predicting default probabilities and implementing trading...
TRANSCRIPT
Predicting Default Probabilities andImplementing Trading Strategies forEmerging Markets Bond Portfolios 1
Stefania CiraoloUniversità di Verona
Andrea BerardiUniversità di Verona
Michele TrovaGruppo Monte Paschi Asset Management SGR, Milan
First draft: September 2001Current version: January 2002
1Stefania Ciraolo and Andrea Berardi: Dipartimento Studi Finanziari, Univer-sità di Verona, via Giardino Giusti 2, 37129 Verona. Michele Trova: Gruppo MontePaschi Asset Management SGR, via San Vittore 37, 20123 Milan. We are gratefulto seminar participants at the 2001 X International Tor Vergata Conference onBanking and Finance for useful comments and suggestions.
Abstract
In this paper we address two main issues: the computation of default prob-ability implicit in emerging markets bond prices and the impact on port-folio risks and returns of di¤erent default probability expectations. Usinga reduced-form model of the Du¢e-Singleton (1999) type, weekly estimatesof default probabilities for US Dollar denominated Global bonds of twelveemerging markets are extrapolated for the sample period 1997-2001. Theestimation of a logit type econometric model shows that weekly changes ofthe default probabilities can be explained by means of some capital marketsfactors. Recursively estimating the logit model using rolling windows of data,out-of-sample forecasts for the dynamics of default probabilities are gener-ated and used to form portfolios of bonds. The practical application of theimpact on portfolio returns of di¤erent default probability expectations pro-vides interesting results, both in terms of testing the ability of a naive tradingstrategy based on model forecasts to outperform a ”customized benchmark”,and in terms of the model ability to actively manage the portfolio risk (eval-uated in terms of VaR) with respect to a constant proportion allocation.
JEL Classi…cation: G33, G15, G11Keywords: Emerging markets, default probabilities, portfolio allocation
1 IntroductionIn the last decade, emerging markets have experienced numerous …nancialcrises. Recent cases include the Asian turmoil of 1997, the 1998 Russiandefault and, more recently, the downgrade of Turkish bonds and the crisisof Argentina, culminated in the January 2002 default. All these events havegiven rise to signi…cant contagion e¤ects among emerging markets.Estimating the default probability implicit in emerging market bond
prices has become extremely important for institutional investors (banksand mutual funds, in particular), given the relatively high weight reachedby these securities in their portfolios. In fact, the high yields o¤ered byemerging markets’ bonds with respect to those obtainable investing in gov-ernment securities make them particularly attractive. Those high returns aremainly explained by credit risk considerations due either to default events(issuer does not pay interest or principal or both) or market losses caused bymore or less frequent downgrading and subsequent bonds’ price volatility.Knowing the degree of con…dence …nancial markets are currently using to
discount a bond issuer’s default is, therefore, at a practical level, extremelyimportant under at least two di¤erent aspects. The …rst one consists in thecomputation of the risk the investor is undertaking over a given horizon.The second one regards the impact on portfolio risks and returns of di¤erentdefault probability expectations. In this paper we address both these issues.First, we extrapolate weekly estimates of default probabilities from a
reduced-form model of the Du¢e-Singleton (1999) type. The empirical workis based on US Dollar denominated Global bonds of twelve emerging marketsfrom February 1997 to July 2001.Then, we show that default probabilities can be predicted by some capital
markets factors (essentially interest rates, exchange rates and credit spreads)and use them as explanatory variables in logit type econometric models forthe prediction of the probability of a market increase/decrease in bond prices.In the …nal part of the paper, we recursively estimate a logit model to
produce out-of-sample forecasts for the probability of observing future appre-ciation/depreciation of the bonds. The practical application of the impact onportfolio returns of di¤erent default probability expectations provides inter-esting results, both in terms of testing the ability of a naive trading strategybased on model forecasts to outperform a ”customized benchmark”, and interms of the model’s capability to actively manage the portfolio risk (evalu-ated in terms of VaR) with respect to a constant proportion allocation.The paper is organised as follows. Section 2 introduces the main features
of emerging markets bonds. Section 3 illustrates the reduced-form valuationmodel used for the pricing of Global bonds. Section 4 describes the data used
1
in the empirical work and section 5 shows the underlying default probabilitiesextrapolated by the estimation of the model. Section 6 provides empiricalevidence on predicting the dynamics of default probabilities and on the per-formances, in terms of returns and VaR measures, of emerging markets bondportfolios built on such forecasts. Finally, section 7 contains some concludingremarks.
2 Emerging markets bondsThe genesis of the EmergingMarkets debt dates back to the early 1970s when,as a consequence of the 1973 oil shock, a great number of commercial banks inalmost all developed countries suddenly found themselves facing the problemof investing sizeable amounts of funds deposited by oil producers. At the sametime, on the other hand, increasing commodity prices led to a worldwideincrease in in‡ation rates and, therefore, to a generalised increase in theaverage level of interest rates. That was the reason that induced the afore-mentioned commercial banks to search for highly pro…table investments.The solution was somehow a ”fait accompli”: investing in those less de-
veloped countries whose fundamentals were improving thanks to the increasein commodity prices, and which were much more worried about the lack offoreign investments than about paying high interest rates. The economicagencies of those less developed countries, believing that commodity priceswould rise forever (remember the predictions that the world would have runout of oil in the 21st century), therefore, begun to borrow sizeable amountsof foreign debt.As commodity prices continued to rise through the early eighties, the loan
demand and supply also continued to rise, inducing a very dangerous cycle:the increase in commodity prices increased the Gross Domestic Product ofa country and therefore its capacity to service additional debt. When bankslooked at this newfound additional capacity they were willing to give loansto Emerging Markets on their demand.Unfortunately, at the beginning of the 1980s, commodity prices began to
fall destroying a sizeable portion of the Emerging Markets richness and leddeveloping countries to face the problem of a growing external debt, in thepresence of a reduced repayment capacity.Since the end of 1986 the average secondary market value of developing
country loans plummeted going from seventy percent (end of 1986) to thirtypercent (end of 1989) and showing a loss of credibility of the issuers in thatperiod. This was due to a growing default rate and a growing appreciationof the default probability among money managers invited to invest in the so
2
called high yield bond markets. The cited loss of credibility had disruptivee¤ects on the Emerging Markets; American banks reduced the share of devel-oping country loans in their portfolios and increased the ratio of bank capitalto such loans. The resulting sharp drop in …nancial resources dedicated todeveloping country loans caused serious problems to countries (such as de-veloping countries) used to repaying debt by contracting further debt. Bythe early 1990s the secondary developing countries bond market had reached(on average) the above-mentioned lows.The persistence of the debt problem forced the U.S. Government to pro-
pose a new debt initiative, the Brady Plan, in March 1989. Essentially, whatthe Plan did was to recognize that a full repayment of the debt for develop-ing countries was no longer a reasonable goal. In particular, it put pressureson commercial and investment banks to concede and to manage some formof debt and debt-service relief and also called for an increase in secondarymarket activity, in order to grant liquidity to these issues.The implementation of the Plan led to the birth of a new kind of Emerging
Market (high yields) bonds: the Brady Bonds.So far, several countries have bene…ted from the program issuing di¤erent
types of Brady Bonds.The majority of debt is from Latin America, withArgentina, Brazil, Mexico and Venezuela covering about 70% - 80% of theoutstanding market.Nowadays the emerging markets’ bonds market is capturing the interest
of both individual and institutional investors because of its uniqueness in atleast two respects: …rst, yields are extremely high, and, second, some issuesare very large and liquid (which was one of Brady’s main aims). Moreover,these features support an active over-the-counter derivatives market, so thatinvestors can take views on country risk, bond spreads or volatility, as well ashedging their own portfolios through the use of customized options and/orfutures.As widely known, developing (as well as developed) countries are used
