Download - presentation-kumhof-stockholm-nov-2-20111
Inequality, Leverage and Crises
Michael Kumhof, International Monetary Fund
Romain Ranciere, International Monetary Fund and Paris School of Economics
The views expressed herein are those of the authors and should not be attributed
to the IMF, its Executive Board, or its management.
1 Introduction
Empirical Motivation: Similarities of the decades preceding the 1929 and 2007
crises
• Sharply increasing income inequality.
• Sharply increasing debt leverage among lower and middle classes.
• Perception of unsustainably high leverage was a key factor causing a large
financial and real crash.
2 Literature on Alternative Causes of the 2007 Crisis• Most recent literature focuses on the final years preceding the crisis:
— Excessive financial liberalization.
— Easy monetary policy.
— Global current account imbalances.
• Rajan (2010), our work: Much of this was simply a manifestation of an
underlying and longer-term dynamics driven by income inequality
— Rajan: Growing inequality created political pressure for easy credit. This
stresses the demand for credit.
— Our work: Growing income inequality simultaneously created
1. Additional demand for credit to sustain living standards of the lower
and middle class.
2. But also additional supply of credit due to the extra income of the
top income group looking for a place to go.
Companion Literature on Causes of Changes in the Income Distribution
• Hacker and Pierson (2011): Government intervention in support of the rich.
• Card, Lemieux and Riddell (2004): Changes in unionization.
• Borjas and Ramey (1995): Role of foreign competition.
• Roberts (2010): Role of jobs offshoring.
• Lemieux, MacLeod and Parent (2009): Increased use of performance pay.
• Lemieux (2006): Increase in the return to post-secondary education.
Source: Statistical Abstract of the United States, U.S. Department of Commerce.
1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 193125
30
35
40
45
50
55
60
23
25
27
29
31
33
35
Private Non Corporate+Trade Debt to GNP
Share of Top 5% in Income Distribution
1920-1931
Per
cen
t
Per
cen
t
Sources: Income shares from Piketty and Saez (2003, updated). Income excludes capital gains. Debt-to-income ratios from Flows of Funds database, Federal Reserve Board. Income excludes capital gains.
1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 200870
80
90
100
110
120
130
140
150
20
22
24
26
28
30
32
34
36
Household Debt to GDP
Share of Top 5% in Income Distribution
1983-2008
Per
cen
t
Per
cen
t
Income Inequality and Household Leverage:(i) Moved up together pre-crisis.(ii) Both pre-1929 and pre-2007.2
Source: Heathcote, Perri and Violante (2010), based on micro-level data from the U.S. Consumer Population Survey. Male annual earnings includes labor income plus two-thirds of self-employment income. Male hourly wages are computed as male annual earnings divided by annual hours. The price deflator used is the Bureau of Labor Statistics CPI-U series, all items.
1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004-40
-30
-20
-10
0
10
20
30
40
50
-40
-30
-20
-10
0
10
20
30
40
50Top Decile of Earnings Distribution
Median Decile of Earnings Distribution
Bottom Decile of Earnings Distribution
Cu
mu
lati
ve
Per
cen
t C
han
ge
Cu
mu
lati
ve
Per
cen
t C
han
ge
Male Annual Earnings by Income Decile:(i) Over 40% cumulative increase for the rich.(ii) Over 30% cumulative decrease for the poor.(ii) 5%-10% cumulative decrease for the median.
Source: Survey of Consumer Finance (triennal), 1983-2007. Debt corresponds to the stock of all outstanding household debt liabilities. Income corresponds to annual income before taxes, including capital gains and transfers, in the year preceding the survey.
1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 200630
50
70
90
110
130
150
30
50
70
90
110
130
150Bottom 95% of the Income Distribution
Top 5% of the Income Distribution
Aggregate Economy
Per
centa
ge
Poin
ts
Per
centa
ge
Poin
ts
Debt to Income Ratios:(i) Lower or flat for the rich.(ii) Sharply higher for the remainder.
Sources: Private Credit to GDP from World Bank Financial Structure Database (real private credit by deposit banksand other financial institutions, relative to GDP). Value Added GDP Share of Financial Sector from Philippon (2008).
1985 1990 1995 2000 200580
100
120
140
160
180
200
220
240
4.5
5.0
5.5
6.0
6.5
7.0
7.5
8.0
8.5
Private Credit to GDP
Value Added GDP Share of Financial Sector
Per
cent
Per
cent
Size of the U.S. Financial Sector:(i) Private Credit to GDP more than doubled.(ii) Banks’ share in GDP more than doubled.
