Latest / The Capitalism and Freedom in the Twenty-First Century Podcast / Technology, AI, Political Economy, and Economic Development

Transcript

Technology, AI, Political Economy, and Economic Development · The Capitalism and Freedom in the Twenty-First Century Podcast

0:00
>> Speaker 1: It's that time of the year.
0:02
Your vacation is coming up.
0:04
You can already hear the beach waves, feel the warm breeze.
0:09
Relax and think about work.
0:13
You really, really want it all to work out while you're away.
0:16
Monday.com gives you and the team that peace of mind.
0:19
When all work is on one platform and everyone's in sync,
0:22
things just flow wherever you are.
0:24
Tap the banner to go to Monday.com.
0:25
[MUSIC] [MUSIC]
0:26
>> Jon Hartley: This is the Capitalism and
0:30
Freedom in the 21st Century podcast
0:36
where we talk about economics,
0:42
markets, and public policy.
0:47
I'm Jon Hartley, your host.
0:49
Today, I'm joined by Simon Johnson, who is the Ronald Kurtz
0:52
professor of entrepreneurship at the MIT Sloan School of Management.
0:56
And was previously, chief economist to the IMF.
0:59
And just co-authored a fascinating newly published book with his MIT
1:03
colleague Daron Acemoglu titled Power and Progress,
1:06
Our One-Thousand Year Struggle Over Technology and Prosperity.
1:10
Welcome, Simon.
1:11
>> Simon Johnson: Thanks for having me.
1:12
>> Jon Hartley: So you're British?
1:14
You were born in the UK, how did you first get interested in economics?
1:18
>> Simon Johnson: Well, I was looking for engineering, but on a grander scale,
1:22
I think.
1:23
And economics captivated me early on as something that has the potential to
1:27
address really, really big issues.
1:28
So I was lucky to fall into it when I did and
1:31
lucky to get the education that I received.
1:33
And I think it's been all fascinating questions since then.
1:37
>> Jon Hartley: You were an MIT economics PhD student,
1:40
certainly the best place to get an economics PhD.
1:44
Were you influenced by anything, say, your childhood?
1:47
I know for some that grew up in the 1970s UK, it was sort of an important experience
1:52
for them, or as they grew up sort of under when Margaret Thatcher was Prime Minister.
1:57
Was any of that formative for you in terms of your sort of outlook and kind of
2:02
global growth in a lot of the topics that you've written on, politics as well?
2:07
>> Simon Johnson: Yes, I think in retrospect, I'm from Sheffield in
2:10
the north of England, a city that was an economic powerhouse.
2:13
My family was involved in making screws on one side and
2:16
making steel on the other side.
2:18
And by the time I was aware of the economy, things had turned downhill.
2:22
And all of the industry there was troubled.
2:25
There was a lot of job losses, people were moving away.
2:27
So I think I'm still trying to understand exactly what went wrong there and
2:33
why we lost so many jobs.
2:34
It does give me a bit of a lens for thinking about economic decline, but also
2:37
recoveries and more sustained growth and innovation and entrepreneurship elsewhere.
2:42
>> Jon Hartley: So I wanna get into your research agenda.
2:45
You've written a lot of highly influential articles in political economy.
2:50
How would you describe your research agenda?
2:52
You've got over 80,000 citations,
2:57
that is quite an accomplishment.
3:02
You co-authored The Colonial Origins of Comparative Development, which is
3:07
both your own most cited paper, as well as Daron Acemoglu's most cited paper.
3:13
You've been at the center of all this groundbreaking research on institutions
3:18
and growth, which has been highly influential for several decades now.
3:22
Certainly had a lot of influence on the World Bank and the IMF and
3:26
their approach to promoting economic growth.
3:30
How would you describe your research agenda in your own words?
3:34
>> Simon Johnson: Well, I think we're looking for the causes of poverty and
3:36
wealth around the world, and many people, of course, have done that before,
3:39
and we're standing on their shoulders.
3:41
But there's a lot of really interesting questions, including,
3:45
after the collapse of colonial empires, when technology in principle,
3:48
could have flowed freely to many places.
3:50
And many places around the world could, in principle and in theory,
3:53
have become much richer, they didn't, so why not?
3:56
What are the actual impediments that these countries face?
3:59
And what could we do in terms of academia or in terms of public policy or
4:04
in terms of the IMF?
4:05
What could we do that would actually help the billions
4:09
of people around the world who wanna live better?
4:12
>> Jon Hartley: And you were chief economist at the IMF,
4:15
right after Anne Krueger?
4:16
Or remind me again, in sort of your time at the IMF,
4:21
what was the big thing happening then?
4:24
>> Simon Johnson: So I became chief economist at the IMF in early 2007,
4:27
I held that job until the end of August 2008.
