Full text of Working Papers (Federal Reserve Bank of Richmond) : Heterogeneous Borrowers in Quantitative Models of Sovereign Default, Working Paper 07-01R

The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.

Working Paper Series

Heterogeneous Borrowers in
Quantitative Models of Sovereign Default

WP 07-01R

Juan Carlos Hatchondo
Federal Reserve Bank of Richmond
Leonardo Martinez
Federal Reserve Bank of Richmond
Horacio Sapriza
Rutgers University

This paper can be downloaded without charge from:
http://www.richmondfed.org/publications/

Heterogeneous Borrowers in Quantitative Models of
Sovereign Default∗
Juan Carlos Hatchondo†

Leonardo Martinez‡

Horacio Sapriza§

Federal Reserve Bank of Richmond Working Paper No. 07-01R

Abstract
We extend the model used in recent quantitative studies of sovereign default, allowing
policymakers of different types to alternate in power. We show that a default episode may
be triggered by a change in the type of policymaker in office, and that such a default is
likely to occur only if there is enough political stability and if policymakers encounter poor
economic conditions. Under high political stability, political turnover enables the model to
generate a weaker correlation between economic conditions and default decisions, a higher
and more volatile spread, and lower borrowing levels after a default episode.
JEL classification: F34, F41.
Keywords: Sovereign Default, Political Risk, Endogenous Borrowing Constraints, MarkovPerfect Equilibrium.

∗

We thank two anonymous referees, our colleagues at the Federal Reserve Bank of Richmond, and seminar
participants at the 2006 Midwest Theory Meetings, the 2006 Midwest Macroeconomics Meetings, the Institute for
International Economic Studies at the University of Stockholm, the 2006 SED, the 2006 Federal Reserve System
Meeting on International Economics, the 2006 Wegmans Conference, the 2007 North American Winter Meetings
of the Econometric Society, the 2008 Federal Reserve System Macroeconomics Meeting, the Universidad Nacional
de Tucumán, and the Pontificia Universidad Católica de Chile for comments and suggestions. In particular, we
thank Cristina Arellano, Marina Azzimonti Renzo, and Martin Bodenstein for helpful discussions of this paper.
We also thank Elaine Mandaleris-Preddy for editorial comments. All remaining mistakes are our own. Any
opinions expressed in this paper are those of the authors and do not necessarily reflect those of the Federal
Reserve Bank of Richmond or the Federal Reserve System.
†
Federal Reserve Bank of Richmond, e-mail: juancarlos.hatchondo@rich.frb.org.
‡
Federal Reserve Bank of Richmond, e-mail: leonardo.martinez@rich.frb.org.
§
Rutgers University, e-mail: hsapriza@andromeda.rutgers.edu.

1

1

Introduction

Business cycles in small emerging economies differ from those in developed economies. Emerging
economies feature interest rates that are higher, more volatile and countercyclical (interest rates
are usually acyclical in developed economies), and have higher output volatility, higher volatility
of consumption relative to income, and more countercyclical net exports.1 Because the interest
rate is highly volatile and countercyclical, a state-dependent interest rate schedule plays an
important role in any model designed to explain the cyclical behavior of quantities and prices in
emerging economies. Some studies assume an exogenous interest rate schedule.2 Others provide
microfoundations for this schedule based on the risk of default.3 This is the approach followed
by recent quantitative models of sovereign default, based on the framework proposed by Eaton
and Gersovitz (1981).4 The present paper introduces political turnover into the second class of
models.
In addition to pure economic variables, political factors are often considered to play a nontrivial role as determinants of defaults.5 Figure 1 illustrates the behavior of the sovereign spread
in Brazil before and after the 2002 election. This behavior is often mentioned as an example of
the importance of political factors as determinants of default decisions. As discussed by Goretti
(2005), the concerns raised because of left-wing candidate Luiz Inacio “Lula” Da Silva’s declarations in favor of a debt repudiation is the most accepted explanation for the sharp increase in the
country spread preceding the Brazilian election. Pimentel and Murphy (2006) explain that more
recently, the elected president of Ecuador—Rafael Correa—declared his intentions to restructure
the country’s debt, which was linked to a decline in sovereign bond prices. Our paper contributes
to the understanding of how political risk may affect borrowing and default decisions.
1

See, for example, Aguiar and Gopinath (2007), Neumeyer and Perri (2005), and Uribe and Yue (2006).
See, for example, Aguiar and Gopinath (2007), Neumeyer and Perri (2005), Schmitt-Grohé and Uribe (2003),
and Uribe and Yue (2006).
3
Tomz and Wright (2007) document 250 sovereign defaults by 106 countries between 1820 and 2004. Some of
the latest episodes are Russia in 1998, Ecuador in 1999, and Argentina in 2001.
4
See, for example, Aguiar and Gopinath (2006), Arellano (2008), Arellano and Ramanarayanan (2006), Bai
and Zhang (2006), Bi (2006), Cuadra and Sapriza (2006, 2008), Eyigungor (2006), Hatchondo and Martinez
(2008), Hatchondo et al. (2006, 2007b), Lizarazo (2005, 2006), and Yue (2005).
5
See, for example, Balkan (1992), Citron and Nickelsburg (1997), Kohlscheen (2003), Meyersson (2006), Reinhart et al. (2003), Moser (2006), Santiso (2003), Sturzenegger and Zettelmeyer (2006), and Van Rijckeghem and
2

2

Elections: second round

Interest rate spread (in %)

22

17

Argentinian Default
Elections: first round

12

7

2
12/3/2001

4/26/2002

9/17/2002

2/8/2003

7/2/2003

2002 - 2003

Figure 1: Elections and sovereign bond spread in Brazil. Source: JP Morgan (EMBI Global)

We introduce a stylized political process into the framework used in recent quantitative studies
on sovereign default. We study a small open economy that receives a stochastic endowment
stream of a single tradable good. The objective of the policymaker in office is to maximize the
expected present value of future utility flows. As in Cole et al. (1995) and Alfaro and Kanczuk
(2005), we assume that two types of policymakers who assign different weights to future utility
flows alternate in power. One type is more patient than the other. The only financial assets
available are one-period non-contingent bonds. These assets are priced in a competitive market
inhabited by a large number of identical risk-neutral lenders. Lenders have perfect information
about the economy’s endowment and the type of policymaker in office. In each period, the
policymaker in office makes two decisions. First, he decides whether to refuse to pay previously
issued debt. Second, he decides how much to borrow or save. The cost of declaring a default is
that the endowment is reduced by a fixed percentage in the following period.
Two types of defaults may be observed in equilibrium. First, a sufficiently low endowment
realization may trigger a default during the spell of a “patient” or “impatient” type. The second
type is a “political default,” which is triggered by a change in the policymaker in power. We find
Weder (2004).

3

that in equilibrium, a political default may only occur when a patient policymaker is replaced
by an impatient policymaker.
We show that a political default is likely to occur only when there is enough political stability.
The intuition behind the role of political stability is the following. The price received by the
government for the bonds issued today incorporates a discount that mirrors the probability of a
default in the following period (recall that the government can only issue one-period bonds). If a
patient type chooses borrowing levels that would lead an impatient type to default next period,
it has to compensate lenders for this contingency, i.e., for the contingency that an impatient type
becomes the decision maker in the following period. If the probability of this contingency is high
enough (political stability is low), it is too expensive for the patient type to choose borrowing
levels that would lead an impatient type to default. In this scenario, the patient type does not
borrow so heavily and, therefore, a change in the government’s type does not trigger a default.
It is shown that our model displays a non-monotonic relationship between the default probability and political stability. On the one hand, for low levels of political stability (i.e., a high
probability of change in the type in power), an increase in political stability increases the probability that a patient government chooses debt levels that would lead an impatient government
to default and, therefore, it increases the default probability. On the other hand, for high levels
of political stability, a patient government is already choosing debt levels that would lead an
impatient government to default. In this case, the default risk premia charged on current bond
issuances already incorporate the probability that the type in power may change tomorrow.
Therefore, a decrease in that probability (i.e., more political stability) decreases the probability
of default.
We find that even in an economy with high political stability, a default does not occur every
time a patient policymaker is replaced by an impatient policymaker. The reason is that a patient
policymaker chooses high issuance volumes (that would lead an impatient policymaker to default)
only after experiencing a stream of sufficiently low endowment realizations during his tenure.
Furthermore, we show that the occurrence of political defaults enables the model to partially
disentangle default decisions from poor economic conditions. Tomz and Wright (2007) report
that even though most default episodes occur in periods of low output (below the trend), the
4

