21:05
TEDSalon Berlin 2014

Thomas Piketty: New thoughts on capital in the twenty-first century

Filmed:

French economist Thomas Piketty caused a sensation in early 2014 with his book on a simple, brutal formula explaining economic inequality: r > g (meaning that return on capital is generally higher than economic growth). Here, he talks through the massive data set that led him to conclude: Economic inequality is not new, but it is getting worse, with radical possible impacts.

- Economist
Thomas Piketty is an economist and professor at the Paris School of Economics. His 2014 book, "Capital in the Twenty-first Century," caused a sensation upon publication. Full bio

It's very nice to be here tonight.
00:12
So I've been working on the history of income
00:14
and wealth distribution for the past 15 years,
00:18
and one of the interesting lessons
00:21
coming from this historical evidence
00:25
is indeed that, in the long run,
00:27
there is a tendency for
the rate of return of capital
00:29
to exceed the economy's growth rate,
00:33
and this tends to lead to
high concentration of wealth.
00:35
Not infinite concentration of wealth,
00:38
but the higher the gap between r and g,
00:40
the higher the level of inequality of wealth
00:43
towards which society tends to converge.
00:46
So this is a key force that
I'm going to talk about today,
00:49
but let me say right away
00:52
that this is not the only important force
00:54
in the dynamics of income
and wealth distribution,
00:57
and there are many other forces that play
00:59
an important role in the long-run dynamics
01:01
of income and wealth distribution.
01:04
Also there is a lot of data
01:06
that still needs to be collected.
01:08
We know a little bit more today
01:09
than we used to know,
but we still know too little,
01:12
and certainly there are
many different processes —
01:15
economic, social, political —
01:17
that need to be studied more.
01:20
And so I'm going to focus today on this simple force,
01:21
but that doesn't mean that other important forces
01:24
do not exist.
01:26
So most of the data I'm going to present
01:28
comes from this database
01:30
that's available online:
01:32
the World Top Incomes Database.
01:33
So this is the largest existing
01:35
historical database on inequality,
01:37
and this comes from the effort
01:39
of over 30 scholars from several dozen countries.
01:41
So let me show you a couple of facts
01:45
coming from this database,
01:47
and then we'll return to r bigger than g.
01:49
So fact number one is that there has been
01:51
a big reversal in the ordering of income inequality
01:53
between the United States and Europe
01:56
over the past century.
01:58
So back in 1900, 1910, income inequality was actually
02:00
much higher in Europe than in the United States,
02:03
whereas today, it is a lot higher in the United States.
02:06
So let me be very clear:
02:09
The main explanation for this is not r bigger than g.
02:10
It has more to do with changing supply and demand
02:13
for skill, the race between education and technology,
02:17
globalization, probably more unequal access
02:20
to skills in the U.S.,
02:24
where you have very good, very top universities
02:25
but where the bottom part of the educational system
02:28
is not as good,
02:30
so very unequal access to skills,
02:31
and also an unprecedented rise
02:33
of top managerial compensation of the United States,
02:35
which is difficult to account for
just on the basis of education.
02:38
So there is more going on here,
02:41
but I'm not going to talk too much about this today,
02:43
because I want to focus on wealth inequality.
02:46
So let me just show you a very simple indicator
02:48
about the income inequality part.
02:51
So this is the share of total income
02:54
going to the top 10 percent.
02:56
So you can see that one century ago,
02:58
it was between 45 and 50 percent in Europe
03:00
and a little bit above 40 percent in the U.S.,
03:04
so there was more inequality in Europe.
03:06
Then there was a sharp decline
03:09
during the first half of the 20th century,
03:11
and in the recent decade, you can see that
03:13
the U.S. has become more unequal than Europe,
03:16
and this is the first fact I just talked about.
03:19
Now, the second fact is more about wealth inequality,
03:22
and here the central fact is that wealth inequality
03:26
is always a lot higher than income inequality,
03:29
and also that wealth inequality,
03:31
although it has also increased in recent decades,
03:34
is still less extreme today
03:36
than what it was a century ago,
03:38
although the total quantity of wealth
03:40
relative to income has now recovered
03:43
from the very large shocks
03:45
caused by World War I, the Great Depression,
03:46
World War II.
03:48
So let me show you two graphs
03:50
illustrating fact number two and fact number three.
