c
?

 


     

c
?

?  
  




c
  ?

  
 

c
 



 
  
 !!
   
 

c


"

 
       
 


c

?## $" % &&'()"   !
 


c

)*
?## $" % && " % &+
 

# , $


# 
#
  ,- .

c
   &#

  ,% 

c
  # ,

" % &&



























How Large is the Outstanding Value of Sovereign Bonds?
jebt issued by governments worldwide is immense. According to the Bank for International
Settlements, at year end 2009 worldwide sovereign debt exceeded $34 trillion, and is
greater than the amount of corporat e bonds outstanding.

Japan and the US dwarf most other borrowers. Together they have about half of all
sovereign debt worldwide. Still, 23 other countries have over $100 billion of debt
outstanding. The other 100+ countries worldwide have a total debt of about $1.4 trillion.

Click on graph for larger image in new window.

Note: This graph shows the sovereign debt in jecember 2007 and jecember 2009.

jue to the recession and increased expenditures to rescue banking systems, total
sovereign debts grew by almost 30% in just two years. Sovereigns became the majority of
worldwide debt. Several countries doubled their debts from 2007 to 2009 (BIS
data).Source: Bank for International Settlements (BIS)

*For the US, figures include public holdings of Treasuries, but not Fannie Mae or Freddie
Mac (about $8.1 trillion year end 2009, per BIS), or the ©intragovernmental holdingsª of
Social Security, Medicare, the Civil Service Retirement Fund, e tc. (about $4.5 trillion year
end 2009, per US Bureau of Public jebt).
When shown as a percent of GjP, the picture looks a bit different. Japan and Italy have
both a large amount of debt in absolute terms, and as a % of GjP.

The United States has a more moderate debt as a % of GjP.The third graph shows the
size of sovereign debt compared to equities and other bonds.

Because of its immense size, sovereign debt is one of the largest risks to the global
financial system. There are many linkages to sovereign debt, including interest rates,
exchange rates, bank debt, and credit default swaps. Many of the potential problems and
risks are surprising, even to those well-versed in their particular area of finance.
Part 2. How Often Have Sovereign Countries jefaulted in the Past?
Sovereign bonds have been defaulting for almost as long as there have been sovereign
bonds. The problems go back many centuries. A good overview created for the IMF is ©The
Costs of Sovereign jefaultª by Eduardo Borensztein and UgoPanizza. Some countries are
©serial defaultersª, with a long history of sovereign defaults. Many have defaulted on
sovereign debt five times or more.

Here is a chart showing the number of countries defaulting each year from 1824 to 2003.
The raw data comes from S&P. Charts were created by the Some Investor Guy.

As you can see, there are some years with no defaults at all, and other years with many.
jefaults tend to come in clusters, and the behavior of lenders often changes substantially
after defaults. In the Volatility Machine, Michael Pettis asserts that sovereign default
contagion follows predictable patterns, and that contagion is primarily due to investo rs in
the first defaulting country also having investments in other countries which are vulnerable.
This is especially the case with leveraged investors.

In the seemingly ©quiet periodª from 1945 to 1959, there was just one sovereign default.
Interestingly, this was also a time with a number of very angry foreign investors. This time
period was the peak of expropriation of foreign assets. There were at least 25
nationalizations and expropriations of foreign assets. Many were by new members of the
Soviet Bloc, and by newly independent colonies (Source: Michael Tomz, Stanford, working
paper).

For you ubernerds who want to see which countries defaulted each year, here they are.
I¨ve broken them down into three periods to make the charts more readable.

1820 to 1920(click on chart for larger image)

1920 to 1980

1981 to 2003
The underlying causes of default (such as rises in interest rates, wars, commodity price
collapses, and simply borrowing too much money) have been diagnosed for many episodes.
Proximate to the default, any of the following six financial changes might occur:

1. Government revenues fall far below history or forecast;

2. Expenses aside from debt service rise far above history or forecast;
3. Interest rates rise substantially; due to inflation, credit spreads, illiquidity, or other causes
4. jemand for bonds suddenly drops or disappears (a sudden stop);
5. Exchange rates move, making payments on foreign denominated bonds much more
expensive (currency risk), and,
6. A government simply decides not to pay, even though it has the capacity to pay
(repudiation).

Paolo Manasse and NourielRoubini studied sovereign default risk and concluded that many
guidelines used for estimating when default was likely did not perform well, primarily
because those guidelines looked at separate risks. For example, total government debt
exceeding 200% of GjP is often used to indicate stress. However, some other
circumstances may make the problems much less severe (like having a growing economy
and no foreign denominated debt). Other factors might make it much worse (like high
inflation).

Part 2B: More on Historic Sovereign jefault Research

Some researchers, especially Reinhart and Rogoff, assert that ©this time is not differentª,
and that rather similar things occur before and after defaults throughout the world and over
a period of many centuries. Politicians might or might not know the his tory. However, one
wonders to what extent bond traders and CjS market participants agree with the
academics.

Within the bond market, CjS market, and the governments themselves, there is also a
chance that someone knows important information regarding the chance of default that
others do not. For example, an official misstating government revenues, reserves, or
borrowing.

There is a very active effort in many countries to convince voters or investors that
©everything will be fineª or ©all debts will be p aid on scheduleª. Often, these claims are
correct. If interest rates on government bonds and liquidity are set based on perception of
risk, the spin might save a sovereign billions of dollars in reduced funding costs. If the
spin loses all credibility, a p articular government official might be regarded as ©The
Baghdad Bob of Bondsª (Term derived from the nickname for Saddam Hussseins¨s
Information Minister, ©Baghdad Bobª).