4:30
So this was the run up to the global financial crisis.
4:32
I'd worked there previously for two years under Raghu Rajan,
4:35
who of course, is a brilliant financial thinker.
4:38
And he pulled the IMF and the part of the IMF that I was in,
4:40
the research department,
4:41
towards thinking about the interaction between finance and macroeconomics.
4:45
And when I got that job, I was Raghu's successor.
4:48
We were immediately, the IMF, my department was in the hot seat in terms of
4:53
understanding what is going on with mortgage backed securities,
4:57
with credit default swap spreads.
4:59
How will this spread around the world?
5:01
What are the sensible interventions that could be made?
5:03
So it was a ringside seat for the events that became, after I left,
5:08
somewhat devastating to millions of people.
5:12
And from that, I took away a lot of questions and a lot of interest in trying
5:16
to resolve those questions with regard to making the financial system safer.
5:21
>> Jon Hartley: You wrote a book, 13 Bankers, really, that whole episode, and
5:26
you've written quite a bit on financial cycles and business cycles as well.
5:33
Now your latest book, Power and Progress,
5:36
it is primarily on economic growth, political economy and technology.
5:42
Could you tell us a little bit about the book and
5:45
the point that you're trying to make with Daron?
5:49
>> Simon Johnson: Yes, I think the book both for Daron and me,
5:51
is a bringing together of work we've done on political economy together and
5:55
work we've done on technology, which tend to be somewhat separate.
5:59
And so we've merged these things.
6:00
And I think it's the political economy of technological choice.
6:03
And it asks the question first and foremost, is technology predestined,
6:06
just something that happens to you?
6:08
And we think the answer is no, there's a lot of choices in there.
6:10
And secondly, is it possible to redirect technological change?
6:14
Can you push it one way or another?
6:16
We see influential individuals doing that, we see companies doing that.
6:20
Can public policy do it?
6:21
Can civil society do it?
6:22
And if so, which direction?
6:24
In which direction would you like to push it?
6:26
>> Jon Hartley: Got it, so there's, these topics, directed technological change,
6:30
distorted technological change.
6:34
To me, it sounds a little bit like the words industrial policy.
6:38
To what degree can policymakers really sort of shape industries in direct
6:43
technology, and what are their limitations like?
6:46
To what degree should they really be involved in, in directing technology?
6:51
Obviously, industrial policy has been
6:53
a controversial set of words in economics for quite some time,
6:58
certainly less controversial in some circles versus others.
7:02
But I'm curious, what is your thesis on topics, directed technological change and
7:07
industrial policy?
7:09
>> Simon Johnson: Well, policymakers definitely can shape technology and
7:12
innovation.
7:13
They do it with the tax code, they do it through advantages for certain industry.
7:17
They do it through the Department of Defense.
7:18
I think a really interesting and still to be totally resolved question is,
7:22
can they tilt it in a way that would be more broadly in the public interest?
7:26
So it's not about capture, it's not about lobbying, but it's about more good jobs,
7:31
for example.
7:31
And we think the answer is yes, we think it's not easy.
7:35
We think it's quite a struggle, requires a fair amount of debate and argument.
7:39
That's one reason we wrote the book, to try and push that debate forward.
7:42
For example,
7:43
if you think about the way that artificial intelligence is being developed now.
7:47
A lot of emphasis on machine intelligence, which is euphemism for replacing workers,
7:52
and replacing them in a way that doesn't raise the marginal worker productivity of
7:57
those who remain employed in the same enterprise.
8:00
We would rather tilt it towards machine usefulness,
8:02
which is a term that we've invented, but we're standing on the shoulders
8:06
of a lot of computer science people who came before us.
8:08
Who say, the point of machine should be meant to augment human capabilities,
8:12
not to replace them, but to augment them.
8:14
I think it's easier to see that in history which way we've gone in various moments.,
8:18
it's easy to talk about the creation of new tasks, stimulating the demand for
8:21
labor, and potentially raising productivity and wages.
8:23
But can that be done in a more deliberate policy driven fashion.
8:26
I think it's a fair question, and that's what now lies before us,
8:30
trying to sort out that question.
8:32
>> Jon Hartley: So there's a popular set of words in sort of the labor economics
8:36
and technological change literature, which is skill bias technological change.
8:42
And in this sort of introduction of new AI tools, things like ChatGPT and
8:48
generative AI, there's some who argue that it's actually gonna
8:53
be skilled jobs that are gonna be severely impacted.
8:57
For example, coders, those in the service economy, those that are writing things, I
9:02
guess, in their service jobs, those jobs, we don't need as many coders anymore.
9:08
We'll have fewer software engineers that will use things like Copilot and
9:12
ChatGPT to code faster.
