Sovereign Debt
East Asian Crisis and Argentina-Brazil
Recent crisis episodes: Some basic facts about the crises in Asia and
Brazil/Argentina
¶ Strong “twin crises” where they erupt
¶ Currencies fixed beforehand
¶ Strong devaluations (potentially in the case of Brazil?)
¶ Poster-child success stories turned “sour”
¶ Financial sector strongly hit
¶ Brazil: different, but why?
Financial Crises Hans-Joachim Voth
Overhead 2
¶ Asian crisis, acc. to Krugman
• No government deficits
• Good growth
• No obvious weakness of the government’s commitment to the
exchange rate
• Run-up and bust of asset prices before crisis hit
• Affects lots of countries in the region, some spillovers, too
¶ How could such a thing happen?
¶ Observations that help to understand the Krugman model
• Afterwards, lots of banks insolvent
• Investments look dodgy
• Focus on corruption etc.
• Growth slumps
• Debate about “Stalinist growth”
Financial Crises Hans-Joachim Voth
Overhead 3
Bhat/$
Financial Crises Hans-Joachim Voth
Overhead 4
Consequences
¶ Ppp-adjusted per capita income in Thailand declined from $8,800 to
$8,300 between 1997 and 2005
¶ in Indonesia it declined from $4,600 to $3,700
¶ in Malaysia it declined from $11,100 to $10,400.
¶ Despite population growth by 30 mio, Indonesia’s economy in 2005
was still smaller than it had been in 1997 (in $ terms).
¶ Over the same period, average world per capita income rose from
$6,500 to $9,300
Financial Crises Hans-Joachim Voth
Overhead 5
Financial Crises Hans-Joachim Voth
Overhead 6
Crisis? What crisis?
Financial Crises Hans-Joachim Voth
Overhead 7
Common characteristics
¶ No big budget deficits
¶ No inflation problem
¶ Some real appreciation
¶ Large current account deficits + big capital inflows
¶ Lending booms
¶ Rise in non-performing loans before the crisis [up to 10 percent]
¶ Low liquidity (few reserves, high foreign short-term debt)
Financial Crises Hans-Joachim Voth
Overhead 8
Sequence of events acc. to Mishkin
1. Financial liberalization
2. Capital inflows
3. Lending booms in a poor institutional environment (minimal
supervision of banks, etc.)
4. Rise in bad loans
5. Exogenous shocks (political in Mexico, bankruptcies in South Korea
and Thailand) – rise in uncertainty
6. Balance sheets of banks + borrowers weaken
Key question: Why is monetary policy less effective in this context
than normally in a developed country?
Underlying problem: Short-term, foreign-denominated debt
Financial Crises Hans-Joachim Voth
Overhead 9
Krugman story
¶ Stage 1: investments + moral hazard
Table 1: Moral hazard and investment decisions
Safe investment Risky investment
Return in good state 107 120
Return in bad state 107 80
Expected return 107 100
Expected return to owner 7 10
Probability 100% 50%-50%
Financial Crises Hans-Joachim Voth
Overhead 10
¶ Moral hazard will lead to probability payoff
• Overinvestment
• Overvaluation [general
equilibrium effect from moral
hazard in table 1 – see right] Bad state 2/3 25
• Value for risk-neutral investor
25*2/3+100*1/3=50
• Intermediaries with moral Good state 1/3 100
hazard will bid until price =100
[assuming, as before, that they
put up no capital of their own]
• Now think 3 period: what is the
market price?
• With uncertain guarantees, i.e.
an unanticipated shock to
confidence?
Financial Crises Hans-Joachim Voth
Overhead 11
Brazil – the crisis that wasn’t
¶ Large, external, dollar-denominated debt to start with
¶ Much of it short-term
¶ Prospects of Lula winning unhinge expectations (April – May 02)
¶ Danger of a self-fulfilling collapse in confidence
• High spreads
• Roll-over crisis
• Fall of the exchange rate (floating since 99)
• Financial collapse
¶ Standard story
• IMF packages
• Fiscal conservatism of Lula government (2% primary budget
surplus, etc.)
Financial Crises Hans-Joachim Voth
Overhead 12
Blanchard paper
¶ Monetary tightening or fiscal tightening to deal with a potential “3rd
generation” debt crisis?