correlation between default decisions and economic conditions is much weaker than that implied
by existing quantitative models of sovereign default (without political turnover). They conjecture
that introducing “political shocks” to the model may help explain the moderate correlation
between output and default decisions found in the data. Our results support their conjecture.
In addition, we find that a distinctive feature of political defaults is that post-default debt
levels are lower than pre-default levels. The observed difficulties in market access after a default
episode—for example, Gelos et al. (2004) and IMF (2002) discuss evidence of a drainage in capital
flows to countries that defaulted—has been used in recent quantitative work on sovereign default
to motivate the assumption that countries are exogenously excluded from capital markets after
a default. In this paper, we do not assume that a defaulting country is exogenously excluded
from capital markets. However, we show that difficulties in market access after a political default
may be triggered by the political turnover that preceded the default. Recall that in our model,
a political default occurs because a patient government is replaced by an impatient one. After a
political default, investors are still willing to lend to the post-default (impatient) government but
under more stringent conditions: The bond price schedule offered to the post-default government
is below that offered to the pre-default (patient) government. We show that this induces impatient post-default governments to choose debt levels that are lower than the pre-default levels.
Furthermore, and consistent with historical evidence, our model predicts that in the periods that
follow a political default, market access improves after the defaulting government loses power.6
On the contrary, if a default is non-political, the model does not predict significant differences
between post-default and pre-default debt levels. Without political turnover, in general, there is
no reason why the optimal debt level chosen by the post-default government should be significantly different from the pre-default levels. We show that post-default debt levels quickly come
back to pre-default levels after a non-political default.
Our simulations show that introducing political turnover improves the ability of the model
to reproduce the high spread (margin of extra yield over U.S. Treasury bonds) paid by emerging
6

A clear example is discussed by Cole et al. (1995): They explain that “the ability of Reconstruction governments in Florida and Mississippi to borrow after the Civil War suggests that the old creditors could not block
new loans once the states’ reputations had been restored by an observable change in regime.”

5

economies. The average spread observed in the politically stable economy is 6.3%. This spread
level is substantially higher than that obtained when political turnover is shut down. In an
economy where all policymakers are patient, the average spread is 0.3%. In an economy where
all policymakers are impatient, the average spread is 1.5%. Given that the mean spread in the
model mirrors the default probability, the previous finding implies that an economy in which
investor-friendly governments alternate in power with less investor-friendly governments displays
a higher default probability than an economy in which governments are never friendly to investors.
It is the alternation in power of different types of policymakers that is crucial for generating a
higher default probability.
It is also shown that the presence of political turnover narrows the gap between the spread
volatility generated by the model and the spread volatility observed in the data. The model
with political turnover and high political stability generates a standard deviation of the spread
of 0.42%. It should be stressed that this measure of spread volatility derives from samples where
only the patient type is in office, and therefore, it is not trivially driven by having two different
policymakers willing to accept different spread levels. The standard deviation obtained when
the model is simulated without political turnover is less than 0.03%. Thus, introducing political
turnover narrows the gap between the spread volatility in the model and the one in the data.7
Mechanically, the presence of political turnover smoothes out the bond price schedule faced by
the government. This helps the model generate a higher and more volatile spread. The shape of
the bond price function plays a key role in the quantitative performance of models of sovereign
default—see, for example, the discussions in Aguiar and Gopinath (2006) and Hatchondo et al.
(2007b). In addition to improving the spread behavior in the simulations, the model with political
turnover is able to replicate other salient features of the macroeconomic performance of emerging
economies.
7

Emerging economies feature relatively high spread volatility. For instance, the standard deviation of the
spread in Argentina between 1993 and 2001 equals 2.51%.

6

1.1

Related literature

This paper is closely related to Cole et al. (1995) and Alfaro and Kanczuk (2005). They study
models of sovereign default with heterogeneous borrowers in which a default occurs if a “patient” policymaker is replaced by an “impatient” policymaker. In contrast to the present paper,
they assume that there is asymmetric information about the government’s type. In order to
simplify the lenders’ learning process and make their models tractable, these studies limit the
government’s ability to choose its borrowing level. The drawback of this approach is that it is
less suitable to study macroeconomic behavior over the business cycle. These papers also focus
on equilibria in which patient policymakers never default and impatient policymakers always
default. We consider a political process similar to the one used by Cole et al. (1995) and Alfaro
and Kanczuk (2005), but ours is embedded in the setup used in recent quantitative models of
sovereign default. This means that the government does not face any restriction when choosing
the optimal borrowing level. We show that a default is likely to be triggered by political turnover
only if there is enough political stability and patient governments encounter poor economic conditions during their tenure. We also show that in politically stable economies, the presence of
political turnover increases the volatility of the spread paid by patient governments.
The models used in recent quantitative studies of sovereign default share blueprints with the
models used in quantitative studies of household bankruptcy—see, for example, Athreya (2002),
Chatterjee et al. (2007a), Chatterjee et al. (2007b), Li and Sarte (2006), and Livshits et al. (2007).
Within the second literature, Chatterjee et al. (2007b) is the study that is most closely related
to this paper. They too analyze a setup with heterogenous borrowers, but our model differs from
theirs in several dimensions. First, we assume that the type of policymaker in power is public
information. They assume that lenders cannot directly observe a borrower’s type. Second, we
allow policymakers to choose debt levels from a continuum. As in Cole et al. (1995) and Alfaro
and Kanczuk (2005), they restrict the set of savings levels available to borrowers to a discrete
and coarse grid (with only three possible asset positions). Third, we assume that borrowers of
different types alternate in power. They assume that the borrower’s discount factor follows a
stochastic process, i.e., the borrower’s type changes over time. We consider ours to be a better

7

representation of political processes that feature the interaction of groups with different objectives
(while theirs is a better representation of borrowers with stochastic preferences). Fourth, we
focus on a case where the impatient type displays a discount factor significantly higher than
zero. They restrict attention to cases where the impatient type is either completely myopic or
nearly myopic—its subjective discount factor is 0.05. In Section 4.5, we show that our model
delivers implausible implications when the impatient policymaker is fully myopic.
Aguiar and Gopinath (2006) study the role of shocks to the endowment growth rate in the
baseline model of sovereign default. Shocks to the endowment growth rate affect future income
profiles, and thus, the willingness to front-load future resources. It is in this sense that these
shocks are similar to the shocks to the borrower’s discount factor studied by Chatterjee et al.
(2007b). Aguiar and Gopinath (2006) explain that introducing shocks to the endowment growth
rate may smooth out the equilibrium bond price function faced by the borrower, and therefore,
they may help generate a higher spread and a higher spread volatility. Our paper introduces a
different mechanism that also delivers smoother price functions.
Amador (2003) and Cuadra and Sapriza (2008) study sovereign default in a setup in which
different types alternate in power. Their types disagree on the optimal allocation of resources
within each period but do not differ in their willingness to pay. Therefore, they receive the same
treatment from lenders. In their environments, political stability affects the equilibrium spread
mainly through its effect on the weight of future utility flows. In contrast, in our setup the two
types differ in their willingness to pay, and political stability affects equilibrium spreads mainly
through the direct effect it exerts on the government’s borrowing opportunities.
The paper proceeds as follows. Section 2 introduces the model. Section 3 presents the
benchmark parameterization. Section 4 discusses the results. Section 5 concludes and suggests
possible extensions.

2

The model

The basic structure of the model studied in the paper follows previous work that extends the
model of sovereign default presented by Eaton and Gersovitz (1981) and studies its quantitative
8

performance. Among these studies, the closest reference to the present paper is Aguiar and
Gopinath (2006). The advantage of their model is that its simplicity allows us to analyze the
effects of political factors more transparently.

2.1

The environment

There is a single tradable good. The economy receives a stochastic endowment stream of y,
where
log(yt ) = (1 − ρ) µ + ρ log(yt−1 ) + εt ,

(1)

|ρ| < 1, and εt ∼ N (0, σǫ2 ).
The government determines the consumption level in the economy, c. The per-period utility
function is of the CRRA type, i.e.,
u (c) =

c1−σ − 1
.
1−σ

The objective of the policymaker in power is to maximize the present discounted value of future
utility flows. As in Cole et al. (1995) and Alfaro and Kanczuk (2005), we assume that two types
of policymakers who assign different weights to future utility flows alternate in power. Patient
policymakers discount future utility flows at a rate βH . Impatient policymakers discount future
utility flows at a rate βL , where βH > βL . At the end of every period, the type of policymaker in
power changes with probability π. The value of π is not state dependent. This simple political
process has the advantage of making our model a parsimonious generalization of the framework
used in recent quantitative studies and, thus, helps gauge the role played by the new mechanism
introduced in this paper.
The policymaker in power makes two decisions every period. First, he decides whether to pay
back previously issued debt. Second, he chooses how much to borrow or save for the following
period. Defaults imply a total repudiation of government debt. The only financial assets available
are one-period non-contingent bonds. Let b denote the bond position at the beginning of the
period. A negative value of b denotes that the country was an issuer of bonds in the previous
period. Each bond delivers one unit of the good in the next period (provided a default is not

