17

Downloaded from https://academic.oup.com/book/35105/chapter/299187205 by Universiteit Leiden - LUMC user on 01 May 2025

Debt, Default, and Interest Rates

There is a long and colorful history of kings and governments defaulting on their
debts. One major reason for defaulting on the outstanding debt is that it cannot
be inflated away. This is true when the debt is denominated in gold or a foreign
currency.

1 . L AT I N A M E R I CA N D E BT C RI S E S

The Latin American Debt Crisis is a name given to the events that unfolded during
the 1980s in Latin America. A variety of countries in the region had borrowed
fairly heavily in the preceding two decades. The sharp rise in oil prices at several
points during the 1970s had contributed to current account imbalances for the
oil importers in this region. At the same time, for others the rise in oil prices had
led to a consumption and investment spree. Between 1975 and 1983, the region’s
external debt went from $75 billion to $315 billion, which meant that the debt
to output ratio stood at 50%. The rise in interest rates in the U.S. in 1979 to try
and reduce inflation (as part of the Volcker disinflation efforts) also contributed
to Latin America’s woes, as did the resulting U.S. economic downturn.
The decade of the 1980s was to play out very negatively in Latin America.
A number of countries temporarily defaulted on their debts before eventually
making some sort of arrangement with their creditors. We turn to discussing two
countries from the region in detail to understand what went on.

Finance and Financial Intermediation: A Modern Treatment of Money, Credit, and Banking.
Harold L. Cole, Oxford University Press (2019). © Harold L. Cole.
DOI: 10.1093/oso/9780190941697.003.0017
218 D E F I C I T S , D E B T, I N F L AT I O N , A N D D E FA U LT

1.1. Argentina

Argentina has a long tradition of struggling to match its government receipts with
its expenditures. This problem became especially severe in the 1980s as many
countries borrowed to offset the impact of high energy prices on their current

Downloaded from https://academic.oup.com/book/35105/chapter/299187205 by Universiteit Leiden - LUMC user on 01 May 2025

accounts. At the end of the 1980s, Argentina reached the crisis point.
Its very high debt-to-GDP ratio put enormous pressure on the government
budget. This led the government to begin printing money at a very fast rate to
meet its spending needs. Figure 17.1 shows the paths of inflation and real output
per capita.1 The high inflation rate wiped out some of the public debt, but could
not wipe it all out since much of it was dollar denominated. In 1983 the dollar-
denominated debt was about 40% of output, and this debt grew substantially.
In February 1989, the austral fell 64% against the dollar and the World Bank
froze lending to Argentina. The combination of inflation, debt overhang, and
uncertainty led to a decline in real GDP, which also lowered tax receipts.
Debt was restructured through a distressed exchange offering whereby the
bondholders received haircuts of approximately 70%. Argentina also instituted
a currency board, which pegged the Argentine peso to the U.S. dollar between
1991 and 2002 in an attempt to eliminate hyperinflation and stimulate economic
growth. However, Argentina never resolved its fiscal problems and the external
debt-to-GDP ratio climbed back up. The International Monetary Fund (IMF),
however, kept lending money to Argentina and postponing its payment sched-
ules. Massive tax evasion and money laundering explained a large part of the
evaporation of funds toward offshore banks.

3200 10800
2800 10200
2400 9600
2010 U.S. Dollars

2000 9000
Percent

1600 8400
1200 7800
800 7200
400 6600
0 6000
1970 1975 1980 1985 1990 1995 2000 2005 2010 2015
–400 5400

Inflation, consumer prices for Argentina (left)

Constant GDP per capita for Argentina (right)

Figure 17.1 Argentina: Inflation and Output.

source: FRED. Data Source: World Bank.

1. World Bank, “Inflation, consumer prices for Argentina.” World Bank, “Constant GDP per Capita
for Argentina. Both retrieved from FRED, Federal Reserve Bank of St. Louis.
Debt, Default, and Interest Rates 219

Late in the 1990s Argentina was hit with a large recession that lowered revenues
and made its debt problems more severe. Spreads on Argentinean bonds over U.S.
Treasuries went from 10% in December 2000 to almost 50% a year later. Because
rates on foreign borrowing were prohibitively high, the government switched to
borrowing from domestic banks: The share of government debt in bank assets

Downloaded from https://academic.oup.com/book/35105/chapter/299187205 by Universiteit Leiden - LUMC user on 01 May 2025

rose from 15.5% to 21.5%. But this made it harder to inflate away the debt. In
October 2001, public discontent with the economic conditions was expressed
in the nationwide election when over 20% of all voters chose to enter so-called
anger votes, returning blank or defaced ballots rather than indicate support for
any candidate.
The government attempted to get out of the crisis through a voluntary debt
exchange, which would avoid a default. However, to induce participation, it
agreed to a 17% interest rate on this debt, which reduced its payments over the
next 4 years by only $12.6 billion. The first exchange was made in June 2001.
When that proved insufficient, a second exchange was tried in November. The
crisis intensified when, on December 5, 2001, the IMF refused to release a US$1.3
billion tranche of its loan, citing the failure of the Argentinean government to
reach previously agreed-upon budget deficit targets, and demanded further bud-
get cuts, amounting to 10% of the federal budget. On December 4, Argentinean
bond yields stood at 34% over U.S. Treasury bonds, and, by December 11, the
spread had jumped to 42%.
Fearing the worst, by the end of November 2001 people began withdrawing
large sums of money from their bank accounts, turning pesos into dollars and
sending them abroad, causing a run on the banks. On December 2, 2001 the
government enacted a set of measures, informally known as the corralito, that
effectively froze all bank accounts for 12 months, allowing for only minor sums of
cash, initially announced to be just $250 a week, to be withdrawn. In January 2002
the government ended the fixed exchange rate of the peso to the dollar. When
default was declared in 2002, foreign investment fled the country, and capital
flows toward Argentina ceased almost completely.
An agreement was eventually reached in 2005 where 76% of the defaulted
bonds were replaced by others, of a much lower nominal value (25–35% of the
original) and at longer terms.