The Research

Paolo Manasse and NourielRoubini studied sovereign default risk and c oncluded that many
guidelines used for estimating when default was likely did not perform well when used in
isolation, primarily because those guidelines looked at separate risks. For example, total
government debt exceeding 200% of GjP is often used to in dicate stress. However, some
circumstances may make the problems much less severe, others might make it much
worse (©Rules of Thumbª for Sovereign jebt Crises , Paolo Manasse and NourielRoubini):
©The analysis has one important, albeit simple, implication for sustainability analysis. It
shows that unconditional thresholds, for example for debt -output ratios, are of little value
per se for assessing the probability of default. One country may be heavily indebted but
have a negligible probability of default, while a second may have moderate values of debt
ratios while running a considerable default risk. Why? Because the joint effects of short
maturity, political uncertainty, and relatively fixed exchange rates make a liquidity crisis in
the latter much more likely than a solvency crisis in the former, particularly if the large
external debt burden goes together with monetary stability, a large current account surpl us,
and sound public finances.ª
Chart source Predicting Sovereign jebt Crises , page 30. For specific countries, default
probabilities from this model could be updated as often as the relevant economic data are
updated.

In the earlier paper for the IMF, they found ( Predicting Sovereign jebt Cr ises, Paolo
Manasse, NourielRoubini, and Axel Schimmelpfennig):

©The empirical evidence suggests that a number of macroeconomic factors predict a debt
crisis and the entry into a debt crisis. Measures of debt ©solvencyª matter: high levels of
foreign debt (relative to a measure of the ability to pay, such as GjP) increase the
probability of a default and entry into default. Measures of illiquidity, particularly short -term
debt (relative to foreign reserves), and measures of debt servicing obligations also matter
in predicting debt crises, consistent with the view that some recent crises had to do with
illiquidity and/or the interaction of illiquidity and insolvency.

Other macroeconomic variables suggested from the analytical literature on debt

sustainability also significantly matter for predicting debt crises: low GjP growth; current
account imbalances; low trade openness; tight liquidity and monetary conditions in the
Group of seven countries; monetary mismanagement (in the form of high inflation); polic y
uncertainty (in the form of high volatility of inflation); and political uncertainty leading to
economic uncertainty (years of presidential elections).ª

Another line of research comes from Carmen Reinhart and Kenneth Rogoff. They have
several relevant works, including This Time is jifferent: Eight Centuries of Financial Crises .
©We find that serial default is a nearly universal phenomenon as countries struggle t o
transform themselves from emerging markets to advanced economies. Major default
episodes are typically spaced some years (or decades) apart, creating an illusion that ©this
time is differentª among policymakers and investors. A recent example of the ©thi s time is
differentª syndrome is the false belief that domestic debt is a novel feature of the modern
financial landscape. We also confirm that crises frequently emanate from the financial
centers with transmission through interest rate shocks and commodit y price collapses.ª

Eight Centuries of Financial Crises also contains quite a bit of data on inflation, which the
authors view as a second way to default (if the bonds are denominated in the issuer¨s
native currency), and exchange rate problems.

Historically, we will see that the average percent of sovereigns in default or restructuring
have sometimes been quite large (source: Eight Centuries of Financial Crises, page 4).

Rogoff also has analysis of what happens with high debt levels when there is not default.
Some readers will enjoy their paper from early 2008, ©Is the 2007 U.S. Sub-Prime Financial
Crisis So jifferent?ª

We will revisit both sets of authors when looking at the indirect effects of default later in
the series. Numerous posters have mentioned Rogoff research showing a possible dropoff
in growth above a threshold of 90% debt to GjP (CR Note: this level is disputed).

Part 3. What are the Market Estimates of the Probabilities of jefault?

There are a number of ways of looking at chances of default and/or expected losses,
including: bond yields vs a low or no default bond in the same currency, credit default
swap prices, bond ratings, and analysis of underlying financial factors.
Bond ratings move more slowly than bond yields or CjS prices. Ratings often are lowered
only after a major problem has been realized and is already incorporated into yields or
CjS prices. While bond prices can be useful, there are an assortment of problems of
trying to extract default probabilities. One is the yield curve, and that for many sovereigns
there aren¨t all that many maturities outstanding. Trying to get a 5 year probability of
default from a dataset including only a 2 year and 20 year maturity presents s ome analytic
problems. Bond prices also have a surprising amount of differences due solely to liquidity.
For example see Longstaff.

©We find a large liquidit y premium in Treasury bonds, which can be more than fifteen
percent of the value of some Treasury bonds. This liquidity premium is related to changes
in consumer confidence, the amount of Treasury debt available to investors, and flows into
equity and money market mutual funds. This suggests that the popularity of Treasury bonds
directly affects their value.ª

Yes, that¨s 15 percent between different US Treasuries and a series of bonds explicitly
guaranteed by the US govt, the 1990s Resolution Trust. Even o n the run and off the run
US treasuries have different yields due to liquidity.

Because it provides daily information for almost all large sovereigns, and calculates
cumulative probabilities of default (CPj), we use data from CMAVision to estimate
sovereign default probabilities.

This chart shows outstanding debt with the (CPj) for each country. jespite it having a
moderate 8.3% probability of default, Japan¨s huge outstanding bond portfolio makes it the
largest contributor to expected sovereign losses . However, it¨s unlikely that any country
would only have a default on a small group of bonds. If Japan defaulted, it is likely that
most or all of its outstanding debt would be restructured (e.g., different interest rate,
extended payment, a haircut on pr incipal).

CPjs from 3/31/10 and 6/30/10 are shown in the next chart. The red bars are Q2, the
orange bars are from Q1. Obviously, some credit default swap prices moved substantially in
those three months, like Greece, Portugal, Spain and Belgium.

As of June 30, 2010, the weighted average expected default rate is 7.4%. When weighted
by value of debt outstanding, CjS pricing worldwide points to 7.4% of it defaulting within 5
years. If the outstanding sovereign debt was still $34 trillion as reported at 12/31/09, that¨s
$2.5 trillion of defaulted debt. If the trend of increased borrowing has continued to $36
trillion at 6/30/10, it¨s about $2.7 trillion of defaulted debt.

