Topics in Latin American Macroeconomics and Development Andy Neumeyer Stanford University and Universidad Torcuato Di Tella
Empirical Evidence on Sovereign Debt
(Neumeyer , Stanford, 2013 )
1 / 39
Introduction
References I
Michael Tomz and Mark L. J. Wright, 2012. "Empirical research on sovereign debt and default," Working Paper Series WP-2012-06, Federal Reserve Bank of Chicago.
I
Carmen M. Reinhart and Kenneth S. Rogoff, 2009. "This Time Is Different: Eight Centuries of Financial Folly," Economics Books, Princeton University Press, edition 1, volume 1, number 8973, November.
I
Cristina Arellano and Ananth Ramanarayanan, 2012. "Default and the Maturity Structure in Sovereign Bonds," Journal of Political Economy, University of Chicago Press, vol. 120(2), pages 187 - 232.
I
Cruces, Juan J. and Trebesch, Christoph: Sovereign Defaults: The Price of Haircuts, American Economic Journal: Macroeconomics, 2013
I
Mark Aguiar and Manuel Amador, Sovereign Debt, prepared for the Handbook of International Economics, Vol. 4
(Neumeyer , Stanford, 2013 )
2 / 39
Introduction
Measuring the Stock of Sovereign Debt I
How should we measure sovereign debt?
I
Typical data reports the face value: undiscounted sum of principal payments (not cash flows) I
I
Problems: I
Principal payments at different dates have the same value
I
Two debts with equal cash flows split differently in principal and interest coupons have different face values
Face value is a relevant concept if there is a default I
Cross default clause: if the sovereign defaults on some of its debt, then that action constitutes a default on other debt even though the sovereign is otherwise current on that debt.
I
Acceleration clause: allow the creditor to accelerate all of the future payments owed to it if one of a set of pre-defined Events of Default takes place (such as a violation of a cross-default provision or a negative pledge clause).
(Neumeyer , Stanford, 2013 )
3 / 39
Introduction
Measuring the Stock of Sovereign Debt I
Market value of debt I
I
Problems: I
Some debts have no market value as they are not traded (official lending, bank lending)
I
Market value of debt depends on default expectations
Informative of recovery values
I
Zero coupon equivalent, ZCE. (Dias, Richmond Wright, 2011): treat all debt payments as a zero coupon bond and sum the face values
I
Risk free present value of promised cash flows (never seen it)
I
Present value of promised cash flows discounted with marginal rate of substitution between present and future consumption for the country (Dias, Richmond, Wright)
(Neumeyer , Stanford, 2013 )
4 / 39
Introduction
(Neumeyer , Stanford, 2013 )
5 / 39
Introduction
How Indebted are Countries?
(Neumeyer , Stanford, 2013 )
6 / 39
Introduction
How Indebted are Countries? Face Values and Zero Coupon Equivalent in Latin America
(Neumeyer , Stanford, 2013 )
7 / 39
Introduction
Maturity and yield spread curves I
Measures of maturity I
Contractual maturity (time to maturity of face value)
I
Macaulay duration: d=
T X i=1
Ci e−rti ti PT −rti i=1 Ci e
where ti is the time to the payment of coupon Ci and rti is a discount factor. I
Duration is a decreasing function of rti
I
Arellano and Ramanarayanan (2012) use market rates on the sovereign defaultable bond
I
Dias, Richmond,Wright (2011) propose rti = 0
I
One could use the risk free rate
(Neumeyer , Stanford, 2013 )
8 / 39
Introduction
Maturity and yield spread curves
I
Measures of the spread curve
I
Given the price ptn of a zero coupon bond with maturity n and a face value of 1 it’s yield, rtn , is derived from n
ptn = e−n rt and the n-period spread at date t is
n stn = rtn − rt,rf
(Neumeyer , Stanford, 2013 )
9 / 39
Introduction
Time Series for Short and Long Spreads Source: Arellano and Ramanarayanan (2012)
(Neumeyer , Stanford, 2013 )
10 / 39
Introduction
Spread Curves Average spread for each country across time for each maturity Yield curves flatten and invert during crises Source: Arellano and Ramanarayanan (2012)
(Neumeyer , Stanford, 2013 )
11 / 39
Introduction
Maturity and Duration of New debt Maturity shortens during crises Source: Arellano and Ramanarayanan (2012)
I
Duration is shorter than maturity
I
Duration is shorter when spreads are high I
I
needs to be separated from the effect of rise in yield
Duration is shorter when yield curve is inverted (Neumeyer , Stanford, 2013 )
12 / 39
Introduction
Are the sovereign spreads default risk?