9:13
I'm curious, what is your view on that?
9:15
Do you think we're sort of in for this reversal where skilled workers
9:19
have done so well for so long over their unskilled, less educated counterparts?
9:24
Do you think AI is gonna trigger a big reversal there?
9:28
Or do you think just broadly, both skilled and
9:30
unskilled work is sort of in jeopardy here?
9:33
>> Simon Johnson: Well, robots replacing unskilled workers as well pretty quickly.
9:37
If you read the novels of Isaac Asimov,
9:40
which I really do advise to everybody [LAUGH].
9:42
Because he had a brilliant imagination of the robot novels in particular,
9:47
what would happen in the 1940s and 1950s when he was writing those.
9:51
The thinking was that robots would take over manual jobs first, and then later,
9:56
as the robots became more evolved, they'd move on to cognitive tasks.
10:00
I think what we've discovered, including the last two years, is the opposite.
10:05
That artificial intelligence is quite good or
10:07
better than humans at some cognitive tasks and really not as good as humans, and
10:12
won't be for a long time at much more manual tasks.
10:15
And that's probably because we've been walking around and
10:18
walking across rough terrain for millions of years, us and our ancestors.
10:23
But abstract thinking is tens of thousands of years old, and we've only been going to
10:27
meetings, writing memos to each other for a few hundred years of the modern soul.
10:31
So I think if you're unskilled, I would call it manual work and cognitive work,
10:35
the pressure is on the cognitive jobs.
10:38
Now within that, there's a lot of variation already developing,
10:42
some occupations, it seems like it's the most highly skilled,
10:46
highly paid people who will benefit.
10:48
But in other places, they're the ones who are gonna get fired and
10:51
then people who own the capital are gonna replace them with much cheaper labor.
10:53
So I think it's absolutely all in flux and
10:56
our main point is it doesn't have to be about replacing workers.
10:59
It can be about augmenting the capabilities, including augmenting
11:02
the capabilities of lower-skilled workers with some sort of cognitive function.
11:06
And that's really interesting because then there could be much more by way of
11:10
productivity gains, individual productivity gains,
11:12
economy wide productivity gains, and wage gains that would filter down.
11:15
So this doesn't have to be a zero-sum game or some people win, some people lose.
11:20
Many more people could win than lose in this instance, but
11:23
that may not be the default course we're on currently.
11:26
>> Jon Hartley: That's fascinating, I'm curious, what do you think about this
11:31
whole argument that we've seen over the past decade, gained,
11:36
I think, some traction in public intellectual circles that,
11:41
this new era of AI is due to unleash massive unemployment?
11:45
And hence we need something like universal basic income to sort of step in and help
11:52
those that are significantly displaced from this technological innovation.
11:59
>> Simon Johnson: Yeah, we're not big fans of Universal Basic Income, UBI,
12:02
primarily because we think people like to work.
12:04
Now it doesn't mean that they have to work 60 hours a week in backbreaking jobs.
12:08
But work seems to be important as a source of income,
12:11
a source of identity, a source of political voice.
12:14
And UBI, I think,
12:15
lets the technology industry off the hook with regard to its impact.
12:19
And I would rather they be on the hook and be responsible and
12:22
not point fingers at other people and say, right,
12:24
it's your job to take care of all these people who've gotten fired.
12:27
I think it's also the case, by the way, that in economy like the US,
12:30
you don't get that much unemployment.
12:32
What you get is people pushed down to very low wage jobs and
12:34
they fall out of the labor force, so labor force participation declines.
12:38
We've certainly seen plenty of that from the digital transformation of the past
12:41
four decades, so I would not put too much weight on a UBI type approach.
12:45
>> Jon Hartley: Interesting, and you think about people like Milton Friedman is
12:49
writing in the early 1960s in Capitalism and Freedom, he famously sort
12:53
of endorsed the negative income tax, sort of a universal basic income.
12:58
I think he famously had some sort of analogy that
13:03
better to pay people something like UBI type payment rather
13:09
than to have them shovel holes and fill them up again.
13:14
I think he had some story about in the Soviet economy,
13:19
there was people sort of that were doing something like this, I guess.
13:24
Do you think that it makes sense to, I guess,
13:28
direct people toward more unproductive types of work in that sense?
13:34
Or do you think that that sort of view has any validity?
13:38
I mean, obviously, it's much more complicated than that, but
13:42
do you think to any degree, these sorts of industrial sort of policy
13:45
type ideas are sort of pushing people toward more unproductive forms of work?
13:50
Where maybe it makes more sense to have something like, I guess,
13:54
a more generous kind of redistribution system.
13:58
Maybe not like UBI, but maybe a more enhanced earning tax credit or
14:02
something like that to respond to this?
