The Great Leveraging

Five Facts and Five Lessons for Policymakers

Alan M. Taylor University of Virginia, NBER & CEPR

International Banking Conference Federal Reserve Bank of Chicago and CEPR 1516 November 2012
1

Fact 1. Crises: Almost forgo4en: now theyre back

Percentage of economies in a nancial crisis 0 10 20 30 40 50 60 70
1800

A long standing problem For DM and EM Exception: 1940 to 1970 period unusually quiescent. Why? Internal or external constraints?

1850 High-income economies

1900

1950

2000

Middle- & low-income economies

Fact 2. Consequences: forgot depressing/deaConary impacts

Evidence-based macroeconomics Event study 14 advanced countries 140 years of data Recessions are painful Those with financial crises are more painful Those with global financial crises are worse still Prewar versus Postwar
Change in real GDP growth rate
4 year windows before/after recession peak

Change in ination rate

4 year windows before/after recession peak

Change in real loan growth rate

4 year windows before/after recession peak

-.005

-.01

-.01

-.02

-.015

-.03

-.02

-.025

-.04

Normal

Global Crisis Crisis

Normal

Crisis

Normal

Global Crisis Crisis

Normal

Crisis

-.08
Normal

-.06

-.04

-.02

Global Crisis Crisis

Normal

Crisis

Prewar Recessions

Postwar Recessions

Prewar Recessions

Postwar Recessions

Prewar Recessions

Postwar Recessions

Fact 3. Extreme leverage: historically unprecedented

Then Age of Money Now Age of Credit How? More leverage Wholesale funding Why? Private actions (recovery from GD/WW2) Government policies (financial liberalizations)
2 0 .5 ratio 1 1.5

1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010 Bank Loans/GDP Bank Assets/GDP Broad Money/GDP

Fact 3. Extreme leverage: banks versus sovereigns

Advanced countries Public debt crisis/ excess?
There is a post 2008 blip
1 .2 .4 ratio .6 .8

Private credit crisis/ excess?

Larger and trending up since 1990

Reversal of ratios striking after 1960

Safe assets?

1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010 Bank Loans/GDP Public Debt/GDP

Fact 4. Global asymmetry: EMs buy insurance, DMs sell it

Post-1990s EMs switch to safer, countercyclical polices and larger buffers Great Reserve Accumulation Unique phase in history? Gold standard Net capital flows Private inflows Official outflows No Lucas paradox?
1000 6000 U.S. $ billions 0 2000 4000 -1000 -500 U.S. $ billion 0

EM Ofcial & Private BOP Account Flows

EM & DM NIIP Account Ofcial Reserves Stocks

500

1980

1990
EM Ofcial Flows

2000

2010

1999q1 2001q1 2003q1 2005q1 2007q1 2009q1

Emerging markets Developed markets

EM Private Capital Flows

Fact 5. Savings glut: short run panic v. long run demography

Dependents as a percentage of working age (2064) 90 50
1970 World

Short term reasons to think the era of cheap capital is over Investment rebound in EM and DM after panic?
Not quite yet!

EM reserve step change completed? DM delevering slow? Longer term reasons? Demography Offsetting/postponing factor Recurrent and ongoing flights to safety in panic Real rate = 0% in Jun 2012!

60

70

80

1980

1990

2000

2010

2020

2030

2040

More developed regions

Less developed regions

Summing up the facts

An Inconvenient Truth Crises just a fact of life in modern finance capitalism? Exception was 195070 with financial repression, regulation, controls. Period of low credit creation. But it was also still a period of high growth. Plus A Series of Unfortunate Events Cheap capital in asymmetric world. Credit boom. Good = productive projects. Bad = risk of boom-bust cycle. Sequel Hunger for safe assets, demographic shifts slow (but coming). Low real rates for now= deflationary shock continues, and credible sovereigns can be funded. Next What lessons for policy in this kind of financial landscape Macro policy / Financial policy
8

Lesson 1: Past private credit growth does contain valuable predicCve informaCon about likelihood of a crisis
1.00 0.00 True positive rate 0.25 0.50 0.75

Schularick and Taylor 2012 AER Credit Booms Gone Bust Use lagged credit growth T5,..,T1 Forecast of a financial crisis {0,1} in year T Ex ante credit boom makes a financial crisis more likely
Beats null (cointoss) Beats narrow or broad money Robust to other controls including macro, interest rates, and stock prices

0.00

0.25

0.50 True negative rate

0.75

1.00

Null, FE only (AUC = 0.640) Money + FE (AUC = 0.686)

Credit + FE (AUC = 0.750) Reference

Lesson 2: External imbalances/public debts are a distracCon Jord, Schularick,

Taylor 2011 IMF Economic Review Financial Crises, Credit Booms, and External Imbalances: 140 Years of Lessons
True positive rate 0.25 0.50 0.75 0.00
0.00

Couldnt it all be down to external imbalances rather than internal?