Before you run out and start shorting sovereigns or panic over your retirement, rem ember
that bondholders seldom lose all of their money on defaulted bonds. Sometimes recovery
rates are quite good. Others, not so much.

Part 4. What are Total Estimated Losses on Sovereign Bonds jue to jefault?

In Part 3, we showed that credit default swaps imply that 7.4% of sovereign debt will
default over the next 5 years. However, the defaulted bonds probably would not lose all of
their value.

In some cases, losses given default are quite small, a few percent. In some restructurings,
only the durations are changed, and a technical default might not result in an actual loss.
At the opposite end of the spectrum, losses on a particular bond could b e 100%,
especially if a sovereign had junior or subordinate debt.
To estimate losses, in addition to the Probability of jefault, we need to estimate Recovery
Rates Given jefault.

Moody¨s has studied sovereign default recoveries in the past several decade s
(Source:Sovereign jefault and Recovery Rates, 1983 -2008). Recovery rates have ranged
from 18% (Russia, 1998) to 95% (jominican Republic, 2005), when measured on all debt
from a particular sovereign at the time of default. The average recovery rate was 50%
when each country was weighted equally, but only 31% when weighted by the face value
of all outstanding bonds from all defaulting issuers.

Applying those recovery rates leads to expected losses of $1.3 to 1.8 trillion , or about 3.7
- 5.1% of outstanding sovereign debt at 12/31/09, and about $100 billion more at 6/30/10.
We now have our baseline estimate of worldwide sovereign default losses.

This might not sound too bad, unless you happen to live in one of the defaulted countries,
or own a lot of that country¨s bonds.

Even at the baseline $1.3 to $1.8 trillion, those losses would be about the size of all
outstanding debt of China, Germany, or France. And, default losses aren¨t the only losses
which could be occurring in your bond portfolio. For example, if Greek bonds default, Irish
interest rates might go up substantially, and Ireland might m erely be in pain rather than in
default. In that case, if you purchased Irish bonds before the crisis, their market value
might drop substantially below what you paid. Hopefully you didn¨t buy them using leverage.
If the bonds aren¨t denominated in your na tive currency, you could be experiencing large
FX losses.

Ubernerd warning. This part of the post contains pictures, but is also pretty technical .

Some posters have asked, ©why not use the CjS prices themselves, which contain
information on both the defa ult probabilities and expected recoveries?ª Well, if the two are
separated, you are able to do certain things in modeling that are impossible to calculate
with them jumbled together. It is easier to compare to historic data which often tracks the
occurrence of default, but not the recovery. It allows you to do frequency -severity
simulations.

For you ubernerds, it also means that you can model and correlate frequency and severity
separately. Would you change your estimate of severity of loss on Greek bonds if you
thought Spain and Italy would also default? jo you think that the probability of default
goes up each time debt needs to be rolled over, but the recoveries given default won¨t
move much? Then the common assumption of 40% recovery rate doesn¨t work w ith your
model.
Here is a chart derived from a large bank¨s presentation on sovereign CjS modeling which
shows just how different your cumulative probability of default could be for a sovereign like
Greece which closed at 811 basis points yesterday (Ju ly 13, 2010). The differences get
pretty big, pretty fast. I¨ve used the same historic range of 18% to 95% recovery rates to
illustrate this. It¨s pretty obvious that the CjS market doesn¨t expect 95% recoveries from
Greece. However, there are many combina tions of probability of default and recovery rate
which are consistent with current pricing.

For those of you who followed the subprime crisis closely, some common CjS modeling
assumptions may sound familiar. My favorite misnomer is ©discount cashflows at the risk-
free rateª. There IS no truly riskfree rate. (CR note: by "riskfree", usually people mean
essentially default risk free like U.S. treasuries. "Some investor guy" is pointing out there is
still interest rate risk).
There might be interest rates on a security incredibly unlikely to default. However, interest
rates on that security will still move around, introducing plenty of risk which has nothing to
do with default. jo you think that commonly used ©riskfreeª rates like 30 day LIBOR or 90
day US Treasuries might fluctuate and diverge in a crisis? Like they did in 2008 and 2009?










O
O
  


There appears to be another market behavior which you might not expect unless you had
seen it before: very strong correlations of securities which don¨t appear to have the same
underlying risks. Take a look at these charts on CjS prices for example. These may be
artifacts of how people are funding, leveraging, or hedging their positions.

O
O
  

The short term correlations are so strong that only events like Hungary¨s ©whoops, we
didn¨t really mean thatª event stand out.

Source: CMA Sovereign Risk Report for Q2 2010

The strong correlations aren¨t just in prices, those correlations also occur in trading volume.
(see chart 4 below) It would not be surprising to find correlation desks at major traders
generating lots of this volume. Who is trading, why, and how, are important in guessing
what might happen if things g o really badly.

Source: CMA CjS Liquidity Study

Part 5A. What Happens If Things Go Really Badly? $15 Trillion of Sovereign jebt in jefault
In order to keep the Really Bad scenario from becoming massively long, it¨s in several
pieces. This one is only on sovereign defaults. There will be more on swaps and banki ng
later in the week.

How likely or unlikely is the ©Really Badª scenario?

The author has built a number of risk simulation models for various purposes. Choosing
scenarios or simulation outputs which show things going really badly involves some
tradeoffs. There is always that small probability of Earth being hit by a huge asteroid
resulting in mass extinction. I usually leave such events out of my models. Wars and
disease are tough items to model, but government financial problems are often relat ed.
More on those in Part 6.

There is a strong tendency among people who are not researchers or modelers to think a
particular bad scenario can¨t or won¨t occur. If you want to get attention from management
or regulators and motivate them to do something , one effective technique is to show what
would happen if something very similar to prior history repeated itself. When they say
©there¨s no way that many sovereigns could default at onceª, you pull out the numbers and
show them that it¨s actually occurred several times in the past.