to issue several types of bonds in di¤erent markets, aimed at covering theoutstanding principal amount of their bank loans and government debt andone or more bonds covering previous interest payments due, as well. Usually,we distinguish between Locally traded bonds (Government, Agencies, andCorporate securities quite illiquid, except for some issues, traded exclusivelyin the local bond markets and rarely present in institutional investors’ port-folios), Brady bonds and Eurobonds (otherwise known as Sovereign Bonds).The two main types of Brady bonds are par and discount bonds, 25 -
30 year registered bullet bonds representing the largest, most common andmost liquid issues in the Brady bond market.Par bonds are issued ”at par” in exchange for the original face value
3
of the rescheduled loans but carry a …xed, below-market interest rate. Asan alternative to the plain vanilla most common case, in some other casesthey also have some additional features complicating the computation of thepresent value of their cash ‡ows (for example the Mexican par bonds have 17series of Value Recovery Rights (VRR) which entitle the holder to additionalpayments linked to the price of oil).Discount bonds, instead, are bonds whose principal is a reduced frac-
tion of the original obligation (typically a 35% reduction is allowed by thecreditors) and that carry a ‡oating interest rate.Both par and discount bonds have generally a principal collateralised
by a U.S. Treasury zero coupon bond (rolling collateral) and the interest iscollateralised by cash deposits maintained at the Federal Reserve Bank ofNew York to cover a speci…c number of coupon payments (usually 12 to 18months).Among Eurobonds, the most liquid (and therefore reliable for our analy-
sis) type of bonds issued by developing countries are the so called GlobalBonds, usually long-term, plain vanilla, uncollateralised bonds, whose cash‡ows are easily computed and discounted at each point in time, the onlyneed being the term structure of risk-free interest rates and coupon paymentdates.Market prices of Brady discount bonds have been used in the empirical
investigation of Izvorski (1998), whereas Eurobonds have been used by Du¢e,Pedersen and Singleton (2000) and Merrick (2001).In this paper, we prefer to use Global bonds to extrapolate default proba-
bilities because of the following reasons: (i) the simplicity of the calculationsinvolved; (ii) the long-term view their pricing is based on; (iii) no assumptionson future term structures, future commodity prices or macroeconomic fun-damentals (implied forward exchange rates, future oil prices, GDP growth,in‡ation rates, etc.) are required; and (last but not least) (iv) the implied de-fault probability measure is somehow ”pure”, as Global bonds are (usually)uncollateralised.
3 The pricing modelSeveral models for the pricing of defaultable bonds have been proposed inthe literature. Usually, three main approaches are distinguished 1.i) Merton’s (1974) option pricing based model, which computes the payo¤
at maturity as the face value of the defaultable bond minus the value of a
1For a recent comprehensive review, see, among the others, Crouhy, Galai and Mark(2000), Gordy (2000), Jarrow and Turnbull (2000), Du¢e and Singleton (2001).
4
put option on the issuer’s value with an exercise price equal to the face valueof the bond.ii) Structural models, which relax one of the unrealistic assumptions of
Merton’s model, that is that default occurs only at maturity of the debt,when the issuer’s assets are no longer su¢cient to face its obligations towardsbondholders. On the contrary, they assume that default may occur at anytime between issuance and maturity of the debt and that default is triggeredwhen the issuer’s assets reach a lower threshold level (Black and Cox (1976),Longsta¤ and Schwartz (1995), Saa-Requejo and Santa-Clara (1997)). Thesemodels generally assume that debtholders, in case of default, get a fractionof debt’s face value back named ”recovery rate”, and that the latter is knowna priori.iii) Reduced-form models, which do not condition default explicitly on
issuer’s value, and therefore are, in general, easier to implement 2. Theyalso di¤er from typical structural models in the degree of predictability ofdefault. In fact, they are considered more general than structural models asthey can easily accommodate defaults coming as sudden surprises (see, forexample, Jarrow, Lando and Turnbull (1997), Du¢e and Singleton (1997,1999), Lando (1998), Schonbucher (1998)). Reduced-form models for pricingsovereign debt have been adopted by Du¢e, Pedersen and Singleton (2000)and Merrick (2001).Our pricing model is a discrete-time version of the Du¢e, Pedersen and
Singleton (2000) model and works as follows.Assuming no arbitrage conditions, the market price of a defaultable asset
should be a function of the default probability term structure, as well as ofthe future cash ‡ows discounted using the current risk-free term structure.Interpreting the coupon bond as a portfolio of zero-coupon bonds, we get
the following expression for the market price of a defaultable bond:
Vt =NXi=1
cti ¢ exp(¡rti ¢ ti) [(1¡ pti) + ± ¢ pti] (1)
where ti = 1; :::; N , indicates the time to i-th maturity, cti the i-th cash‡ow, rti the risk-free interest rate for the i-th maturity, pti = 1; :::; N , theprobability that default occurs between times ti¡1 and ti and ± the recoveryratio.Following Izvorski (1998) and Du¢e, Pedersen and Singleton (2000), we
assume that pti is constant over all time maturities and changes only forthe e¤ect of changes in the term structure of risk-free interest rates, or for
2This is particularly true when referred to sovereign debt.
5
the e¤ect of macroeconomic and/or political events that investors discountthrough prices.We also assume that, once a country defaults on some issues, just a frac-
tion (the recovery ratio) of both coupon and principal will be paid for bythe issuer in all the subsequent payments. This is a restrictive hypothesis,which does not allow us to account for the possibility that economic con-ditions could improve in the future and, therefore, the country repaymentcapacity be re-established. However, in our opinion, this drawback is a verymarginal one; in fact, under the hypothesis that a default occurs prior tomaturity, after the moratorium period, a new bond could be issued for anamount equivalent to the recovery ratio and for a maturity corresponding tothe old one. In this case, the old bondholder will freeze its …nancial situationalong the lines outlined above.A third assumption concerns the recovery ratio, which we assume to be
known and constant. Some more sophisticated models infer it from the his-torical recovery rates (those observed during past defaults) for identicallyrated issuers, some others describe the recovery ratio by means of a ran-dom variable. We believe that this is not a drastic drawback, since one canalso extrapolate default probabilities conditionally on di¤erent measures ofrecovery ratio.Given these assumptions, the following equilibrium relationship between
the market price of a defaultable bond and its expected cash ‡ows can bederived:
Vt =NXi=1
cti ¢ exp(¡rti ¢ ti)h(1¡ p)i + ±
³1¡ (1¡ p)i
´i(2)
where p = pti , i = 1; :::; N .Given the term structure of risk-free interest rates, the bond price and
the recovery rate, the equation above can be solved with respect to the prob-ability of default p. In our application, all computations are carried outconditionally on the recovery rate parameter ±. In evaluating the bonds,we adopt a conservative hypothesis and …x it equal to 20%, based on bondmanagers’ experience 3.From equation (2) we can recover, for each emerging market, the default-
able term structure:
yti = rti +1
ti¢ ln
h(1¡ p)i + ±
³1¡ (1¡ p)i
´i(3)
where yti is the credit-risky interest rate for the i-th maturity.
3See, for example, Xu and Nencioni (2000) on J.P.Morgan practice.
6
4 The dataIn the empirical work, we consider long-run (usually 10 to 30 years to matu-rity at the time of issuance) US Dollar denominated Global bonds of twelveemerging markets, namely Argentina, Brazil, Colombia, Mexico, Russia,Venezuela, Panama, South Africa, Turkey, China, Philippines and SouthKorea. The main features of these issues are reported in table 1.The sample contains weekly market values of Global bond prices (mean
of bid and ask quotes), ranging from 14 February 1997 to 27 July 2001 4.As we consider only US Dollar denominated emerging markets Global
bonds, USD Libor and swap rates are used to …t the risk-free term structurein correspondence of the payments dates. As for the Libor rates, we useall maturities between 1 month and 12 months, whereas for the swap rateswe include all maturities between 2 and 10 years and the 15, 20 and 30years maturities. The risk-free term structure of interest rates is obtainedestimating a two-factor version of the Cox, Ingersoll and Ross (1985) modelusing a maximum likelihood - Kalman …lter technique.