03:52
So first, if you look at the level of wealth inequality,
03:55
this is the share of total wealth
03:59
going to the top 10 percent of wealth holders,
04:02
so you can see the same kind of reversal
04:05
between the U.S. and Europe that we had before
04:07
for income inequality.
04:10
So wealth concentration was higher
04:12
in Europe than in the U.S. a century ago,
04:15
and now it is the opposite.
04:17
But you can also show two things:
04:19
First, the general level of wealth inequality
04:21
is always higher than income inequality.
04:25
So remember, for income inequality,
04:27
the share going to the top 10 percent
04:30
was between 30 and 50 percent of total income,
04:32
whereas for wealth, the share is always
04:36
between 60 and 90 percent.
04:39
Okay, so that's fact number one,
04:41
and that's very important for what follows.
04:43
Wealth concentration is always
04:45
a lot higher than income concentration.
04:46
Fact number two is that the rise
04:48
in wealth inequality in recent decades
04:52
is still not enough to get us back to 1910.
04:55
So the big difference today,
04:59
wealth inequality is still very large,
05:01
with 60, 70 percent of total wealth for the top 10,
05:03
but the good news is that it's actually
05:06
better than one century ago,
05:08
where you had 90 percent in
Europe going to the top 10.
05:10
So today what you have
05:13
is what I call the middle 40 percent,
05:15
the people who are not in the top 10
05:17
and who are not in the bottom 50,
05:19
and what you can view as the wealth middle class
05:21
that owns 20 to 30 percent
05:23
of total wealth, national wealth,
05:26
whereas they used to be poor, a century ago,
05:28
when there was basically no wealth middle class.
05:31
So this is an important change,
05:34
and it's interesting to see that wealth inequality
05:35
has not fully recovered to pre-World War I levels,
05:40
although the total quantity of wealth has recovered.
05:43
Okay? So this is the total value
05:47
of wealth relative to income,
05:49
and you can see that in particular in Europe,
05:51
we are almost back to the pre-World War I level.
05:53
So there are really two
05:57
different parts of the story here.
05:59
One has to do with
06:01
the total quantity of wealth that we accumulate,
06:02
and there is nothing bad per se, of course,
06:05
in accumulating a lot of wealth,
06:07
and in particular if it is more diffuse
06:08
and less concentrated.
06:11
So what we really want to focus on
06:13
is the long-run evolution of wealth inequality,
06:15
and what's going to happen in the future.
06:18
How can we account for the fact that
06:20
until World War I, wealth inequality was so high
06:22
and, if anything, was rising to even higher levels,
06:26
and how can we think about the future?
06:29
So let me come to some of the explanations
06:32
and speculations about the future.
06:36
Let me first say that
06:38
probably the best model to explain
06:40
why wealth is so much
06:42
more concentrated than income
06:44
is a dynamic, dynastic model
06:46
where individuals have a long horizon
06:49
and accumulate wealth for all sorts of reasons.
06:52
If people were accumulating wealth
06:54
only for life cycle reasons,
06:57
you know, to be able to consume
06:59
when they are old,
07:01
then the level of wealth inequality
07:03
should be more or less in line
07:05
with the level of income inequality.
07:07
But it will be very difficult to explain
07:09
why you have so much more wealth inequality
07:11
than income inequality
07:13
with a pure life cycle model,
07:15
so you need a story
07:16
where people also care
07:18
about wealth accumulation for other reasons.
07:20
So typically, they want to transmit
07:22
wealth to the next generation, to their children,
07:24
or sometimes they want to accumulate wealth
07:28
because of the prestige, the
power that goes with wealth.
07:29
So there must be other reasons
07:32
for accumulating wealth than just life cycle
07:34
to explain what we see in the data.
07:36
Now, in a large class of dynamic models
07:38
of wealth accumulation
07:42
with such dynastic motive for accumulating wealth,
07:44
you will have all sorts of random,
07:47
multiplicative shocks.
07:50
So for instance, some families
07:51
have a very large number of children,
07:53
so the wealth will be divided.
07:55
Some families have fewer children.
07:57
You also have shocks to rates of return.
07:59
Some families make huge capital gains.
08:01
Some made bad investments.
08:03
So you will always have some mobility
08:05
in the wealth process.
08:07
Some people will move up,
some people will move down.
08:08
The important point is that,
08:11
in any such model,
08:12
for a given variance of such shocks,
08:13
the equilibrium level of wealth inequality
08:16
will be a steeply rising function of r minus g.
08:18