People who don¨t do statistics are usually drawn to individual stories or scenarios. I look for
scenarios which are informative about risks that might be mitigated in some way. It¨s
frustrating for everyone involved if they be lieve what you are presenting, but then can¨t do
anything about the risks or other problems.

So, out of the multitude of potential scenarios, I have settled upon one which is really bad,
but doesn¨t involve asteroids, mass extinctions, or apes taking over . It is consistent with
prior bad episodes of sovereign debt default.

Here is the Really Bad scenario. It¨s not a worst possible scenario. It is more like the Long
jepression or the Great jepression reoccurring under 2010 conditions.

In the Really Bad scenario, 45% of the countries with large outstanding sovereign debts
are in default within a 2 -3 year period.
Click on graph for larger image in new window.

As we saw in Part 2B, levels between 40% and 50% of sovereigns in default have been
reached five times in the last two centuries, (1830s, 1840s, 1880s, 1930s, 1940s, source:
Eight Centuries of Financial Crises, page 4).

So, who defaults? A simple method is to choose the 45% of countries with large sovereign
debts (over $50 billion) that currently have the highest cumulative probability of default.
They are assumed to default in the same order as implied by their cumulative probability of
default at 6/30/10 from CMA: Greece, Argentina (again), Portugal, Ireland, Spain, Italy,
Turkey, Indonesia, Belgium, South Africa, Thailand, South Korea, Poland, Brazil, Mexico and
Malaysia.

This involves about $5.6 trillion of debt in default, about 16% of all soverei gn debt. If
historic trends repeat themselves, it all happens within about two years of the first default
(Greece), and 11 home currencies are involved. At the low end recovery rate of 31% of
face value, there are about $3.8 trillion of losses. This is abo ut 2-3 times the amount
currently embodied in credit default swap pricing which we calculated in Part 4 ($1.3 -1.8
trillion).

But then, since this is a really bad scenario, Japan defaults too. This might occur because
of a global economic slowdown, a rise in general risk premiums and interest rates raising
Japan¨s debt service (this could take longer, Japan¨s average maturity is 5 -6 years),
Japan¨s banking system being affected by defaults elsewhere in the world, lack of political
will to make reforms, or several other mechanisms.

For those of you who say Japan¨s default is an incredibly unlikely event, note the following.
Many officials have a long term concern about Japan. If there are many other sovereign
defaults, the time frame could speed up consider ably.

1. Prime Minister Naoto Kan said ©We cannot sustain public finance that overly relies on
issuing bonds," Kan told parliament in his first policy speech. As we can see in the euro
zone confusion that started from Greece, there is a risk of default if the growing public
debt is neglected and if trust is lost in the bond market."
2. In the last year, the price of Japanese CjS have tripled. CMAVision estimates a 5 year
probability of default of 8.3% from their CjS prices at June 30, 2010.

3. Japan has the highest debt to GjP ratio of any developed country. It is expected to
reach 225% of GjP in 2010 (Source: The O utlook for Financing Japan¨s Public jebt, IMF
2010)

4. ©Japan is in danger of a slow motion sovereign debt crisis 5 to 10 years from

now,warns Kenneth Rogoff, professor of public pol icy & economics at Harvard University.ª

5. The IMF modeled the effect of sovereign defaults elsewhere on Japan. They found
©Sovereign debt crisis and increased risk premium. Assessing the macroeconomic
implications of a sharp rise in government bond yiel ds, this scenario assumes an increase
in the risk premium by 100 basis points for the U.S. and 200 basis points for the euro
area without offsetting fiscal policy. If the risk premium in Japan remains unaffected,
growth would slow through the export and ex change rate channel by between 0.5 to 1
percentage points in 2010. If, on the other hand, the risk premium in Japan also increases
(by 100 basis points), growth could fall by as much as 2 percentage points. Given
depressed demand, deflation would worsen by about 0.5to 1 percentage point below the
baseline in 2011.ª (source: IMF, Japan Staff Report for the 2010 Article IV Consultation,
June 2010).

Japan has over a quarter of all sovereign debt outstanding worldwide. It¨s default would be
bigger than all of the others in this scenario combined.

In ou
In our r Really Bad scenario, another $9.7 trillion in sovereign debt goes into default. The
total debt in default reaches $15.3 trillion, and almost half of all outstanding sovereign debt
is in default. The losses are $10.5 trillion at the low end recovery rate of 31%.

At the low end, losses are $7.5 trillion (50% of face value). Of course, recoveries in some
countries will be higher than others, but you get the general idea.

Part 5B. What Happens If Things Go Really Badly? More Things Can Go Badly: Credit
jefault Swaps, Interest Swaps and Options, Foreign Exchange

In Part 5A, I showed a Really Bad scenario consistent with some very bad historic default
rates for sovereign debt. That produced an estimate of $15.3 trillion of def aulted sovereign
debt, with $7.5 to $10.5 trillion of losses.

In today¨s post, we look at the effects of sovereign default on credit default swaps, interest
rate swaps and options, and currency exchange contracts.
Sovereign risk can make its way into over the counter markets in many ways, including:

A. For credit default swaps, the sovereigns are often the ©reference entityª. In other
words, market participants are often buying and selling insurance related to a potential
future default of a country on its debt.

B. Sovereign bonds are a major form of collateral for over the counter trades. The
collateral itself is often marked to market. Thus, movements in interest rates often changes
the value of highly rated sovereign bonds posted as collate ral, and a participant might post
more or fewer bonds as collateral as a result. However, sovereign bonds are typically
subject to larger haircuts if downgraded, and might not be accepted at all if rated lower
than BB-.

C. Sovereigns ARE the particip ants in a large number of contracts. Yes, many
governments and/or their central banks or sovereign wealth funds are active participants in
OTC markets. The author¨s impression is that sovereign participation in the CjS market is
low, but participation in i nterest rate and FX markets is considerably higher.