14:04
>> Simon Johnson: Yeah, I haven't seen anybody really proposing truly
14:07
unproductive work.
14:09
I mean, yeah, there's a question of how much you subsidize the construction of
14:12
new chip fabs, for example, in the US.
14:14
And there's a very interesting debate about how much more expensive it
14:17
is to build such fabs in the US versus in Taiwan, for example.
14:19
The Taiwanese people say, it's four times expensive, my expert friends at MIT say,
14:23
it's 20% more expensive.
14:25
So we're gonna find out.
14:26
>> Jon Hartley: What's sort of a worker shortage thing with that as well.
14:29
>> Simon Johnson: Well, so-.
14:29
>> Jon Hartley: That's one of the complaints that I think TSMC and
14:32
some of these chip companies have with [INAUDIBLE].
14:35
>> Simon Johnson: Yes, they do say that, when you drill down into it, it's so
14:39
interesting, so they use PhD qualified engineers on the shop floor in Taiwan, and
14:43
it's not clear to everyone that that's a super productive use of PhDs.
14:47
And there may be a malleability of labor and
14:50
a willingness of highly skilled people to do mundane or even routine tasks.
14:54
In Taiwan, you also hear that about South Korea compared to the United States.
14:58
So I think we are attempting to find out what kind of training is needed,
15:02
where those people will come from.
15:04
These are good jobs in the fabs, who's gonna get the job?
15:07
How much education do you need to be effective in the clean rooms and so on?
15:11
Very, very interesting and important questions, but I'm pretty optimistic in
15:16
this country, which is a big country with a lot of people who wanna work hard,
15:20
I think we will find plenty of talented people.
15:23
Back to just Friedman for a second, I think that the question that we
15:28
grapple with is first of all, are we taxing labor too much?
15:32
Payroll taxes, raising the cost of labor at a time when people are thinking,
15:35
machines versus labor, which one do I wanna go for?
15:37
I think that that's an important issue.
15:39
And I think also that we have a lot of issues around the care economy,
15:42
around how much are you willing to pay out of the public purse for
15:45
home health aides, who take care of people who have really terrible health problems?
15:50
You don't wanna send them to hospital, that's way more expensive, somebody has to
15:53
take care of them and their families, can't afford to do it by themselves.
15:56
How much are you willing to pay?
15:57
Can you pay a living wage?
15:58
And if we don't, who really goes in into that work?
16:01
So I wouldn't call that at all unproductive work, but
16:04
I think it is a massive question of how we handle additional longevity,
16:08
how we handle many extra years of health,
16:10
but also some ill health at the end of life, those are tough questions.
16:14
On the whole tax side of things, it seems like in recent years there's
16:18
obviously been a lot of interest in taxing wealth.
16:22
There's been a lot of big push from folks like Emmanuel Sieyès, Thomas Piketty,
16:27
Gibraltar Zucman on sort of things like a wealth tax.
16:30
And I guess, more in the sort of inequality frame than perhaps it's
16:35
certainly related to the sort of AI and
16:38
gains to capital discussion that we're talking about now.
16:42
But I'm curious, there's been people like Bill Gates have talked about,
16:46
having a robot tax, what do you think about that idea?
16:49
So I do think that tax code is a bit too tilted towards encouraging machines versus
16:54
hiring labor.
16:55
So I think that can be redressed in various ways, but
16:57
in contrast to those people who are arguing for redistribution.
17:01
So let the productive process do its thing, look at the distributed outcomes,
17:06
and then do some tax and subsidy welfare payments if you don't like the outcomes.
17:10
We are much more about tilting or
17:12
redirecting technological progress to change the outcomes that way.
17:15
And that's primarily,
17:16
because I don't think a place like the US is ever gonna do a lot of redistribution.
17:20
And I think the nature of work and the kinds of jobs you get,
17:22
that's important in and of itself.
17:24
And redistribution, of course,
17:25
doesn't address that just gives you a bit more money for the work that you've done.
17:29
So thinking about how to redirect technological progress,
17:32
that's our main agenda in this book.
17:34
>> Jon Hartley: Got it, but sort of on this question of massive unemployment,
17:38
I guess the unemployment rate is between three and a half percent and 4% currently.
17:43
You don't really see there being like a massive surge in unemployment,
17:48
say, anytime soon, or?
17:50
>> Simon Johnson: Well, [LAUGH] I think my former colleague and
17:52
good friend Mike Musa, who sadly departed us, left us, liked to say,
17:57
his main advice for anybody who worked at the IMF was never forecast a number and
18:01
a date in the same forecast, right?
18:04
So look, I think the ChatGPT is a wake up call and there's a speed of change in some
18:09
of these cognitive tasks, the capabilities of AI, that is disconcerting.