1.2. Mexico

The decade of the 1990s looked to be a bright one for Mexico. It had recovered
from the Latin American debt crises of the 1980s and the “lost decade” in which
growth was negligible. Inflation seemed under control. Following a series of
liberalizing reforms, investment funds poured into Mexico in the early 1990s. On
220 D E F I C I T S , D E B T, I N F L AT I O N , A N D D E FA U LT

top of this, Mexico had entered into a major trade deal with the U.S. The North
American Free Trade Agreement (NAFTA) took effect in 1994 and would reduce
trade barriers between the two countries to a new low. There were a couple of dark
clouds on the horizon, but at the time, they did not seem that alarming. The influx
of foreign investment also served to finance a very large current account deficit.

Downloaded from https://academic.oup.com/book/35105/chapter/299187205 by Universiteit Leiden - LUMC user on 01 May 2025

These outcomes are depicted in Figure 17.2.
The Mexican exchange rate system featured a crawling peg in which govern-
ment intervention kept the exchange rate within a narrow band. The top of the
band was supposed to rise gradually, generating a gradual depreciation against the
dollar. However, the real exchange rate, which we can define as P/(P*E), where P
is the domestic price level, P* is the foreign price level, and E is the exchange rate,
was rising fairly sharply. This is referred to as a real appreciation of the peso. This
real appreciation meant that the real price of Mexican goods was rising relative to
the U.S., damaging exports, which contributed to the current account deficits.
Investor confidence collapsed in 1994 after a rebellion in the south and the
assassination of the ruling party’s presidential candidate. Capital began to exit
rapidly. The central bank intervened to support the peso and the government
repeatedly raised rates. As a result, the Mexican central bank lost close to $11
billion in reserves during a four-week period. This is a large loss given that the
central bank had started with a bit under $30 billion in reserves in February 1994.
The Mexican government borrowed using both domestically denominated
debt, called Cetes, and foreign-indexed debt, called Tesobonos. The fact that
Tesobonos were foreign indexed meant that operationally they worked like
foreign-denominated debt. Before the crises in late 1994, about 75% of
government debt was domestically denominated. By May the rate on one-month
government domestically denominated debt had risen to 16.4%, from 9.5% just

6 6
Billions

Billions

5 4
Component Sub-periods

Component Sub-periods
US Dollars, Sum Over

US Dollars, Sum Over

4 2

3 0

2 –2

1 –4

0 –6
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

Financial Account: Liabilities: Direct Foreign Investment in Reporting Country for Mexico (left)
Current Account Balance: Total Trade of Goods for Mexico (right)

Figure 17.2 Mexico: Foreign Direct Investment and the Current Account.
source: FRED. Data Source: Organization for Economic Co-operation and
Development.
Debt, Default, and Interest Rates 221

3 months earlier. In response to this, the government switched to financing itself

using foreign-denominated debt. This foreign-denominated debt could not be
inflated away, which helped to promote investor confidence in it.
By November 1994, the Mexican government was primarily financing itself
using 90 day Tesobonos. Domestically denominated debt was only 25% of total

Downloaded from https://academic.oup.com/book/35105/chapter/299187205 by Universiteit Leiden - LUMC user on 01 May 2025

debt, while dollar-indexed debt was 70% of total debt. This change in the com-
position of the debt reduced the cost of debt service significantly because the
interest rate on this debt was substantially below the rate on domestic debt.
However, this change in composition also tied the government’s hands, since it
could not devalue (and use inflation) as a means of lowering the real value of its
outstanding debts.
Another feature of the change in the composition of the government’s debt was
that the average maturity was now fairly short; the government was auctioning off
Tesobonos in weekly auctions. As the maturity shrank and the dependence on the
Tesobono grew, and the amount being auctioned off each week rose sharply.
The pressure on the government increased as its reserves fell further. This led
the government to devalue the peso, which took the form of a sharp increase
in the top of the crawling peg. The initial response to the devaluation was positive.
The Tesobono auction following the devaluation was fairly successful, with $416
million in bonds sold at an 8.61% interest rate, only a 38 basis points increase.
The Cetes auction the next day had an average yield of 16.22%, only a 142 basis
points increase from the prior week. However, one troubling aspect was that the
government had intended to auction off $600 million in Tesobonos.
The pressure on government reserves continued as people continued exchang-
ing pesos for dollars. The central bank lost $4.5 billion in reserves in a single day,
and reserves fell below $6 billion. This led the government to abandon its attempts
to peg the currency on December 22. The response by the markets was swift and
negative. The secondary market interest rate on Cetes rose to 24.5% and the peso
fell sharply (40%) against the dollar.
The Tesobono auctions did not go well either. At the December 27 auction,
where the government sought to auction $416 million in bonds, only $28 million
was sold at an average yield of 10.23%. The next day’s Cetes auction also did
not go well, with bids well below the amount offered, and the yield jumped way
up to 31.41%. This drying up of financial access for the Mexican government
exacerbated investor concern over the $10 billion in Tesobono debt maturing in
the first quarter of 1995.
By January 1995 the Mexican government was clearly in a state of crisis and,
unless there was a massive shift in investor confidence, was going to be forced
to default on many of its maturing debt claims. The U.S. initially organized an
$18 billion dollar line of credit to aid the Mexicans on January 2. But this failed
to stem the problem. The amount of Tesobonos sold in the first two auctions
222 D E F I C I T S , D E B T, I N F L AT I O N , A N D D E FA U LT

in January were quite small, while the yield rose to close to 20%. In February,
the U.S. approved a $20 billion loan to Mexico on favorable terms with delayed
repayment. The financial crisis passed and Mexico eventually repaid the U.S.2