OTC markets are huge. Their notional values are over $600 trillion, and exceed all other
forms of investment combined (source: BIS Quarterly Review, June 2010 ). juring the 2008
crisis, the market values of CjS, interest rate derivatives, and FX contracts was about the
same as all outstanding sovereign debt, or all outstanding equities.

There are also exchange -traded interest and FX futures and options, about $67 trillion in
2009 (source: BIS Quarterly Review, June 2010, page 126 ).

Unlike stocks and bonds, the total market values for the OTC rise during times of turmoil.
That¨s partly because when interest rates or FX move from their values when the contracts
were entered into, one party to the transaction now has a positive mark to market value
for the swap (the other counterparty has a negative mark to market value).
From mid 2007 to the end of 2008, the market values of FX and interest rate derivatives
tripled, even though their notional values moved by less than 9%. The market value of CjS
went up by a factor of seven, on an almost unchanged notional value .

Credit jefault Swaps.

The next chart shows the ratios of gross and net notional to current outstanding bonds.
Unlike the case for many corporate borrowers, for sovereigns it¨s rarely the case that even
gross notional CjS values exceed sovereign bonds outstanding. It¨s even rarer for net
notional to exceed outstanding sovereign bonds.

Good to know that CjS are unlikely to make sovereign default much worse, right? Well,
these are recent CjS numbers, but take note. The ratios of CjS to bonds outsta nding are
higher for those countries believed to be in the most trouble. If sovereign financial
conditions got much worse, the volume of CjS could go up considerably. In our Really
Bad scenario, I assume the amount of CjS outstanding on sovereigns triples, and about
half of it pays off. Currently there are about $2 trillion of sovereign CjS outstanding. Let¨s
assume it goes to $6 trillion of notional, that half of that pays off due to the insolvency of
reference entity (the country the CjS is written on), o r $3 trillion. That produces $3 trillion
of payments, but there is a lot of netting. Even with some counterparty problems, net
losses will probably not go over $1 trillion. Let¨s assume $500 billion to $1 trillion of net
payment and losses due to counterpa rty problems. Not too bad next to $7.5 to $10.5
trillion of losses on defaulted bonds.

Interest Rate and FX jerivatives.

Here the situation is much different. The total notional value of OTC interest rate and
derivatives is $500 trillion (half a quadril lion). It is many times the size of the sovereign
bond market.

Interest rate and FX derivatives typically start out at inception with a zero market value.
Over time, as interest rates move, a contract acquires a market value. Sometimes, those
market values are very large. In the 2008 distress, market values rose to $32 trillion from
$11 trillion just 18 months before (Source: BIS, Table 19). Much of that market value is
collateralized. Much of it is not, especially for sovereigns.

juring the 2008 market turmoil, all kinds of typical relationships between rates changed.
Significantly, there were no sovereign defaults during that period. A sovereign default would
have made this period even more volatile.

juring late 2008, huge amounts of collateral were s ucked into the over the counter
markets. With numerous sovereign defaults, this would likely happen again, probably on a
bigger scale. There are several major sovereign -related risks:

1. Sovereigns are frequently counterparties, and many of them typi cally don¨t post any
collateral. According to the ISjA Margin Survey 2010, exposures to ©sovereign governments
and supra-national institutions tend to have the lowest collateralization levels©. Only 25% of
what would normally be collateralized is collatera lized if the counterparty is a sovereign.
Thus, a highly-rated sovereign, its central bank, or another quasi -sovereign entity could
accumulate very large risk positions with no collateral. For example, a AAA -rated sovereign
might slide to lower ratings and eventually to default, leaving an immense amount of
collateral and derivatives payments due.

joes that story sound familiar? It¨s what happened to AIG. AIG went from AAA rated in
early 2005 to being majority owned by the Federal Reserve in its first bail out in September
2008.

AIG had some poor risk management decisions, aside from selling mostly unhedged credit
default swap protection. ISjA asserts that
©Apparently, AIG relied excessively on a credit risk model that did not adequately account
for both the sharp decline in the mortgage market and a downgrade of AIG's credit rating.
It has been said that AIG would never have needed government assistance if not for its
investments in credit default swaps. Perhaps that is true. But it would be more accurate to
state that AIG would never have needed government assistance if had not so heavily
exposed itself to mortgage backed securitiesª
(ISjA, AIG and Credit jefault Swaps, 2009 ).

Good to know that sovereigns don¨t provide unhedged credit guarantees, right? Among
many other guarantees, it is common for sovereigns to guarantee bank deposits.

If interest rates or exchange rates moved quickly near the time a sovereign was
downgraded or defaulted , a number of in-the-money swaps could be anywhere from
worrisome to worthless to the other counterparty. Cash is an acceptable form of collateral,
so some sovereigns who control their own currencies might print money, though printing
really large amounts of it could have unintended consequences. Countries on the Euro
don¨t have the choice of printing their own money. Often, the native currency is not the
one the OTC contract is written in, so any money printed might have to be converted to
dollars, euros, or yen.

2. Sovereign bonds are a major form of collateral. As downgrades occur, haircuts

become larger, meaning that for every $100 million of collateral required, a counterparty
might have to post $112 million of bonds with a BB rating. Then, when t he bonds go
below BB- (typical Basel II collateral requirements ), it wouldn¨t be acceptable at all.

3. Sovereign distress and default would likely cause a breakdown between hist oric
relationships and term structures. The relationships between short and long term rates in a
particular currency could move to a strong preference toward short term debt, with long
term rates going up. The relationships of interest rates in different c urrencies could move
considerably. In Part 5C on banks, we¨ll also discuss how LIBOR departed very substantially
from its usual relationships in 2008, and in a period of distress this could easily happen
again.

4. Many interest rates for OTC contracts are directly set based on a sovereign
interest rate, or are strongly influenced by them. Thus, a contract between two banks
might be based on a rate which becomes very volatile, or unglued from its historic
relationship.