18:13
Because we know, particularly in contact the US, we can handle a lot of different
18:17
shocks thrown at us, but there is a speed of adjustment issue.
18:20
So I don't think we're gonna face mass unemployment,
18:22
I do think there's gonna be pressure on some jobs that were previously good jobs.
18:26
And I do worry that we may not be creating, that's automation and
18:29
that's a natural part of economic development.
18:31
The offset and what we were good at in the early 20th to mid 20th century was
18:35
creating a lot of new tasks.
18:36
So that people had jobs, were employed and
18:39
we could absorb all the people who moved out of agriculture,
18:43
all the jobs that were eliminated when Henry Ford automated car production.
18:48
We didn't lose jobs in the car business,
18:50
we gained jobs in the peak phase of that transition, but it takes some time.
18:55
Electricity was adopted over 30 or 40 years, electrification of factories and
18:59
production.
19:00
AI seems to be coming at us a lot faster than that, so
19:03
I think we need to step up our game in terms of response time.
19:06
>> Jon Hartley: And I guess in this adjustment, I guess at some point some
19:10
people will leave the workforce, but maybe some of those people will be
19:15
retrained to manage the robots or manage the AI is, I guess, the hope in that.
19:20
I guess,
19:21
pivot a little bit to the political economy sort of discussion around AI.
19:27
So I'm curious, what do you think about when we talk
19:32
about AI in the context of totalitarian countries like
19:37
China where they're using AI for, I guess, suppression of its people.
19:44
I'm curious, there's a sort of idea that was promoted by Milton Friedman
19:49
that's partially related to the work that Jerome and yourself have done.
19:54
Milton Friedman's argument,
19:55
this is like the first chapter of capitalism and freedom.
19:58
And the idea is that, growth would cause democracy or
20:00
economic freedom would cause political freedom.
20:03
Jerome and his co-authors, the sort of institutional framework, I think,
20:08
has also made some arguments that democracy also causes economic growth.
20:12
This thing with China, I think hasn't quite played out,
20:16
it's still not a politically free society, but
20:19
it has experienced quite a bit of growth in recent decades that may be fading now.
20:24
But I think something that Friedman acknowledged later in life was that what
20:29
he had argued in capitalism and freedom was not playing out with China.
20:33
That becoming a much more prosperous society did not ultimately cause
20:38
it to become democratic, and to this day it's not democratic.
20:43
How do you think things like AI are sort of disrupting these traditional processes
20:47
of growth and democracy as being something that we would get together and
20:50
some sort of indigenous relationship?
20:52
How do you see that impacting that symbiotic relationship?
20:55
>> Simon Johnson: Well, I think there's been a problem there for a while,
20:58
I remember for some reason there's vivid image of Milton Friedman in Hong Kong.
21:01
I think it was his PBS special, talking about capitalism and
21:04
freedom exactly this way.
21:05
I remember thinking, well, that's good, all we need to do is get growth,
21:08
then we come more like Hong Kong.
21:09
And first I went to Hong Kong probably in the early 1990s, I was quite impressed,
21:13
but I think all that have not been recently.
21:15
But all the stories from Hong Kong, including right now,
21:18
are quite discouraging in terms of the way in which freedom, in any Friedman sense or
21:23
any sense, has been suppressed, and it's become an oppressive place.
21:27
So I think that's a wake up call.
21:28
Now is AI a technology for liberation, a technology for self expression,
21:33
or is it a technology for surveillance and suppression?
21:37
And the answer is, yes, it's both right, and it depends on how you use it.
21:41
So I think that the new split in the world is going to be between countries that
21:44
are more like us and are gonna put, I think we'll have to put a lot of
21:48
guardrails around surveillance, for example, and that's a good thing.
21:51
And then there'll be other countries that put no guardrails around surveillance with
21:54
regard to state surveillance, certainly.
21:56
And they'll be following China, and
21:58
they will be following a certain line of technology.
22:00
And if you think about, being authoritarian is obviously a set of
22:04
policies, and it's implemented by a technology.
22:07
So how much does it cost you to be an authoritarian or
22:09
run an authoritarian regime?
22:11
Well, I think the AI, as developed by the Chinese,
22:13
is gonna make it a lot cheaper to be authoritarian.
22:16
So if you think in terms of an incentive framework,
22:18
we're gonna have more authoritarians, and they're gonna last for longer.
22:22
Facilitated by the same technology that I hope, I believe we will use to strengthen
22:26
freedom and true liberty in the United States.
22:28
>> Jon Hartley: One of the interesting things I think about this book is that it
22:32
really talks about technology in a sort of more political sense.
22:38
And it's not usually a frame that we typically, as economists or
22:42
casual readers, really think about technology.