Downloaded from https://academic.oup.com/book/35105/chapter/299187205 by Universiteit Leiden - LUMC user on 01 May 2025

2 . T H E E U C R I SI S

At the end of 2010, investors became nervous about several European countries
with substantial budget deficits, large debt levels, and slowing economic growth.
This was especially true when a large amount of the debt was in foreign hands,
as in Greece and Portugal. The result was a sharp increase in credit spreads on
government bonds and risk insurance on CDS relative to safer countries like
Germany.
Each European country took a different route to get into the crisis: Greece was
Europe’s fastest growing country, but it also ran large deficits. Then the economic
slowdown hit its shipping and tourism industries very hard; this in turn led its
debt-to-GDP ratio to grow rapidly. Ireland also ended up with a large debt-to-
GDP ratio, but this was not because of government over-spending; rather it was
because of the government’s insurance of Irish banks that speculated and lost
on the local real estate bubble. Portugal’s problem has been blamed on a large
government sector and excessive public employment. Italy had a small deficit but
a large public debt and the decades-long slowdown in growth had made this debt
burden seem too large. However, Italy’s debt has a long maturity and is largely
domestically held. In the next section we focus on Greece since it has attracted so
much attention from the international financial community.

2.1. Greece

Greece joined the EU in 1981, ushering in a period of sustained strong growth;

see Figure 17.3. The beginning of the end came in 2006 when growth started to
slow. Eventually the situation turned very negative, and the level of GDP dropped
sharply. By 2014, all of the growth from 2000 had been erased. One of the major
sources of the economic decline was a sharp drop in the employment rate; this can
also be seen in Figure 17.3. Employment dropped by more than 10 percentage
points, a huge fall.3

2. For more on the Mexican crisis, see J. Whitt, “The Mexican Peso Crisis,” in the Federal Reserve
Bank of Richmond’s Economic Review, vol. 81, no. 1, 1996.
3. Sources for figure are the University of Groningen and the University of California, Davis,
Real GDP at Constant National Prices for Greece [RGDPNAGRA666NRUG]. Organization for
 Debt, Default, and Interest Rates 223

340,000 63.00
Millions of 2011 U.S. Dollars
320,000 61.25
300,000 59.50
280,000 57.75

Percent
260,000 56.00
240,000 54.25
220,000 52.50

Downloaded from https://academic.oup.com/book/35105/chapter/299187205 by Universiteit Leiden - LUMC user on 01 May 2025

200,000 50.75
180,000 49.00
160,000 47.25
1980 1985 1990 1995 2000 2005 2010

Real GDP at Constant National Prices for Greece (left)

Employment Rate: Aged 15–64: All Persons for Greece (right)

Figure 17.3 Output and Employment in Greece.

source: FRED. Data Source: OECD, University of Gronigen.

190 –3.0
180 –4.5
170 –6.0
160 –7.5
Percent of GDP

Percent of GDP
150 –9.0
140 –10.5
130 –12.0
120 –13.5
110 –15.0
100 –16.5
90 –18.0
1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 2015

General government gross debt for Greece (left)

Cash surplus/deficit (% of GDP) for Greece (right)

Figure 17.4 Government Debt and Deficits in Greece.

source: FRED. Data Source: International Monetary Fund, World Bank.

In Figure 17.4, one can see how the economic decline led to a worsening of the
government’s fiscal situation. The government’s cash surplus, which is a measure
of the excess of revenue over expenditures, was negative to begin with, reflecting
the deficits the government ran during the growth years. However, the downturn
led the country to drop sharply into the red.4 The Greek government (in response
to pressure from the ECB [European Central Bank] and European Commission)
did reduce the extent of the negative numbers after 2009, but it’s worth noting
here how large the budget as a whole was during this period. In response to the
fall in output and the large deficits, the debt-to-output ratio rose sharply from its

Economic Co-operation and Development, Employment Rate: Aged 15–64: All Persons for Greece
[LREM64TTGRQ156S]. Retrieved from FRED, Federal Reserve Bank of St. Louis.
4. Note that this measure seems to be preferred over the standard official measure—the primary
surplus—because “exceptional” expenditure charge-offs were allowed here.
224 D E F I C I T S , D E B T, I N F L AT I O N , A N D D E FA U LT

already very high level. The truth is that it takes a large amount of fiscal rectitude
to pay off debt-to-output ratios of 150%.5 However, a recovery in employment
and output would go a long way toward fixing the fiscal situation.
If we look at the fundamental sources of government expenditure pressure, we
see that Greece has a very generous pension system with a fairly early retirement