5 Extracting default probabilitiesFigure 1 shows the implied risk-neutral default probabilities estimated forthe countries considered in the sample. Table 2 contains some summarystatistics 5.As expected, the dynamics of the estimated risk-neutral probabilities re-
‡ect the evolution of both the political and macroeconomic situations ofthe di¤erent countries during a four years period (1997-2001) including deep…nancial and economic crisis, such as the Asian …nancial turmoil (1997), Rus-sia’s default (1998), the Turkish crisis (2001), Argentina’s recession (2001),as well as subsequent exceptional recoveries.The sample initially exhibits rather stable dynamics, with low levels of
implied default probability, which encouraged a low risk aversion attitude.Over this period investors commit the same error made almost twenty yearsbefore: they believe that emerging economies would not experience di¢cul-ties in repaying their external debt, and that they would bene…t from de-creasing world interest rates, therefore continuing to improve their economicfundamentals.
4The source of the data is Thompson Financial (formerly Datastream Ltd.).5Solving equation (2), which is non-linear with respect to p, requires the compilation
of a dedicated GAUSS program, involving the use of an ad-hoc version of the Newton-Raphson algorithm. See Trova (2000).
7
Unfortunately, this ”view” was undermined by disruptions in the balanceof payments, disequilibria in the gross domestic product growth (e.g. a realestate bubble in most of Asian countries), stock exchange bubbles and extra-ordinary overvalued real e¤ective exchange rates in several less developedcountries, all of which culminated in the outbreak of the Asian …nancialcrisis from mid 1997 onward (the mentioned outbreak dates back to July1997 when Thailand’s Monetary Authorities abandoned the pegged exchangerate system, leaving the Thai Baht to ‡oat against major exchange rates).As a result, investors all over the world re-evaluated the risk implied in
keeping Emerging Markets securities in their portfolios. This led to a di¤usedpanic reaction in late 1997, when World’s …nancial markets witnessed to apanic selling of Emerging Markets bonds and stocks in favour of safer assetsin developed countries (‡ight-to-quality e¤ect).During this period default probabilities for all the countries in our sample
increased (even if not too dramatically according to our estimates). Argenti-na’s implied risk-neutral probability of default almost doubled passing froma semi-annual 2% to almost 4%; Brazil’s more than doubled, increasing from2% to 5%; Mexico’s and Venezuela’s increased by 1% and 2.4%, while a lessdramatic impact of the mentioned crisis is seen in Ecuador 6 and Colombia.After this critical period, another period of relative calm preceded the
outbreak of a more dramatic and widespread Emerging Markets crisis in1998.Beginning in June 1998 with Ecuador (whose currency - the Nuevo Sucre
- was devalued in order to increase exports and thereby mitigate its severebalance of payments imbalances), a new …nancial turmoil (the most severesince the ”Great Depression” in the 1930s in peace periods), together withfears of a worldwide credit crunch following the failure of Long Term CapitalManagement (LTCM) in the United States and Long Term Credit Bank(LTCB) in Japan, infected the Emerging Markets, including Latin America.In July 1998 Venezuela, forced by a decrease in exports (due to decreasing
oil prices, one of the most privileged safe harbours in bad times like those),devalued its currency, the Bolivar, (again) in order to increase exports. Asa result, a general sell-o¤ occurred …rst in Latin America and then spreadto other Emerging Markets, especially to Russia, whose banking system’sfragility allowed for successful speculative attacks against the Ruble, by thatway forcing the subsequent default (declared in august 1998), even if limitedto the domestic debt.
6Although we do not include Ecuador among the twelve countries in our sample, weestimate default probabilities for this country (using the 11.25% Global bond expiring inApril 2002) for the period preceding default.
8
In this period, semi-annual risk-neutral default probabilities reached theirpeaks: 8.3% for Argentina, 14% for Brazil, 8.8% for Colombia, 16% forEcuador, 6.1% for Mexico and 26.2% for Venezuela (the most dramaticallyhit country). For this period it is virtually impossible to obtain a convergencefor our algorithm for the Russian bond if we consider recovery rates greaterthan 2%.History continues with another period of ”calm before (and after in this
case) the storm”. Another …nancial crisis was getting ready to a¤ect LatinAmerica once again: the Brazilian crisis of December 1998 - February 1999.The currency (the Real) was devalued after speculative attacks based onthe …nancial and …scal fragility of the Brazilian economy. As a result, thedefault probability peaked (even if at lower levels than the preceding ones)once more.As other countries, including Brazil, recovered late in 1999, this last crisis
was the ”beginning of the end” for Ecuador whose implied probability con-tinued to increase up to 45% in October, when default was …nally declared.The critical situation of the …nancial sector (especially the banking sector,
with its huge amount of non-performing loans, and its links, both directand indirect, with the political forces governing the country), the high levelof the in‡ation rate, the political uncertainty, and the unmatched requestsformulated by the World Bank and the International Monetary Fund forcritical but necessary reforms, were the motivations leading Turkey in thesevere crisis dated spring 2001. The markets, worried about the capabilityof the Turkish economic authorities to deal with the disappointing politicaland economic situation, and about the capability of the Turkish governmentto roll out the increasing public debt as well, increased the risk appreciationversus Turkish bonds, as shown by the time path of the default probabilityestimated by our model. After the recovery, subsequent to new loans bythe IMF (granted as counterpart of a plan of huge reforms involving thebanking sector as well as the new exchange rate management and the settingof in‡ation target policies), the re-appreciation of the risk implied by thepresence of Turkish bonds in institutional investor’s portfolios (early summer2001) was probably due (besides some doubts concerning the e¤ectiveness ofthe planned reforms) to the ”contagion e¤ect” induced by the outbreak ofthe Argentine crisis.Argentina’s troubles began as an e¤ect of the recession glooms involving
the most important trading (and supporting) partner: the United States,as well as of the continuous political uncertainty concerning the governmentof the country itself. The bond markets, once again, called for higher riskpremiums (and therefore higher implied default probabilities), starting fromJune 2001 and spreading from the domestic market to the Eurobonds and the
9
Brady Bonds markets. On the other hand, the International Monetary Fund(once again) called for more structural reforms involving the currency peg (a‡oating exchange rate could allow for a sort of ”amortization” of some of thenegative e¤ects of the current account situation, and the decreasing growthin gross domestic product, allowing for an upswing of the export sector), thebanking system and more stability in the political situation 7.Looking at the interrelation between default probabilities estimated for
the di¤erent countries, we observe that a ”regional contagion” e¤ect holds.Table 3 shows the high correlation degree, calculated over the entire sampleperiod, between the estimated risk-neutral default probabilities across coun-tries belonging to the same economic region. This is particularly true for themost economically homogeneous region in our sample, Latin America, withcorrelations around 70-80%.Applying a principal components analysis to the default probabilities es-
timated for the twelve countries in the sample, we can show that just twofactors can explain almost 70% of the total variability of default probabilities.Looking more closely at the distributional properties of implied risk-
neutral default probabilities, in particular observing the distances of min-imum and maximum values from the sample means and medians in table2, we can notice that the shape of such distributions looks far from beingsymmetric and is characterised by right fat tails. This intuition is reinforcedby the estimation of a non-parametric empirical probability density functionof the estimates. A normal kernel has been used to obtain the probabilitydensity functions (pdf’s) plotted in Figure 2.Using the estimated default probabilities along with the …tted US risk-
free term structure, we can exploit equation (3) of the reduced-form model torecover, for each country, the implicit term structure of credit spreads, whichre‡ects market’s medium-long term expectations about bonds default prob-ability. Figure 3 shows the time series estimates for some countries, whereastable 4 contains some summary statistics on the whole sample. In general, weobserve relatively ‡at term structures, which can become very steep duringhigh volatility periods and downward sloping in the weeks following the endof a crisis.