And intuitively, the reason why the difference
08:23
between the rate of return to wealth
08:26
and the growth rate is important
08:28
is that initial wealth inequalities
08:29
will be amplified at a faster pace
08:32
with a bigger r minus g.
08:34
So take a simple example,
08:36
with r equals five percent and g equals one percent,
08:38
wealth holders only need to reinvest
08:41
one fifth of their capital income to ensure
08:43
that their wealth rises as fast
08:46
as the size of the economy.
08:49
So this makes it easier
08:51
to build and perpetuate large fortunes
08:52
because you can consume four fifths,
08:54
assuming zero tax,
08:56
and you can just reinvest one fifth.
08:58
So of course some families
will consume more than that,
08:59
some will consume less, so there will be
09:02
some mobility in the distribution,
09:04
but on average, they only need to reinvest one fifth,
09:05
so this allows high wealth inequalities to be sustained.
09:08
Now, you should not be surprised
09:12
by the statement that r can be bigger than g forever,
09:14
because, in fact, this is what happened
09:18
during most of the history of mankind.
09:20
And this was in a way very obvious to everybody
09:22
for a simple reason, which is that growth
09:25
was close to zero percent
09:27
during most of the history of mankind.
09:29
Growth was maybe 0.1, 0.2, 0.3 percent,
09:31
but very slow growth of population
09:34
and output per capita,
09:36
whereas the rate of return on capital
09:38
of course was not zero percent.
09:40
It was, for land assets, which was
09:42
the traditional form
09:44
of assets in preindustrial societies,
09:46
it was typically five percent.
09:48
Any reader of Jane Austen would know that.
09:50
If you want an annual income of 1,000 pounds,
09:54
you should have a capital value
09:57
of 20,000 pounds so that
09:58
five percent of 20,000 is 1,000.
10:00
And in a way, this was
10:03
the very foundation of society,
10:05
because r bigger than g
10:06
was what allowed holders of wealth and assets
10:09
to live off their capital income
10:14
and to do something else in life
10:16
than just to care about their own survival.
10:19
Now, one important conclusion
10:22
of my historical research is that
10:24
modern industrial growth did not change
10:26
this basic fact as much as one might have expected.
10:29
Of course, the growth rate
10:32
following the Industrial Revolution
10:34
rose, typically from zero to one to two percent,
10:35
but at the same time, the rate of return
10:40
to capital also rose
10:42
so that the gap between the two
10:43
did not really change.
10:46
So during the 20th century,
10:47
you had a very unique combination of events.
10:49
First, a very low rate of return
10:52
due to the 1914 and 1945 war shocks,
10:54
destruction of wealth, inflation,
10:57
bankruptcy during the Great Depression,
10:59
and all of this reduced
11:01
the private rate of return to wealth
11:03
to unusually low levels
11:05
between 1914 and 1945.
11:07
And then, in the postwar period,
11:09
you had unusually high growth rate,
11:11
partly due to the reconstruction.
11:14
You know, in Germany, in France, in Japan,
11:16
you had five percent growth rate
11:18
between 1950 and 1980
11:20
largely due to reconstruction,
11:23
and also due to very large demographic growth,
11:25
the Baby Boom Cohort effect.
11:27
Now, apparently that's not going to last for very long,
11:29
or at least the population growth
11:32
is supposed to decline in the future,
11:33
and the best projections we have is that
11:36
the long-run growth is going to be closer
11:40
to one to two percent
11:42
rather than four to five percent.
11:43
So if you look at this,
11:45
these are the best estimates we have
11:48
of world GDP growth
11:50
and rate of return on capital,
11:51
average rates of return on capital,
11:54
so you can see that during most
11:56
of the history of mankind,
11:57
the growth rate was very small,
11:58
much lower than the rate of return,
12:00
and then during the 20th century,
12:02
it is really the population growth,
12:04
very high in the postwar period,
12:06
and the reconstruction process
12:09
that brought growth
12:10
to a smaller gap with the rate of return.
12:12
Here I use the United Nations population projections,
12:15
so of course they are uncertain.
12:18
It could be that we all start
12:21
having a lot of children in the future,
12:22
and the growth rates are going to be higher,
12:24
but from now on,
12:27
these are the best projections we have,
12:28
and this will make global growth
12:31
decline and the gap between
12:33
the rate of return go up.
12:36
Now, the other unusual event
12:38
during the 20th century
12:41
was, as I said,
12:42
destruction, taxation of capital,
12:44
so this is the pre-tax rate of return.
12:46
This is the after-tax rate of return,
12:49
and after destruction,
