risk - Harvard Kennedy School

Download Report

Transcript risk - Harvard Kennedy School

Portfolio Balance
Lecture 24 assumption:
exchange risk is the only important risk.
Lecture 25 assumption:
default risk is important.
API-120 - Prof.J.Frankel
Recap of L24
• Questions
– How can we allow for effects of risk?
• Currency risk.
• Country risk.
– How can we allow for effects of debt even if it is not monetized?
• Effects of budget deficits &
•
current account deficits.
• Key parameters
– Risk-aversion, ρ
– Variance of returns, V
– Covariances among returns, Cov.
• Home bias in portfolio holdings
API-120 - Prof.J.Frankel
Evidence of home bias in US holdings
Equity shares are from 1997 comprehensive survey of US residents’ holdings of foreign securities.
Bias column ≡ 1 minus (foreign equity share / world market share).
If US investors held foreign securities in proportions equal to those in the world equity market benchmark, bias would = 0.
From: G.Baekert & R.Hodrick, Intl. Fin.Management, 2004, Table 14.16, Panel A
Country
Share in U.S.
Equity Portfolio
US
89.90
UK
1.82
Japan
1.14
France
0.71
Canada
0.59
Germany
0.54
Italy
0.35
Netherlands 0.89
Switzerland 0.52
Sweden
0.32
Bias
0.79
0.88
0.75
0.75
0.85
0.76
0.55
0.79
0.72
Source: Based on Table 1, in Ahearne, Griever & Warnock (2002).
Country
Share in U.S.
Equity Portfolio
Spain
0.21
Australia
0.26
Hong Kong 0.23
Mexico
0.29
Brazil
0.26
India
0.05
China
0.02
Taiwan
0.04
Russia
0.07
South Africa 0.08
Bias
0.83
0.79
0.87
0.56
0.76
0.91
0.98
0.97
0.87
0.92
International diversification has risen
Proportion of foreign bonds + equities in total equity +
bond portfolios of residents in the reported countries.
From a 2002 UBS Asset Management study.
Source: Baekert & Hodrick, op.cit., Table 14.16, Panel B
1991
US
4%
Japan
12%
The Netherlands 12%
UK
23%
Switzerland
11%
Australia
14%
Sweden
4%
2000
11%
27%
62%
26%
21%
19%
25%
Home bias in equity holdings.
Most equities are held by domestic residents.
Home bias in equity holdings has slowly declined
Lecture 25: Country Risk
One lesson of portfolio diversification theory:
A country borrowing too much drives up the expected rate of
return it must pay. The supply of funds is not infinitely elastic.
-- especially for developing countries.
The portfolio-balance model can be very general (menu of assets).
•
In Lecture 24, we considered a special case relevant
especially to rich-country bonds: exchange risk is the only risk.
•
Some modifications are appropriate for developing country debt.
API-120 - Prof.J.Frankel
API-120 - Prof.J.Frankel
EM sovereign spreads
Bpblogspot.com
↑ Spreads shot up in 1990s crises,
• and fell to low levels in next decade.↓
Spreads rose again in Sept.2008 ↑ ,
• esp. on $-denominated debt
• & in E.Europe.
WesternAsset.com
World Bank
API-120 - Prof.J.Frankel
What determines spreads?
EMBI is correlated with risk perceptions
risk off
“risk on”
Laura Jaramillo & Catalina Michelle Tejada, IMF Working Paper, 2011
API-120 - Prof.J.Frankel
The portfolio balance model
can be applied to country risk
Demand for assets issued by various countries f:
x i, t = Ai
+ [ρV]i -1
Et (r ft+1 – r dt+1) ;
Now the expected return Et (r ft+1) subtracts from i ft
the probability of default times loss in event of default.
Similarly, the variances & covariances
factor in risks of loss through default.
When perceptions of risk are high, sovereign spreads must be high
for investors to absorb given supplies of debt.
API-120 - Prof.J.Frankel
In developing countries:
The view
from
the South
• Domestic country is usually assumed a debtor, not a creditor.
• It must pay a premium as compensation for default risk.
•
Debt to foreigners was usually $-denominated (before 2000).
• Then, expected return = observed spread between interest
rate on the country’s loans or bonds & risk-free $ rate,
minus expected loss through default
-- instead of rp.