I
Emerging market bond yield spreads exhibit significant co-movement
I
Global factors like the return to the US stock market, yields on risky corporate debt and VIX volatility indices explain a large fraction of the movement in spreads.
I
Holders of risky sovereign are compensated for aggregate risk in addition to idiosyncratic default risk
(Neumeyer , Stanford, 2013 )
13 / 39
Introduction
Sovereign default I
Definition: default occurs when the legal terms of the debt contract are violated or when the sovereign tenders an exchange offer with less favorable terms than the original issue
I
Questions: I
How often does default occur?
I
Do rich countries default?
I
Does default occur in bad times?
I
How long do defaults last?
I
How big are haircuts
I
Do defaults coincide with banking crises
I
Are defaults more frequent when maturities are shorter?
(Neumeyer , Stanford, 2013 )
14 / 39
Introduction
How Often Do Countries Default? Source: Tomz and Wight (2012) 176 countries since 1820
(Neumeyer , Stanford, 2013 )
15 / 39
Introduction
How Often Do Countries Default? Source: Reinhart and Rogoff (2009) 66 countries and 8 centuries (for independent countries)
(Neumeyer , Stanford, 2013 )
16 / 39
Introduction
How Often do Countries Default? I
Tomz and Wright I
251 defaults by 107 entities
I
Most frequent (serial defaulters). Ecuador, Mexico, Uruguay, Venezuela (+ 8 episodes)
I
Ecuador and Honduras more than 120 years in default (each)
I
Largest Defaults. Greece 2012 - EUR200bn, Argentina 2002 - USD132bn, Russia 1918.
I
Longest defaults Russia 1918 (lasted 69 years), Greek default in 1826 lasted 53 years Unconditional probability of default is 1.8% used in calibrated models
I
I
I
It makes a difference to condition by level of debt and by country
I
It makes a difference how default episodes are aggregated
Median time of defalts is 6.5 years
(Neumeyer , Stanford, 2013 )
17 / 39
Introduction
Do Rich Countries Default? Source: Reinhart and Rogoff (2009)
(Neumeyer , Stanford, 2013 )
18 / 39
Introduction
Do Rich Countries Default? Source: Reinhart and Rogoff (2009)
(Neumeyer , Stanford, 2013 )
19 / 39
Introduction
Do Rich Countries Default? Source: Reinhart and Rogoff (2009)
(Neumeyer , Stanford, 2013 )
20 / 39
Introduction
Do Rich Countries Default? Source: Reinhart and Rogoff (2009)
(Neumeyer , Stanford, 2013 )
21 / 39
Introduction
Do Countries Default in Bad Times? Source: Tomz and Wright (2009) I
incomplete market models (e.g Arellano with cost of default increasing in y ). Default in bad times
I
“Limited commitment” models temptation to default when output is high and persistent Tomz and Wright (2007) weakly negative relation between default and output
I
I I
Loans to sovereigns from private foreign creditors Sovereigns default when output is below HP trend only 60% of the time. I I I
I
Robust to linear trend? other measures of income: taxes, exports Non-linear? . Defaults twice as likely when GDP is more than 7% below trend than during other periods
Reinhart-Rogoff: output costs are larger when accompanied by domestic default.
