime cause for the

46:34

people that are victims of crime?

46:36

So what research gets funded, for example?

46:38

It's much easier to get funded

46:40

studying problems that face girls than problems

46:42

that face boys. Richard Reeves has

46:44

done a great job of highlighting the

46:46

problems that face boys. Maybe

46:48

the funders will start to listen and

46:51

fund research on that problem. Let's

46:54

talk about AI because this is

46:56

a place where a few years ago

46:58

you wrote, I don't think AI

47:00

is going to be that big a

47:02

deal, that it's going to be

47:04

a kind of technological improvement that will

47:06

be kind of rolled out and

47:09

won't be that disruptive. Now

47:11

you seem to be much more

47:13

bullish on AI. Can you explain why?

47:15

Yeah. So first of all, I

47:17

still have the same view I had

47:19

before, which is our core problem

47:22

we have now is that we have

47:24

too little AI, not too much. Productivity

47:27

growth in some ways is a measure

47:29

of how many things people used to

47:31

do that machines can now do. And

47:33

it's not nearly high enough. I'd like

47:35

to see it much higher. So the

47:37

biggest problem we have is how to

47:39

have more AI. On the

47:41

disruptive side in terms of jobs,

47:43

I also continue to think that AI

47:45

is going to make new types

47:47

of jobs. that it's only going to

47:49

replace parts of some types of

47:51

jobs, that it's going to make us

47:53

richer, so we're going to want

47:55

to consume more services, and all of

47:57

that. So in broad

47:59

strokes, I still think the same

48:01

thing. I'm just much more optimistic,

48:03

and I'm much more optimistic because

48:05

they made breakthroughs at a pace

48:08

that a lot of the experts

48:10

in this area didn't think they

48:12

would. There's still a question of

48:14

how that gets adapted and used

48:16

in the workplace, how long that

48:18

process takes. So I've shifted towards

48:20

a more positive direction. Because there's

48:22

more of it and it seems

48:24

to be speeding up. Yeah. Now,

48:26

another piece, though, of AI that

48:29

I think has been really interesting to

48:31

me is how you think about

48:33

the digital giants and how concerned are

48:35

you about competition in the digital

48:37

space. I continue to be concerned. I

48:39

think antitrust enforcement is important because

48:41

I think competition is the source of

48:44

innovation. But what's really surprised me

48:46

is that the existing incumbents weren't the

48:48

ones that were able to dominate

48:50

in AI. Why does that surprise you,

48:52

though? Isn't that always the case,

48:54

that Microsoft comes along and knocks? knocks

48:57

off IBM even as they have

48:59

a kind of complementary relationship with

49:01

them and whatnot. Yeah, I agree.

49:03

It's happened over and over and

49:05

over again. So maybe this is

49:08

like the opposite of Lucy and

49:10

the football or something. But my

49:12

worry was that we were in

49:14

a place where AI depended so

49:16

much on data and so much

49:18

on scale and that nobody other

49:20

than Google and you

49:22

know, Facebook, Meta would have that

49:24

scale and those resources. And instead,

49:27

what we've seen is it turns

49:29

out that the algorithm matters a

49:31

lot too. And so open AI

49:33

and anthropic and perplexity, you know,

49:35

came to some degree from nowhere.

49:37

They got funding from some of

49:40

the big companies. DeepSeek came

49:42

from even less, probably more than

49:44

they claimed. Does that give you pause

49:46

about like, oh, you know, the

49:48

Chinese communists are going to somehow be

49:50

running everything? On

49:52

AI, I think there's some motivation

49:54

that comes from geopolitics, but boy,

49:57

is it hard to contain innovations

49:59

there. And on foundational models, it's

50:01

hard to be more than six

50:03

months ahead of the next person. And

50:06

so microchips is much

50:08

easier to maintain a

50:10

five, 10 -year lead

50:12

on than it is

50:15

on these algorithms. As

50:17

a final question, Talking

50:19

about AI and the fear

50:21

that we're going to wake

50:24

up tomorrow and whole industries

50:26

will be gone and there

50:28

won't be any jobs left

50:30

for certain types of people.

50:32

Your former boss, Barack Obama,

50:34

famously talked about how ATMs

50:36

had wiped out bank teller

50:38

jobs. This is a kind

50:40

of classic example of where

50:42

people think of technology as

50:44

something that robs people of

50:46

work. You

50:48

know, that certainly is not the case

50:51

even in the retail banking industry. Why

50:53

are we prone to that error

50:55

again and again, that a new

50:57

technology that actually makes things more

51:00

productive is going to put people

51:02

out of work? Be

51:04

clear. I think you got it right.

51:06

But what Obama said is people expected ATM

51:08

to take teller jobs. But instead, we

51:10

ended up with more bank branches because of

51:12

ATMs. And so you ended up with

51:14

more bank tellers, not Or people in the

51:17

banking industry. Yeah. So, you know.

51:19

We have hundreds of years of experience

51:21

with this. You know, no one could

51:23

have imagined what would happen in a

51:25

world where most people weren't working on

51:27

farms. What will they all do? What

51:29

else is there to do? Well, they'll

51:31

go to factories, but now they're not

51:33

in factories. Where have they gone? So

51:35

just our imagination isn't that great, but

51:37

it also doesn't need to be that

51:39

great because that's the beauty of the

51:42

market. Everyone's out there, the businesses, the

51:44

individuals trying to find new things, trying

51:46

to start a new business, trying to

51:48

figure out a new job. You

51:50

know, the central planners don't figure that

51:52

out. It's, you know, sort of a

51:54

Hayekian spontaneous order. And it's worked out

51:56

pretty well for a long time now.

51:59

So that would be my best guess.

52:01

So it will. But then, you know,

52:03

is there a stronger argument to say

52:05

to people that like, hey, you know,

52:07

in the past it all worked out? Or,

52:11

you know, trust the

52:13

Hayekian extended order and decentralization?

52:15

Well, look, you

52:18

know, there's also disruption. I mean,

52:20

the Luddites weren't wrong. These

52:22

machines were bad for the Luddites.

52:24

They were good for society as a

52:26

whole, but not for them. And

52:28

so part of why people are nervous

52:30

is because they don't understand what's

52:32

going to happen. And if only somebody

52:34

could explain it to them, they'd

52:36

be thrilled. Part of why they're nervous

52:38

is because they do understand. And

52:40

they understand that it won't be good

52:42

for them. And it's hard

52:44

to make the longer term

52:46

positive sum case. for the world.

52:49

But if you constantly make

52:51

whatever decision is loss minimizing over

52:53

the next minute, but has,

52:55

you know, costs that grow over

52:57

time, you know, you'll end

52:59

up more like Argentina over time

53:01

and less like the United

53:04

States. Okay. Well, without getting into

53:06

a discussion of Javier Millet,

53:08

who's remaking Argentina, we're going to

53:10

leave it there. Jason Furman,

53:12

thanks for talking to Riva. It's

53:14

been great talking to you.