7Estimates for Argentina’s default probability in the period August 2001 - January2002 are presented in the Appendix.
10
6 Predicting the dynamics of default proba-bilities for portfolio trading strategies
In the previous section, we have estimated historical default probabilitiesfrom market prices of emerging markets bonds. For the bonds we are con-sidering, we can observe that to an increase (decrease) in default probabilitycorresponds a decrease (increase) in the market price of the bond one weekahead, as the average correlation between lagged default probabilities andbond prices is -0.8.This means that predicting default probabilities, or, at least, the direction
of default probabilities, can provide useful insights about future movementsin bond prices. This would obviously represent a relevant information forportfolio allocation.In this section, we develop a forecasting model for the probability of ob-
serving an increase/decrease in future default probabilities, which is based onthe use of frequently observed …nancial variables. The forecasts are then usedto implement e¢cient trading strategies for portfolios of emerging marketsbonds.Because of the availability of data on the explanatory variables, the analy-
sis in the following is restricted to seven countries, which are representative ofdi¤erent economic regions: Argentina, Brazil, Mexico, Russia, Turkey, SouthKorea and Philippines.First, for each country, we regress estimated default probabilities against
some signi…cant …nancial variables, such as interest rates, bond indices andexchange rates. More precisely, as explanatory variables, we use lagged values(up to two weeks) of short and long term interest rates in local currencies,J.P.Morgan and Lehman Brothers local indexes, log changes in exchangerates and interest rate spreads calculated with respect to US rates.As shown in table 5, the …nancial variables seem to contain a signi…-
cant amount of information for future default probabilities, with adjustedR-squared of the regressions around 75%, on average.We then use this evidence to build a logit type model for the prediction
of the probability of a market downgrading or upgrading of Global bonds.From now on, we use the term downgrading to indicate either an increase inbonds’ default probability or a decrease in bonds’ price. Similarly, we usethe term upgrading to indicate a decrease in bonds’ default probability oran increase in bonds’ price. In fact, as said above, an increase (decrease) indefault probability is almost equivalent to a decrease (increase) in the marketprice of the bond.The dependent variable in the logit model assumes either value 1 for
11
positive weekly changes of the estimated default probability or value 0 fornon-positive weekly changes. As explanatory variables, we use lagged valuesof the …nancial variables included in the regression analysis above, that is,short and long term interest rates in local currency, J.P.Morgan and LehmanBrothers local indexes, log changes in exchange rates and interest rate spreadswith respect to US rates.The model generally provides accurate predictions both for market down-
grading and upgrading of bonds. Table 6 shows the percentage of correctin-sample predictions for one week ahead default probabilities and changesin bond prices. We notice that, with the only exception of South Korea, inabout 75% of the cases the model correctly predicts future movements indefault probabilities. As regards future changes in bond prices, the statis-tics are relatively satisfactory for all countries, except for South Korea andRussia (only in the downgrading case).The …nal step of our empirical investigation consists in using the logit
speci…cation to produce out-of-sample forecasts for the dynamics of defaultprobabilities. In this case, we recursively estimate the model using win-dows of three years (Argentina, Mexico, Brazil) or one year (Russia, Turkey,Philippines, South Korea) of weekly data.At each point in time, we generate one-step-ahead forecasts for the prob-
ability of having a bond market up/downgrading and use them to simulatetrading strategies for portfolios of emerging market bonds.The simulations are carried out for the investment period 1 September
2000 - 27 July 2001 (48 weeks) assuming a starting equally weighted portfolioof $1,000,000 Global bonds.Portfolios of di¤erent bonds are considered and the following naive trading
strategy is applied 8:- upgrading signal (the probability of a decreasing default probability
forecasted by the model is greater than 60%): position increased by $100,000dollars;- downgrading signal (the probability of an increasing default probability
forecasted by the model is greater than 60%): position closed;- no clear signal (the probability of a decreasing/increasing default proba-
bility forecasted by the model is between 40% and 60%): position unchanged;- minimum investment required to re-open a position on a bond: $100,000
dollars;- borrowing and lending at the USD risk-free 1 week Libor rate.We apply the strategy to 22 di¤erent portfolios: one containing all the
8This strategy has been implemented after several constructive discussions with nu-merous traders.
12
bonds of the seven countries and 21 formed combining the seven bonds taking…ve at a time.Table 7 shows that the simple active portfolio strategy, which is based
on the signals derived from the out-of-sample forecasts obtained by the logitmodel for the probability of an upgrading/downgrading of the bond over thenext week, provides quite satisfactory results, especially in comparison withthe buy & hold strategy and the J.P.Morgan - Lehman Brothers benchmarksrecalculated for the countries in the portfolios.In general, along the sample period considered in the simulations, emerg-
ing markets bonds have not performed particularly well, as the average re-turns on the benchmarks and the buy & hold portfolios are negative: ¡1:67%and ¡9:41%, respectively. Instead, the active portfolio strategy always pro-duces positive returns, with an average value around 5:7%. Moreover, thevariability of returns among the di¤erent portfolios is much lower (0:97%)than in the benchmarks (4:16%) and buy & hold (3:79%) case 9.The naive portfolio strategy based on the out-of-sample forecasts for de-
fault probabilities is also ‡exible enough to control for the risk of the portfolio.Figure 4 shows that a VaR measure at the 95% con…dence level, calculatedusing the J.P.Morgan RiskMetrics’s methodology, satis…es the capital re-quirements for the 7-bond portfolio formed applying the active strategy (thisis true also for all the 5-bond portfolios). We observe that the active strategyprovides a more suitable VaR measure than the buy & hold portfolio. In fact,it gives rise to only two breaks along the 48 weeks considered (4.2%), whereasthere are …ve breaks in the case of the buy & hold strategy. Moreover, asit adapts to new market conditions, the VaR measure in the active strategycase is less conservative and imposes, on average, lower capital requirementsthan in the buy & hold case.
7 ConclusionIn this paper we have addressed two main issues: the computation of defaultprobabilities implicit in emerging markets bonds prices and the impact onportfolio risks and returns of di¤erent default probability expectations.
9In our calculations, we have not explicitly considered transaction costs. We observe adecrease of about 1% in the returns of the active portfolio strategy when a 10 basis pointspercentage of transaction costs is introduced. However, we believe that the returns intable 7 are not over-estimated, especially if we take the point of view of a relatively largeemerging market fund. In fact, we use the mean of bid and ask quotes for bond prices,which means that, on average, transaction costs are already included in the buying andselling prices used to implement the active strategy.
13
First, using a reduced-form model for the pricing of defaultable bonds,we have extracted default probabilities from Global bonds market prices oftwelve countries. The estimated default probabilities re‡ect quite closely ac-tual crisis observed in the market over the sample period comprised betweenFebruary 1997 and July 2001.Then, using logit type econometric models, we have shown that weekly
changes of the estimated probabilities could be predicted by means of somecapital markets factors (interest rates, exchange rates and credit spreads).Finally, we have used recursive estimates of a logit model to produce
out-of-sample forecasts for appreciation/depreciation of the Global bonds.The application of a naive portfolio strategy, based on the out-of-sample
forecasts of the logit model for the probability of a weekly up/downmovementin the market value of the bonds, has provided quite satisfactory results,both in terms of returns and in terms of portfolio risk management. This isparticularly true when we compare them with those obtained by a buy & holdstrategy and the J.P.Morgan - Lehman Brothers benchmarks recalculated forthe countries in the portfolios.