(Neumeyer , Stanford, 2013 )
22 / 39
Introduction
Do Countries Default in Bad Times? Source: Tomz and Wright (2009)
(Neumeyer , Stanford, 2013 )
Mean Deviation from trend (%) In first year of default In periods of default In last year of default In first periods of no default
-1.6 -1.4 -1.2 0.2
Country years below trend (%) In first year of default In periods of default In last year of default In first periods of no default
61.5 56.2 58.8 47.2
23 / 39
Introduction
Output and Inflation during Domestic and External Defaults Source: Reinhart and Rogoff (2009)
(Neumeyer , Stanford, 2013 )
24 / 39
Introduction
Output during Domestic and External Defaults Debt and GDP Source: Reinhart and Rogoff (2009)
(Neumeyer , Stanford, 2013 )
25 / 39
Introduction
Output during Domestic and External Defaults Frequency of occurrence (%) of growth rate between t − 3 and t. Source: Reinhart and Rogoff (2009)
(Neumeyer , Stanford, 2013 )
26 / 39
Introduction
Taxes during Domestic and External Defaults Source: Reinhart and Rogoff (2009)
(Neumeyer , Stanford, 2013 )
27 / 39
Introduction
Inflation during Domestic and External Defaults Default and Price levels Source: Reinhart and Rogoff (2009)
(Neumeyer , Stanford, 2013 )
28 / 39
Introduction
Argentina
(Neumeyer , Stanford, 2013 )
29 / 39
Introduction
Inflation and Output during Domestic and External Defaults I
Default is more negatively correlated with output for domestic than for external default
I
For external default the correlation is weak
I
For domestic defaults output falls before the default and the trough is at the time of default
I
Private bank credit falls after defaults (Gennaioli, Martin, Rossi (2012)
I
Inflation rises before and after domestic defaults
I
Questions: I I
What is the counterfactual? Are these output shocks transitory or permanent?
(Neumeyer , Stanford, 2013 )
30 / 39
Introduction
Exclusion, Settlement and Default Spreads after Defaults I
How long do sovereigns defaults last? I
I
length = time to settlement
How big are haircuts (creditor losses in defaults)? I
What happens to the face value of debt? I
May go up?
I
What happens to sovereign spreads after default?
I
How long does it take for sovereigns to borrow again?? I
I
Inactivity may not be exclusion if not borrowing is a choice of the borrower.
Are haircuts, post-default spreads and new loans related?
(Neumeyer , Stanford, 2013 )
31 / 39
Introduction
How long do Defaults Last? Source: Reinhart and Rogoff (2009)
(Neumeyer , Stanford, 2013 )
32 / 39
Introduction
How big are haircuts? I
Defaults usually end with a swap of the old debt in default for new debt. I
the new debt may not be sustainable (Greece 2012)
I
We measure creditor losses at the time of the debt restructuring
I
Measuring haircuts (1) HM = 1 −
I I I
Present value of new debt (rt ) Face value old debt
Formula used by financial market participants Correct if acceleration clauses are triggered Some times there are defaults without triggering acceleration clauses (Greece 2012)
(Neumeyer , Stanford, 2013 )
33 / 39
Introduction
How big are haircuts? Cruces Trebesch (2012)
I
Measuring haircuts (2) HSZ = 1 −
I
What rt should be used? I I
A fixed rate, say 10% Yield to maturity on new bonds on the first date of trading after restructuring (Sturzenegger-Settlemeyer) I
I
I
Present value of new debt (rt ) Present value of old debt (rt )
sometimes unavailable
Impute interest rate based on low grade US corporate bond prices, credit ratings and sovereign spreads for each credit rating (Cruces-Trebesch).
Cruces Trebesch (2012) I
180 deals in 68 countries between 1978 and 2010
(Neumeyer , Stanford, 2013 )
34 / 39
Introduction
Sovereign Debt Restructurings with Foreign Private Creditors: 1978-2010 Source: Cruces-Trebesch (2012)
(Neumeyer , Stanford, 2013 )
35 / 39
Introduction
Sovereign Debt Restructurings with Foreign Private Creditors: 1978-2010 Source: Cruces-Trebesch (2012)
(Neumeyer , Stanford, 2013 )
36 / 39
Introduction
Haircuts in Sovereign Debt Restructurings 1978-2010 Source: Cruces-Trebesch (2012)
Circles represent size of default (Neumeyer , Stanford, 2013 )
37 / 39
Introduction
Haircuts and re-entry to credit markets Source: Cruces-Trebesch (2012)
Kaplan-Meier Survival Functions for Duration of Credit Market Inactivity (Neumeyer , Stanford, 2013 )
38 / 39
Introduction
Haircuts and post restructuring spreads Source: Cruces-Trebesch (2012)
(Neumeyer , Stanford, 2013 )
39 / 39