Having defaulted or distressed sovereigns as counterparties could cause cascading defaults.

jowngrades of sovereign debt being used as collateral could cause a massive rush for
debt which is still highly rated, and cash. Remember that not only will traders have to
replace downgraded deb t with other collateral, downgraded debt will lose a big portion of
its value and is affect traders¨ balance sheets, and the entire system will need much more
collateral.

Total Current OTC Collateral, Current Credit Risk

At year end 2009, there was only about $3.2 trillion of collateral posted, on over $600
trillion of notional value contracts, ISjA Margin Survey 2010.

With $22 trillion of gross market value of derivatives at year end 2009, BIS shows ©gross
credit exposureª of $3.7 trillion (Gross market values after taking into account legally
enforceable bilateral netting agreements), BIS Quarterly Review, June 2010Much of the half
trillion of uncollateralized exposure is sovereigns. While I have no reason to doubt the BIS
data collection on this item, the $3.7 number is not what it might seem to be.

Stress Cases

If you were to assume that there was only $3.7 trillion of credit exposure in the OTC
derivatives market, you would be greatly underestimating.

Why? 1. The BIS number assumes that all expected netting actually occurs. These are
bilateral contracts. It is not a given that netting will function properly w hen there are some
big defaulted counterparties and quickly rising collateral requirements. 2. The BIS number
would be more appropriately titled ©gross credit exposure at current market valuesª. If
market values change wildly due to movements in interest r ates and FX, the market values
of the OTC contracts will rise considerably. Vastly more collateral and/or credit exposure
would now be booked at the new prices.

How much more collateral? Between mid 2007 and late 2008, the market value of OTC
derivatives rose by $21 trillion (source: BIS). The ©gross credit exposureª rose by $2 trillion.

A couple of analogies may help. If a bank looks at a line of credit for a borrower, it has
two different risks. First, the risk that the borrower won¨t repay part of all of what they
have currently borrowed. Second, the risk that they will borrow considerably more, even
max out the credit line, and then not pay that amount. The second one is an extreme
stress case.

Another example is even more informative for OTC derivati ves. Credit default swaps are
similar to insurance in many ways. What many people miss is that interest rate and FX
derivatives are also similar to insurance in many ways. They are much like the way that
property insurers transfer risk around between vario us companies. Writing homeowners
coverage in an area prone to hurricanes involves taking in risk. It is then sliced, diced,
moved around in various ways, and mixed in with risk from other areas and hazards. Of
course, regulators do tests for adequacy of ca pital.

In an analysis which sounds like you could simply substitute ©financial crisisª for ©hurricaneª,
Rawle King wrote,
©as development increased in coastal areas, a catastrophi c hurricane could result in huge
government outlays for disaster assistance and present insurers with significant financial
hazards, such as the risk of insolvency, a rapid reduction of earnings and statutory surplus,
forced asset liquidation to meet cash needs, and ratings downgrade ... for the very highest
layers of catastrophe risk, the government (and consequently the taxpayer) is now, by
default, the insurer of last resort.ª

©Insurers were caught off-guard by the large losses associated with Hurricane Andrew
because of significant errors in actuarial estimates of potential hurricane related losses.
Prior to Hurricane Hugo in 1989, the insurance industry never suffered any loss over $1
billion from a single hurricane. Further, most insurance industry ex perts estimated the
probable maximum loss (PML) for a single hurricane in the United States at between $8
and $10 billion, and that such an event would occur only once in a century. Hurricane
Andrew took insurers and forecasters by total surprise. In hindsight, because of the lull in
hurricane activity during the 1970s and 1980s, insurance policies were underpriced and
insurers accepted far more hurricane exposure than could be supported by their capital
resources (including reinsurance).ª

That was published several months before Hurricane Katrina.

The low hurricane period of the 1970s and 80s has a close analog in finance, with similar
results. The Great Moderation lasted from the late 1980s to 2007. In a speech in 2004,
Fed Governor Ben Bernanke said

©One of the most striking features of the economic landscape over the past twenty years or
so has been a substantial decline in macroeconomic volatility ...

I have argued today that improved monetary policy has likely made an important
contribution not only to the reduced volatility of inflation (which is not particularly
controversial) but to the reduced volatility of output as well. Moreover, because a change
in the monetary policy regime has pervasive effects, I have suggested that some of the
effects of improved monetary policies may have been misidentified as exogenous changes
in economic structure or in the distribution of economic shocks. This conclusion on my pa rt
makes me optimistic for the future, because I am confident that monetary policymakers will
not forget the lessons of the 1970s.ª

In the Really Bad scenario, we have passed the Great Moderation, and are currently sitting
in the relative calm between Hurricane Hugo and the much larger Hurricane Andrew (3x
Hugo), or the even larger Hurricane Katrina (6x Hugo).

In our Really Bad scenario, there isn¨t enough collateral to cover the gross market
values. If collateral was required on all positions of all traders and the market values
reached 3x their 2008 values, gross market value would be about $100 trillion dollars. That
would exceed all sovereign debt, all corporate debt and all outstanding equities. However,
even in a dysfunctional market, there is a lot of netting.

There might not be enough collateral even if netting. In recent years, the ratio of ©gross
credit exposureª to ©gross market valueª has been about 14 -22% (BIS data). Let¨s say that
there is a need for a maximum of $20 trillion of collatera l given the net positions. While
there is probably enough collateral worldwide to cover $20 trillion, it would undoubtedly
move market prices for acceptable assets. It would also be difficult to buy or borrow that
much in a short time period.

It¨s not just collateral. juring a crisis, huge amounts of money would be changing hands.
Many interest rate and FX contracts would have their scheduled payments.

An estimate of OTC losses.

