Long-run Trends in Wealth Inequality
and the Role of Inheritance
Daniel Waldenström
Uppsala University
Winter School in Inequality and Social Welfare, Canazei, 2012
1
Outline of talk
1. Introduction: Role of wealth in inequality trends
2. Wealth concentration over the path of development
–
Cross-country evidence, 1774-2006
–
The case of Sweden, 1873-2006
3. Inheritance − the role of "old wealth"
4. Concluding remarks
2
(Part of the) Income/Wealth relationship
Savings
(sY)
Wealth
(Level,
distribution)
Inheritance
Income
(Level,
distribution)
Capital income
(rW)
Other processes
(opportunity, insurance etc)
3
Capital income share in total income,
Sweden, top decile, 1912-2004
Capital share of total income (excl. capital gains)
70%
60%
50%
40%
30%
20%
10%
0%
1912
1922
1932
P90-99
1942
1952
P99-99.9
1962
1972
P99.9-99.99
1982
1992
2002
P99.99-100
4
Wealth matters to income mobility:
Intergenerational transmission in Sweden
Björklund, Roine and Waldenström (2011), “Intergenerational Top Income Mobility in
Sweden – Capitalist Dynasties in the Land of Equal Opportunity?”,
5
Wealth matters to income mobility:
Intergenerational transmission in Sweden
1.0
Income
0.8
FinAss
3.5
3.0
2.5
0.6
2.0
0.4
1.5
1.0
0.2
Educ
0.0
Non-cog
0.5
0.0
IQ
-0.2
P90-95
P95-99
Björklund, Roine and Waldenström (2011)
P99-99.9
P99.9-100
-0.5
6
Link between inequality and development
• Large literature − no consensus
– Equalization view (Stiglitz 1969; Becker & Tomes; Loury 1981)
– Disequalization view (Ray; Mookherjee & Ray 2003)
– History matters (Banerjee & Newman 1993; Galor and Zeira 1993)
• Kuznets (1953, 1955): Structural change
– Shift from agericultural (trad.) to industrial (modern) sector
generates inverted U-like relationship
– Also: concentration of capital and savings boost inequality
Inequality
Time
7
Is there a Kuznets Curve in income inequality?
30
Kuznets
inverted-U?
Share of total income (%)
25
20
USA
UK
Canada
France
Australia
Netherlands
Sweden (incl. CG)
Sweden (excl. CG)
15
10
5
0
1903
1913
1923
1933
1943
1953
1963
1973
1983
1993
2003
What about wealth inequality?
8
Long-run trends in
wealth concentration:
Cross-country evidence
Ohlsson, H, J. Roine and D. Waldenström (2008), “Long-Run Changes in the
Concentration of Wealth: An Overview of Recent Findings”, in Davies, J.B. (ed.),
Personal Wealth from a Global Perspective, Oxford, Oxford University Press. Link
9
Starting point
• Main question: Is there a long-term linkage between
development and inequality? (cf. Kuznets)
• Further questions:
– Common vs. specific trends across countries
– Potential heterogeneity within the top wealth decile
• This study:
– Reviews recent empirical findings of others
– Presents new evidence on Nordic countries
10
Measurement and data
• Estate data
– Incentives to minimze tax - but also to divide fairly
– Homogenous source over time
– Problems: Sample size, W of deceased (mortality multipl.)
• Wealth tax data
– Large samples; Relatively homogenous over time
– Problems: tax incentives, asset valuation, excludes items
• Survey data
– Covers most asset items; Scattered points; Short history
– Problems: sample size, response rates, top-coding
11
Measurement and data
• Wealth concept: Net worth
– Real + financial assets less debts (excl. pensions, human
cap.)
