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.

3 Stylized Facts

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.

CanadaFranceItaly

JapanNetherlands

Sweden

Switzerland

2

4

6

8

10

12

Change in Income Share of Top 5% (x-axis) and

Change in CA Balance (y-axis)

Australia

Germany

New Zealand

Portugal

Spain

United LKingdom

United States

R² = 0.6321

-8

-6

-4

-2

0

-2.00 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00