Appendix: 2002 Argentina’s default
What we have seen in the last six months in Argentina (and numericallyevaluated with a continuously growing estimate of default probability by ourmodel) is nothing else than the consequence of what many of the world’s topeconomists have said for the past two years, i.e. Argentina was caught ina vicious downward spiral that would lead to political unrest and economiccollapse.What led Argentina on the edge of the abyss is now clear. Argentina was
hanging on too long to a currency regime that linked the value of the Argen-tine peso to the US dollar, with a …xed exchange rate of one Argentine pesoper US dollar. For a long time the system worked wonders to help Argentinain the control of its notorious hyperin‡ation. But, unfortunately, in the pastthree years, the peg to the rising U.S. dollar made Argentine products tooexpensive on world markets, throwing the country into recession.As in other similar cases, for a while the government, the companies and
the households were able to protect themselves from some of the negativeconsequences of a shrinking economy by borrowing increasing amounts ofmoney from its own banks and pension funds, from foreigners and eventuallyfrom the International Monetary Fund.But the story is widely known: when the economy did not turn around
as expected (on the contrary, the situation got even worse due to the men-
14
tioned glooms on the world’s economy), lenders became more cautious anddemanded ever-rising interest rates. The higher rates drained even moremoney from the economy, causing still more unemployment (up to 18% inNovember 2001) and slower growth.Over the period from October 2001 to December 2001, Argentina’s eco-
nomic ministry launched a desperate plan to avoid o¢cial devaluation anddefault. Government workers and pensioners were required to take cuts inpay and bene…ts of about 10%, and at the same time banks and pensionfunds holding government bonds were required to exchange them for newbonds paying lower interest rates. Strict limits were placed on how muchmoney Argentines could take out of their bank accounts each week. TheArgentine peso began trading uno¢cially at a 30%, 40%, and …nally 50%discount against the dollar.The consequences of this situation were really dramatic; from a social
point of view, violent popular riots took place with tragic (human and eco-nomic) e¤ects, while on the other side, under the …nancial markets’ pointof view, the desperate attempt of Argentina’s economic authorities to ferrythe economy to a safe harbour was perceived as an uno¢cial declaration ofdefault and consequently the prices of Argentine bonds fell to their histor-ical lows, with the consequent upswing of the implied default probabilitiesincreasing from about 10% in early October to more than 30% at the end ofthe year (according to our estimates), as shown in …gure 5.A …rst step towards new economic equilibria for Argentina (supported by
further loans by the International Monetary Fund and the World Bank whichshould be discussed around the middle of February 2002) could be the o¢cialpeso devaluation (30%) of January 7th, 2002; this could help to come back topositive economic growth rates (or at least less dramatic negative rates) bymeans of the international trade channel, and perhaps the …rst step towardsthe structural reforms required by the world’s economic agents to recovertrust in Argentina’s future, surely this will resolve in the beginning of ratherhard times under the economic point of view for Argentines.
References
Black, F. and J. Cox (1976), ”Valuing corporate securities: some e¤ectsof bond indenture provisions”, Journal of Finance, 31, 351-367.Cox, J.C., J.E. Ingersoll and S.A. Ross (1985), ”A theory of the term
structure of interest rates”, Econometrica, 53, 385-407.Crouhy, M., D. Galai and R. Mark (2000), ”A comparative analysis of
current credit risk models”, Journal of Banking and Finance, 24, 59-117.
15
Du¢e, D., L.H. Pedersen and K.J. Singleton (2000), ”Modeling sovereignyield spreads: a case study of russian debt”, working paper, Stanford Uni-versity.Du¢e, D. and K. Singleton (1997), ”An econometric model of the term
structure of interest rate swap yields”, Journal of Finance, 52, 1287-1323.Du¢e, D. and K. Singleton (1999), ”Modeling term structures of default-
able bonds”, Review of Financial Studies, 12, 687-720.Du¢e, D. and K. Singleton (2001), Credit risk for …nancial institutions:
management and pricing, Graduate School of Business, Stanford University.Gordy, M.B. (2000), ”A comparative anatomy of credit risk models”,
Journal of Banking and Finance, 24, 119-149.Izvorski, I. (1998), ”Brady bonds and default probabilities”, International
Monetary Fund, working paper 98/16.Jarrow, R., D. Lando and S. Turnbull (1997), ”A markov model for the
term structure of credit risk spreads”, Review of Financial Studies, 10, 481-523.Jarrow, R. and S. Turnbull (2000), ”The intersection of market and credit
risk”, Journal of Banking and Finance, 24, 271-299.Lando, D. (1998), ”Cox processes and credit-risky securities”, Review of
Derivatives Research, 2, 99-120.Longsta¤, F. and E. Schwartz (1995), ”A simple approach to valuing risky
…xed and ‡oating debt”, Journal of Finance, 50, 789-819.Merrick, J.J. (2001), ”Crisis dynamics of implied default recovery ratios:
evidence from Russia and Argentina”, Journal of Banking and Finance, 25,1921-1939.Merton, R. (1974), ”On the pricing of corporate debt: the risk structure
of interest rates”, Journal of Finance, 29, 449-470.Saa-Requejo, J. and P. Santa-Clara (1997), ”Bond pricing with default
risk”, working paper, Anderson School of Management, University of Cali-fornia Los Angeles.Schonbucher, P.J. (1998), ”Term structure modelling of defaultable bond-
s”, Review of Derivatives Research, 2, 161-192.Trova, M. (2000), ”Emerging markets, Brady bonds and default proba-
bilities: a portfolio selection approach”, working paper, Intesa Asset Man-agement, Milan.Xu, D. and F. Nencioni (2000), ”Introducing the J.P.Morgan implied de-
fault probability model: a powerful tool for bond valuation”, working paper,JPMorgan, New York.
16
Table 1Main features of Global bonds
This table shows the main features of the Global bonds included in thesample. The sample contains weekly market values of prices ranging from 14February 1997 to 27 July 2001. All bonds pay the coupon semi-annually andare not collateralised.
Issuer Maturity Coupon
Argentina 19 Sept. 2027 9.75%Brazil 15 May 2027 10.125%Colombia 15 Feb. 2027 8.375%Mexico 15 May 2026 11.50%Panama 30 Sept. 2027 8.875%Venezuela 15 Sept. 2027 9.25%Russia 24 July 2018 11.00%
South Africa 19 May 2009 9.125%Turkey 15 June 2009 12.375%China 22 Oct. 2027 7.30%
Philippines 15 Jan. 2019 9.875%South Korea 15 Apr. 2008 8.875%
17
Table 2Summary statistics on estimated default probabilities
This table shows summary statistics on estimated default probabilities.These are weekly estimates from equation (2), where we consider long-termUS Dollar denominated Global bonds (see table 1) for a sample period rang-ing between 14 February 1997 and 27 July 2001. USD Libor (all maturitiesbetween 1 month and 12 months) and swap rates (all maturities between 2and 10 years and the 15, 20 and 30 years maturities) are used to …t the risk-free term structure in correspondence of the payments dates. The risk-freeterm structure of interest rates is obtained estimating a two-factor versionof the Cox, Ingersoll and Ross (1985) model using a maximum likelihood -Kalman …lter technique.Values are expressed in percentage terms.