• Wealth owners:
– Wealth tax data: Households (tax units; >18yrs)
– Estate data: Adults (18 yrs + ); deceased
– Survey data: Households (everyone in dwelling)
– Historically, households difficult to define
12
Measurement and data
Main methodology:
• Compute shares of total household wealth (using
Pareto interpolation) that goes to the top 1%,
top10% etc of all potential wealth holders
• The shares of both wealth and population are
defined in relation to reference totals
– Reference total for wealth: All personal wealth in the
economy (not only taxed wealth)
– Reference total for the population: All potential tax units
(not just those who file tax returns)
13
French wealth concentration, 1807-1994
60
50
Share of total wealth (%)
P99-100
40
30
20
P95-99
P99.9-100
10
P90-95
0
1800
1840
1880
1920
1960
2000
14
U.K. wealth concentration, 1774-2001
Lindert (2000)
100
90
Atkinson et al
IRS (2006)
P95-100
80
Share of total wealth (%)
70
P99-100
60
50
40
30
P95-99
20
10
England and Wales (1740-1937)
U.K. (1938-)
0
1740
1760
1780
1800
1820
1840
1860
1880
1900
1920
1940
1960
1980
2000
15
U.S. wealth concentration, 1774-2001
Lindert
(2000)
Shammas (1993)
45
Kopzcuk & Saez (2004)
Wolff (1987, etc)
40
P99-100 (adults)
Share of total wealth (%)
35
30
P95-99 (households)
25
20
P99-100 (households)
15
P90-95 (households)
10
1770
1790
1810
1830
1850
1870
1890
1910
1930
1950
1970
1990
16
Swiss wealth concentration, 1913-1997
50
P99-100
45
Share of total wealth (%)
40
35
30
P95-99
25
20
P99.9-100
15
10
P90-95
5
1910
1920
1930
1940
1950
1960
1970
1980
1990
2000
17
Danish wealth concentration, 1789-1996
60
Share of total wealth (%)
50
P99-100
40
30
P95-99
20
10
0
1780
P99.9-100
P90-95
1800
1820
1840
1860
1880
1900
1920
1940
1960
1980
2000
18
Norwegian wealth concentration, 1789-2002
50
45
P99-100
Share of total wealth (%)
40
35
30
25
20
P95-99
P99.9-100
15
10
P90-95
5
0
1780
1800
1820
1840
1860
1880
1900
1920
1940
1960
1980
2000
19
Swedish wealth concentration, 1800-2002
70
Share of total wealth (%)
60
50
40
30
20
10
0
1800
1840
P99-100 (wealth tax based)
P90-99 (SCB, 1975-2000)
P99-100 (Spånt, 1920-75)
1880
1920
P90-99 (wealth tax based)
P99-100 (estate based)
P90-99 (Spånt, 1920-1975)
1960
2000
P99-100 (SCB, 1975-2000)
P90-99 (estate based)
20
Cross country top percentile (P99-100)
70
60
Share of total wealth (%)
Denmark
50
Sweden
UK*
Norway
France
Switzerland
40
30
USA
20
10
1740
1760
1780
1800
1820
1840
1860
1880
1900
1920
1940
1960
1980
2000
21
Cross-country "next four" (P95-99)
45
UK*
Share of total wealth (%)
40
35
Switzerland
30
France
25
Denmark
USA
Norway
20
Sweden
15
1740
1760
1780
1800
1820
1840
1860
1880
1900
1920
1940
1960
1980
2000
22
Heterogeneity between countries and within the top
Period:
Fractile:
≈1780-1914
1914-2000
Top1%
Next 4%
Top1%
Next 4%
Increase
Flat
Decrease
Flat
-
-
Flat
Flat
UK
Increase
Decrease
Decrease
Flat
US
Increase
Flat?
Decrease
Flat?
Denmark
Decrease
Flat
Decrease
Flat
Norway
Decrease
Increase
Decrease
Decrease
Sweden
Flat
Flat
Decrease
Decrease*
France
Switzerland
23
Summing up cross-country evidence
1. Industrialization: no uniform impact on wealth
inequality
– Inequality increased in the US, UK (top1%), France
– Flat (or decreasing) inequality in Nordic countries
– Role of country size? Timing of industrialization?
2. Other factors matter (20th experience)
–
Geopolitical shocks, Crises, Redistribtion
3. Little support for Kuznets-type "inverse-U"
– Trends rather look like an inverse-J (or inverse-S)
24
Wealth concentration in
Sweden, 1870-2006
Roine, J. and D. Waldenström (2009), “Wealth Concentration over the Path of
Development: Sweden, 1873–2006”, Scandinavian Journal of Economics, 111(1),
151–187. Link
25
Questions
• Closer at changes in wealth concentration
• Evolution during episodes:
– Industrial take-off (1870-1910)
– Rise of the welfare state (educational reforms, home
ownership, redistribution) (1930-1970)
– Globalization and deregulation (1980- )
• Role of offshore wealth
26
Top wealth decile in Sweden
100
Top 10%
Share of total household wealth (%)
90
80
70
60
50
40
1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000
27
Wealth distribution in Sweden, 1870-2005
Industrial take-off
Education, welfare state,
home ownership
60
Share of total household wealth (%)
50
Top 1%
40
30
Top 10-1%
20
10
Bottom 90%
0
1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000
28
Interwar era: High-wage earners
accumulate wealth
40
35
Wealth shares (%)
30
25
20
15
10
5
0
1911
1916
1921
1926
P99.9-100
1931
P90-99
1936
1941
1946
1951
P99-99.9
29
Wealth distribution in Sweden, 1870-2005
Industrial take-off
Education, welfare state,
home ownership
Globalization,
deregulation...