12:52
and this is what brought
12:53
the average rate of return
12:55
after tax, after destruction,
12:57
below the growth rate during a long time period.
12:59
But without the destruction,
13:01
without the taxation, this
would not have happened.
13:03
So let me say that the balance between
13:05
returns on capital and growth
13:08
depends on many different factors
13:11
that are very difficult to predict:
13:13
technology and the development
13:15
of capital-intensive techniques.
13:17
So right now, the most capital-intensive sectors
13:19
in the economy are the real estate sector, housing,
13:22
the energy sector, but it could be in the future
13:26
that we have a lot more robots in a number of sectors
13:29
and that this would be a bigger share
13:32
of the total capital stock that it is today.
13:34
Well, we are very far from this,
13:36
and from now, what's going on
13:38
in the real estate sector, the energy sector,
13:40
is much more important for the total capital stock
13:42
and capital share.
13:44
The other important issue
13:45
is that there are scale effects
in portfolio management,
13:47
together with financial complexity,
13:49
financial deregulation,
13:52
that make it easier to get higher rates of return
13:53
for a large portfolio,
13:56
and this seems to be particularly strong
13:57
for billionaires, large capital endowments.
14:00
Just to give you one example,
14:02
this comes from the Forbes billionaire rankings
14:04
over the 1987-2013 period,
14:08
and you can see the very top wealth holders
14:11
have been going up at six, seven percent per year
14:14
in real terms above inflation,
14:17
whereas average income in the world,
14:19
average wealth in the world,
14:22
have increased at only two percent per year.
14:23
And you find the same
14:26
for large university endowments —
14:28
the bigger the initial endowments,
14:29
the bigger the rate of return.
14:32
Now, what could be done?
14:34
The first thing is that I think we need
14:35
more financial transparency.
14:38
We know too little about global wealth dynamics,
14:40
so we need international transmission
14:44
of bank information.
14:46
We need a global registry of financial assets,
14:47
more coordination on wealth taxation,
14:50
and even wealth tax with a small tax rate
14:52
will be a way to produce information
14:55
so that then we can adapt our policies
14:57
to whatever we observe.
15:00
And to some extent, the fight
15:02
against tax havens
15:04
and automatic transmission of information
15:05
is pushing us in this direction.
15:07
Now, there are other ways to redistribute wealth,
15:09
which it can be tempting to use.
15:11
Inflation:
15:14
it's much easier to print money
15:16
than to write a tax code, so that's very tempting,
15:17
but sometimes you don't know
what you do with the money.
15:19
This is a problem.
15:22
Expropriation is very tempting.
15:23
Just when you feel some people get too wealthy,
15:25
you just expropriate them.
15:27
But this is not a very efficient way
15:29
to organize a regulation of wealth dynamics.
15:30
So war is an even less efficient way,
15:33
so I tend to prefer progressive taxation,
15:36
but of course, history — (Laughter) —
15:38
history will invent its own best ways,
15:41
and it will probably involve
15:42
a combination of all of these.
15:44
Thank you.
15:46
(Applause)
15:48
Bruno Giussani: Thomas Piketty. Thank you.
15:50
Thomas, I want to ask you two or three questions,
15:55
because it's impressive how you're
in command of your data, of course,
15:57
but basically what you suggest is
16:01
growing wealth concentration is kind of
16:05
a natural tendency of capitalism,
16:06
and if we leave it to its own devices,
16:08
it may threaten the system itself,
16:12
so you're suggesting that we need to act
16:14
to implement policies that redistribute wealth,
16:16
including the ones we just saw:
16:19
progressive taxation, etc.
16:21
In the current political context,
16:22
how realistic are those?
16:24
How likely do you think that it is
16:26
that they will be implemented?
16:28
Thomas Piketty: Well, you know, I think
16:30
if you look back through time,
16:31
the history of income, wealth and taxation
16:33
is full of surprise.
16:35
So I am not terribly impressed
16:37
by those who know in advance
16:40
what will or will not happen.
16:41
I think one century ago,
16:43
many people would have said
16:45
that progressive income taxation would never happen
16:46
and then it happened.
16:48
And even five years ago,
16:50
many people would have said that bank secrecy
16:52
will be with us forever in Switzerland,
16:54
that Switzerland was too powerful
16:56
for the rest of the world,
16:58
and then suddenly it took a few U.S. sanctions
16:59
against Swiss banks for a big change to happen,
17:02
and now we are moving toward
17:05
more financial transparency.