Country No. obs. Mean Median St. Dev. Max Min
Argentina 203 3.54 3.34 1.13 9.42 2.04Brazil 217 4.35 3.73 2.05 15.18 2.24Colombia 233 2.88 2.95 1.20 8.40 0.88Mexico 233 2.04 1.98 0.84 5.74 0.70Panama 202 2.21 2.11 0.56 5.06 1.25Venezuela 203 5.64 4.43 4.53 45.13 2.03Russia 120 11.26 5.70 13.04 76.00 3.83
South Africa 116 1.41 1.44 0.31 2.25 0.85Turkey 110 3.54 3.55 1.07 6.22 2.02China 197 1.13 1.07 0.40 2.39 0.27
Philippines 134 2.90 2.85 0.81 4.67 1.51South Korea 173 1.31 0.79 1.20 6.71 0.35
18
Table 3”Regional contagion” e¤ect
This table shows the correlation between estimated default probabilitiesof twelve emerging markets bonds over the sample period ranging between14 February 1997 and 27 July 2001.
Country Brazil Colombia Mexico Panama Venezuela
Argentina 0.68 0.65 0.50 0.63 0.63Brazil 0.77 0.84 0.82 0.79Colombia 0.35 0.68 0.71Mexico 0.83 0.73Panama 0.78
Country S.th Africa Turkey China Philippines S.th Korea
Russia 0.52 0.12 -0.39 -0.33 0.45S.th Africa 0.29 -0.23 0.19 0.59Turkey 0.53 0.43 0.29China 0.65 0.28
Philippines -0.12
19
Table 4Estimated term structures of credit spreads
This table shows average estimated credit spreads for di¤erent maturities.These are implicitly derived from estimated default probabilities and USrisk-free term structure. The sample period varies for each country andranges between 14 February 1997 and 27 July 2001. Values are expressed inpercentage terms. Standard deviation in parentheses.
Country No. obs. 1 year 5 year 10 year
Argentina 203 3.090 3.050 2.984(1.017) (0.980) (0.917)
Brazil 217 3.846 3.763 3.637(1.924) (1.794) (1.607)
Colombia 233 2.505 2.478 2.433(1.068) (1.037) (0.994)
Mexico 233 1.758 1.747 1.727(0.736) (0.720) (0.699)
Panama 202 1.901 1.889 1.868(0.488) (0.476) (0.462)
Venezuela 203 5.198 4.868 4.502(5.429) (3.749) (2.473)
Russia 120 8.053 7.499 6.676(5.987) (5.083) (3.801)
South Africa 116 1.212 1.211 1.204(0.268) (0.265) (0.262)
Turkey 110 3.094 3.059 2.995(0.938) (0.918) (0.878)
China 197 0.965 0.964 0.959(0.342) (0.340) (0.337)
Philippines 134 2.518 2.496 2.457(0.716) (0.699) (0.676)
South Korea 173 1.125 1.116 1.102(1.056) (1.031) (0.998)
20
Table 5Predictive ability of …nancial variables for default probabilities
This table shows the predictive ability of …nancial variables for defaultprobabilities. Estimated default probabilties are regressed against laggedvalues (up to two weeks) of short and long term interest rates in local cur-rencies, J.P. Morgan and Lehman Brothers local indexes, log changes inexchange rates and interest rate spreads calculated with respect to US rates.The sample period varies for each country and ranges between 14 February1997 and 27 July 2001.
Country No. of obs. R2
Argentina 203 0.64Brazil 217 0.65Mexico 233 0.79Russia 120 0.70Turkey 110 0.89
Philippines 134 0.74South Korea 173 0.80
21
Table 6Percentage of correct in-sample predictions
This table shows the percentage of correct in-sample predictions of amarket downgrading (weekly increase in default probability or decrease inbond price) or upgrading (weekly decrease in default probability or increasein bond price) in the underlying bonds. These are obtained estimating alogit model, where the dependent variable assumes either value 1 for posi-tive weekly changes of default probability or value 0 for non-positive weeklychanges. As explanatory variables, short and long term interest rates in localcurrencies, J.P. Morgan and Lehman Brothers local indexes, log changes inexchange rates and interest rate spreads calculated with respect to US ratesare used. The sample period varies for each country and ranges between 14February 1997 and 27 July 2001. Values are expressed in percentage terms.
Change in def. prob. Change in bond priceCountry No. obs. Downgr. Upgr. Downgr. Upgr.
Argentina 203 75 75 79 70Brazil 217 78 74 76 67Mexico 233 76 75 62 69Russia 120 80 80 53 75Turkey 110 78 79 67 71
Philippines 134 77 72 70 65South Korea 173 61 79 58 56
22
Table 7Return on di¤erent investment strategies
This table shows the returns given by di¤erent investment strategies onportfolios of Global bonds. Simulations are carried out for the period 1 Sep-tember 2000 - 27 July 2001 (48 weeks) assuming that the portfolios are ini-tially equally weighted. The active portfolio strategy is based on the buy/sellsignals provided by the out-of-sample forecasts for the dynamics of defaultprobabilities obtained through the logit model. An initial investment of$1,000,000 is assumed. On the active strategy, a minimum investment re-quirement of $100,000 is imposed to re-open a position. No constraints onborrowing and lending at the USD 1 week risk-free Libor rate are imposed.All values are in percentage terms. Bonds used to form portfolios are those oftable 1 for Argentina (A), Brazil (B), South Korea (K), Mexico (M), Philip-pines (P), Russia (R) and Turkey (T).
