WEO-26.05.09

Download Report

Transcript WEO-26.05.09

World Economic Outlook
April 2009
http://www.imf.org/external/pubs/ft/weo/2009/01/index.htm
Chapter IV
How Linkages Fuel The Fire:
The Transmission of Financial Stress
from Advanced to Emerging Economies
Prepared by:
Ravi Balakrishnan, Stephan Danninger, Selim Elekdag,
and Irina Tytell
with support from Stephanie Denis
and Murad Omoev
Background: unprecedented financial
stress in advanced economies.
(GDP weighted number of countries with stress index one stdev above trend)
United States and Canada
United Kingdom
Western Europe
Japan and Australia
1.0
October 1987 stock
market crash
Nikkei crash, DBL
bankruptcy, and
Scandinavian
banking crisis
U.S. banking
stress
LTCM
collapse
Dot-com
crash
0.8
US
Corporate
crisis
0.6
ERM Crisis
0.4
0.2
1981
85
89
93
97
2001
05
0.0
Q1: 3
09
Questions
1. How severe is the current crisis in emerging
economies (EMs)?
2. How strong is the transmission of financial
stress from advanced to emerging economies?
And what is the role of banking sector stress?
3. Can policies help mitigate stress transmission?
4
Key Findings


Financial stress in EMs has surpassed peaks of Asian
crises.
Financial stress transmits strongly and rapidly to EMs

Almost one-for-one response within 1-2 months.

Transmission stronger with closer financial & trade links.

Bank lending linkages main channel in current crisis.

Recovery of capital flows likely to be slow. Past banking
sector stress led to protracted declines in capital flows to
EMs.

During a global crisis, higher fiscal and current account
balances offer little insulation.
5
Outline
1. Financial Stress Index
2. Transmission of Stress
3. Impact on Capital Flows
6
Emerging economies have
experienced severe exchange market
turmoil…
Exchange Rates
(month-to-month percent change)
20
15
Foreign Reserves
(month-to-month percent change)
12
Emerging Asia
Latin America
Emerging Europe and Middle East and Africa
Line 4
8
4
10
0
5
-4
Emerging Asia
Latin America
0
-8
Emerging Europe and Middle East and Africa
Line 4
-5
2007
2008
Jan. 2009
2007
2008
-12
Jan. 2009
7
…and have seen sovereign spreads
widen and equity markets fluctuate.
JPMorgan EMBI Stripped Spreads
(basis points)
Stock Index Volatility
700
4
Emerging Asia
Latin America
Emerging Europe and Middle East and Africa
Line 4
Emerging Asia
600
500
Latin America
Emerging Europe and Middle East and
Africa
3
2
400
1
300
0
200
-1
100
0
2007
2008
Jan. 2009
2007
2008
-2
Jan. 2009
8
1. Financial Stress Index
Composite index comprises 5 components:
EM-FSI = EMPI (exchange market pressure index)
+ sovereign spreads
+ banking sector β
+ stock returns
+ stock volatility