• Denominator for Debt:
More relevant than world wealth is
the country’s GDP or X. Why? Earnings determine ability to repay.
•
Supply-of-lending-curve slopes up: when debt is large
investors fear default & build a country risk premium into i.
API-120 - Prof.J.Frankel
• The spread may rise steeply when Debt/GDP is high.
Stiglitz: it may even bend backwards, due to rising risk of default.
Supply of
funds from
world
investors
i
b
API-120 - Prof.J.Frankel
≡ Debt/GDP
Eichengreen & Mody (2000):
Spreads charged by banks
on emerging market loans
are significantly:
• lower if the borrower
generates more business for
the bank,
but
• higher if the country has:
-----
high total ratio of Debt/GDP,
rescheduled in previous year
high Debt Service/X, or
unstable exports; and
• reduced if it has:
-- a good credit rating,
-- high growth, or
-- high reserves/short-term debt
API-120 - Prof.J.Frankel
For some years after a restructuring, the defaulter
may be excluded from access to international finance.
Estimated from
67 restructurings,
1980-2009
Juan
Cruces &
Christoph
Trebesch,
2013,
“Sovereign
Defaults:
The Price
of Haircuts,”
AEJ: Macro,
Fig.5, p. 111.
For some years after a restructuring,
the defaulter has to be paid higher interest rates,
especially if creditors had to take a big write-down (“haircut”).
Estimated,
1993-2010
especially the 1st 5 years
Cruces & Trebesch, 2013, “Sovereign Defaults: The Price of Haircuts,” Fig.3.
API-120 - Prof.J.Frankel
Why don’t debtor countries default more often,
given absence of an international enforcement mechanism?
1.
Common answer: They want to preserve their
creditworthiness, to borrow again in the future.
Kletzer & Wright (2000), Amador (2003), Aguiar & Gopinath (2006), Arellano (2008), Yue (2010).
But: • “Defaulters don’t seem to bear much of a penalty for long”:
Eichengreen (1987), Eichengreen & Portes (2000), Arellano (2009), Panizza, Sturzenegger & Zettelmeyer (2009).
• “Not a sustainable repeated-game equilibrium”: Bulow-Rogoff (1989).
2. Cynical answer: Finance Ministers want to remain members
in good standing of the international elite.
3. Best answer (perhaps): Defaulters may lose access
to international banking system, including trade credit.
Loss of credit disrupts production, even for export.
Theory: Eaton & Gersovitz (RES 1981, EER 1986). Evidence: Rose (JDE 2005).
API-120 - Prof.J.Frankel
Debt dynamics:
Definition of sustainability: a steady or falling debt/GDP ratio
𝐷𝑒𝑏𝑡
𝑏 ≡
𝑌
𝑑𝑏
𝑑𝑡
=>
𝑑𝑏
𝑑𝑡
=
where Y ≡ nominal GDP.
𝑑 𝐷𝑒𝑏𝑡/𝑑𝑡
𝑌
−
𝐷𝑒𝑏𝑡 𝑑𝑌
𝑌2 𝑑𝑡
=
𝑇𝑜𝑡𝑎𝑙 𝐹𝑖𝑠𝑐𝑎𝑙 𝐷𝑒𝑓𝑖𝑐𝑖𝑡
𝑌
=
𝑃𝑟𝑖𝑚𝑎𝑟𝑦 𝐷𝑒𝑓𝑖𝑐𝑖𝑡 + (𝑖 𝐷𝑒𝑏𝑡)
𝑌
= 𝑑 +
= 𝑑 +
−
𝐷𝑒𝑏𝑡 𝑑𝑌/𝑑𝑡
𝑌
𝑌
𝑖 𝑏 − 𝑏𝑛
𝑖 − 𝑛 𝑏.
− 𝑏𝑛
where n 
nominal growth rate.
where d  Primary Deficit / Y.
=> Debt ratio explodes if d > 0 and i > n (or r > real growth rate).
API-120 - Prof.J.Frankel
db
= d + (i - n) b.
dt
where
Debt
b
Y
n  nominal growth rate, and d  primary deficit / Y .
Debt dynamics line
shows the relationship between b and (i-n), for db/dt = 0.