Downloaded from https://academic.oup.com/book/35105/chapter/299187205 by Universiteit Leiden - LUMC user on 01 May 2025

age. At the same time, life expectancy has risen, the fertility rate has fallen, so the
share of the elderly in the population has risen. This implies that honoring these
pension promises will be a substantial burden for the government going forward.
The debt problem was compounded by nervousness in the bond markets and
a sharp increase in the yields that investors required to buy Greek debt. This
is illustrated in Figure 17.5, where long-term bond yields are plotted for both
Greece and Germany.6
Because both countries’ bonds are denominated in euros, the difference in
yields should just reflect the difference in the extent of default risk and the premia
attached to this difference. One thing that is striking in the figure is how close
the yields were in the two countries before the crisis, and how the gap widened
during the crisis. The difference in yields alone may not fully reflect Greece’s

Long-Term Government Bond Yields: 10-year

30
25
20
Percent

15
10
5
0
2000 2002 2004 2006 2008 2010 2012 2014 2016 2018
–5
Greece Germany

Figure 17.5 Bond Yields: Greece vs. Germany.

source: FRED. Data Source: Organization for Economic Co-operation and
Development.

5. Sources are International Monetary Fund, General government gross debt for Greece
[GGGDTAGRC188N], Eurostat, Gross Domestic Product for Greece [CPMNACSCAB1GQEL],
and World Bank, Cash surplus/deficit (% of GDP) for Greece [CASHBLGRA188A]. Retrieved
from FRED, Federal Reserve Bank of St. Louis.
6. Organization for Economic Co-operation and Development, Long-Term Government Bond
Yields: 10-year: Main (Including Benchmark) for Greece [IRLTLT01GRM156N], and Orga-
nization for Economic Co-operation and Development, Long-Term Government Bond Yields:
10-year: Main (Including Benchmark) for Germany [IRLTLT01DEM156N]. Retrieved from
FRED, Federal Reserve Bank of St. Louis.
Debt, Default, and Interest Rates 225

credit difficulties. In October 2009 Greece experienced the first of a series of

ratings downgrades. As a result, Greece found it increasingly difficult to borrow
from private creditors and turned to official sources.
The Greek government came up with its first austerity package in February
2010, followed shortly (in March) by a second austerity package. There was a

Downloaded from https://academic.oup.com/book/35105/chapter/299187205 by Universiteit Leiden - LUMC user on 01 May 2025

further series of austerity packages, culminating in the seventh austerity package
in July 2013. In April, Greece and the EU leaders agreed to the first bailout
package, for 110 billion euros over three years.
By June 2011 Moody’s had reduced Greece’s rating to Caa, and Standard &
Poors (S&P) had reduced it to their lowest rating. In October 2011 investors in
Greek debt agreed to a write-down of 50%. In June 2015 Greece missed a payment
to the IMF. This put Greece at least temporarily in default to the IMF. The list
of countries that have defaulted on IMF loans is very small: Cuba, Nicaragua,
Vietnam, and so on.

3 . MO D E L I N G DE FA U LT

In this section I develop a simple model of sovereign default which we can use to
think through the various episodes that have just been discussed.7

3.1. Consequences of Defaulting

In the corporate world, debt contracts are enforced by courts. A corporation can
be sued and the courts can force it to hand over its assets, restructure its debts, or
shut down and liquidate its remaining assets. For a government, there is no analog
to bankruptcy court, and enforcing claims is more problematic:

(1) Few sovereign assets are located in foreign countries and a sovereign
state cannot commit to hand over its assets.
(2) There are legal principles, called sovereign immunity, which claim that
sovereigns cannot be sued in foreign courts without their consent.

7. A much more sophisticated infinite horizon model with quantitative applications can be found
in M. Aguiar, S. Chatterjee, H. Cole and Z. Stangebye, “Quantitative Models of Sovereign Debt
Crises,” in Handbook of Macroeconomics, 2016, vol. 2. Another source is M. Augiar, “Sovereign Debt,”
Handbook of International Economics, 2014, vol. 4.
226 D E F I C I T S , D E B T, I N F L AT I O N , A N D D E FA U LT

3.1.1. Sovereign Immunity

During the 19th century sovereigns were protected by absolute sovereign immu-
nity which held that sovereigns could not be sued in foreign courts without their
consent. After WWII a more limited view of sovereign immunity took hold in
which the sovereign can be sued for commercial activity. (The U.S. advanced

Downloaded from https://academic.oup.com/book/35105/chapter/299187205 by Universiteit Leiden - LUMC user on 01 May 2025

this view to potentially control Soviet economic activity within its borders.)
Bond issuance and payments are considered commercial activity. Moreover,
sovereign bonds now typically include waivers of sovereign immunity. Central
bank assets—including international reserves—are typically immune to this
waiver. However, changes in the law since the 1970s have made it easier for
holdout creditors to receive judgment claims.

3.2. Costs of Default

The costs of default are thought to include

(1) loss of access to credit markets;

(2) various trade disruptions coming from creditors’ ability to seize
payments and a loss of access to trade credit;
(3) drops in output due to trade disruption, the need to adjust the
government’s fiscal stance, and use overall uncertainty with respect to
future government behavior; and
(4) possibly an impact on the domestic banking system and domestic
bond market.