Components standardized (variance weighted &
demeaned)
FSI robust to other weighting schemes (principal
components analysis)
Data will be made publicly available
9
For example, FSI in Russia
10
Emerging Europe
Russia
Other Emerging Economies
Securities markets
Bank beta
8
10
Sovereign spreads (JPMorgan EMBI)
8
Exchange market pressure
Financial stress index
6
6
4
4
2
2
0
0
-2
-2
-4
2007
08
Jan. 09
-4
2007
08
Jan. 09
10
Recent stress above previous peaks.
Comoves with advanced economy stress.
(GDP weighted levels of financial stress indices)
12
6
Emerging economies (left scale)
5
10
Advanced economies (right scale)
4
8
3
6
2
4
1
2
0
0
-1
-2
-2
-4
-3
-6
1996
97
98
99
2000
01
02
03
04
05
06
07
Q1:
09
11
2. Transmission of Stress
Disentangle stress transmission:
global or country-specific effects?
Financial Stress
Emerging Economies
Common factors
Country-specific factors
Commodity prices
Global output
Global interest
rates
Vulnerabilities
Economic characteristics
Financial linkages
Trade linkages
Advanced Economies
Financial Stress
13
Vulnerability Indicators
Current Account Balance
(percent of GDP)
Fiscal Balance
(percent of GDP)
15
15
Emerging Asia
Emerging Europe
CIS
Latin America
MENA
Sub-saharan Africa
10
Emerging Asia
Emerging Europe
CIS
Latin America
MENA
Sub-saharan Africa
10
5
5
0
0
-5
-5
-10
-10
-15
-15
1980
83
86
89
92
95
98
01
04
07
1980
83
86
89
92
95
98
01
04 14 07
How do trade and financial links
affect stress transmission?
1. Through actual or expected capital outflows to
advanced economies (L channel)
Through actual or expected losses on foreign
investments in advanced economies (A channel)
2. Through actual or expected declines in
exports to advanced economies
3. Through 2nd round effects / spillovers
From emerging back to advanced economies
Among emerging economies
15
Advanced and emerging economies
have become more closely linked…
Assets of Advanced Economy Banks
in Emerging and Developing
Economies
(percent of advanced economies’ GDP)
Portfolio Exposures of Advanced
Economies to Emerging and
Developing Economies
(percent of advanced economies’ GDP)
15
15
10
Western Europe
Western Europe
United States and Canada
United States and Canada
Japan and Australia
Japan and Australia
10
5
5
0
0
1987
91
95
99
2003
07
1997
99
2001
03
05
07
16
…with bank links between emerging
and western Europe the strongest.
Liabilities to Advanced
Economy Banks as of 2007
(percent of emerging economies’ GDP)
Portfolio Exposure to Advanced
Economies as of 2007
(percent of emerging economies’ GDP)
60
60
50
Japan
Japan and Australia
Western Europe
Western Europe
United States and Canada
United States and Canada
50
40
40
30
30
20
20
10
10
Africa
MENA
Latin
America
CIS and
Russia
Emerging
Europe
Emerging
Asia
Africa
MENA
Latin
America
CIS and
Russia
Emerging
Europe
Emerging
Asia
0
0
17
Stress Transmission:
Empirical strategy
Estimate stress transmission coefficient (βi) and its
determinants:
EM-FSIit = f (AE-FSIt , Global factorst , Xit )
EM-FSI...... Emerging economy financial stress index
AE-FSI….. Advanced economy financial stress index
Two approaches
 Two-stage approach (Forbes and Chinn 2004)
 Annual panel estimation
18
The Forbes-Chinn Approach—Stage 1:
Gauging the transmission (betas)

For each emerging economy i estimate
EMFSI it   i    ( icl AEFSI tcl 
c l  0,1
cl
c

D
AEFSI
 i 
t l ) 
 1,2
gl
g

GF
  i t l  i EMFSIit 1   it
g l  0,1


SBIC says 1 month lag: rapid transmission!
Betas allowed to vary according to:



Periods: past stress (July 1998 – June 2003) and current
crisis (July 2007 onwards) in advanced economies
Regions: US and Canada, Western Europe, Japan and
Australia
GF = changes in world industrial production,
aggregate commodity price index, 3-month libor
19
The Forbes-Chinn Approach—Stage 2:
Understanding variation in betas

Betas are computed as
ic  ic 0  ic1  ic 0i
ci  (ic 0  ci0 )  (ic1  ci1 )  ( ic 0  ci0 )i

Use betas as dependent variables and estimate
c
ci  i   k FLikc lTLilc   m X im
 ci
k


l
m
Include banking, portfolio, and FDI as Financial
Linkages (FL) and a Trade Linkage (TL) and add
country (and advanced region) effects
FL are measured as liabilities (and assets, when
available) and TL are measured as exports to each
advanced region relative to GDP
20
Stress transmission to emerging
economies is large but it varies…
Country specific β of stress co-movement with advanced economy stress
1.4
1.2
1.0
Mean co-movement (ßi)
0.8
0.6
0.4
0.2
Turkey
Chile
Colombia
Korea
Philippines
Malaysia
South Africa
Mexico
Brazil
Peru
Thailand
Morocco
Poland
Argentina
Egypt
Pakistan
Hungary
China
0.0
21
…by country and time period.
Country specific β of stress co-movement with advanced economy stress
China
1.0
450 line
South Africa
Chile
0.8
Peru
Mexico
Korea
0.5
Colombia
Philippines
Turkey
Malaysia
Brazil
0.3
Thailand
Morocco
0.0
Poland
Latest stress (July 2007 onwards)
1.3
Hungary
Argentina
1.5
-0.3
-0.3
0.0
0.3
0.5
0.8
1.0
Past stress (July 1998 – June 2003)
1.3
1.5
22
Financial linkages matter for stress
transmission, especially via banks