4 A Theoretical Model to Explain The Data• Economy consists of two separate household groups, the top income group(“investors”) and the lower and middle class (“workers”).
• Economy experiences a highly persistent decrease in the income bargainingpowers of the lower and middle class.
• Response of the top income group (top 5% of incomes):1. Higher consumption.2. Higher physical (equity) investment.3. Much higher financial investment = recycling gains back to lower andmiddle class as loans.
• Response of the lower and middle class (bottom 95% of incomes):1. Lower consumption, but consumption drops by less than income.2. Much higher borrowing from the top income group = higher leverageover decades.
• Result: Higher financial fragility⇒ risk of financial crisis⇒ eventual crash.
Baseline Scenario• Highly persistent decrease in workers’ bargaining power.
• Financial and real crisis in year 30.
0 10 20 30 40 50−8
−6
−4
−2
0
Bargaining Power
% d
evi
atio
n
0 10 20 30 40 50
−6
−4
−2
0
2Real Wage
% d
evi
atio
n
0 10 20 30 40 50−1
0
1
2
Return to Capital
pp
de
via
tion
0 10 20 30 40 50
0
10
20
Investors‘ Consumption
% d
evi
atio
n
0 10 20 30 40 500
5
10
15
Investors‘ Physical Investment
% d
evi
atio
n
0 10 20 30 40 500
20
40
60
80
Investors‘ Loans
% d
evi
atio
n
0 10 20 30 40 50
−6
−4
−2
0
Workers‘ Consumption
% d
evi
atio
n
0 10 20 30 40 5060
80
100
120
140Workers‘ Debt−to−Income Ratio
leve
l in
%
0 10 20 30 40 500
1
2
3Crisis Probability
leve
l in
%
Baseline Scenario• Highly persistent decrease in workers’ bargaining power.
• Financial and real crisis in year 30.
0 10 20 30 40 50−8
−6
−4
−2
0
Bargaining Power
% d
evi
atio
n
0 10 20 30 40 50
−6
−4
−2
0
2Real Wage
% d
evi
atio
n
0 10 20 30 40 50−1
0
1
2
Return to Capital
pp
de
via
tion
0 10 20 30 40 50
0
10
20
Investors‘ Consumption
% d
evi
atio
n
0 10 20 30 40 500
5
10
15
Investors‘ Physical Investment
% d
evi
atio
n
0 10 20 30 40 500
20
40
60
80
Investors‘ Loans
% d
evi
atio
n
0 10 20 30 40 50
−6
−4
−2
0
Workers‘ Consumption
% d
evi
atio
n
0 10 20 30 40 5060
80
100
120
140Workers‘ Debt−to−Income Ratio
leve
l in
%
0 10 20 30 40 500
1
2
3Crisis Probability
leve
l in
%
Baseline Scenario• Highly persistent decrease in workers’ bargaining power.
• Financial and real crisis in year 30.
0 10 20 30 40 50−8
−6
−4
−2
0
Bargaining Power
% d
evi
atio
n
0 10 20 30 40 50
−6
−4
−2
0
2Real Wage
% d
evi
atio
n
0 10 20 30 40 50−1
0
1
2
Return to Capital
pp
de
via
tion
0 10 20 30 40 50
0
10
20
Investors‘ Consumption
% d
evi
atio
n
0 10 20 30 40 500
5
10
15
Investors‘ Physical Investment
% d
evi
atio
n
0 10 20 30 40 500
20
40
60
80
Investors‘ Loans
% d
evi
atio
n
0 10 20 30 40 50
−6
−4
−2
0
Workers‘ Consumption
% d
evi
atio
n
0 10 20 30 40 5060
80
100
120
140Workers‘ Debt−to−Income Ratio
leve
l in
%
0 10 20 30 40 500
1
2
3Crisis Probability
leve
l in
%
An Improved Scenario:Orderly Debt Restructuring
• Highly persistent decrease in workers’ bargaining power, as before.
• Financial crisis in year 30, but real crisis is mostly avoided.