While this is not a market value estimate like Part 5A, in the Really Bad scenario, defaulted
sovereigns, banks, and other counterparties might find themselves unable to make a
substantial part of their payments, or provide collateral. The IMF has discussed ©the
systemic risk associated with cascading counterparty failuresª in M aking Over-the-Counter
jerivatives Safer.

jespite some large efforts by academics, governments, and others to estimate stress case
losses, there are a myriad of modeling issues and choices. For the Really Bad scenario, I
use a similar assumption to the losses on sovereign bonds. 45% of the counterparties have
problems providing collateral or making their derivative payments as due. Losses to their
counterparties are 50 -69% of what is owed. That results in another $4.5 to $9.0 trillion of
losses on interest rates and FX, and $0.5 to 1.0 trillion on CjS. So far, we¨re at $12.5 to
20.5 trillion of losses in the Really Bad scenario

Such large losses require that in addition to the sovereigns themselves, a number of large
OTC traders would default. The largest traders include banks and other institutions which
have already received government rescues or substantial additional capital from
governments. The Really Bad scenario assumes that solvent governments do not cover, or
cannot cover, many of those traders¨ OTC losses. NourielRoubini andMichaelPettis have
interesting views on this. jue to OTC losses and direct losses on sovereign bonds, in our
Really Bad scenario many of those firms would be in bankruptcy, conservatorship,
receivership, or liquidation. As a result, massive changes would occur in the OTC market,
and financial systems as a whole.

What can be done? Can we just print money? Turn unacceptable collateral into acceptable?
Issue IOUs? Can governments change the terms of the OTC contracts? How about
austerity, would that help?

Full jisclosure: Some Investor Guy owns no credit default swaps, interest rate derivatives,
or FX contracts. This series is not investment advice nor an offer to buy or sell anything.
Unless cited to other sources, opinions are the author¨s. Like jus t about any other analyst,
he could be wrong. He might be right, but he is not psychic. He is not a regulator, nor is
he running for any elected office. If a regulator or researcher wants to talk about the
series, email CR. He will forward it to Some Investor Guy.

Sovereign jebt Part 5C. Some Policy Options, Good and Bad

Financial institutions, governments, and central banks can be creative in a crisis. Policy
options fall into three broad groups: options for deeply distressed or defaulted sovereigns;
options for creditors; and regulatory policies. In this part, we discuss the first two.
Regulatory options will come in a later part.

To some extent, we¨ve seen something similar to the Really Bad sovereign default situation
before in the 2008 crisis. There have also been several episodes of very bad sovereign
default over the past 200 years.

In our Really Bad scenario, about half of the sovereign debt is in default, and more would
be downgraded so that it isn¨t useful as collateral in many circumstances. In 2008, central
banks from many countries provided cash and guarantees to help with the crisis. What
happens when many of those governments and central banks are themselves in trouble?
That too has happened before, coincidentally often shortly after banking crises.

Click on graph for larger image in new window.

Source: Banking Crises: An Equal Opportunity Menace, Reinhart and Rogoff .

The authors claim that

©Banking crises dramatically weaken fiscal positions in both groups, with government
revenues invariably contracting, and fiscal expenditures often expanding sharply. Three
years after a financial crisis central government debt increases, on average, by ab out 86
percent. Thus the fiscal burden of banking crisis extends far beyond the commonly cited
cost of the bailouts.ª
Options for jistressed or jefaulted Sovereigns . What will sovereigns do? Some of the more
common or reasonable options which are available to sovereigns who are in distress or at
high risk of default include:

1. jialing for jollars (or Euros, Yen, or other currencies). Just like a college student who
is out of money, many countries will call their friends, relatives, and business associates .
They will be looking for gifts, subsidies, loans, guarantees, and cosigners. Sometimes,
having a big letter of credit and some liquidity buys enough time to actually fix things.

Greece called the EU and the IMF for help. We will probably see more of th at from other
countries. Many people are surprised to find that usually most of the loan commitments
made by the IMF are not drawn upon.

2. The most common option, for those in or near default is to negotiate with creditors .
This could involve combinations of reducing principal, extending maturity, and reducing
interest rates. For some sovereigns, there might be new debt in a different currency, like
jrachma, or Lira instead of Euros. For an interesting and detailed overview of options for
Greece, see How to Restructure Greek jebt, by Buchheit and Gulati.

3. Print money to pay debts. Not all countries can do this. Members of the euro can¨t
unilaterally print money. Greece can¨t just print its own money. Japan can. A large number
of countries can print their own money, but most of their debt is in d ollars or euros. For
example, many eastern European countries can print, but have debts in euros.

If they print, they have to convert that money to another currency. Historically, inflation has
not generally been a method of preventing defaults, but a s often occurred at the same
time or shortly afterward. (chart source: This Time is jifferent: A Panoramic View of Eight
Centuries of Financial Crises , Reinhart &Rogoff).

4. Print money, sell assets, or borrow from others to buy back your debts at a large
discount. No need to default.
Usually governments try to calm bond market fears to keep their borrowing costs low.
However, if many owners of your debt are very worried, are scrambling for cash, or are
willing to part with bonds priced at 50 cents on the dollar, it is the author¨s opinion that
you should strongly consider taking the deal. Those are likely to be medium to long term
bonds, and you could cut your deb t service substantially for years, or even decades.

Ford did a version of this in 2009 and paid less than half of face value. Here is
anoverview. Fabulous idea in the right circumstances (even though one of the rating
agencies completely missed the point and thought it was equivalent to selective default,
they clued in later). The hard part of doing this is usually finding enough cash to buy back
debt at a deep discount. Ford had plenty of cash, always a good thing to have in a
crisis.

If you happen to be a government with some very marketable assets lying around (e.g.,
gold, oil, a pile of hard currency in a sovereign wealth fund) and you are not in default,
this could work out quite well. Brazil is a country that has successfully done this, and
gotten a rating upgrade .

Governments with no distress have also bought back bonds. In 2000, the US Treasury was
running a surplus (remember that? Seems so long ago), and bought back billions in 30
year bonds.