FL are jointly significant determinants of betas

In past stress, FL are also significant individually, but not
when put together

In current crisis, FL via banks is most significant: higher
bank liabilities of emerging to western Europe (by 35%
of GDP) → higher stress transmission from western to
emerging Europe (by a factor of 1)

TL is significant in past stress, but not in current
crisis

Individual country characteristics are insignificant
(i.e. no indirect effects)
23
The Panel Approach:
Effects of country vulnerabilities

Estimate for 18 countries over 1997-2008:
EMFSIit  i   AEFSIt   X it   AEFSIt  X it   GFt   it

Examine direct effects on stress


Explore indirect (interaction) effects


Higher current account and fiscal balances help lower
financial stress, but effects are fairly small in magnitude
Interaction effects turn out not to be significant
Evaluate contributions of explanatory variables to
the variation in the financial stress index
24
Individual country policies offer
little insulation from global stress.
Changes in Emerging Economies’ Financial Stress Index in periods of stress and calm
Emerging economies’ FSI
Contributions from:
Advanced economies’ FSI
Global Factors
Openness
Current account balance
Fiscal balance
Stress
Calm
Foreign reserves
Residual
-0.6
-0.4
-0.2
0.0
0.2
0.4
0.6
0.8
21.0
25
3. Impact on Capital Flows
Case study: after banking crises
capital flows recover only slowly
U.S. banking sector of the 1980s
 Capital flows to Latin America
Japanese Banking crisis of the 1990s
 Capital flows to Emerging Asia
Both cases illustrate long and protracted withdrawals
from EMs by advanced economy banks concerned
27
U.S. banks pulled out relatively more
from all EMs after the debt crisis
Consolidated Bank Claims on Latin America
(percent of destination region’s GDP)
16
Consolidated Bank Claims on Emerging
and Other Developing Economies
(percent of destination region’s GDP)
Japan
United States
Japan
Western Europe
United Kingdom
Western Europe
United Kingdom
United States
14
16
14
12
12
10
10
8
8
6
6
4
4
2
2
0
0
1983
84
85
86
87
88
89
90
1983
84
85
86
87
88
89
90
28
Japanese banks massively withdrew
from East Asia in the late 1990s
Japanese Bank Claims on East
and Offshore Asia
(percent of destination region’s GDP)
Offshore Asia (left scale)
200
East Asia (right scale)
175
Advanced Economy Bank Claims
on East Asia
(percent of destination region’s GDP)
9
United States
United Kingdom
20
8
European Banks
Japan
18
16
7
150
14
6
125
12
5
100
10
4
8
75
3
50
6
2
4
25
1
2
0
0
0
1983
86
89
92
95
98
2001
04
07
1993
95
97
99
2001
03
05
07
29
Past crises presaged declines in capital
inflows and large output losses in EMs.
Net flows, percent of GDP (left scale)
Real output index (right scale)
Trend output (right scale)
Latin America Post 1980 Debt Crisis
(weighted average)
Emerging Asia Post 1997 Asian Crisis
(weighted average)
12
220
12
220
10
200
10
200
8
180
8
180
6
160
6
160
4
140
4
140
2
120
2
120
0
100
0
100
-2
80
-2
80
-4
60
-4
60
-6
40
-6
1970
74
78
82
86
90
40
1990
94
98
02
06 30
Summary & Conclusions

Crisis is of an unprecedented magnitude in advanced
economies.

Financial stress has spread rapidly to all EM regions but
more so to economies with strong financial linkages to
advanced economies.

Strong macroeconomic policies did not insulate
economies because the shock was too large. But they
may limit the real implications of the crisis.

A protracted decline in capital flows to EMs is likely.

The key question going forward is what policies can help
economies emerge rapidly from this crisis.
31
Policy Messages

Strong and rapid stress transmission and little protection
through strong domestic policies argue for coordinated
policy action against global crises.

Important to avoid second round of deleveraging through
both official support for EMs and clean-up of bank
balance sheets in advanced economies.

Beyond near-term: need to improve multilateral insurance
systems to reduce risks for EMs from global financial
integration (e.g. the new IMF FCL).
32