4 . M O DE L OF S O V E R E I G N BO R RO W I N G A N D DEFAULT

The government’s budget constraint can be written as

qt Bt + Mt − Mt−1 + Pt Tt = 𝛿t Bt−1 + Pt Gt + Trt ,

• qt Bt is the revenue from selling pure discount bonds

• Mt − Mt−1 is the revenue from money creation
• Pt T t are tax revenues
• Bt−1 are the bonds coming due
• Pt Gt is government expenditure on goods and services
• Trt are government transfers
Debt, Default, and Interest Rates 227

4.1. Default Decision

Since repayment is, in some sense, optional with sovereign debt, the government
must weigh whether to fully repay and the various ways it could choose not repay.
The choices are:

Downloaded from https://academic.oup.com/book/35105/chapter/299187205 by Universiteit Leiden - LUMC user on 01 May 2025

(1) fully repay;
(2) try to inflate the debt away;
(3) try to negotiate down its payments; or
(4) default, and possibly negotiate the extent of repayment down the road.

Here we focus on options 1 and 4.

Assume that there are two periods, t = 1, 2 and the government must decide
whether to repay its debts in each period. Assume that if it fails to repay, then the
country’s output will fall by a fraction 𝛾 henceforth and that it will lose access
to credit markets. (We could modify our model by not having the fall of 𝛾 last
forever, but instead just for a couple of years.)
The country’s output is stochastic, with Y t ∈ Y and the probability is Π(Y 1 )
in the first period and Π(Y 2 |Y 1 ) in the second. Assume also that the government
cannot adjust the tax rate 𝜏 very readily, so we will take it to be fixed.
Strategic models are often solved by backward induction, and that is what we
are going to do here. The steps are:

(1) Compute the optimal default rule in t = 2,

• first assuming no prior default,
• then assuming prior default
(2) Compute the second period payoff and bond pricing function
• Compute the payoff when default both has and has not has
already occurred.
• Compute the bond pricing function both when default has and
has not has already occurred.
(3) Compute optimal t = 1 behavior and bond pricing if we want to
default in t = 1.

4.1.1. Period 2 Default Decision - Never Defaulted Before

In the second period, the government’s revenue is

𝜏Y2 if it doesn’t default, and

𝜏𝛾Y2 if it does.
228 D E F I C I T S , D E B T, I N F L AT I O N , A N D D E FA U LT

The government budget constraint then implies that

G2 = 𝜏Y2 − B2 if it doesn’t default, and

G2 = 𝜏𝛾Y2 if it does

Downloaded from https://academic.oup.com/book/35105/chapter/299187205 by Universiteit Leiden - LUMC user on 01 May 2025

and private consumption is

C2 = (1 − 𝜏)Y2 if it doesn’t default, and

C2 = (1 − 𝜏)𝛾Y2 if it does.

If we assume that the government’s payoff in period 2 depends upon both gov-
ernment spending and private consumption,

V(G2 ) + U(C2 ),

then the government should default if the payoff from doing so is greater than
repaying, or

V(𝜏𝛾Y2 ) + U((1 − 𝜏)𝛾Y2 ) ≥ V(𝜏Y2 − B2 ) + U((1 − 𝜏)Y2 ).

Clearly, the government should never default if (1 − 𝛾)𝜏Y 2 > B2 . Moreover,

the benefit to defaulting is increasing in B2 and decreasing in Y 2. There will exist a
cut off level of debt B2 (Y 2 ) such that the government will default if B2 > B2 (Y 2 ),
and repay otherwise.
To get a better sense of things, I have plotted a “typical” case in Figure 17.6.
In the figure we have plotted a line that shows the combinations of Y 2 and B2 at
which the government is just indifferent between defaulting and not. We assume
that when the government is indifferent it chooses to repay. So 𝛿 = 1 along the
line. 𝛿 is also equal to 1 at all points below the line, and 𝛿 = 0 at all points above
the line.

Default

Don’t Default

Output Y

Figure 17.6 Default Set.

Debt, Default, and Interest Rates 229

4.1.2. Period 2 Default Decision: Already Defaulted

Since the government has already defaulted, its second period revenue is 𝜏𝛾Y 2 ,
irrespective of whether it defaults or not today. But then its consumption is

G2 = 𝜏𝛾Y2 − B2 if it doesn’t default, and

Downloaded from https://academic.oup.com/book/35105/chapter/299187205 by Universiteit Leiden - LUMC user on 01 May 2025

G2 = 𝜏𝛾Y2 if it does

and private consumption is C2 = (1 − 𝜏)𝛾Y 2 whether it defaults or not.

So, the government should always default if B2 > 0.

4.1.3. The Second Period Payoff

The second period payoff to the government (assuming it did not default in the
first period) is given by

W(B2 , Y2 ) =
𝛿(B2 , Y2 ) [V(𝜏Y2 − B2 ) + U((1 − 𝜏)Y2 )]
+ (1 − 𝛿(B2 , Y2 )) [V(𝜏𝛾Y2 ) + U((1 − 𝜏)𝛾Y2 )] .

This payoff will be decreasing in B2 and increasing in Y 2 .

Note that the option of defaulting is acting like partial insurance for the
government, since it prevents both government and private consumption from
falling too far if Y 2 is small relative to B2 .

4.1.4. The Lending Decision

We will assume that there are a large number of competitive lenders who buy
the government’s first period debt. We assume that debt issuance happens after
the first period default decision, so these lenders need to know about default
only in t = 2.
Let 𝛿(B2 , Y 2 ) denote the government’s default decision, where

𝛿(B2 , Y2 ) = 1 if B2 ≤ B2 (Y2 ), and

𝛿(B2 , Y2 ) = 0 otherwise.