• Monetary tightening normally works because (a) it cools demand, reducing
imports (b) increases demand for the country’s currency – real
appreciation
• However, the sign of channel (b) unclear
– Could work in the normal way, by offering higher interest rate to depositors
– Alternatively, if the cost of servicing internal debt goes up too much, it increases probability
of default
– This may lead to real depreciation
• Paper estimates default probabilities over time
– Simplest case: risk-neutral investors
– With risk-averse investors, need to estimate log S=log p + a θ*+u where S is the spread, p
is the default probability, and θ is a measure of risk aversion
– Proxies θ with BAA spread in the US
• Then want to explain its determinants:
– Capital flows equation: raising i changes e
– Default risk relationship: default probability depends on debt, which depends on the
exchange rate next period
Financial Crises Hans-Joachim Voth
Overhead 13
Financial Crises Hans-Joachim Voth
Overhead 14
Financial Crises Hans-Joachim Voth
Overhead 15
Financial Crises Hans-Joachim Voth
Overhead 16
Financial Crises Hans-Joachim Voth
Overhead 17
¶ Net result already clear:
• Capital flow relation unclear,
not tightly estimated
• Debt default equation with a
clear effect
• Net effect depends on size of
coefficients
• D=debt level
• μ=share of foreign-
denominated debt
• θ=risk appetite
• Net effect – 100 bp extra Selic
translates into a depreciation of
258 bp of the Real
Financial Crises Hans-Joachim Voth
Overhead 18
An alternative story
¶ Turning points in Brazilian spreads accidentally related to political
events
¶ Global risk appetite changed dramatically
¶ Brazil was first unlucky, then lucky
¶ Fiscal tightening had nothing to do with the escape from the debt trap
Financial Crises Hans-Joachim Voth
Overhead 19
Addenddum
Polls indicate Lula’s favoritism
2800 1100
1000
2300
900
EMBI+ Brazil (bps)
800
1800
EMBI+ (bps)
700
1300
EMBI+
600
500
800
400
300
Brazil EMBI+ 300
dic-01
abr-02
jun-02
ago-02
oct-02
dic-02
abr-03
jun-03
ago-03
oct-03
dic-03
feb-02
feb-03
feb-04
Brazil important for EMBI+ (average of EM bonds)…
Financial Crises Hans-Joachim Voth
Overhead 20
Addenddum
Polls indicate Lula’s favoritism
2800 800
700
2300
600
EMBI+ Brazil (bps)
1800
EMBI+ (bps)
EMBI+ (simple average*) 500
1300
400
800
300
300
Brazil EMBI+ 200
dic-01
jun-02
ago-02
oct-02
dic-02
jun-03
ago-03
oct-03
dic-03
abr-02
abr-03
feb-02
feb-03
feb-04
Simple average of EMBI+ countries excluding Argentina, Brazil and Ecuador
Financial Crises Hans-Joachim Voth
Overhead 21
Addenddum
1000 Polls indicate Lula’s favoritism 1100
900 1000
900
800
US Corporate High Yield
US Corporate High Yield (bps)
800
700
700
EMBI+ (bps)
600
600
500
500
400 EMBI+ 400
300 300
dic-01
abr-02
jun-02
ago-02
oct-02
dic-02
abr-03
jun-03
ago-03
oct-03
dic-03
feb-02
feb-03
feb-04
Brazil important for EMBI+, but for US HY Corporate bonds??
Financial Crises Hans-Joachim Voth
Overhead 22
Addenddum
Polls indicate Lula’s favoritism
1000 100
900 80
US Corporate AAA Yield
US Corporate AAA Yield (bps)
800 60
700 40
EMBI+ (bps)
600 20
500 0
400 EMBI+ -20
300 -40
dic-01
abr-02
jun-02
ago-02
oct-02
dic-02
abr-03
jun-03
ago-03
oct-03
dic-03
feb-02
feb-03
feb-04
And for US AAA Corporate bonds??
Financial Crises Hans-Joachim Voth
Overhead 23
Key lessons learned
¶ Changes in the global price of risk may be driving a large part of
“contagion” around the globe
¶ One factor behind higher co-movements today than in the 19th
century
¶ Arguably driven by the structure of international markets, role of retail
investors
¶ Role of foreign currency debt
• Crucial in East Asian crisis, Argentina, Brazil
• Origins of exposure – original sin? Or weak institutions?
• 19C history suggests – market liquidity for currency crucial
• Recent trends suggest that it may be easier to cure than previously
thought
Financial Crises Hans-Joachim Voth
Overhead 24