Portfolios Active portfolio Benchmark Buy&holdA, B, K, M, P 4:58 0:24 ¡9:45A, B, K, M, R 6:75 0:85 ¡7:41A, B, K, M, T 5:34 ¡6:37 ¡13:73A, B, K, P, R 5:81 0:59 ¡8:21A, B, K, P, T 4:41 ¡6:61 ¡14:53A, B, K, R, T 6:58 ¡6:08 ¡12:48A, B, P, R, T 5:64 ¡6:77 ¡14:34A, B, M, P, R 5:82 0:11 ¡9:26A, B, M, P, T 4:41 ¡7:05 ¡15:58A, B, M, R, T 6:58 ¡6:53 ¡13:54A, K, M, P, R 6:73 3:22 ¡5:19A, K, M, P, T 5:32 ¡4:18 ¡11:51A, M, P, R, T 6:56 ¡4:32 ¡11:32A, K, P, R, T 6:56 ¡3:87 ¡10:27A, K, M, R, T 7:49 ¡3:61 ¡9:47B, K, M, P, R 5:06 8:43 ¡1:24B, K, M, P, T 3:65 0:64 ¡7:56B, M, P, R, T 4:89 0:54 ¡7:37B, K, P, R, T 4:89 1:00 ¡6:32B, K, M, R, T 5:82 1:26 ¡5:52K, M, P, R, T 5:80 3:58 ¡3:30
A, B, K, M, P, R, T 6:44 ¡1:70 ¡9:41Average 5:69 ¡1:67 ¡9:41
Standard deviation 0:97 4:16 3:79
23
FIGURE 1 (Panel A)
Estimated default probabilities
Argentina
0.0%
1.0%
2.0%
3.0%
4.0%
5.0%
6.0%
7.0%
8.0%
9.0%
10.0%
12-S
ep-9
7
21-N
ov-9
7
30-J
an-9
8
10-A
pr-9
8
19-J
un-9
8
28-A
ug-9
8
6-N
ov-9
8
15-J
an-9
9
26-M
ar-9
9
4-Ju
n-99
13-A
ug-9
9
22-O
ct-9
9
31-D
ec-9
9
10-M
ar-0
0
19-M
ay-0
0
28-J
ul-0
0
6-O
ct-0
0
15-D
ec-0
0
23-F
eb-0
1
4-M
ay-0
1
13-J
ul-0
1
Brazil
0.0%
2.0%
4.0%
6.0%
8.0%
10.0%
12.0%
14.0%
16.0%
6-Ju
n-97
15-A
ug-9
7
24-O
ct-9
7
2-Ja
n-98
13-M
ar-9
8
22-M
ay-9
8
31-J
ul-9
8
9-O
ct-9
8
18-D
ec-9
8
26-F
eb-9
9
7-M
ay-9
9
16-J
ul-9
9
24-S
ep-9
9
3-D
ec-9
9
11-F
eb-0
0
21-A
pr-0
0
30-J
un-0
0
8-Se
p-00
17-N
ov-0
0
26-J
an-0
1
6-A
pr-0
1
15-J
un-0
1
Colombia
0.0%
1.0%
2.0%
3.0%
4.0%
5.0%
6.0%
7.0%
8.0%
9.0%
14-F
eb-9
7
25-A
pr-9
7
4-Ju
l-97
12-S
ep-9
7
21-N
ov-9
7
30-J
an-9
8
10-A
pr-9
8
19-J
un-9
8
28-A
ug-9
8
6-N
ov-9
8
15-J
an-9
9
26-M
ar-9
9
4-Ju
n-99
13-A
ug-9
9
22-O
ct-9
9
31-D
ec-9
9
10-M
ar-0
0
19-M
ay-0
0
28-J
ul-0
0
6-O
ct-0
0
15-D
ec-0
0
23-F
eb-0
1
4-M
ay-0
1
13-J
ul-0
1
Mexico
0.0%
1.0%
2.0%
3.0%
4.0%
5.0%
6.0%
7.0%
14-F
eb-9
7
25-A
pr-9
7
4-Ju
l-97
12-S
ep-9
7
21-N
ov-9
7
30-J
an-9
8
10-A
pr-9
8
19-J
un-9
8
28-A
ug-9
8
6-N
ov-9
8
15-J
an-9
9
26-M
ar-9
9
4-Ju
n-99
13-A
ug-9
9
22-O
ct-9
9
31-D
ec-9
9
10-M
ar-0
0
19-M
ay-0
0
28-J
ul-0
0
6-O
ct-0
0
15-D
ec-0
0
23-F
eb-0
1
4-M
ay-0
1
13-J
ul-0
1
FIGURE 1 (Panel B)
Estimated default probabilities
Venezuela
0.0%
5.0%
10.0%
15.0%
20.0%
25.0%
30.0%
35.0%
40.0%
45.0%
50.0%
12-S
ep-9
7
21-N
ov-9
7
30-J
an-9
8
10-A
pr-9
8
19-J
un-9
8
28-A
ug-9
8
6-N
ov-9
8
15-J
an-9
9
26-M
ar-9
9
4-Ju
n-99
13-A
ug-9
9
22-O
ct-9
9
31-D
ec-9
9
10-M
ar-0
0
19-M
ay-0
0
28-J
ul-0
0
6-O
ct-0
0
15-D
ec-0
0
23-F
eb-0
1
4-M
ay-0
1
13-J
ul-0
1
Panama
0.0%
1.0%
2.0%
3.0%
4.0%
5.0%
6.0%
19-S
ep-9
7
28-N
ov-9
7
6-Fe
b-98
17-A
pr-9
8
26-J
un-9
8
4-Se
p-98
13-N
ov-9
8
22-J
an-9
9
2-A
pr-9
9
11-J
un-9
9
20-A
ug-9
9
29-O
ct-9
9
7-Ja
n-00
17-M
ar-0
0
26-M
ay-0
0
4-A
ug-0
0
13-O
ct-0
0
22-D
ec-0
0
2-M
ar-0
1
11-M
ay-0
1
20-J
ul-0
1
Russia
0.0%
10.0%
20.0%
30.0%
40.0%
50.0%
60.0%
70.0%
80.0%
16-A
pr-9
9
21-M
ay-9
9
25-J
un-9
9
30-J
ul-9
9
3-Se
p-99
8-O
ct-9
9
12-N
ov-9
9
17-D
ec-9
9
21-J
an-0
0
25-F
eb-0
0
31-M
ar-0
0
5-M
ay-0
0
9-Ju
n-00
14-J
ul-0
0
18-A
ug-0
0
22-S
ep-0
0
27-O
ct-0
0
1-D
ec-0
0
5-Ja
n-01
9-Fe
b-01
16-M
ar-0
1
20-A
pr-0
1
25-M
ay-0
1
29-J
un-0
1
South Africa
0.0%
0.5%
1.0%
1.5%
2.0%
2.5%
14-M
ay-9
9
18-J
un-9
9
23-J
ul-9
9
27-A
ug-9
9
1-O
ct-9
9
5-N
ov-9
9
10-D
ec-9
9
14-J
an-0
0
18-F
eb-0
0
24-M
ar-0
0
28-A
pr-0
0
2-Ju
n-00
7-Ju
l-00
11-A
ug-0
0
15-S
ep-0
0
20-O
ct-0
0
24-N
ov-0
0
29-D
ec-0
0
2-Fe
b-01
9-M
ar-0
1
13-A
pr-0
1
18-M
ay-0
1
22-J
un-0
1
27-J
ul-0
1
FIGURE 1 (Panel C)
Estimated default probabilities
China
0.0%
0.5%
1.0%
1.5%
2.0%
2.5%
3.0%
24-O
ct-9
7
2-Ja
n-98
13-M
ar-9
8
22-M
ay-9
8
31-J
ul-9
8
9-O
ct-9
8
18-D
ec-9
8
26-F
eb-9
9
7-M
ay-9
9
16-J
ul-9
9
24-S
ep-9
9
3-D
ec-9
9
11-F
eb-0
0
21-A
pr-0
0
30-J
un-0
0
8-Se
p-00
17-N
ov-0
0
26-J
an-0
1
6-A
pr-0
1
15-J
un-0
1
Philippines
0.0%
0.5%
1.0%
1.5%
2.0%
2.5%
3.0%
3.5%
4.0%
4.5%
5.0%
8-Ja
n-99
19-M
ar-9
9
28-M
ay-9
9
6-A
ug-9
9
15-O
ct-9
9
24-D
ec-9
9
3-M
ar-0
0
12-M
ay-0
0
21-J
ul-0
0
29-S
ep-0
0
8-D
ec-0
0
16-F
eb-0
1
27-A
pr-0
1
6-Ju
l-01
South Korea
0.0%
1.0%
2.0%
3.0%
4.0%
5.0%
6.0%
7.0%
8.0%
10-A
pr-9
8
19-J
un-9
8
28-A
ug-9
8
6-N
ov-9
8
15-J
an-9
9
26-M
ar-9
9
4-Ju
n-99
13-A
ug-9
9
22-O
ct-9
9
31-D
ec-9
9
10-M
ar-0
0
19-M
ay-0
0
28-J
ul-0
0
6-O
ct-0
0
15-D
ec-0
0
23-F
eb-0
1
4-M
ay-0
1
13-J
ul-0
1
Turkey
0.0%
1.0%
2.0%
3.0%
4.0%
5.0%
6.0%
7.0%
25-J
un-9
9
30-J
ul-9
9
3-Se
p-99
8-O
ct-9
9
12-N
ov-9
9
17-D
ec-9
9
21-J
an-0
0
25-F
eb-0
0
31-M
ar-0
0
5-M
ay-0
0
9-Ju
n-00
14-J
ul-0
0
18-A
ug-0
0
22-S
ep-0
0
27-O
ct-0
0
1-D
ec-0
0
5-Ja
n-01
9-Fe
b-01
16-M
ar-0
1
20-A
pr-0
1
25-M
ay-0
1
29-J
un-0
1
FIGURE 2 (Panel A)