9

declared). The cost of declaring a default is that the endowment in the following period is
reduced by a fraction λ.
There is a continuum of risk-neutral lenders. Each lender can borrow or lend at the riskfree rate r. Lenders have perfect information regarding the economy’s endowment and the
government’s type.
The equilibrium bond price is determined as follows. First, the government announces how
many bonds it wants to issue. Second, lenders offer a price for the bonds. Finally, the government
sells bonds to the lenders who offered the highest price. Let qjd (b′ , y) denote the equilibrium
price offered for each bond when the government issues a total of −b′ bonds. The subindex
j ∈ {H, L} indicates the government’s type. The subindex d is equal to 1 if the government
has defaulted in the current period and is equal to 0 if it has not. The price qjd (b′ , y) does not
depend on default decisions in previous periods because we assume that the effect of a default
on output only lasts for one period and lenders are forward looking.
The following equation summarizes the government’s budget constraint in a given period:
c + qjd (b′ , y) b′ = (1 − hλ) y + (1 − d) b,
where h denotes the credit history. The variable h takes a value of 1 if a default was declared in
the previous period and takes a value of 0 otherwise.
The timing in the model is summarized in Figure 2. At the beginning of the period, there
is a type-j policymaker in power. He observes the endowment realization and decides whether
to pay back previously issued debt. If the decision is to pay back, the policymaker issues an
amount −b′j0 (bt , yt , ht ) of bonds. If the decision is to default, the policymaker issues an amount
−b′j1 (yt , ht ) of bonds. Before the beginning of period t + 1, the type of policymaker in power
changes with probability π and does not change with probability 1 − π. In the period following
a default, the economy suffers an output loss of λ percent.
When deciding whether to default, the policymaker in power compares two continuation
values, Vj1 (y, h) and Vj0 (b, y, h). The former denotes the value function of a policymaker of
type j if a default is declared in the current period, the current endowment realization is y, and
the credit history is h. The second expression denotes the value function of a policymaker of
10

Issue b’j0 (bt , yt , ht )
Type j is replaced
t +1

Pay back bt
Type j continues

t +1

t
β j is in yt is realized
power

Type j is replaced
t +1

Default on bt
Type j continues

t +1
Issue

b’j1 (yt , ht )
Output loss
of λ %

Figure 2: Order of events in period t.

type j if a default is not declared in the current period, the current endowment realization is y,
and the credit history is h.
Let Vj (b, y, h) denote the value function of a policymaker of type j at the beginning of a
period when he is in power. Let Wj (b, y, h) denote the value function of a policymaker of type j
at the beginning of a period when he is not in power. Since the behavior of the two types may
differ, the value functions Vj and Wj do not need to coincide. The optimal borrowing decision of
a government that has defaulted in the current period solves the following dynamic programming
problem:



 u (y (1 − hλ) − qj1 (b′ , y) b′ ) +

h R
i
,
Vj1 (y, h) = max
R
b′  β

′
′
′
′
′
′
π Wj (b , y , 1)F (dy | y) + (1 − π) Vj (b , y , 1)F (dy | y)
j

(2)

where F (·) denotes the cumulative distribution function for y ′. The value function of a policymaker of type j when he has decided to pay back its debt is obtained from the following Bellman
equation:

11



 u (y (1 − hλ) + b − qj0 (b′ , y) b′ ) +

h R
i
. (3)
Vj0 (b, y, h) = max
R
b′  β

′
′
′
′
′
′
π
W
(b
,
y
,
0)F
(dy
|
y)
+
(1
−
π)
V
(b
,
y
,
0)F
(dy
|
y)
j
j
j

The function Vj (b, y, h) is computed as follows:

Vj (b, y, h) = max{Vj1 (y, h) , Vj0 (b, y, h)}.
Let dj (b, y, h) denote the optimal default decision of a policymaker of type j, namely,

 1 if V (y, h) > V (b, y, h)
j1
j0
dj (b, y, h) =
 0 if V (y, h) ≤ V (b, y, h) .
j1

(4)

(5)

j0

The value function of a policymaker of type j when he is not in office depends on the optimal

behavior of the other type, denoted by −j. In the set of states where type −j finds it optimal to
default (d−j (b, y, h) = 1), the value function of a policymaker of type j when he is not in office
is given by





Wj (b, y, h) = u y (1 − hλ) − q−j1 b′−j1 (y, h) , y b′−j1 (y, h) +
(6)
h R
i
R
βj π Vj (b′−j1 (y, h) , y ′, 1)F (dy ′ | y) + (1 − π) Wj (b′−j1 (y, h) , y ′, 1)F (dy ′ | y) ,
where b′−j1 (y, h) denotes the optimal saving behavior of a policymaker of type −j if a default has
been declared in the current period. Similarly, when type −j is in power and finds it optimal to
pay back previously issued debt (when states are such that d−j (b, y, h) = 0), the value function
of a policymaker of type j is given by




Wj (b, y, h) = u y (1 − hλ) + b − q−j0 b′−j0 (b, y, h) , y b′−j0 (b, y, h) +
(7)
h R
i
R
βj π Vj (b′−j0 (b, y, h) , y ′, 0)F (dy ′ | y) + (1 − π) Wj (b′−j0 (b, y, h) , y ′, 0)F (dy ′ | y) ,
where b′−j0 (b, y, h) denotes the optimal savings of a policymaker of type −j in a period where a
default is not declared.
Bond prices satisfy the lenders’ zero profit condition. That is, bond prices are given by
12



Z
Z
1
′
′
′
′
′
′
′
′
qjd (b , y) =
1 − π d−j (b , y , h ) F (dy | y) − (1 − π) dj (b , y , h ) F (dy | y) .
1+r
′

(8)

Bond prices depend on the credit history that the future government will inherit (h′ ), which
is determined by the default decision in the current period (d). Recall that defaulting in the
current period decreases future output and, therefore, affects future default decisions. Prices also
depend on the type of policymaker in power because the latter conveys information about the
probability distribution over future types and, therefore, it affects the probability distribution
over next-period default decisions.

2.2

Equilibrium Concept

We focus on differentiable Markov-perfect equilibria. Krusell and Smith (2003) show that, typically, there is a problem of indeterminacy of Markov-perfect equilibria in an infinite-horizon
economy. In order to avoid this problem, we analyze the equilibrium that arises as the limit of
the finite-horizon economy equilibrium.
Definition 1 A Markov-perfect equilibrium is characterized by
1. a set of value functions Vj (b, y, h), Vj1 (y, h), Vj0 (b, y, h), and Wj (b, y, h) for j = L, H,
2. a set of savings rules b′j0 (b, y, h) and b′j1 (y, h), and a default decision rule dj (b, y, h) for
j = L, H,
3. and a bond price function qjd (b′ , y) for j = L, H,
such that
(a) Vj (b, y, h), Vj1 (y, h), Vj0 (b, y, h), and Wj (b, y, h) satisfy the system of functional equations
(2)-(8);
(b) the default policy dj (b, y, h) and the savings rules b′j0 (b, y, h) and b′j1 (y, h) solve the
dynamic programming problem specified by equations (2)-(8);
(c) the bond price function qjd (b′ , y) satisfies the lenders’ zero profit condition implicit in
equation (8).
13

2.3

Comparison with existing models

The environment presented above departs from the baseline model of sovereign default used in
recent quantitative studies in three dimensions. First, we do not assume that countries can be
exogenously excluded from capital markets after a default episode.8 The observed difficulties in
market access after a default episode has been used by most recent quantitative work on sovereign
default to motivate the assumption that countries are exogenously excluded from capital markets
after a default. However, the exogenous exclusion assumption is controversial on several grounds.
First, it appears to be at odds with the existence of competitive international capital markets
(assumed in this type of model). Wright (2005) discusses how in the past three decades, the
sovereign debt market has become more competitive and explains how an increase in competition (number of creditors) may diminish the creditors’ ability to coordinate; see also Wright
(2002).9 Second, empirical studies suggest that once variables such as the quality of policies and
institutions are used as controls, market access is not significantly influenced by previous default
decisions; see, for example, Eichengreen and Portes (2000), Gelos et al. (2004), and Meyersson
(2006). This suggests that it may very well be that difficulties in market access observed after
a default episode respond to the same factors that triggered the default decision itself. In this
paper, defaults and difficulties in market access that follow a default can be jointly explained by
political turnover.
The second difference between our model and previous studies is that in this paper the
output loss triggered by a default is realized in the period after the default, whereas in previous
studies the output loss lasts for a stochastic number of periods. Our assumption is motivated by
tractability reasons. Hatchondo et al. (2007b) explain that the assumption that the output loss
only takes place in one period allows us to abandon the exclusion assumption without increasing
the dimensionality of the state space. In addition, Hatchondo et al. (2007b) show that the
predictions of the model with exclusion are not affected by whether the output loss occurs in
the period after the default or for a stochastic number of periods. The output-loss assumption
8
Hatchondo et al. (2007b) show that the model delivers a slightly higher equilibrium default probability without
exclusion.
9
Athreya and Janicki (2006) and Cole et al. (1995) raise a similar point.