0 10 20 30 40 50−8
−6
−4
−2
0
Bargaining Power
% d
evi
atio
n
0 10 20 30 40 50
−6
−4
−2
0
2
Real Wage
% d
evi
atio
n
0 10 20 30 40 50−1
0
1
2
Return to Capital
pp
de
via
tion
0 10 20 30 40 50
0
5
10
15
20
Investors‘ Consumption
% d
evi
atio
n
0 10 20 30 40 500
5
10
15
Investors‘ Physical Investment
% d
evi
atio
n
0 10 20 30 40 500
50
100Investors‘ Loans
% d
evi
atio
n
0 10 20 30 40 50
−4
−2
0
Workers‘ Consumption
% d
evi
atio
n
0 10 20 30 40 5060
80
100
120
140Workers‘ Debt−to−Income Ratio
leve
l in
%
0 10 20 30 40 500
1
2
3Crisis Probability
leve
l in
%
A Much More Sustainable Scenario:Restoration of Workers’ Bargaining Power
• Highly persistent decrease in workers’ bargaining power, as before.
• But in year 30 workers’ bargaining power is restored to its original level.
• Financial and real crisis is thereby avoided.
0 10 20 30 40 50−8
−6
−4
−2
0
Bargaining Power
% d
evia
tion
0 10 20 30 40 50
−6
−4
−2
0
2
4
Real Wage
% d
evia
tion
0 10 20 30 40 50
−1
0
1
2
Return to Capital
pp d
evia
tion
0 10 20 30 40 50
0
10
20
Investors‘ Consumption
% d
evia
tion
0 10 20 30 40 500
5
10
15
Investors‘ Physical Investment
% d
evia
tion
0 10 20 30 40 500
20
40
60
80
Investors‘ Loans
% d
evia
tion
0 10 20 30 40 50
−4
−2
0
Workers‘ Consumption
% d
evia
tion
0 10 20 30 40 5060
80
100
120
140Workers‘ Debt−to−Income Ratio
leve
l in
%
0 10 20 30 40 500
1
2
3Crisis Probability
leve
l in
%
Leverage Comparison Across Scenarios• Orderly debt restructuring can help in the short run,
but with inequality unchanged debt starts to trend up again.
• Restoration of workers’ bargaining power
puts leverage on a sustained downward trend.
0 10 20 30 40 5060
80
100
120
140
leve
l in %
Baseline
Orderly Debt Restructuring
Restoration of Workers‘ Bargaining Power
• Discussion: How Can This Policy Be Implemented?
1. Higher Pre-Tax Wages through Higher Bargaining Power:
— Strengthening collective bargaining rights?
— Difficulties: Wage competition from China and other countries.
— Payoffs: Avoiding further crises.
2. Higher After-Tax Wages through Lower Taxes:
— Switch from labor income taxes to other taxes?
— Difficulties: Higher capital income taxes would drive investment else-
where.
— Ways Out? Taxes on rents (land, natural resources, financial sector).
5 Summary• Empirical Link in 1929 and 2007: Higher income inequality⇒ higher lever-
age ⇒ large crises.
• Theoretical Model:— Key shock: Decrease in workers’ bargaining powers over incomes =
smaller “share of the pie”.— Key mechanism: Recycling of investors’ income gains back to workers
as loans.
• Conclusion:— Only an improvement of workers’ bargaining power leads to a sustained
reduction in crisis probability.— Solutions to financial fragility that leave bargaining power (or alterna-
tively taxation) untouched run into the problem that investors’s surplus
funds will keep pushing loans and therefore crisis probability higher.
6 Is Government Debt a Separate Issue?• Not really.
• A significant share of government debt has just been another (indirect) way
for the lower and middle classes to borrow from the top income group.
• Much spending was on governmental programs that went to the majority,
while much of the resulting debt is held by the top income group.
• In other words, problems of high government debt have an important income
distribution dimension.
• Major exception: Government debt held by foreigners.
Financial Asset Shares of the Top 5% Income Group
1990 1995 2000 200560
65
70
75
80
85
90
60
65
70
75
80
85
90Direct Bond Holdings Share (in %)
1990 1995 2000 200540
45
50
55
60
65
70
40
45
50
55
60
65
70Mutual Funds Holdings Share (in %)
1990 1995 2000 200532
34
36
38
40
42
32
34
36
38
40
42Retirement Accounts Share (in %)
2
7 How About Foreign Debt?
• Empirical regularities for major economies:
— More inequality almost always accompanied by CA deterioration.
— Major exception: China.
• Explanation in general:
— Workers borrow from both domestic and foreign investors.
— Capital account surplus implies current account deficit.
• Explanations for China: Chinese workers face borrowing constraints, so Chi-
nese investors deploy their savings overseas.