17:07
So I think it's not that difficult
17:08
to better coordinate politically.
17:13
We are going to have a treaty
17:15
with half of the world GDP around the table
17:17
with the U.S. and the European Union,
17:20
so if half of the world GDP is not enough
17:22
to make progress on financial transparency
17:24
and minimal tax for multinational corporate profits,
17:27
what does it take?
17:31
So I think these are not technical difficulties.
17:33
I think we can make progress
17:36
if we have a more pragmatic
approach to these questions
17:38
and we have the proper sanctions
17:41
on those who benefit from financial opacity.
17:43
BG: One of the arguments
17:46
against your point of view
17:47
is that economic inequality
17:49
is not only a feature of capitalism
but is actually one of its engines.
17:50
So we take measures to lower inequality,
17:54
and at the same time we lower growth, potentially.
17:57
What do you answer to that?
17:59
TP: Yeah, I think inequality
18:01
is not a problem per se.
18:02
I think inequality up to a point
18:04
can actually be useful for innovation and growth.
18:06
The problem is, it's a question of degree.
18:09
When inequality gets too extreme,
18:11
then it becomes useless for growth
18:14
and it can even become bad
18:17
because it tends to lead to high perpetuation
18:18
of inequality over time
18:21
and low mobility.
18:23
And for instance, the kind of wealth concentrations
18:25
that we had in the 19th century
18:28
and pretty much until World War I
18:31
in every European country
18:33
was, I think, not useful for growth.
18:35
This was destroyed by a combination
18:37
of tragic events and policy changes,
18:39
and this did not prevent growth from happening.
18:41
And also, extreme inequality can be bad
18:43
for our democratic institutions
18:47
if it creates very unequal access to political voice,
18:49
and the influence of private money
18:51
in U.S. politics, I think,
18:53
is a matter of concern right now.
18:55
So we don't want to return to that kind of extreme,
18:58
pre-World War I inequality.
19:01
Having a decent share of the national wealth
19:03
for the middle class is not bad for growth.
19:07
It is actually useful
19:10
both for equity and efficiency reasons.
19:11
BG: I said at the beginning
19:14
that your book has been criticized.
19:16
Some of your data has been criticized.
19:18
Some of your choice of data sets has been criticized.
19:19
You have been accused of cherry-picking data
19:22
to make your case. What do you answer to that?
19:24
TP: Well, I answer that I am very happy
19:26
that this book is stimulating debate.
19:28
This is part of what it is intended for.
19:31
Look, the reason why I put all the data online
19:33
with all of the detailed computation
19:37
is so that we can have
an open and transparent
19:38
debate about this.
19:41
So I have responded point by point
19:42
to every concern.
19:44
Let me say that if I was to rewrite the book today,
19:46
I would actually conclude
19:49
that the rise in wealth inequality,
19:51
particularly in the United States,
19:53
has been actually higher
than what I report in my book.
19:55
There is a recent study by Saez and Zucman
19:57
showing, with new data
20:00
which I didn't have at the time of the book,
20:02
that wealth concentration in the U.S. has risen
20:04
even more than what I report.
20:06
And there will be other data in the future.
20:08
Some of it will go in different directions.
20:10
Look, we put online almost every week
20:12
new, updated series on the
World Top Income Database
20:17
and we will keep doing so in the future,
20:19
in particular in emerging countries,
20:21
and I welcome all of those who want to contribute
20:24
to this data collection process.
20:27
In fact, I certainly agree
20:29
that there is not enough
20:32
transparency about wealth dynamics,
20:33
and a good way to have better data
20:35
would be to have a wealth tax
20:37
with a small tax rate to begin with
20:39
so that we can all agree
20:41
about this important evolution
20:43
and adapt our policies to whatever we observe.
20:45
So taxation is a source of knowledge,
20:48
and that's what we need the most right now.
20:50
BG: Thomas Piketty, merci beaucoup.
20:53
Thank you.
TP: Thank you. (Applause)
20:55

▲Back to top

About the Speaker:

Thomas Piketty - Economist
Thomas Piketty is an economist and professor at the Paris School of Economics. His 2014 book, "Capital in the Twenty-first Century," caused a sensation upon publication.

Why you should listen
Is the global economy accelerating toward a future that’s incompatible with democracy? In this provocative talk about inequality and wealth, economist Thomas Piketty provides new context for his groundbreaking book, Capital in the Twenty-First Century.
More profile about the speaker
Thomas Piketty | Speaker | TED.com