14

intends to capture the disruptions in economic activity caused by a default decision. It has
been argued that a government default decreases private financing and, thus, it reduces output.
Using micro-level data, Arteta and Hale (2008) find that sovereign debt crises are systematically
accompanied by a large decline in foreign credit to domestic private firms. This may be the
case because a sovereign default may signal to investors a higher risk of expropriation or bad
economic conditions, and therefore, it may reduce firms’ net worth and their ability to borrow;
see Sandleris (2006) and the references therein. IMF (2002), Kumhof (2004), and Kumhof and
Tanner (2005) discuss how financial crises that lead to severe recessions follow sovereign defaults.
Similarly, Kaminsky and Reinhart (1999) show that debt devaluations in developing countries
tend to cause banking problems. Kobayashi (2006) presents a model in which a shock that
disturbs the payments system causes a decrease in aggregate productivity. Mendoza and Yue
(2007) study the link between default risk and output.
Finally, and more importantly, we allow for the possibility that the composition of the government or the distribution of power among government officials changes over time. This possibility
is embedded in the assumption that policymakers with different time preferences alternate in
power. It should be stressed that the change of the type in power does not need to be caused
by an election. For instance, as discussed by Moser (2006) and Santiso (2003), the turnover of
finance ministers is often linked to changes in desired policies.

3

Parameterization

The model is solved numerically using value function iteration and interpolation.10 Table 1
presents the parameter values of our benchmark parameterization. We assume a coefficient of
relative risk aversion of 2. A period in the model refers to a quarter. The risk-free interest
rate is set equal to 1%. The parameter values that govern the endowment process are chosen
so as to mimic the behavior of GDP in Argentina from the fourth quarter of 1993 to the third
10

The model is solved using Chebychev collocation. The algorithm finds four value functions, conditional on
the type being in office or not, and conditional on a default having been declared today or not. We use 15
polynomials on the asset space and 10 on the endowment level. The results do not change significantly when
more polynomials are used. We also verified that the conclusions presented in the paper are not altered when we
solve the model using discrete state space.

15

Risk aversion

σ

2

Interest rate

r

1%

Output autocorrelation coefficient

ρ

0.9

Standard deviation of innovations

σǫ

2.7%

Mean log output

µ

(-1/2)σǫ2

Output loss

λ

8.3%

Higher discount factor

βH

0.9

Lower discount factor

βL

0.6

Table 1: Parameter values.

quarter of 2001. In Section 4.4, we compare our simulation results with macroeconomic data
from Argentina during this period. The parameterization of the output process would essentially
not change if a longer sample period were considered. Aguiar and Gopinath (2006) use a similar
parameterization of the endowment process to mimic the behavior of GDP in Argentina between
1983 and 2000.
The value of λ is taken from Hatchondo et al. (2007b). When λ = 8.3%, the present value
of the cost of defaulting is similar to the cost implied by the process of output loss assumed in
Aguiar and Gopinath (2006) (a loss of 2% per period for an average duration of 10 periods).
As in previous studies, high impatience is necessary to generate default in equilibrium. For
instance, Aguiar and Gopinath (2006) choose a discount factor of 0.8. Here, we choose a higher
discount factor for the patient policymaker and a lower discount factor for the impatient policymaker.11
11

The difference between the two discount factors could be calibrated to match differences in the actions
of different types of government. For instance, as explained in Section 4, the differences between the average
borrowing level chosen by patient and impatient governments in the model is a direct consequence of the assumed
difference between discount factors. However, the presence of other factors that affect debt levels makes it difficult
to identify the extent to which differences in debt levels result from differences in political circumstances. For
example, differences in debt levels chosen by Argentina before and after its default are not only explained by
differences in political factors but also by the large devaluation that occurred in Argentina, as well as by changes
in other economic conditions. Section 4.5 discusses how our results depend on the assumed difference between
the two discount factors.

16

The degree of political stability is left as a free parameter. We consider two economies that
are identical except for the degree of political stability. In the “stable” economy, the probability
that the current government type is replaced (π) is set equal to 1.5%. This implies an average
tenure in office of 16 years independently of the type. In the “unstable” economy, the value of π
is set equal to 2.5%. The latter implies an average tenure in office of 10 years. We also compare
the equilibrium in these economies with the equilibrium in economies without political turnover.

4

Results

We show that political defaults may occur in equilibrium and are likely to occur only if there is
enough political stability and if patient policymakers encounter poor economic conditions during
their tenure. Furthermore, it is shown that when the degree of political stability is high, the
presence of political turnover enables the model to generate (i) the moderate correlation between
economic conditions and default decisions documented by Tomz and Wright (2007), (ii) lower
borrowing levels after a default episode, and (iii) a higher and more volatile spread, even when
we focus on samples in which only the patient type is in office.

4.1

Political risk, default risk, and political stability

Can changes in political circumstances trigger sovereign defaults? Cole et al. (1995) and Alfaro
and Kanczuk (2005) present models in which this is the case. We add to their insight by showing
that a sovereign default is likely to be triggered by political turnover only if there is enough
political stability in the economy. Furthermore, even in an economy with high political stability,
a political default occurs only when patient policymakers encounter sufficiently poor economic
conditions during their tenure.
In order to explain how political stability may affect the relationship between political risk
and default risk, we solve the model using the parameterization in Table 1.12 We simulate the
12

Bilson et al. (2002) define political risk as “the risk that arises from the potential actions of governments and
other influential domestic forces, which threaten expected returns on investment.” In our environment, default is
the government’s action that affects the return obtained by lenders and, for a given debt level, political risk is
low (high) when a patient (impatient) policymaker is in power.

17

model for 750,000 periods (500 samples of 1,500 observations each).
We find that in the economy with high political stability (π = 1.5%), 96% of the changes
from a patient to an impatient government trigger a default. The remaining 4% correspond to
situations in which patient governments do not encounter poor economic conditions during their
tenure. The result is different in the unstable economy. When π = 2.5%, a change in type from
βH to βL triggers a default only 4% of the time.
In what follows, we explain the link between the degree of political stability and the frequency
with which political defaults occur in equilibrium. We show that in a politically stable economy
it is more likely that a patient policymaker chooses debt levels that would lead an impatient
policymaker to default. First, we explain that impatient policymakers choose to default on lower
debt levels than do patient policymakers (Figures 3 and 4). Second, we show how differences in
the default rules of the two types and the presence of political turnover affect the bond price menu
faced by patient policymakers (Figure 5). Third, we show how the shape of the objective function
of the patient policymaker depends on the bond price menu, which helps in understanding how
the degree of political stability affects the optimal borrowing decision of a patient type (Figure 6).
We illustrate how a patient policymaker may find it optimal to choose debt levels that can trigger
a political default in the politically stable economy, but may find those debt levels suboptimal in
the politically unstable economy. Finally, we show that the patient type only chooses a borrowing
level that would lead an impatient type to default if the endowment realization is sufficiently low
(Figure 7).
In order to illustrate the mechanics of the model using two-dimensional charts, it is necessary
to hold the values of some variables fixed. For that reason, Figures 3-7 rely on specific assumptions
about the current default decision, the previous period default decision, the current endowment
level, and/or the debt level at the beginning of the period. The behavior described in the figures
would not be substantially different under alternative assumptions about those variables.
We first show that there are combinations of debt levels and endowment realizations at which
an impatient type defaults and a patient type does not default. Furthermore, for combinations of
debt levels and endowment realizations such that a patient government would choose to default,
an impatient government would also choose to default. Figure 3 shows the value functions of
18

patient and impatient policymakers when they have defaulted and when they have not defaulted.
The value functions under default do not depend on b because a default resets the value of initial
debt to zero. The increasing curves illustrate that the expected utility of both policymakers
decreases with debt when they do not default. The scales of the left and right vertical axes of
Figure 3 were chosen so that the lines describing the value function under default of the two types
overlap. The fact that the line describing the value function under no default of a patient type
lies above the line describing the value function under no default of an impatient type implies
that there are debt levels (between 0.058 and 0.079) at which an impatient type defaults and a
patient type does not default. This is also illustrated in Figure 4, which describes the optimal
default decision of both types of policymakers.

−10

−2.5

Value functions for β = 0.9

Value functions for β = 0.6

No default, β = 0.6
Default, β = 0.6

−10.1

−2.6
−0.12

No default, β = 0.9
Default, β = 0.9
−0.1

−0.08

−0.06
b

−0.04

−0.02

0

Figure 3: Value functions in a stable economy. The figure corresponds to the case where the government
inherits a good credit history (h = 0), and the endowment realization coincides with the unconditional
mean of the distribution.