60
Share of total household wealth (%)
50
Top 1%
40
30
Top 10-1%
20
10
?
Bottom 90%
0
1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000
30
What has happened since 1980?
• Sweden in the 1980s and 1990s:
– High taxes on wealth, inheritance, property
– Booming financial markets
– Liberalized capital account (since 1989)
• Potential impact: Capital flight
– Private wealth (and wealth owners) leaves Sweden
– Wealth is moved to closely held companies
• If so, what is the effect on the wealth distribution?
31
Wealth concentration in Sweden since 1950
(official series)
50
45
Wealth sahre (%)
40
Bottom 90%
35
30
25
Top 1%
20
15
1950
1960
1970
1980
1990
2000
32
Wealth concentration in Sweden since 1950
(adjusted for foreign wealth of Swedish hh's)
50
45
Wealth sahre (%)
40
Bottom 90%
35
30
incl. foreign wealth
25
Top 1%
20
15
1950
1960
1970
1980
1990
2000
33
How estimate the wealth of the rich?
• We add wealth to the top percentile (≈50,000 hh's)
• Three additions:
1. Household wealth abroad
•
•
Accumulated mismatches in Balance of Payment
Unaccounted savings in the Financial Accounts
2. Wealth of super-rich Swedes in Sweden
•
Unlisted wealth; Source: Rich Lists in magazines 1983-
3. Wealth of super-rich Swedes abroad
•
Unlisted wealth; Source: Rich Lists in magazines 1983-
34
Accumulated BoP mismatch as share of GDP
Balance of Payments mismatch, six countries
70%
60%
50%
Switzerland
40%
30%
20%
10%
USA
0%
Denmark
Finland
-10%
-20%
Sweden
Norway
-30%
1975
1980
1985
1990
1995
2000
2005
35
Comparing top wealth percentile: Sweden vs USA
45
40
Share of total wealth (%)
+ Rich Swedes abroad
+ BoP error, dom.
and for. super-rich
35
USA
+ Rich Swedes at home
30
+ BoP mismatch
25
20
15
1945
Official series
1955
1965
1975
1985
1995
2005
36
Summing up: Swedish wealth inequality trends
1. Industrial take-off had a small effect on wealth
concentration
– But possibly larger on the composition of the wealthy ...
2. 20th century: Equalization
– Before 1950 the top 1% lost to the medium-rich
– 1950-1980 rise of "popular wealth"
– After 1980, wealth compression halts
3. Adding estimates of foreign wealth and large
domestic unlisted fortunes reverses trend 198037
The Importance of Old Wealth:
Estimating Long-Run Flows of
Inheritance
Ohlsson, H., J. Roine and D. Waldenström (2012), “The Role of Inheritance in
Sweden, 1860–2010”, mimeo.
Waldenström (2012), “Household Wealth in Sweden, 1810–2010”, mimeo.
38
How important is inheritance?
• Next question: How important are inheritances
relative to other sources of income?
• Households get rich from two sources:
– they save out of their income (new wealth)
– they receive transfers from others (old wealth)
• Conventional story: Old wealth mattered in the past,
today new wealth (growth-based) is all that matters.
• But is this really true...?
39
Annual inheritance flow as share of
national income: France, 1820-2008
Source: Piketty, Thomas (2011), “On the Long-Run Evolution of Inheritance:
France 1850–2050”, Quarterly Journal of Economics 126(3), 1071–1131.
40
Annual inheritance flow as share of
national income: Sweden, 1860-2010
0.16
0.14
0.12
0.1
0.08
0.06
0.04
0.02
0
1860
1880
1900
1920
Economic inheritance flow
1940
1960
1980
2000
Fiscal inheritance flow
41
Basic approach
• Piketty (2011): the importance of inheritance
depends on ratio between economic growth rate g
and returns to capital r.