In Figure 17.6 we have graphed a version of the default rule 𝛿(B2 , Y 2 ) with
the second period debt level on the y-axis and the second period output level on
the x-axis. Along the line, the country is just indifferent between defaulting and
not, while below the line, the level debt level has fallen relative to the indifferent
level, and hence the country now prefers to repay, while above the line the reverse
is true. The set of outcomes above the line is called the default set, and those
230 D E F I C I T S , D E B T, I N F L AT I O N , A N D D E FA U LT

below are called the repayment set. The lenders will take these two sets as given.
The borrower will choose its debt level B2 in the first period. Given this, the
probability that the lenders will be repaid is the probability of getting an outcome
level at or above the cut-off for the repayment set. Since this output level is
increasing in the debt level, the likelihood of repayment is decreasing, because

Downloaded from https://academic.oup.com/book/35105/chapter/299187205 by Universiteit Leiden - LUMC user on 01 May 2025

it will take a more fortunate outcome to induce the government to repay.
If lenders are risk neutral and have a discount rate of R−1 the price of a
government bond today is

1
q= ∑ 𝛿(B2 , Y2 )Π(Y2 |Y1 ). (19)
R Y
2

This price is the probability of repayment times the lenders’ discount factor. At
this price, the lenders just break even in expectation. Of course ex post they either
enjoy a high return when the government repays or a complete loss when the
government defaults. The high return is supposed to compensate them for the ex
ante risk of default.
The price of government debt in (19) will depend on the amount the govern-
ment borrows through 𝛿. The realization of the period 2 output level Y 2 implicitly
impacts on the determination of q through both through 𝛿 and the probability
of Y 2 . But q does not depend upon Y 2 directly since it is not known in period.
The price can depend upon period 1 output, Y 1 to the extent that it affects the
probabilities w.r.t. Y 2 tomorrow. This dependence is direct since Y 1 is known in
period 1. The price q will be declining in B2 . Since a high output level Y 2 will lead
the government to choose not to default, if a higher Y 1 increases the probability
of a high Y 2 tomorrow, then q will be increasing in Y 1 . (If the reverse is true, then
it will be decreasing in Y 1 .) And, if Y 2 is independent of Y 1 , it will not depend on
current output, only current borrowing. To reflect all these possibillities, denote
this price by q(B2 , Y 1 ).
The fact that the price that it gets for its debt will fall when increasing its
issuance leads to a higher likelihood of default is an important component of
sovereign borrowing. In this respect sovereign borrowing is like the problem of
a monopolist who understands that the more she tries to sell, the lower the price
of her good. The government here is a monopolist too, in that while there may be
other debt products in the marketplace, their debt is special in that their default
decision directly affects its payoff.

4.1.5. The Lending Decision—One Key Caveat

If the lenders anticipate that the government is going to default in period 1 then,
as we noted earlier, the government will default for sure tomorrow. This implies
that 𝛿(B2 , Y 2 ) = 0 for all possible Y 2 draws, and hence
Debt, Default, and Interest Rates 231

q = 0 if government will default in t = 1.

In Figure 17.7 I have graphed the price schedule for both the case in which the
country had not defaulted in the prior period (solid line) and the case when it had
(dotted line). The price schedule when it had not previously defaulted starts at the

Downloaded from https://academic.oup.com/book/35105/chapter/299187205 by Universiteit Leiden - LUMC user on 01 May 2025

level implied by the risk-free discount rate 1/R, and stays at this level so long as
the debt level is low enough that the probability of default is 0. This corresponds
to being below the point where the default indifference line hits the y axis in
figure 17.7. As the debt rises above this intersection point, then the country will
optimally choose to default at the lowest output level. Now default will occur with
positive probability, and the price schedule turns down. With higher debt, the
likelihood of default is higher still and the price schedule keeps going down, until
it hits a price of 0, at which point default is occurring for sure. This corresponds
to such a high debt level that the country is defaulting in Figure 17.7. For all debt
levels above this point the price is 0. In the prior default price schedule, the price
is 0 for any positive amount of debt issuance.
A key aspect of the price schedules the country faces is that they depend upon
the country’s prior default and current borrowing decisions. This is something the
country will optimally choose to internalize in deciding how much debt to issue.

4.1.6. Borrowing Decision

The payoff to the government if it does not default in period 1 is given by

max V(𝜏Y1 − B1 + q(B2 , Y1 )B2 ) + U((1 − 𝜏)Y1 )

B2

+ 𝛽E {W(B2 , Y2 )|Y1 } ,

where the government is optimizing over B2 . Note that if it chooses too large a
value of B2 then 𝛿 2 = 0 for sure and q = 0. Hence, there is a natural cap on the

q
No Prior Default
1/R

Prior Default
0

Borrowing

Figure 17.7 Price Schedules:

Regular and Prior Default.
232 D E F I C I T S , D E B T, I N F L AT I O N , A N D D E FA U LT

amount it can borrow. Today’s income affects the payoff both directly and through
the expectations about Y 2 .

4.1.7. Default Decision

Countries normally borrow in order to help repay debt that is coming due. This

Downloaded from https://academic.oup.com/book/35105/chapter/299187205 by Universiteit Leiden - LUMC user on 01 May 2025

opens up the possibility of borrowing and then defaulting. What limits this is
that debt is normally coming due throughout the year, and the government holds
regular debt auctions to roll over the debt piece-by-piece. To account for this in an
annual model, we will assume that when a country defaults, it only gets a fraction
𝛾 of the new debt that will be issued that year.