22:45
It seems to me like it's only been recently where we think about the politics
22:50
of big tech and censorship, and obviously, that's a very big topic right now.
22:55
But traditionally, it's usually been, great, we have iPhones now or
22:58
we have computers now, this is great, we can do things faster.
23:02
Traditionally, at least in my own lifetime,
23:05
technology had a relatively sort of positive, sort of connotation with it.
23:11
I'm just like,
23:12
what are your favorite examples from your book about this sort of struggle,
23:17
where elites have used sort of technology in some sort of way that may not have
23:21
been to promote broad based prosperity as much as it could have otherwise?
23:26
I'm curious, or
23:27
what are some of the examples in the book that you sort of outlined?
23:31
>> Simon Johnson: Well, I think the most awful example, the cotton gin, right?
23:34
Which was developed right at the end of the 18th century,
23:38
right after American independence.
23:40
It made it easier to process upland cotton, it meant you could run cotton
23:43
plantations away from the east coast, across the deep south.
23:45
And what happened, of course, was enslaved people were moved from a very harsh life
23:49
on the east coast into a much worse conditions across the deep south.
23:53
And that became the mainstay of the slave economy, and that lasted and
23:58
remained with extremely harsh conditions until the Civil War.
24:03
And of course, that was also facilitated and supported or driven by
24:06
industrialization in Britain, which was about cotton and cotton textiles.
24:10
And they were buying a lot of the raw cotton from the American South.
24:13
So I think that it's actually rather pervasive throughout history that
24:16
some people gain and other people lose from the deployment of technology.
24:20
Much more unusual and with the sort of holy grail we're looking for
24:24
here is the sort of Henry Ford experience, where Henry Ford automates car production,
24:28
brings electricity to the factory.
24:30
He replaces a lot of workers, but he also generates a vast number of new tasks.
24:35
And there's a class of people, call them, and
24:38
we've adopted the term manager engineers.
24:40
White collar workers who emerge to plan this to organize it,
24:44
to run all the infrastructure around car production.
24:48
And that's a lot of people, and that is a lot of money that's made and
24:52
very high wages that are paid through sharing.
24:55
I don't think Henry Ford was that altruistic,
24:57
I think he was quite paternalistic.
24:58
But he faced some countervailing power, including trade unions, and
25:01
he wanted to preempt that with high wages.
25:03
And he also had to at some point negotiate with the unions and
25:07
have collective bargains.
25:08
But that was in a context of the company did well, his family did well,
25:12
the whole auto industry did well.
25:13
So finding that kind of win-win approach, win-win solution,
25:17
we did a lot of that before, during, and after World War II.
25:21
After 1980, it's become much more unusual.
25:25
>> Jon Hartley: I mean, it's fascinating,
25:27
I guess, in the past few decades that we've had these productivity gains, but
25:32
we haven't quite had real wage gains to sort of follow that.
25:36
Do you have a favorite explanation of why that's not
25:40
following as much historically as it's been the case,
25:45
that real wage gains are generally fall productivity gains?
25:50
There's a lot of, I feel like,
25:52
different narratives that are told about why this is the case.
25:55
>> Simon Johnson: Right, so real wage has obviously risen for some people,
25:58
more skilled people.
26:00
It's for the lower, less skilled, less educated people, that they haven't risen.
26:04
I think the main contenders are automation and globalization.
26:07
And Geron's work with Pascual Restrepo, says that automation is 60,
26:11
70% of the explanation.
26:12
I think even if you lowered it a little bit and said, well, I think globalization,
26:15
the China shock is a bigger part of it.
26:17
You have to remember that there's a technology dimension to globalization,
26:21
including communication, including computers being applied to trade.
26:25
And so if you look at this, I would say the direct effect of technology change and
26:29
automation.
26:30
And the indirect effect through globalization and
26:32
the shifts in where you put certain kinds of jobs, including manufacturing jobs.
26:36
I think you're looking at 70 to 80% being attributable one way or
26:40
another to the way in which technology has changed.
26:42
I do think that attitudes among management have also changed.
26:45
I've looked at a lot of literature what management were writing,
26:48
what executives were saying,
26:49
what people who led industrial organizations were saying in the 1920s.
26:52
They were nowhere near as confrontational and
26:54
antagonistic to workers as more recently.
26:56
They were much more patting themselves in the back and saying, hey,
27:00
capitalism is a win-win.
27:01
In fact, Herbert Hoover is a perfect example of exactly this.
27:04
And that's why the Republicans were so strong in control of the presidency in
27:09
the 1920s in the US, because they were seen as the men.
27:12
They were men of business, right, and that everybody could gain from that.
27:16
So there was a shared prosperity moment.
27:18
Okay, a particular rock was hit in terms of the Great Depression, but
27:22
there was also recovery from that.
