Monthly Economic Monitoring of Ukraine No 231, April 2024
Michael Kumhof, Stockholm, 2 nov
1. Inequality, Leverage and Crises
Michael Kumhof, International Monetary Fund
Romain Ranciere, International Monetary Fund and Paris School of Economics
2. The views expressed herein are those of the authors and should not be attributed
to the IMF, its Executive Board, or its management.
3. 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.
4. 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.
5. 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.
7. 1920-1931
60 35
55 33
50
31
45
Percent
Percent
29
40
27
35
25
30
Private Non Corporate+Trade Debt to GNP
Share of Top 5% in Income Distribution
25 23
1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931
Source: Statistical Abstract of the United States, U.S. Department of Commerce.
1983-2008
150 36
Household Debt to GDP
140 Share of Top 5% in Income Distribution 34
130 32
120 30
Percent
Percent
110 28
100 26
90 24
80 22
70 20
1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008
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.
Income Inequality and Household Leverage:
(i) Moved up together pre-crisis.
(ii) Both pre-1929 and pre-2007.
2
8. 50 Top Decile of Earnings Distribution 50
Median Decile of Earnings Distribution
40 40
Bottom Decile of Earnings Distribution
Cumulative Percent Change
Cumulative Percent Change
30 30
20 20
10 10
0 0
-10 -10
-20 -20
-30 -30
-40 -40
1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004
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.
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.
9. 150 150
Bottom 95% of the Income Distribution
Top 5% of the Income Distribution
130 Aggregate Economy 130
Percentage Points
Percentage Points
110 110
90 90
70 70
50 50
30 30
1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006
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.
Debt to Income Ratios:
(i) Lower or flat for the rich.
(ii) Sharply higher for the remainder.
10. 240 8.5
Private Credit to GDP
220 Value Added GDP Share of Financial Sector 8.0
200 7.5
180 7.0
Percent
Percent
160 6.5
140 6.0
120 5.5
100 5.0
80 4.5
1985 1990 1995 2000 2005
Sources: Private Credit to GDP from World Bank Financial Structure Database (real private credit by deposit banks
and other financial institutions, relative to GDP). Value Added GDP Share of Financial Sector from Philippon (2008).
Size of the U.S. Financial Sector:
(i) Private Credit to GDP more than doubled.
(ii) Banks’ share in GDP more than doubled.
11. 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 bargaining
powers 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 and
middle 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 leverage
over decades.
• Result: Higher financial fragility ⇒ risk of financial crisis ⇒ eventual crash.
13. - Investors consume more.
- Investors invest more in equity.
- Investors make more loans.
Baseline Scenario
• Highly persistent decrease in workers’ bargaining power.
• Financial and real crisis in year 30.
Bargaining Power Real Wage Return to Capital
0 2
−2 0 2
pp deviation
% deviation
% deviation
−4 −2 1
−6 −4 0
−6 −1
−8
0 10 20 30 40 50 0 10 20 30 40 50 0 10 20 30 40 50
Investors‘ Consumption Investors‘ Physical Investment Investors‘ Loans
20
80
15
% deviation
% deviation
% deviation
10 60
10
40
0 5 20
0 0
0 10 20 30 40 50 0 10 20 30 40 50 0 10 20 30 40 50
Workers‘ Consumption Workers‘ Debt−to−Income Ratio Crisis Probability
140 3
0
120
% deviation
level in %
level in %
−2 2
100
−4
1
80
−6
60 0
0 10 20 30 40 50 0 10 20 30 40 50 0 10 20 30 40 50
14. Baseline Scenario
• Highly persistent decrease in workers’ bargaining power.
• Financial and real crisis in year 30.
Bargaining Power Real Wage Return to Capital
0 2
−2 0 2
pp deviation
% deviation
% deviation
−4 −2 1
−6 −4 0
−6 −1
−8
0 10 20 30 40 50 0 10 20 30 40 50 0 10 20 30 40 50
Investors‘ Consumption Investors‘ Physical Investment Investors‘ Loans
20
15
80 - Workers'
% deviation
% deviation
% deviation
10
10
60 leverage
Workers 40
increases.
0 5 20
reduce 0
- This
0
consumption. 10
0 20 30 40 50 0 10 20 30 40 50 0 10 20 30 40 50
increases the
Workers‘ Consumption Workers‘ Debt−to−Income Ratio Crisis Probability
140 3 probability of
0
120 crises.
% deviation
level in %
level in %
−2 2
100
−4
1
80
−6
60 0
0 10 20 30 40 50 0 10 20 30 40 50 0 10 20 30 40 50
15. 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. Real wage
Bargaining Power Real Wage Return to Capital collapse at
0 2
0 2
crisis time is
−2
pp deviation
% deviation
−4 −2 1 % deviation now very much
−6 −4 0 smaller.
−6 −1
−8
0 10 20 30 40 50 0 10 20 30 40 50 0 10 20 30 40 50
The drop in
Investors‘ Consumption Investors‘ Physical Investment Investors‘ Loans leverage at
100
20 15 crisis time is
% deviation
% deviation
% deviation
15
10 therefore
10 50
5 5 much more
0
0 0
substantial.
0 10 20 30 40 50 0 10 20 30 40 50 0 10 20 30 40 50
Workers‘ Consumption Workers‘ Debt−to−Income Ratio Crisis Probability
But
0
140 3
immediately
120
afterwards
% deviation
level in %
level in %
2
−2
100
1 leverage
−4 80
starts rising
60 0
0 10 20 30 40 50 0 10 20 30 40 50 0 10 20 30 40 50 again.
16. 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.
Bargaining Power Real Wage Return to Capital
Recovery in
0 4
2
real wage
2
gives workers
pp deviation
−2
% deviation
% deviation
0 1
−4
−2 0 the means to
−6 −4 −1
−8 −6 service their
0 10 20 30 40 50 0 10 20 30 40 50 0 10 20 30 40 50
debts.
Investors‘ Consumption Investors‘ Physical Investment Investors‘ Loans
20
15 80
% deviation
% deviation
% deviation
10 10
60 Leverage
40
0 5
20
therefore goes
0 0 on a sustained
0 10 20 30 40 50 0 10 20 30 40 50 0 10 20 30 40 50
Workers‘ Consumption Workers‘ Debt−to−Income Ratio Crisis Probability
downward
140 3
path.
0 120
% deviation
level in %
level in %
2
−2 100
1
80
−4
60 0
0 10 20 30 40 50 0 10 20 30 40 50 0 10 20 30 40 50
17. 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.
140
120
level in %
100
80
Baseline
Orderly Debt Restructuring
Restoration of Workers‘ Bargaining Power
60
0 10 20 30 40 50
18. • 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).
19. 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.
20. 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.
21. Financial Asset Shares of the Top 5% Income Group
Direct Bond Holdings Share (in %) Mutual Funds Holdings Share (in %) Retirement Accounts Share (in %)
90 90 70 70 42 42
85 85 65 65 40 40
80 80 60 60
38 38
75 75 55 55
36 36
70 70 50 50
65 65 45 45 34 34
60 60 40 40 32 32
1990 1995 2000 2005 1990 1995 2000 2005 1990 1995 2000 2005
2
22. 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.
23. Change in Income Share of Top 5% (x-axis) and
Change in CA Balance (y-axis)
12
10
Switzerland
8
6
Sweden
4
Netherlands
Japan
France Canada
2 Italy
Germany
0
-2.00 0.00 2.00 Spain 4.00 6.00 8.00 10.00 12.00 14.00
Australia
-2 New Zealand
United LKingdom
-4 United States
Portugal
-6
R² = 0.6321
-8