5. Austerity. Austerity is often useful for countries in moderate distress, especially if they
could balance their budget with debt restructuring. For countries who have already
defaulted, austerity is commonly part of emerging from default. Austerity is not useful if lot
of money is required quickly. That could easily happen when maturing debt needs to be
rolled over, and no one wants to buy it.

It could also happen on interest rate or FX derivatives. In a crisis, swap payments and/or
collateral could be rising quickly, and would be due on short time horizons. A few months
at the most, and sometimes a few hours.

And there are some slightly edgier options. Some of these might be fully justified in certain
circumstances. Others might just be ways of disguising default, seizure, or stupidity.

6. jefault only on the foreign debt. It is common for a country to issue multiple types of
debt, often with one class of debt mostly owned by residents, and another class mostly
owned by foreigners.
jomestic financial interests, voters, and even graffiti taggers often understand the
difference, and often put pressure on governments to ©stick it to the foreign investorsª.
Oddly, the opposite often occurs, with domestic debt being devalued along with the
domestic currency, and foreign debt remaining in dollars or euros.
Photo source.

7. Seize assets. Why would a debtor seize assets? Three common answers: because they
can; because they feel they ©have toª; or because they fee l they have been wronged. One
example is Argentina in 2001, which expropriated $3.1 billion of pension assets in exchange
for Treasury notes in jec 2001, and then defaulted anyway in Jan 2002. This chart shows
a timeline of how things went wrong.

8. Claim fraud and try not to pay, especially for OTC derivatives. The possible rationales
are endless. An example case from Italy is here. Prosecutor ©Robledo alleges the London
units of the four banks misled Milan on the economic advantage of a financing package
that included the swaps and earned 101 million euros in hidden fees.

He also claims the banks violated U.K. securities rules by f ailing to inform Milan in writing
that for the swap deal the city was a counterparty to the lenders rather than a customer.
Banks abiding by the rules of the Financial Services Authority are required to shield
customers from conflicts of interest and provide them with clear and fair information that
isn¨t misleading.ª The prosecutor seized assets from the banks equal to their share of the
alleged profit.

9. Claim that contracts were invalid in the first place . An example of this isIceland,
©Iceland¨s lenders may lose as much as $4.3 billion, equivalent to a third of the economy,
after a court last month found that some f oreign loans were illegalª. There may also be
sovereigns claiming that some official was not authorized to enter into particular loans or
derivatives.

10. Claim the people who entered into certain contracts were trying to cause default or
distress for the
sovereign. Germany made this illegal earlier this year.

11. jepose, subpoena, or even arrest some of the counte rparties. A number of countries
have rather different legal traditions than the US, the EU, or Japan, and lots of their jails
aren¨t up to Western standards either. Your typical counterparty doesn¨t have these kinds of
powers. Once a regulator or prosecuto r starts digging, they might find nothing. Or, they
might find something extremely damaging. Even charges unrelated to cheating a sovereign
could carry some serious jail time, like bid -rigging, money laundering, or tax evasion.

There are also some recent examples of central bankers or bank CEOs being arrested,
which might affect whether the sovereigns would honor contracts they entered into (Kosovo
Central Banker, Nigerian banks and customers, multibillion forged claims on UAE Central
Bank on two separate occasions). Here, here and here.

12. Cut loose sovereign wealth funds , quasi national organizations, and other borrowers or
CjS participants which are separate legal entities. jepen ding on ©moral obligationsª or
©implied guaranteesª is always risky. Investors in Fannie Mae and Freddie Mac bonds won
their bets that the US government would step in. This will not necessarily be the case
elsewhere.

13. And of course, do positive spin, even lie about financial condition. ©If a government
wants to cheat, it can cheat,ª said Garry Schinasi, a veteran of the International Monetary
Fund¨s capital markets surveillance unit, which monitors vulnerability in global capital
markets. Source.

Options for Creditors. For creditors, there are some interesting options. Common ones
include:

1. Reduce trade credit, try to get trade sanctions. These are very common measures.

2. Press for higher taxes, business-friendly regulations, austerity. These are straight out of
the traditional IMF playbook (now slightly modified).

3. Get your government to bail you out, perhaps by buying defaulted bonds above market
price. Another possibility for partial recovery is getting changes to the tax code, such as
allowing losses to be carried back, or offset against other forms of income.

4. Seize local assets. There is a long history of this, and sovereigns with extensive foreign
assets could be at risk. Recently there was an unsuccessful suit by a creditor to seize
assets of Argentina held in the US to compensate for defaulted Argentina bonds.

Somewhat rarer approaches include:

5. Convince your own government to sail gunboats into the harbor of the capital. Sell
tickets to the news media. Venezuela in 1913 was a prominent example. Two discussions
of how prevalent this method of collection once was or wasn¨t areat here and here. This has
been rare to nonexistent since World War I. However, looking at the earlier experience, it
would be a bad idea to already have major military powers upset at you and default on
your bonds. Killing or kidnapping foreign citizens was espe cially likely to get a response.
Having large oil or mineral reserves might also make a country a target.

6. Turn the borrower into part of another country. Yes, there are sometimes sovereign
consolidations due to finances. This happens more often at the municipal level, but in the
early 1930s Newfoundland went from semi -independent status to being part of Canada.
Much of the reason was Canada¨s willingness to pay most of Newfoundland¨s debt. More
details here. While this might occur somewhere at the country level in the Really Bad
scenario, it¨s more likely to happen with municipalities, provinces, or states.

jisclosures: At a prior job Some Investor Guy had a number of automakers as cli ents. He
owns some Ford bonds, some other corporate bonds, and some municipals. He owns no
foreign sovereigns, because he is expecting prices to fall and yields to rise on many of
them. He owns no gold, has very little in equities, some real estate, and a lot of cash.
Make your own decisions on what your portfolio should be. This is disclosure, not
investment advice. Some Investor Guy has no remote doomstead or bomb shelters, but
thinks everyone should have a week¨s worth of water and food. There are plenty of
disasters where those could be useful. He does sometimes talk to regulators and
lawmakers.