Even with a primary surplus (d<0),
if i is high (relative to n), then b is on explosive path.
i
db/dt=0
range of explosive debt
range of
declining Debt/GDP ratio
n1
0
ius
b
API-120 - Prof.J.Frankel
Debt dynamics, continued
• It is best to keep b low to begin with,
especially for “debt-intolerant countries.”
• Otherwise, it may be hard to stay on the stable path
• if
– i rises suddenly,
• due to either a rise in world i* (e.g., 1982, 2015?), or
• an increase in risk concerns (e.g., 2008);
– or n exogenously slows down.
• Now add the upward-sloping supply of funds curve.
• i includes a default premium,
which probably depends in turn on db/dt.
• => It may be difficult or impossible to escape the unstable path
– without default, write-down, or restructuring of the debt,
– or else inflating it away,
• if you are lucky enough to have borrowed in your own currency.
API-120 - Prof.J.Frankel
Debt dynamics, with inelastic supply of funds
i
Greece
2012
range of explosive debt
range of
declining Debt/GDP
Ireland
2012
n1
ius
0
API-120 - Prof.J.Frankel
b
explosive debt path
API-120 - Prof.J.Frankel
Professor Jeffrey Frankel, Kennedy School, Harvard University
Appendix 1: Debt dynamics graph,
with possible unstable equilibrium
Supply of
funds line
i
Initial debt
dynamics line
{
sovereign
spread
iUS
Debt
b
Y
API-120 - Prof.J.Frankel
(1) Good times.
Growth is strong.
db/dt = 0, or if > 0 nobody minds. Default premium is small.
(2) Adverse shift. Say growth n slows down. Debt dynamics
line shifts down, so the country suddenly falls in the range db/dt>0.
=> gradually moving rightward along the supply-of-lending curve.
(3) Adjustment. The government responds by a fiscal
contraction, turning budget into a surplus (d<0). This shifts the
debt dynamics line back up. If the shift is big enough, then once
again db/dt=0.
(4) Repeat. What if there is a further adverse shift?
E.g., a further growth slowdown (n↓) in response to the higher i & budget
surplus. => b starts to climb again.
But by now we are into steep part of the supply-of-lending curve.
There is now substantial fear of default => i rises sharply.
The system could be unstable….
API-120 - Prof.J.Frankel
Appendix 2: The blurring of lines between debt
of advanced countries and developing countries
• 1) Since the crisis of the euro periphery began in Greece
in 2010, we have become aware that “advanced”
countries also have sovereign default risk.
• 2) Since 2000, Emerging Market Countries have
increasingly been able to borrow in their own
currencies, so their debt carries currency risk (not just
default risk).
API-120 - Prof.J.Frankel
1) Country creditworthiness is now inter-shuffled
“Advanced” countries (Formerly) “Developing” countries
AAA Germany, UK
Singapore, Hong Kong
AA+
US, France
AA
Belgium
Chile
AAJapan
China
A+
Korea
A
Malaysia, South Africa
ABrazil, Thailand, Botswana
BBB+ Ireland, Italy, Spain
BBB- Iceland
Colombia, India
BB+
Indonesia, Philippines
BB Portugal
Costa Rica, Jordan
B
Burkina Faso
SD
Greece
S&P ratings 2012
,
API-120 - Prof.J.Frankel
Feb.2012 updated 8/
Spreads for Italy, Greece, & other Mediterranean members
of € were near zero, from 2001 until 2008
and then shot up in 2010
Market Nighshift Nov. 16, 2011
API-120 - Prof.J.Frankel
2) The end of Original Sin:
After 2000, Emerging Markets successfully issued more debt
in their own local currencies (LC), instead of $-denominated (FC).
Fig. 2 from Jesse Schreger & Wenxin Du
“Local Currency Sovereign Risk,” HU, March 2013
API-120 - Prof.J.Frankel
Many developing country governments increasingly borrow
in terms of local currency rather than foreign.
International Monetary Fund, 2014
API-120 - Prof.J.Frankel
Turkey is able to borrow in local currency (lira),
but has to pay a high currency premium to do so.
{
Total premium on
Turkey’s lira debt
over US treasuries
Pure default risk premium on lira debt
Schreger & Du, 2013, “Local Currency Sovereign Risk,” HU, 2013, Fig. 5
API-120 - Prof.J.Frankel
{