27:23
And Eisenhower is another Republican who was regarded as a practical man, okay?
27:28
He was a military general, but he was also pro-business, and
27:31
the business of America was doing business.
27:34
>> Jon Hartley: It's also interesting to think about the 1920s,
27:37
even though we think it was at the time of prosperity.
27:39
I mean, it's also the time when you have zoning laws became legal in the US, and
27:43
the whole immigration system is sort of created as we know it today.
27:47
Sort of two of the largest sort of barriers to entry were created
27:51
during that time.
27:52
It's a very interesting sort of time period in general.
27:55
Things that probably didn't have much great effects till much, much later on.
27:58
But it's interesting, too, on points of disruption, too.
28:01
It's also, I think, a regional thing as well, where automation certainly impacts
28:05
the automobile industry, which is very much rooted in the Midwest.
28:09
Whereas the China shock,
28:10
it impacts textiles, which very much impacts the Appalachians.
28:13
So two very different geographical areas with two different regional labor markets
28:18
being disrupted in different ways.
28:20
Do you have any thoughts on the Luddites?
28:22
This whole term, it's sort of a famous term, is used in a derogatory
28:27
sense toward folks that maybe spout left leaning tendencies.
28:32
But it's very relevant, and probably one of the historical examples.
28:36
I would think of first, where it's this famous story of, and
28:39
it may be apocryphal, I'm not totally sure.
28:41
But it's this idea that people were destroying these textile machines,
28:45
I think, in the 1800s.
28:47
Do you think that's kinda a relevant example here of, and
28:52
I'm not sure if it's an apocryphal story or not.
28:56
I think a famous person associated with that as well,
28:58
I forget the name of the person.
28:59
>> Simon Johnson: Yeah, well, Lord Byron gave some great speeches about it, and
29:02
that's in our book.
29:02
Yeah, the Luddites were not apocryphal, Ned Ludd.
29:05
Whether there was something called Ned Ludd, that is questionable-.
29:08
>> Jon Hartley: Definitely.
29:08
>> Simon Johnson: Perhaps not, but look at that, so the Luddites were,
29:11
depending on how you read the history.
29:13
They were either skilled workers who were threatened
29:16
by the arrival of textile factories and automated looms, so
29:19
they were losing the weaving jobs that they had before.
29:22
Or I would actually say they were independent entrepreneurs, because by any
29:26
standard, like IR's definition today, they had their own equipment, they worked at
29:30
home, they controlled their own hours, right, they had a lot of autonomy.
29:33
And actually, either they weren't being offered jobs in the factory, or
29:38
if they had to come into the factory, they had to take on relatively routine tasks,
29:42
relatively unskilled tasks.
29:44
Which were supervising the machinery that took care of what they'd previously done
29:49
with skill in an artisanal fashion.
29:51
So I think that if you see them as independent entrepreneurs who
29:55
are threatened by the growth of big business,
29:57
[LAUGH] people could be a little more sympathetic.
30:00
But I think, honestly, also, stop trying to prevent automation and
30:04
prevent the elimination of jobs by machines is tilting at windmills.
30:09
However, it doesn't mean that's the only thing that can happen or should happen.
30:13
And in some phases of industrial development, including when the railways
30:17
came to Britain, about the same time, or slightly after the big Luddite,
30:21
that created a lot of new tasks, a lot of new jobs.
30:23
And a lot of jobs in which the railways wanted to pay premium wages to their
30:26
workers, because they wanted them to take care of safety on the railways.
30:29
Actually, that was an absolute key issue,
30:31
be very responsible in a relatively independent way.
30:33
So I think, technology can and has been developed in ways that generate new
30:38
tasks and a lot of new tasks at the same time as it automates existing work.
30:43
After 1980, we had less task creation than we really needed.
30:48
The demand for relatively unskilled labor has been weak,
30:51
we haven't been able to turn that around.
30:53
And there is a concern that AI will worsen that problem or maybe layer
30:56
on some other versions of that problem now and I don't think it has to be that way.
31:00
I think we could push, and the industry could certainly push to
31:05
develop algorithms and approaches that would complement human ability.
31:11
So don't replace humans, make them more productive, augment their capabilities.
31:16
And we would say, to the extent that raises marginal work of productivity,
31:20
paid them a high wage.
31:22
>> Jon Hartley: I wanna get back to this sort of broad topic
31:26
of growth in institutions.
31:29
Obviously, it's a highly related political economy topic here.
31:33
Can you explain which institutions matter?
31:36
For me, I've always been told that government policy and politics matter for
31:41
economic growth.
31:42
However, we wanna describe that, you call it institutions.
31:46
But I think the key question has always been, which institutions or
31:49
policies matter?
31:50
And I think this really summarizes a lot of your work,
31:54
as well as a lot of Jerome's work.