The optimal default strategies affect the shape of the bond price function faced by the policymaker in office. Figure 5 describes how the bond price menu faced by a patient type differs
across different economies: the politically stable economy, the politically unstable economy, and
the economy without political turnover. The price received by the government for the bonds
issued today incorporates a discount that mirrors the probability of a default in the following
19

1.2
1.15

Output

1.1
1.05
1
0.95
0.9
0.85
−0.12

−0.1

−0.08

−0.06
−0.04
Bond position

−0.02

0

Figure 4: Default regions in a stable economy. The graph corresponds to the case where the government
inherits a good credit history (h = 0). The dark area shows combinations of bond positions and
endowment realizations at which an impatient type defaults and a patient type does not default. The
gray area shows combinations of bond positions and endowment realizations at which both types default.
The white area shows combinations of bond positions and endowment realizations at which neither the
patient nor the impatient type defaults.

period. This explains why the price functions in economies with political turnover display three
steps. The first step corresponds to “low” issuance volumes. At these volumes, the debt issued
is sufficiently low that the government will almost surely pay it back in the following period,
regardless of the type in power. When the issuance level is within this range of values, investors
charge the risk-free rate (bond prices are high). The second step corresponds to “intermediate”
issuance levels. In this range of issuance values, the debt issued is such that an impatient policymaker would default in the next period if he becomes the decision maker, while a patient
policymaker would pay the debt back if he remains in office. For these issuance volumes, the
default premia (spread) charged by lenders coincides with the probability of a change in the government type. Finally, the third step corresponds to “high” issuance volumes. At these volumes,
investors realize that the government will almost surely default tomorrow, regardless of the type
in power. Therefore, they offer a zero price for the bonds issued today (this cannot be seen in
Figure 5 because of the scale of the vertical axis).

20

Figure 5 shows that the higher the degree of political stability (i.e., the lower π), the better
the price at which a patient government can issue “intermediate” borrowing levels. As explained
above, when a patient type chooses these borrowing levels, he compensates lenders for the contingency that an impatient type (who would default on “intermediate” debt levels) becomes the
decision maker in the following period. Thus, if the probability of this contingency decreases
(i.e., if π becomes lower), bond prices at “intermediate” borrowing levels increase. In addition,
Figure 5 shows that in the economy in which there is no alternation in power and all governments
are patient, the bond price menu does not feature the intermediate step.
1

π = 1.5%

0.995
1
1+r

0.99

Bond price

0.985
0.98
1−π
1+r

0.975
0.97

1 − π̄
1+r

0.965
0.96
π = 2.5%
No turnover

0.955

0.95
−0.08 −0.075 −0.07 −0.065 −0.06 −0.055 −0.05 −0.045 −0.04
b’

Figure 5: Bond price faced by a patient policymaker in the economies with high and low political stability and in economies without political turnover. The figure considers the case in which the endowment
realization coincides with the unconditional mean of the distribution and in which the government has
not defaulted today.

The shape of the government’s objective function is closely related to the shape of the bond
price function. Figure 6 shows the objective function of a patient policymaker as a function of
the current issuance volume. Typically, the objective function is not globally concave.13 For
“low” borrowing levels, the objective function is increasing with respect to the issuance volume
(−b′ ). For issuance values in the transition from the first to the second step of the bond price
13
We use a nonlinear optimization routine to find the optimal borrowing level. The initial guess is found using
a global search procedure.

21

function, the objective function decreases with the issuance level. This accounts for the right local
maximum. Once the second step of the bond price function is reached, the objective function
becomes increasing again. Finally, as the issuance volume moves from the second to the third step
of the bond price function, the objective function declines. This explains the left local maximum.
The objective function remains low after the third step has been reached, so the optimal issuance
volume coincides with one of the two local maxima. The increasing portions of the objective
function are mostly explained by the difference between the rate at which future utility flows are
discounted and the interest rate. The decreasing portions of the objective function are mostly
explained by the decline in the bond price implied by an increase in the issuance volume.

Objective function when π = 2.5%

Objective function when π = 1.5%

−10.03

−10.04
π = 2.5%

−10.04

−0.08

π = 1.5%

−0.07

−0.06

−0.05

−0.04
b’

−0.03

−0.02

−0.01

−10.05
0

Figure 6: Objective function of a patient policymaker in the stable and unstable economy. The graph
considers the case where the government has decided not to default, the initial endowment realization
coincides with the unconditional mean of the endowment process, the government inherits a good credit
history, and the initial bond position takes a value of -0.0715, which is within the range of values
observed in the simulations of the stable economy.

Figure 6 illustrates how the optimal issuance volume may be at “intermediate” borrowing
levels (left local maximum) in the politically stable economy and at “low” borrowing levels
(right local maximum) in the politically unstable economy. Recall that as the degree of political
instability decreases (π is lower), the patient type pays a lower spread at intermediate borrowing
levels, which makes those levels more attractive. This is the mechanism by which a politically
22

stable economy induces patient policymakers to be willing to choose intermediate borrowing
levels and makes political defaults more likely.
0.08
b = −0.073
0.07

Borrowing (−b’)

0.06
0.05
0.04
0.03
0.02
0.01
0

0.85

0.9

0.95

1
1.05
Current endowment

1.1

1.15

1.2

Figure 7: Optimal bond issuance of a patient type that inherits a good credit history and a debt level
of −0.073 in an economy with π = 1.5%. The issuance decision is computed after the optimal default
decision has been made.

Finally, we show that even in the politically stable economy, a political default occurs only
when patient policymakers encounter relatively poor economic conditions. Figure 7 shows that
patient policymakers choose “low” borrowing levels if the current endowment realization is sufficiently high. Thus, political turnover does not trigger a default when it occurs after a patient
policymaker encounters good economic conditions. Figure 7 also shows that a patient policymaker only chooses “intermediate” borrowing levels after intermediate current endowment
realizations. At the left end of the graph, a patient type has low borrowing needs given that it
has no debt to roll over: the endowment level is so low that the patient type finds it optimal to
default and reset the debt level to zero.

4.2

The correlation between default and output

Using a historical data set with 169 sovereign default episodes, Tomz and Wright (2007) report
a weak correlation between economic conditions and default decisions. They find that 38% of
23

default episodes in their sample occurred in years when the output level in the defaulting country
was above the trend value. They argue that the baseline model of sovereign default (without
political turnover) is ill-suited to replicate this weak correlation. In order to illustrate this, we
shut down political turnover and simulate the economy with patient governments only. We find
that only 3% of default episodes occur in periods where the output level is above its long-run
mean. The fraction increases to only 7% in an economy where all policymakers are impatient.
Tomz and Wright (2007) suggest that the inability of the baseline model to generate the weak
correlation observed in the data may be a result of the lack of political turnover in the model. We
show that introducing political turnover may weaken the correlation between default and output
generated by the model but only when there is enough political stability. In our simulations of
the economy with high political stability, 38% of the default episodes take place in periods where
output is above its long-run mean. This percentage decreases to 7% in the politically unstable
economy.

4.3

Distinctive features of debt and spread behavior around political
defaults

In this section, we describe properties of political defaults in the politically stable economy (where
political defaults are likely to occur). We find two distinctive features of political defaults in our
simulations of this economy.
First, if a default is political, post-default issuance levels are lower than pre-default issuance
levels. On the contrary, if a default is non-political, the baseline model of sovereign default does
not predict significant differences between post-default and pre-default issuance levels. This is
illustrated in Figure 8, which shows that issuance levels before and after a non-political default
are similar (in default periods, though, debt issuances are low because after defaulting, the
government does not have any debt to roll over). In general, in the absence of political turnover,
there is no reason why the optimal debt level chosen by the government after a default should
differ from pre-default levels.
Around a political default, post-default issuance levels are lower than pre-default issuance
24

0.09

Gross issuance / output

0.08
0.07
0.06
0.05
0.04
0.03
0.02
−4

Political default
Non−political default
−3

−2

−1

0

1

2

3

4

Period

Figure 8: Mean issuance level before and after a default episode in the politically stable economy.
Non-political defaults are only considered if a patient policymaker is in office. Period 0 is the default
period.

levels, because impatient policymakers choose lower issuance levels than patient policymakers—
recall that a political default occurs when a patient policymaker is replaced by an impatient
policymaker. Figure 9 helps one understand why impatient policymakers choose lower issuance
volumes. It illustrates how bond price schedules faced by impatient policymakers display three
steps as do the schedules faced by patient policymakers. If the current policymaker is impatient, it
is highly likely that the next-period policymaker will also be impatient. Given that the impatient
type defaults on “intermediate” debt levels, lenders demand a high discount for “intermediate”
borrowing levels, which explains why the intermediate step in Figure 9 is close to zero. The fact
that the bond price for intermediate issuance levels is significantly lower than the bond price for
“low” issuance levels leads impatient governments to choose “low” issuance levels. As explained
before, in politically stable economies, patient governments often choose intermediate issuance
levels. That is, our theory predicts that even though impatient governments assign more weight
to current utility flows, they may decide to borrow less than patient governments.
The second distinctive feature of political defaults is that post-default equilibrium spreads are
lower than pre-default spreads. Before a political default, patient governments choose “interme-