Non-parametric estimates of default probabilities pdf's
Argentina
0.0
10.0
20.0
30.0
40.0
50.0
60.0
-0.0
1
-0.0
1
0.00
0.00
0.00
0.01
0.01
0.02
0.02
0.03
0.03
0.04
0.04
0.05
0.05
0.06
0.06
0.07
0.07
0.08
0.08
0.09
0.09
0.10
0.10
Brazil
0.0
5.0
10.0
15.0
20.0
25.0
30.0
35.0
40.0
-0.0
448
-0.0
359
-0.0
271
-0.0
183
-0.0
094
-0.0
006
0.00
82
0.01
71
0.02
59
0.03
47
0.04
35
0.05
24
0.06
12
0.07
00
0.07
89
0.08
77
0.09
65
0.10
53
0.11
42
0.12
30
0.13
18
0.14
07
0.14
95
0.15
83
0.16
71
Colombia
0.0
5.0
10.0
15.0
20.0
25.0
30.0
35.0
40.0
-0.0
341
-0.0
283
-0.0
225
-0.0
168
-0.0
110
-0.0
052
0.00
05
0.00
63
0.01
21
0.01
78
0.02
36
0.02
94
0.03
51
0.04
09
0.04
66
0.05
24
0.05
82
0.06
39
0.06
97
0.07
55
0.08
12
0.08
70
0.09
28
0.09
85
0.10
43
Mexico
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
-0.0
242
-0.0
201
-0.0
159
-0.0
117
-0.0
076
-0.0
034
0.00
08
0.00
50
0.00
91
0.01
33
0.01
75
0.02
17
0.02
58
0.03
00
0.03
42
0.03
83
0.04
25
0.04
67
0.05
09
0.05
50
0.05
92
0.06
34
0.06
75
0.07
17
0.07
59
FIGURE 2 (Panel B)
Non-parametric estimates of default probabilities pdf's
Panama
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
80.0
90.0
100.0
-0.0
077
-0.0
048
-0.0
019
0.00
10
0.00
39
0.00
67
0.00
96
0.01
25
0.01
54
0.01
82
0.02
11
0.02
40
0.02
69
0.02
98
0.03
26
0.03
55
0.03
84
0.04
13
0.04
42
0.04
70
0.04
99
0.05
28
0.05
57
0.05
85
0.06
14
Venezuela
0.0
5.0
10.0
15.0
20.0
25.0
-0.0
940
-0.0
772
-0.0
604
-0.0
435
-0.0
267
-0.0
099
0.00
70
0.02
38
0.04
07
0.05
75
0.07
43
0.09
12
0.10
80
0.12
48
0.14
17
0.15
85
0.17
54
0.19
22
0.20
90
0.22
59
0.24
27
0.25
95
0.27
64
0.29
32
0.31
01
Russia
0.0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
-0.2
323
-0.1
956
-0.1
589
-0.1
222
-0.0
855
-0.0
489
-0.0
122
0.02
45
0.06
12
0.09
79
0.13
46
0.17
13
0.20
80
0.24
47
0.28
14
0.31
81
0.35
48
0.39
15
0.42
82
0.46
49
0.50
16
0.53
83
0.57
50
0.61
17
0.64
84
South Africa
0.0
20.0
40.0
60.0
80.0
100.0
120.0
-0.0
052
-0.0
037
-0.0
021
-0.0
006
0.00
10
0.00
25
0.00
41
0.00
56
0.00
72
0.00
87
0.01
03
0.01
18
0.01
34
0.01
49
0.01
65
0.01
80
0.01
96
0.02
12
0.02
27
0.02
43
0.02
58
0.02
74
0.02
89
0.03
05
0.03
20
FIGURE 2 (Panel C)
Non-parametric estimates of default probabilities pdf's
Turkey
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
80.0
-0.0
059
-0.0
032
-0.0
004
0.00
23
0.00
51
0.00
78
0.01
06
0.01
33
0.01
61
0.01
88
0.02
15
0.02
43
0.02
70
0.02
98
0.03
25
0.03
53
0.03
80
0.04
08
0.04
35
0.04
62
0.04
90
0.05
17
0.05
45
0.05
72
0.06
00
China
0.0
20.0
40.0
60.0
80.0
100.0
120.0
140.0
160.0
180.0
-0.0
099
-0.0
081
-0.0
064
-0.0
046
-0.0
028
-0.0
010
0.00
07
0.00
25
0.00
43
0.00
60
0.00
78
0.00
96
0.01
14
0.01
31
0.01
49
0.01
67
0.01
85
0.02
02
0.02
20
0.02
38
0.02
55
0.02
73
0.02
91
0.03
09
0.03
26
Philippines
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
-0.0
089
-0.0
059
-0.0
030
0.00
00
0.00
29
0.00
58
0.00
88
0.01
17
0.01
46
0.01
76
0.02
05
0.02
34
0.02
64
0.02
93
0.03
22
0.03
52
0.03
81
0.04
10
0.04
40
0.04
69
0.04
98
0.05
28
0.05
57
0.05
86
0.06
16
South Korea
0.0
10.0
20.0
30.0
40.0
50.0
60.0
-0.0
429
-0.0
369
-0.0
309
-0.0
250
-0.0
190
-0.0
130
-0.0
070
-0.0
011
0.00
49
0.01
09
0.01
69
0.02
29
0.02
88
0.03
48
0.04
08
0.04
68
0.05
27
0.05
87
0.06
47
0.07
07
0.07
67
0.08
26
0.08
86
0.09
46
0.10
06
12-Sep-97
06-Mar-98
28-Aug-98
19-Feb-99
13-Aug-99
04-Feb-00
28-Jul-00
19-Jan-01
13-Jul-01
13
5 7 9
0
1
2
3
4
5
6
7
8
9
FIGURE 3 (Panel A)Argentina term structure of credit spreads
14-Feb-97
25-Jul-97
02-Jan-98
12-Jun-98
20-Nov-98
30-Apr-99
08-Oct-99
17-Mar-00
25-Aug-00
02-Feb-01
13-Jul-01
14
710
0.00
1.00
2.00
3.00
4.00
5.00
6.00
FIGURE 3 (Panel B)Mexico term structure of credit spreads
04-Jun-99
24-Sep-99
14-Jan-00
05-May-00
25-Aug-00
15-Dec-00
06-Apr-01
27-Jul-01
1
47
10
0
5
10
15
20
25
30
FIGURE 3 (Panel C)Russia term structure of credit spreads
24-Oct-97
10-Apr-98
25-Sep-98
12-Mar-99
27-Aug-99
11-Feb-00
28-Jul-00
12-Jan-01
29-Jun-01
13
5 7 9
0.00
0.50
1.00
1.50
2.00
2.50
FIGURE 3 (Panel D)China term structure of credit spreads
FIGURE 4 (Panel A)VaR active strategy 7-bond portfolio
-120000
-100000
-80000
-60000
-40000
-20000
0
1-Sep-00
19-Oct-00
6-Dec-00
23-Jan-01
12-Mar-01
29-Apr-01
16-Jun-01
3-Aug-01
VaR (95%)actual loss
VaR (95%) "breaks"
FIGURE 4 (Panel B)VaR buy & hold 7-bond portfolio
-90000
-80000
-70000
-60000
-50000
-40000
-30000
-20000
-10000
0
1-Sep-00
19-Oct-00
6-Dec-00
23-Jan-01
12-Mar-01
29-Apr-01
16-Jun-01
3-Aug-01
VaR (95%)
actual loss
VaR (95%) "breaks"
FIGURE 5Argentina's default probability
0%
5%
10%
15%
20%
25%
30%
35%
12-Sep-97
12-Dec-97
12-Mar-98
12-Jun-98
12-Sep-98
12-Dec-98
12-Mar-99
12-Jun-99
12-Sep-99
12-Dec-99
12-Mar-00
12-Jun-00
12-Sep-00
12-Dec-00
12-Mar-01
12-Jun-01
12-Sep-01
12-Dec-01
Asia
Argentina
TurkeyEcuadorBrazil
Russia