31:57
Can you explain what your taxonomy of institutions Is and
32:02
what ones are most important?
32:04
>> Simon Johnson: Well, I think a central institution on the economic side is
32:07
property rights, and that has to be supported by political rights.
32:11
And if you go back to the situation in Hong Kong, for example,
32:14
which had very strong property rights,
32:15
limited political rights under the British, but they were definitely there.
32:19
Those political rights have been stripped away, and
32:21
I think the property rights are under great pressure as a result.
32:25
But I would also say that, if you think about big technological transformations
32:29
and think about what happens in those transformations.
32:32
There's often a phase in which things that were regarded as common property and
32:37
shared or worked in a common way and maybe work quite productively,
32:41
like common land and medieval Britain.
32:43
There is an episode or part of the modernization of Britain in which that
32:47
land is turned into private property.
32:50
The communal rights are stripped out and they're not well protected.
32:53
The private property is then accumulated, becomes a big source of wealth, and
32:57
that wealth is used politically to take more common land into private property.
33:02
I think the same thing is happening now with data.
33:04
There was a lot of data we put on the Internet, a lot of data we allowed other
33:08
people to look at and to use through social media, for example.
33:11
Those data are now being used to train algorithms in ways that were not being
33:14
compensated.
33:15
They didn't ask for permission.
33:16
You may not like, I would predict that you wouldn't like some of the surveillance
33:21
outcomes that are gonna come from pictures that you and others posted on Facebook.
33:25
And that, I think, is exactly analogous to that taking of common property
33:30
without permission or forcing it through in the enclosure movement in Britain
33:35
before the industrial revolution.
33:37
And I think that recognizing and respecting property
33:40
rights on the Internet, including the rights to the data that you've created and
33:44
the images that you've put there, that's actually really important.
33:48
And if we could get some recognition of that, and we're already behind the curve.
33:52
There isn't sufficient legislation, for example, on that,
33:55
and there won't be any time soon.
33:56
But if we could address that, I think that would also help us a lot with regards to
34:00
what we can ask from tech companies, what we can expect from them,
34:03
and maybe some of the political consequences also.
34:05
>> Jon Hartley: So this is like the Coase theorem for data.
34:09
I guess the idea is that we should be compensated at some level for
34:14
the data that these companies are selling to some degree?
34:18
>> Simon Johnson: Yes, I mean, I don't think the amount of compensation, I don't
34:21
think it's worth that much, I don't think it's gonna enable you to retire.
34:24
But I do think recognizing that it's your property and you should have a say in how
34:28
it's used, and you should be able to say, no, this is my data.
34:31
And you're not allowed to use it for workplace surveillance that I don't
34:35
believe is good for me or for other people around me.
34:38
And you can only do that if those property rights are properly defined and protected.
34:45
We won't be able to protect them individually cuz we're too small relative
34:48
to big companies.
34:49
But if we could form data unions on whatever basis,
34:51
that then becomes an interesting conversation.
34:54
>> Jon Hartley: So that you think is really scoped for
34:57
like a regulatory authority.
34:59
I guess something like the FTC or a body like that in the US,
35:03
at least that could I guess, impose some sort of changes on big tech companies?
35:09
>> Simon Johnson: Well, you would certainly need legislation to assert that
35:12
there are such things as ownership over data and to put those restrictions on it.
35:16
Would it then need to run through regulatory body?
35:18
Would you need to grant that regulatory power to someone?
35:21
Yes, probably, although, I think I would prefer a body that was very focused
35:26
on protecting consumers and didn't have multiple tasks.
35:30
I think when you give these regulators, you say, worry about market structure,
35:33
worry about conduct and other things.
35:36
And worry about data protection.
35:37
That is not necessarily gonna be their top priority.
35:41
>> Jon Hartley: That's fascinating, it's so amazing, I think about just all these
35:46
huge questions around technology, political economy, growth, institutions.
35:52
This has been an amazing discussion.
35:54
Thank you so much for joining us today, Simon.
35:56
>> Simon Johnson: Thank you.
35:57
>> Jon Hartley: Today, our guest was Simon Johnson, who is the Ronald Kurtz professor
36:01
of entrepreneurship at the MIT Sloan School of Management,
36:04
was previously chief economist at the IMF.
36:06
He just co-authored a newly published book with his MIT colleague Daron Acemoglu,
36:11
Power and Progress, Our One-Thousand Year Struggle Over Technology and Prosperity.
36:16
I highly recommend you check it out at your local bookstore.
36:19
This is the Capitalism and Freedom in the 21st century podcast,
36:22
where we talk about economics, markets and public policy.
36:25
I'm Jon Hartley, your host, thanks so much for joining us.
36:27
[MUSIC]