25

1

Bond price

0.8

0.6

0.4

0.2

π = 2.5%

π = 1.5%
0
−0.08 −0.075 −0.07 −0.065 −0.06 −0.055 −0.05 −0.045 −0.04
b’

Figure 9: Bond price faced by an impatient policymaker in economies with high and low political stability. The graph considers the case in which the endowment realization coincides with the unconditional
mean of the distribution and in which the government has defaulted today.

diate” issuance levels and, therefore, they pay “intermediate” spreads. After a political default,
impatient governments choose “low” issuance levels and, therefore, they pay “low” spreads. That
is, our theory predicts that more investor-friendly governments pay higher spreads in equilibrium
than less investor-friendly governments. On the contrary, if a default is non-political, post-default
spreads are similar to pre-default spreads. In our model, if there is no political turnover, lenders
have no reason to offer a different treatment to a country that has defaulted in the past and,
in general, there is no reason for the equilibrium government’s behavior after the default to be
significantly different from its pre-default behavior. Figure 10 shows that equilibrium spreads
before and after a default are consistent with the equilibrium issuance levels presented in Figure 8. Subsection 4.6 argues first that there is evidence that suggests that the 2001 default in
Argentina was influenced by political factors, and second, that the behavior of debt and spread
around a political default in our model are not inconsistent with what was observed in Argentina
before and after the 2001 default.

26

0.07
0.06
0.05

Spread

0.04
0.03
0.02
0.01
0
−0.01
−4

Political default
Non−political default
−3

−2

−1

0
Period

1

2

3

4

Figure 10: Mean spread before and after a default episode in the politically stable economy. Nonpolitical defaults are only considered if a patient policymaker is in office. Period 0 is the default period.

4.4

Political turnover and business cycles statistics

Recent quantitative studies of sovereign default focus on the ability of the model to replicate
features of the macroeconomic behavior of emerging economies before a default. In this section,
we discuss how the predictions of the baseline model are affected when political turnover is
introduced. In order to illustrate the role of political turnover, we simulate economies in which
the type of policymaker in power does not change (π = 0), and we compare these simulations
with those of economies with political turnover.
We extract from the simulations samples of 32 consecutive periods because we want to compare data generated by the model with Argentine data from the fourth quarter of 1993 to the
third quarter of 2001. We identify this period as a time of relatively low political risk (see Section
4.6), and we compare data from this period with data from simulation samples in which only
patient policymakers are in office. We also present simulation results for an economy in which
all policymakers are impatient, and for samples with political turnover. Results do not change
substantially if we focus on longer samples.
If t is the first period of one of our samples, we check that a default does not occur between

27

periods t − 2 and t + 31. This makes our samples free from the effects of within-sample default episodes and, therefore, comparable with the Argentine sample (the Argentine government
honored its debt between the fourth quarter of 1993 and the third quarter of 2001).
From the simulations of economies without political turnover, we extract the first 500 samples
before a default. From the simulations of economies with political turnover, we present two sets of
results. First, we extract the first 500 samples that satisfy the following criteria: (i) the beginning
of the sample is determined by a change of the policymaker in power, from an impatient type
to a patient type, (ii) the patient type remains in office for 32 consecutive periods, and (iii) in
the first period after the end of the sample, an impatient policymaker gains power and declares
a default. We also present unconditional results using the first 500 samples before any type of
default episode and without imposing any restriction on the type of policymaker in power.
Table 2 reports our findings. The second column of the table presents the moments in
the data.14 The reported moments are chosen so as to evaluate the ability of the model to
replicate distinctive business cycle properties of emerging economies. Relative to developed
economies, emerging economies feature interest rates that are high, volatile, and countercyclical;
high volatility of consumption relative to income; and more countercyclical net exports. The
trade balance (T B) is expressed as a fraction of output (Y ), and the interest rate spread (Rs )
is expressed in annual terms. The logarithm of income and consumption are denoted by y
and c, respectively. All series are HP filtered with a smoothing parameter of 1, 600. Standard
deviations are denoted by σ and are reported in percentage terms; correlations are denoted by
ρ. The third column of the table reports the average business cycle moments in an economy
inhabited by patient policymakers only. The fourth column reports the average moments in
an economy inhabited by impatient policymakers only. The fifth and sixth columns report the
average business cycle moments when π = 1.5% and π = 2.5%, respectively, before a political
default, when there is no political turnover within the samples. The last column presents the
unconditional moments when π = 1.5%.
14

The output, consumption, and trade balance data are obtained from the finance ministry of Argentina. The
spread series before the first quarter of 1998 is obtained from Neumeyer and Perri (2005), and from the EMBI
Global after that.

28

σ(y)

Data βH only βL only π = 1.5%
3.17
3.14
3.09
3.08

π = 2.5% Unconditional
3.24
3.01

σ(c)

2.98

3.21

3.15

3.16

3.29

3.08

σ (T B/Y )

1.35

0.16

0.14

0.34

0.19

0.22

σ (Rs )

2.51

0.03

0.02

0.42

0.61

0.44

ρ (c, y)

0.97

1.00

1.00

0.99

0.99

1.00

ρ (T B/Y, y)

-0.69

-0.45

-0.45

-0.19

-0.35

-0.35

ρ (Rs , y)

-0.65

-0.94

-0.95

-0.69

-0.79

-0.68

ρ (Rs , T B/Y )

0.56

0.71

0.70

0.50

0.43

0.38

E(Rs )

7.4

0.3

1.5

6.3

0.8

5.3

Table 2: Business cycle statistics. The second column is computed using data from Argentina from
1993 to 2001. The remaining columns summarize the simulation results: Each cell reports the mean
value of the moment using 500 samples of 32 periods.

Note that the data moments reported in Table 2 display the same qualitative features that
are obtained using a longer sample period or using data from other emerging markets; see, for
example, Aguiar and Gopinath (2007), Neumeyer and Perri (2005), and Uribe and Yue (2006).
The exception is that in the table, the reported volatility of consumption is slightly lower than
the volatility of income.
Table 2 shows that only when there is enough political stability, the presence of political
turnover improves the ability of the model to generate high spreads. The average spread observed
in the politically stable economy is 6.3%, compared with 0.8% in the unstable economy. As
explained above, in the politically unstable economy, patient governments frequently choose
debt levels that are too low to lead impatient governments to default. Therefore, lenders ask for
a small spread when they purchase bonds.
The average spread delivered by the model in the stable economy is substantially higher than
the one obtained when political turnover is shut down. In the economy where all policymakers
are patient, the average spread is 0.3%. In the economy where all policymakers are impatient,
the average spread is 1.5%. Since the mean spread in the model mirrors the default probability,
29

an economy in which investor-friendly governments alternate in power with less investor-friendly
governments has a higher default probability than one where governments are never friendly
to investors. It is the presence of political turnover—and not the low discount factor of the
impatient type—that is crucial for generating a higher default probability.
Mechanically, political turnover enables the model to generate a higher average spread because
it smoothes out the bond price function, i.e., political turnover makes the bond price less sensitive
to the issuance volume. With a smoother price function, a given decrease in the bond price
allows for a larger increase in the issuance level. This makes the choice of lower bond prices more
attractive and induces the government to pay a higher spread.15
Table 2 illustrates the non-monotonic relationship between the mean spread and the degree
of political stability. The third column can be interpreted as an extreme case where π = 0 and
the patient policymaker is in power in the first period. Starting from this case, a decrease in the
level of political stability to π = 1.5% increases the mean spread from 0.3% to 6.3%. A further
decrease in the level of political stability to π = 2.5% decreases the mean spread from 6.3% to
0.8%.
There are two forces through which changes in π affect the mean spread. First, an increase
in π increases the spread paid at intermediate borrowing levels. Second, an increase in π makes
intermediate borrowing levels less attractive and, therefore, less frequent in equilibrium. The
first effect is dominant when π is sufficiently low, i.e., when the discount in the bond price at
intermediate borrowing levels is sufficiently low. For a low π, the government chooses intermediate borrowing levels often, and small changes in π have a direct effect on the spread paid at
these borrowing levels. The second effect is dominant when π is sufficiently high and, therefore,
intermediate borrowing levels are chosen less frequently.
The previous discussion illustrates that in our model, the degree of political stability affects
spreads mainly by changing the shape of the bond price. This is not the case in the environments
studied by Amador (2003) and Cuadra and Sapriza (2008). In their models, political stability
affects spreads by changing the value of future utility flows.
15

In the economy with low political stability, a smoother price function does not increase the mean spread
because intermediate debt levels are rarely observed in equilibrium.