– With r > g, old wealth accumulates faster than new wealth
is created
• Let
–
–
–
–
–
B = Aggregate inheritance flow
Y = national income,
W = aggergate private wealth
m = mortality rate
μ = ratio aveW of deceased/aveW of the living.
42
Basic approach
• Then, compute the inheritance flow over income as:
B/Y = μ ∙ m ∙ W/Y
• In a dynastic model, heirs save a fraction g/r of the
return on their inerited wealth, making the wealthincome ratio W/Y stationary.
• Steady-state bequest flow B/Y = (W/Y)/H
– H = generation length
• If W/Y = 400% and H = 30, then B/Y = 13%.
43
B/Y = μ ∙ m ∙ W/Y (France)
AveW of decendents/living (μ): France
Mortality rate (m) in France
Wealth-Income ratio (W/Y): France
Inheritance flow/Income (B/Y): France
Source: Piketty (2011)
44
Constructing W, Y, m, μ for Sweden
• Y: National income data exist (disposable less so)
• m: Decendents over population. Data exist.
• μ: AveW of decendents/living. Difficult. Estimated
from estate and wealth tax sources.
– Gift-correction: +15%
• W: Real and financial assets less debts. No solid
annual data prior to 1970 (1950). Construct new
series back to early 19th century...
45
Constructing W: Real assets
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
1810
1850
1890
Agricultural net assets (farmland, forestry, livestock)
1930
1970
Real estate (buildings, housing)
2010
Consumer durables
46
Constructing W: Financial assets
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
1810
1850
1890
1970
1930
Bank deposits and currency
Shares and mutual funds
Other claims
Private pension savings
2010
Bonds
47
Constructing W: Debt
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
1810
1850
1890
Informal liabilities (Other debt)
1930
State
1970
2010
Financial sector
48
Debt components up to 1970.
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
1810
1830
1850
1870
1890
1910
1930
Commercial banks
Savings banks
Mortgage associations, Credit companies
Insurance companies
State
Informal liabilities (Other debt)
1950
49
Asset composition
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
1810
1850
1890
Real assets
1930
1970
2010
Financial assets
50
Swedish household wealth
over GDP, 1810-2010
3
2.5
2
1.5
1
0.5
0
1810
1850
Real assets / GDP
1890
1930
Financial assets / GDP
1970
2010
Debts / GDP
51
Wealth-Income ratio (W/Y): Sweden
700%
600%
500%
400%
300%
200%
1810
1830
1850
1870
1890
1910
1930
1950
1970
1990
2010
52
B/Y = μ ∙ m ∙ W/Y (Sweden)
AveW of decendents/living (μ): Sweden
300%
Mortality rate (m) in Sweden
2.5%
250%
2.0%
200%
1.5%
150%
1.0%
100%
0.5%
50%
0%
1860
1880
1900
1920
1940
1960
1980
0.0%
1860
2000
Wealth-Income ratio (W/Y): Sweden
1880
1900
1920
1940
1960
1980
2000
Inheritance flow/Income (B/Y): Sweden
0.16
700%
0.14
0.12
600%
0.1
500%
0.08
0.06
400%
0.04
300%
0.02
0
200%
1810
1830
1850
1870
1890
1910
1930
1950
1970
1990
2010
1860
1880
1900
1920
Economic inheritance flow
1940
1960
1980
Fiscal inheritance flow
2000
53
So, does inheritance have a role?
1. Evidence from France and Sweden suggests that:
– Inheritance was important up to WWI
– After 1950, inheritance did not matter (small W/Y)
– Today, inheritance flows are again become more significant
(W/Y is growing)
2. Piketty's main lesson:
– "with r > g (say, r=4-5% vs g=1-2%), then wealth coming
from the past is being capitalized faster than growth, and
inherited wealth dominates self-made wealth"
54
4. Concluding remarks
1. Before industrialization, wealth was highly unequal
2. Wealth equalized during 20th century
–
But not due to Kuznets-curve
–
Instead: Political shifts (education, redistribution) and
Exogenous shocks (wars, crises)
3. Current globalization and modest growth rates in
Western world suggest increased future role of
wealth and inheritance
55