4.1.8. The First Period Default Decision

We have already computed the payoff to not defaulting, given that the government
chooses the optimal level of borrowing,

max V(𝜏Y1 − B1 + q(B2 , Y1 )B2 ) + U((1 − 𝜏)Y1 )

B2

+ 𝛽E {W(B2 , Y2 )|Y1 }.

In this problem, the government will be trading off the benefits of smoothing gov-
ernment consumption in the face of revenue shocks, and the potential increase in
the costs of borrowing.
The payoff if the government defaults is given by

V(𝜏𝛾Y1 + 𝛾q(B2 , Y1 )B2 ) + U((1 − 𝜏)𝛾Y1 )

+ 𝛽E {V(𝜏𝛾Y2 ) + U((1 − 𝜏)𝛾Y2 )} .

The government will choose to default or not in period 1, depending upon

which payoff is larger.
The model we have constructed features something like a solvency-based default.
This occurs when the amount the government has coming due is simply too big
for it to ever possibly repay. This is what happened in Grenada after it was hit by
a hurricane.
Even if the government can repay, it may choose to default instead. These
defaults of choice will generally be triggered by low output realizations. This is
because the cost of defaulting in terms of lost output (1 − 𝛾)Y is increasing in
Y, while the amount to be repaid is independent of Y.
In addition, an increase in the likelihood of a default decreases the current price
of the new government debt, which, in turn, makes default more attractive. So
there is a natural feedback loop that makes the government not want to default
for high output levels and want to default for low ones.
Debt, Default, and Interest Rates 233

Defaults of choice are much more common than defaults of solvency.

Argentina is an example of defaults of choice.
The model also allows for liquidity crisis types of default, in a manner somewhat
similar to the Diamond-Dybvig model of banks. To see this, assume that no one
would lend to the country in period 1 for fear that the government would default

Downloaded from https://academic.oup.com/book/35105/chapter/299187205 by Universiteit Leiden - LUMC user on 01 May 2025

on these new loans. Then, the government would have to suddenly repay all of the
outstanding debt coming due.
This changes our payoffs from not defaulting to

V(𝜏Y1 − B1 ) + U((1 − 𝜏)Y1 )

+ 𝛽E {W(0, Y2 )|Y1 },

and from defaulting to

V(𝜏𝛾Y1 ) + U((1 − 𝜏)𝛾Y1 ) + 𝛽E {V(𝜏𝛾Y2 ) + U((1 − 𝜏)𝛾Y2 )}.

The sudden need to drastically cut government consumption in order to repay the
country’s debts can make defaulting either more attractive or simply necessary.
This can be true even when the country would have repaid under the standard
equilibrium we considered first. Such a liquidity crisis outcome can potentially
explain contagion effects and outcomes like the Mexican debt crisis.
There are two potential price schedules in the case of a liquidity crisis. The
first corresponds to the case in which the debt level is high enough that the
country would optimally choose to default today in the event of a crisis. This price
schedule looks like the prior-default price schedule in Figure 17.7. The other price
schedule corresponds to the case in which the government would not default
today even if it could not borrow. In this case, the government can credibly say
it will not default today, and hence, its price schedule looks like the regular no-
prior-default schedule in Figure 17.7.
When the country’s debt level is high enough and its output level is low enough
that a crisis is possible, we refer to the country as being “in the crisis zone.” By this
we mean that fundamentals are such that a crisis can occur, but not necessarily
that it will occur. Whether a crisis occurs depends upon which of the two possible
equilibrium outcomes will be selected, something very hard to say much about
concretely.8

8. For more on sovereign debt and sovereign debt crises, a good source is Aguiar and Amador’s
“Sovereign Debt: A Review,” which was published in the Handbook of International Economics,
vol. 4, 2014. See also the entry “Self-Fulfilling Crises” in Wikipedia.
234 D E F I C I T S , D E B T, I N F L AT I O N , A N D D E FA U LT

4.2. Default Worries and Economic Activity

Thus far, we have taken output of the country to be exogenous, but there is
an important interaction here. What if agents expect the economy to be less
productive if the country defaults? This can be especially important with respect

Downloaded from https://academic.oup.com/book/35105/chapter/299187205 by Universiteit Leiden - LUMC user on 01 May 2025

to the incentives to invest.
To see why this is the case, consider the payoff to a firm from investing in an
extra unit of capital. Assume that it takes the following fairly standard form

FK (K, L)

if the government does not default and

𝛾FK (K, L)

where 𝛾 < 1, if it does default. This reduction in the return is coming from the
fall in overall productivity that one expects to see following a default. Then, the
expected payoff today is given by

1 𝛿(B2 , Y2 ) [FK (K, L)]

∑[ ] Π(Y2 |Y1 ).
R Y +(1 − 𝛿(B2 , Y2 )) [𝛾FK (K, L)]
2

Note that the higher the probability of default, the lower the expected return on
investing in more capital. Thus, concerns about future sovereign default can lower
investment today and output tomorrow. This in turn can make it more likely that
the government will end up defaulting.
So-called sudden stops refer to cases in which a country is growing rapidly
while experiencing large amounts of capital inflows. Often, these inflows mean
that either there is a lot of private borrowing going through the banking system,
or government borrowing, or both. Often the price of nontradeable goods like
houses and wages can rise sharply. At some point fears about a possible default
seem to kick in and this all can reverse sharply. This sharp reversal is the sudden
stop, and it can lead to defaults and financial crises. The simple extension that we
have just considered shows how fears of default can lead to a sudden reduction
in investment. This sudden reduction in investment is an important element of
these sudden stops.9