30

In addition, Table 2 shows that the presence of political turnover enables the model to generate
higher spread volatility. In the economy with high political stability, the standard deviation
of the spread is 0.42%. In contrast, we find a standard deviation of 0.03% in the economies
without political turnover. The higher spread volatility is also a consequence of a smoother bond
price function. It should be stressed that this higher volatility is not driven by the presence of
policymakers of different types (who are willing to pay different spread levels) in the simulation
samples. Recall that this higher standard deviation is computed using samples in which only
patient policymakers are in power.
The last column of Table 2 shows that, in the politically stable economy, the unconditional
business cycle moments are similar to the ones obtained when only pre-political-default samples
without within-sample political turnover are considered (fifth column). This is the case because
most defaults in the stable economy are political defaults. In addition, the table shows that the
introduction of political turnover does not significantly affect other business cycle moments.16

4.5

Equilibrium with a completely myopic type

In this section, we study the stable (π = 1.5%) and unstable (π = 2.5%) economies assuming
the impatient type is completely myopic (βL = 0). This corresponds to the case studied in Cole
et al. (1995), and Alfaro and Kanczuk (2005). Chatterjee et al. (2007b) consider cases where the
impatient type is fully myopic and nearly myopic (βL = 0.05).
We find that when the impatient type is fully myopic, our model generates predictions that are
strongly counterfactual. Spreads paid by impatient governments are too high. For instance, the
maximum annualized spread paid by an impatient government in the politically stable (unstable)
economy is 2, 062, 878, 145% (267, 244, 563%). Even though completely myopic types always
16
As in the benchmark case without heterogeneity, one may think that the debt levels generated by the model
are low (between 5% and 7% of quarterly output). In part, this is because we do not assume that a defaulting
economy is excluded from capital markets; see Hatchondo et al. (2007b). Furthermore, simplifying assumptions
limit the ability of the model to generate higher debt levels. For instance, it is assumed that governments cannot
save and borrow at the same time, that defaults imply a total repudiation of debt, and that all the debt is held
by foreigners. Moreover, there are costs of defaulting that are not present in the model; see Hatchondo et al.
(2007a) for a discussion of defaulting cost. Since our model builds on the baseline framework that has been used
in recent quantitative studies and also shares the same parameterization, it generates debt levels comparable to
the ones in those studies.

31

default, a completely myopic government may be offered a positive bond price, because with
positive probability, the next-period default decision is going to be made by a patient type—who
would choose to honor its debt. Since the impatient type only cares about current consumption,
it is optimal for it to issue bonds even when these bonds are sold at a large discount.
We also find that when βL = 0, patient governments often prefer “intermediate” issuance
levels over “low” issuance levels even in the politically unstable economy. This is the case
because the extra borrowing that can be obtained by paying the intermediate spread level is
large—i.e., the length of the intermediate step in the bond price function faced by patient types
is large. Thus, it pays off to choose intermediate spread levels as this enables patient governments
to increase the borrowing level by a relatively large amount. This illustrates how, in general,
the larger the difference between βH and βL (the more politically polarized is the economy), the
more likely patient governments choose debt levels that would lead impatient governments to
default. When βL = 0, for any borrowing level, a patient government would have to pay at least
the intermediate spread level.

4.6

The 2001 default in Argentina

In this subsection, we present a brief case study of the 2001 default episode in Argentina. We
first argue that political circumstances around this default episode are similar to the ones in the
political defaults of our model. Then, we show that the behavior of the spread and the debt level
in Argentina around the default episode are not inconsistent with the predictions of our theory.
Let us discuss first why the default episode in Argentina can be interpreted as a political
default. Recall that we defined political defaults as episodes triggered by political turnover.
Political turnover preceded the default in Argentina. After president De La Rua resigned on
December 20, 2001, Congress appointed Rodriguez Saa as the interim president on December
23, 2001. As discussed by Sturzenegger and Zettelmeyer (2006), the next day, Rodriguez Saa
announced the suspension of all payments on debt instruments.
In addition, in our model, a political default occurs when an investor-friendly government is
replaced by a government that is less friendly to investors. Everything else being equal, political

32

risk for investors (default risk) is higher (lower) when impatient (patient) policymakers are in
power. Thus, our model predicts that—everything else being equal—pre-political-default levels
of political risk for investors are lower than post-political-default levels. In order to gauge whether
such a change in political risk for investors occurred in Argentina, we look at the composite index
of political risk for investors constructed by the International Country Risk Guide (ICRG). This
index of political risk is one of the three components of the overall risk index constructed by the
ICRG (the other two indexes are the financial risk index and the economic risk index). The index
of political risk is supposed to reflect political risk only, independently from economic risk and
financial risk (which are captured by the other two indexes).17 Thus, the index of political risk
does not necessarily mirror default risk. In fact, we will explain that default risk in Argentina
was presumably higher (the spread was higher) when political risk was lower.
We compare the mean value of the index of political risk in the eight years prior to the default
date with the mean value between the default date and June 2006. The index of political risk
for investors is such that a higher value indicates less political risk. The pre-default mean value
of the index is 74.4, and the post-default value is 64.3.18 If we identify the higher (lower) values
of the index with an “investor-friendly” (less friendly) type being in power, the variations of the
index of political risk in Argentina suggest that the default episode in Argentina is characterized
by the type of regime change our model indicates is necessary for political defaults to occur.
That is, in Argentina, post-default governments were considered more risky for investors than
pre-default governments.19
17

The ICRG indexes are widely used in empirical studies. See, for example, Erb et al. (1996), Erb et al. (2003),
Bilson et al. (2002), Bekaert and Harvey (1996), Bekaert et al. (2007a), and Bekaert et al. (2007b). These studies
provide a more thorough discussion of the indexes.
18
The higher political risk after the default does not seem to be a mere consequence of the default itself. Other
countries do not exhibit such large increase in political risk after defaulting. For instance, when we repeat the
same exercise conducted for Argentina for the recent defaults in Ecuador (1999), Pakistan (1999), Russia (1998),
and Uruguay (2003), we find that (i) in Ecuador, the pre-default mean value of the index of political risk for
investors is 60.7, and the post-default value is 57.0; (ii) in Pakistan, the pre-default value is 51.4, and the postdefault value is 47.3; (iii) in Russia, the pre-default value is 56.7 (considering data since April 1992), and the
post-default value is 60.7; and (iv) in Uruguay, the pre-default value is 72.1, and the post-default value is 72.4.
19
Of course, higher values of the index of political risk cannot be mapped exactly with more patient governments
in our model. The index is a noisy measure of the government type. For instance, while our stylized model assumes
that the government stability parameter (π) is constant, the index may be affected by changes in government
stability. However, the decrease in the average value of the index after the default is indicative of a regime change.

33

Furthermore, Argentina exhibits a relatively high degree of political stability, a feature that
this paper identifies as necessary for the occurrence of political defaults. In the ten years prior
to the default episode, the index of political risk is higher than the average post-default index.
This indicates that it is plausible that investor-friendly governments in Argentina may have been
perceived as stable during the 1990s.
As explained in Section 4.3, our theory predicts that a distinctive feature of political defaults
is that post-default debt levels are lower than pre-default levels. This is consistent with what
we observed in Argentina. The per capita public external debt in Argentina expressed in 2006
U.S. dollars ranged from $2084 to $2801 between March 1994 and December 2001. It was $2659
in December 2001. After the default, it was $1311 in March 2006 (without considering $121 of
arrears).
Our theory predicts that another distinctive feature of political defaults is that post-default
spreads are lower than pre-default spreads. This seems consistent with what we observed in
Argentina. As reported in Table 2, the average annualized spread for Argentina between the
fourth quarter of 1993 and the third quoter of 2001 is 7.4%. The average spread for Argentina
between the third quarter of 2005 and the second quarter of 2007 is 3.4%.20 Thus, as our theory
predicts, governments that were considered to be “less friendly” to investors paid lower spreads
in Argentina.

5

Conclusions and extensions

This paper introduces a stylized political process into the framework used in recent quantitative studies of sovereign default. We show that a default episode is likely to be triggered by
political turnover only if there is enough political stability in the economy and investor-friendly
governments encounter sufficiently poor economic conditions during their tenure. The presence
of political turnover enables the model to generate (i) a moderate correlation between economic
conditions and default decisions, (ii) lower borrowing levels after a default episode, and (iii) a
20

We do not consider the spread between the default episode in 2001 and the debt exchange in 2005, because
during that period, the spread is computed using the price of defaulted bonds and, therefore, does not reflect the
interest rate at which Argentina borrowed over that period.

34

higher and more volatile spread.
Our stylized political process simplifies the analysis and enables us to provide a transparent
description of the mechanism behind our results. Allowing for a richer political process could
improve the performance of the model. For example, if the probability of political turnover
reacts to changes in output or consumption, the model may be able to generate a higher spread
volatility.