Add current account (%GDP) to the forecast system and run a horse race. As a policymaker, which signal has more valuable information about incipient financial fragility?

1.00

0.25

0.50 True negative rate

0.75

1.00

Null, FE only (AUC = 0.641) CA + FE (AUC = 0.685) Reference

Credit + FE (AUC = 0.750) Credit + CA + FE (AUC = 0.767)

Not CA/GDP Same result holds for public debt/GDP

0.00
0.00

True positive rate 0.25 0.50 0.75

1.00

0.25

0.50 True negative rate

0.75

1.00

Null, FE only (AUC = 0.638) Pub. debt + FE (AUC = 0.645) Reference

Credit + FE (AUC = 0.745) Credit + pub. debt + FE (AUC = 0.747)

10

Lesson 3: AQer a credit boom, expect a more painful recession, normal or nancial-crisis
Credit boom before v lost output afterwards Jord, Schularick, Taylor 2012 When credit bites back Larger credit boom ex ante correlates with deeper recessions in each case
In addition to the larger credit boom making more painful financial crisis case more likely to occur

coef = -.50628744, se = .17115664, t = -2.96

Level of log real GDP per capita in year 1-5, orthogonal component -40 -20 0 20

Normal recessions
Level of log real GDP per capita in year 1-5, orthogonal component

Financial recessions

-40

-30

-20

-10

10

coef = -.76393738, se = .22300767, t = -3.43

-4

-2

-2

Excess credit in prior expansion, orthogonal component

Excess credit in prior expansion, orthogonal component

11

Lesson 4: In a nancial crisis with large run-up in private sector credit, mark down growth/inaCon more
Credit boom before v other outcomes Jord, Schularick, Taylor 2012 When credit bites back Larger credit boom ex ante correlates with deeper recessions in each case
In addition to the larger credit boom making more painful financial crisis case more likely to occur Also depressing for investment and inflation outcomes

Cumulative Change From the Start of the Recession

Real GDP per capita

10

Real Investment per capita

10 20

CPI Prices

Financial
-5

-20

-10

-10

-30

0
30

5
.2

5
.1

10

15

Normal

Real Lending per capita

Govt. Short-term Interest Rates

Govt. Long-term Interest Rates

20

10

-.2

-.4

-.2

-.1

12

Lesson 5: In a nancial crisis with large public debt, and large run-up in private sector credit mark down growth/ inaCon even more
JST, work in progress Zero reference = no treatment Blue = normal recession after +1% extra credit/ GDP ppy treatment Red = financial recession after +1% extra credit/ GDP ppy treatment Lt gray = Blue line path as public debt/GDP vary from 0% to 100% Dk gray = Red line path public debt/GDP vary from 0% to 100%
2

Change in growth rate in each year after start of recession

Real GDP per Capita

Financial: Debt/GDP = 0% Normal: Debt/GDP = 0%

Percent -1

Normal: Debt/GDP = 50% Financial: Debt/GDP = 50% Normal: Debt/GDP = 100%

-3

-2

Financial: Debt/GDP = 100%

-4

3 Years

Normal recession: Debt/GDP = 50% Normal recession: Debt/GDP = 0100%

Financial crisis recession: Debt/GDP = 50% Financial crisis recession: Debt/GDP = 0100%

13

Summing up the lessons

Pre-crisis prevention Central bank complacency, with two obvious and key failures Inflation targeting not enough, unable to avert credit boom/bust crisis
Didnt we know this already from history of the gold standard, etc.?

Not having well thought out banking supervision/resolution, LOLR regime

Ditto

Both failures present with a vengeance in the Eurozone with amplification factors Post-crisis response What will the path look like? Worse than people thought/think Massive deflationary shock; CBs beware of premature tightening ECB rate rise in 2011? Fed/BoE/BoJ responses more accomodative, but large headwinds also

14

The Great Leveraging

Five Facts and Five Lessons for Policymakers

Alan M. Taylor University of Virginia, NBER & CEPR

International Banking Conference Federal Reserve Bank of Chicago and CEPR 1516 November 2012
15