9. See Guillermo A. Calvo, “Capital Flows and Capital-Market Crises: The Simple Economics of
Sudden Stops” Journal of Applied Economics, (1998), vol. 1, no. 1 (November) for more on sudden
stops.
Debt, Default, and Interest Rates 235

4.3. Default and Maturity

The maturity structure of the government’s debt can have an important impact
on the government incentives with respect to its policy choices. Remember that
the longer the maturity, the easier it is to inflate away the outstanding debt. With

Downloaded from https://academic.oup.com/book/35105/chapter/299187205 by Universiteit Leiden - LUMC user on 01 May 2025

10- or 20-year-old debt, a mild inflation will do the job of inflating half or more
of the debt away. As a result of this, many countries find borrowing using long
term debt very expensive, especially during any kind of crisis period. This can
lead them to shorten the maturity structure of their debt. But this shortening can
have an important impact on their incentives to default.
To discuss this, consider a simple two-period model: Assume output is con-
stant at Y; that the government has an outstanding debt balance of B; and that
the gross interest rate is 1.
Case 1: Assume the government has B/2 units of debt coming due today and B/2
units tomorrow and is planning to pay off its debt each period.

• Then, its payoff from repaying looks like

[V(𝜏Y − B/2) + U((1 − 𝜏)Y)] [1 + 𝛽].

• This payoff is independent of whether anyone would lend to the country

in the first period so no lending is being planned.

Case 2: Assume that the government has B units coming due today and is
planning on paying off half of its debt and rolling over the other half.

• Then, its payoff from repaying, given that it can roll over at a gross
interest rate of 1, is

[V(𝜏Y − B/2) + U((1 − 𝜏)Y)] [1 + 𝛽].

• But if there is a liquidity crisis and no one will lend to the country in the
first period, this payoff becomes

V(𝜏Y − B) + 𝛽V (𝜏Y) + U((1 − 𝜏)Y) [1 + 𝛽]

The sudden need to pay off all of the maturing debt can make defaulting much
more attractive. Something like this is what happened in Mexico.
236 D E F I C I T S , D E B T, I N F L AT I O N , A N D D E FA U LT

4.4. International Rescue and Incentive Problems

In the old days, bondholders were largely on their own. One counter example to
this is when the British intervened in Egypt in the 1880s in part to protect British
bondholders. During the 19th century, Egypt was formally part of the Ottoman

Downloaded from https://academic.oup.com/book/35105/chapter/299187205 by Universiteit Leiden - LUMC user on 01 May 2025

Empire and the ruling dynasty in Egypt had borrowed and spent large sums on
various infrastructure projects. These projects were at least partially intended
to bolster military control of the country, and the government itself was quite
corrupt. Perhaps as a result of all this, the expenditures did not end up translating
into a large increase in output. This made the country’s economic situation poor,
and the paying off the debt to European lenders difficult. When various militant
actors in the government, in particular the military, threatened to overthrow the
Khedive (who was ruling the country), the Europeans intervened. In the end,
Europeans took over the running of many aspects of the country (including the
Treasury), the debt was forgiven, and the Europeans took control of the Suez
Canal in return.10
Things have changed a lot since those colonial days. Now, there are interna-
tional agencies like the IMF that can rush to the rescue. But this leads to the
following question for the bondholders: Can we get some of our money back from
the IMF? Often, the initial phase of an IMF intervention includes loans that help
the government roll over its debt on more favorable terms. This typically leads
to private lenders being paid off and private loans being replaced by official loans
from the IMF and the World Bank, among other non-private entities. The IMF
can also help the country negotiate the terms of its outstanding liabilities, often
leading to a reduction in the amount that the country has to pay, and lengthening
the horizon over which it can make these payments.
If British investors in Egyptian bonds during the 19th century anticipated that
the British government would intervene to protect their interests, then they also
realized that they were more likely to be paid off. Hence, they were more eager
to lend to Egypt than to a government in which the British government was
less interested and hence less likely to intervene. In a similar manner, if investors
believe that the IMF will come in and bail out the country, at least for a while, the
magnitude of the investors’ losses in the event of a crisis will be smaller.
To see how this works, assume that investors think this bailout will reduce
their losses by 50% in the event of a default. Then, their bond price function for
government debt changes from (19) to

10. See the discussion in Wikipedia under “History of Egypt under the British” for more informa-
tion on this period. For more amusing reading about this period, see the Mamur Zapt detective
novels by Michael Pearce.
Debt, Default, and Interest Rates 237

1
q= ∑ [𝛿(B2 , Y2 ) + (1 − 𝛿(B2 , Y2 ))/2] Π(Y2 |Y1 ),
R Y
2

This reduces the price reduction from default risk. In effect, default is being
subsidized by the IMF. This subsidy then encourages governments to take these

Downloaded from https://academic.oup.com/book/35105/chapter/299187205 by Universiteit Leiden - LUMC user on 01 May 2025

risks that can raise the likelihood of default because lenders are now willing to
lend to them on better terms than before if they do so. The moral hazard aspect
of IMF interventions has been a topic of concern for a long time. It was also
an important issue for the EU and was the reason for requirements on fiscal
discipline within the EU that sought to limit deficit spending.
Downloaded from https://academic.oup.com/book/35105/chapter/299187205 by Universiteit Leiden - LUMC user on 01 May 2025