35

References
Aguiar, M. and G. Gopinath, “Defaultable debt, interest rates and the current account,” Journal
of International Economics 69 (2006), 64–83.
Aguiar, M. and G. Gopinath, “Emerging markets business cycles: the cycle is the trend,” Journal
of Political Economy 115 (2007), 69–102.
Alfaro, L. and F. Kanczuk, “Sovereign debt as a contingent claim: a quantitative approach,”
Journal of International Economics 65 (2005), 297–314.
Amador, M., “A Political Economy Model of Sovereign Debt Repayment,” mimeo, Stanford
University, 2003.
Arellano, C., “Default Risk and Income Fluctuations in Emerging Economies,” American Economic Review 98 (2008), 690–712.
Arellano, C. and A. Ramanarayanan, “Default and the term structure in sovereign bonds,”
mimeo, University of Minnesota, 2006.
Arteta, C. and G. Hale, “Sovereign Debt Crises and Credit to the Private Sector,” Journal of
International Economics 74 (2008), 53–69.
Athreya, K., “Welfare Implications of the Bankruptcy Reform Act of 1999,” Journal of Monetary
Economics 49 (2002), 1567–1595.
Athreya, K. B. and H. P. Janicki, “Credit exclusion in quantitative models of bankruptcy: does
it matter?,” Economic Quarterly 92 (2006), 17–50.
Bai, Y. and J. Zhang, “Financial integration and international risk sharing,” mimeo, Michigan
University, 2006.
Balkan, E., “Political Instability, Country Risk and Probability of Default,” Applied Economics
24 (1992), 999–1008.

36

Bekaert, G. and C. R. Harvey, “Emerging Markets Finance,” Journal of Empirical Finance 10
(1996), 46–58.
Bekaert, G., C. R. Harvey and C. Lundblad, “Political Risk, Financial Risk and Economic Risk,”
Review of Financial Studies 20 (2007a), 1783–1831.
Bekaert, G., C. R. Harvey, C. Lundblad and S. Siegel, “Global Growth Opportunities and Market
Integration,” The Journal of Finance 62 (2007b), 1081–1137.
Bi, R., “Debt Dilution and the Maturity Structure of Sovereign Bonds,” mimeo, University of
Maryland, 2006.
Bilson, C., T. Brailsford and V. Hooper, “The Explanatory Power of Political Risk in Emerging
Markets,” International Review of Financial Analysis 11 (2002), 1–27.
Chatterjee, S., D. Corbae, M. Nakajima and J.-V. Rı́os-Rull, “A Quantitative Theory of Unsecured Consumer Credit with Risk of Default,” Econometrica 75 (2007a), 1525–1589.
Chatterjee, S., D. Corbae and J.-V. Rı́os-Rull, “Credit Scoring and Competitive Pricing of Default Risk,” mimeo, University of Texas at Austin, 2007b.
Citron, J. and G. Nickelsburg, “Country Risk and Political Instability,” Journal of Development
Economics 25 (1997), 385–392.
Cole, H., J. Dow and W. English, “Default, settlement, and signaling: lending resumption in a
reputational model of sovereign debt,” International Economic Review 36 (1995), 365–385.
Cuadra, G. and H. Sapriza, “Sovereign default, terms of trade, and interest rates in emerging
markets,” working paper 2006-01, Banco de Mexico, 2006.
Cuadra, G. and H. Sapriza, “Sovereign default, interest rates and political uncertainty in emerging markets,” Journal of International Economics (2008), forthcoming.
Eaton, J. and M. Gersovitz, “Debt with potential repudiation: theoretical and empirical analysis,” Review of Economic Studies 48 (1981), 289–309.
37

Eichengreen, B. and R. Portes, “Debt Restructuring With and Without the IMF,” mimeo, UC
Berkeley, 2000.
Erb, C., C. R. Harvey and T. Viskanta, “Political Risk, Financial Risk and Economic Risk,”
Financial Analysts Journal 52 (1996), 28–46.
Erb, C., C. R. Harvey and T. Viskanta, “Expected Returns and Volatility in 135 Countries,”
Journal of Portfolio Management 10 (2003), 3–56.
Eyigungor, B., “Sovereign Debt Spreads in a Markov Switching Regime,” mimeo, UCLA, 2006.
Gelos, G., R. Sahay and G. Sandleris, “Sovereign Borrowing by Developing Countries: What
Determines Market Access?,” working paper 04/221, IMF, 2004.
Goretti, M., “The Brazilian currency turmoil of 2002: a nonlinear analysis,” International Journal of Finance and Economics 10 (2005), 289–306.
Hatchondo, J. C. and L. Martinez, “Long-duration Bonds and Sovereign Defaults,” working
paper 08-2, Federal Reserve Bank of Richmond, 2008.
Hatchondo, J. C., L. Martinez and H. Sapriza, “On the cyclicality of the interest rate in emerging
economies: solution methods matter,” mimeo, Federal Reserve Bank of Richmond, 2006.
Hatchondo, J. C., L. Martinez and H. Sapriza, “The economics of sovereign default,” Economic
Quarterly 93 (2007a), 163–197.
Hatchondo, J. C., L. Martinez and H. Sapriza, “Quantitative Models of Sovereign Default and
the Threat of Financial Exclusion,” Economic Quarterly 93 (2007b), 251–286.
IMF, “Sovereign Debt Restructurings and the Domestic Economy Experience in Four Resent
Cases,” Policy development and review department, 2002.
Kaminsky, G. and C. M. Reinhart, “The Twin Crises: The Causes of Banking and Balance of
Payments Crises,” American Economic Review 89 (1999), 473–500.

38

Kobayashi, K., “Payment uncertainty, the division of labor, and productivity declines in great
depressions,” Review of Economic Dynamics 9 (2006), 715–741.
Kohlscheen, E., “Constitutions Rule,” mimeo, IIES, 2003.
Krusell, P. and A. Smith, “Consumption-savings decisions with quasi-geometric discounting,”
Econometrica 71 (2003), 365–375.
Kumhof, M., “Fiscal crisis resolution: taxation versus inflation,” mimeo, Stanford University,
2004.
Kumhof, M. and E. Tanner, “Government Debt: A Key Role in Financial Intermediation,”
working paper 05/57, IMF, 2005.
Li, W. and P.-D. Sarte, “U.S. consumer bankruptcy choice: The importance of general equilibrium effects,” Journal of Monetary Economics 53 (2006), 613–631.
Livshits, I., J. MacGee and M. Tertilt, “Consumer Bankruptcy: A Fresh Start,,” American
Economic Review 97 (2007), 402–418.
Lizarazo, S., “Sovereign risk and risk averse international investors,” mimeo, ITAM, 2005.
Lizarazo, S., “Contagion of financial crises in sovereign debt markets,” mimeo, ITAM, 2006.
Mendoza, E. and V. Yue, “A Solution to the Default Risk-Business Cycle Disconnect,” mimeo,
New York University, 2007.
Meyersson, E., “Debt intolerance and institutions, an empirical investigation into the institutional
effect of sovereign debt,” mimeo, Stockholm University, 2006.
Moser, C., “The Impact of Political Risk on Sovereign Bond Spreads. Evidence from Latin
America,” mimeo, University of Mainz, 2006.
Neumeyer, P. and F. Perri, “Business cycles in emerging economies: the role of interest rates,”
Journal of Monetary Economics 52 (2005), 345–380.
39

Pimentel, L. and H. Murphy, “Ecuador’s Bonds Fall on Correa’s Restructure Pledge, Oil Prices,”
Bloomberg (2006), november 17.
Reinhart, C., K. S. Rogoff and M. A. Savastano, “Debt Intolerance,” Brookings Papers on Economic Activity 34 (2003), 1–74.
Sandleris, G., “Sovereign Defaults: Information, Investment and Credit,” mimeo, Johns Hopkins
University, 2006.
Santiso, J., The Political Economy of Emerging Markets: Actors, Institutions and Crisis in Latin
America (New York / London, Palgrave, 2003).
Schmitt-Grohé, S. and M. Uribe, “Closing small open economy models,” Journal of International
Economics 61 (2003), 163–185.
Sturzenegger, F. and J. Zettelmeyer, “Defaults in the 90s,” mimeo, Universidad Torcuato Di
Tella, 2006.
Tomz, M. and M. L. J. Wright, “Do Countries Default in “Bad Times”?,” Journal of the European
Economic Association 5 (2007), 352–360.
Uribe, M. and V. Yue, “Country spreads and emerging countries: Who drives whom?,” Journal
of International Economics 69 (2006), 6–36.
Van Rijckeghem, C. and B. Weder, “The Politics of Debt Crises,” mimeo, University of Mainz,
2004.
Wright, M. L. J., “Reputations and Sovereign Debt,” mimeo, Stanford University, 2002.
Wright, M. L. J., “Coordinating Creditors,” American Economic Review Papers and Proceedings
95 (2005), 388–392.
Yue, V., “Sovereign default and debt renegotiation,” mimeo, NYU, 2005.

40