Credit Default Swaps In Emerging Markets

I. Correlation Between Recovery Value and Probability of Default .........................................3

II. An Alternate Methodology: The Cheapest-to-Deliver Bonds for Argentina and Brazil .......4

III. Implication of CDS Spreads for Distressed Emerging Markets ...........................................6

Figures

1. Credit Default Swap in Practice .............................................................................................3

2. Argentine Bonds at Default ...................................................................................................4

3. Default Probabilities using Cheapest-to-Deliver Bonds .........................................................5

4. BrazilЎЇs CDS Spreads in 2002 ................................................................................................5

References...................................................................................................................................7

- 3 -

An Alternative Methodology for Proxying Recovery Value in Credit Default Swap Contracts

In times of distress when a country loses access to markets, there is evidence that credit default swap

(CDS) spreads are a leading indicator for sovereign risk than the EMBI+ sub-index for the country.

However, it is not easy to discern the variables that determine the level of CDS spreads in Emerging

Markets (EM); traders only quote the CDS spreads and not the inputs that are required to calculate

such spreads. This note provides some evidence from Argentina and Brazil that reveals inconsistency

between theory and practice in pricing CDS spreads in EM. This note suggests an alternate

methodology that links CTD (cheapest to deliver) bonds to recovery values assumed in CDS

contracts.

I. CORRELATION BETWEEN RECOVERY VALUE AND PROBABILITY OF DEFAULT

The assumptions underlying the correlation between recovery value r, and the probability of default p

is key to understanding CDS spreads. At a theoretical level, it is clear that CDS spreads are a joint

function of the r and the p where both r and p are determined jointly; in other words, the correlation

between r and p is not zero.2 Discussions with traders in EM reveal that in practice r is usually fixed

at roughly 20 cents of the face (or par) value. Mathematically, using r as a constant will result in the

correlation of r and p to be zero. Also, empirically evidence from the corporate literature has shown

(not surprising) that r is inversely correlated to p. The correlation between r and p cannot be zero in

EM when theory and corporate literature suggest that r and p are negatively correlated.

Figure 1. Credit Default Swap in Practice

Source: Bloomberg, L.P.

Figure 1 shows CDS spreads (s) =f (r, p) and explains how CDS works in 3 components:

r is the recovery value at default, p is the probability of default; and s is

2 Mathematically, the correlation between two variables, where one is a constant, is zero.

- 4 -

the premium per dollar of notational protection from time zero to T that solves the following

equation:

Where

p(t) = risk-neutral default probability;

r = expected recovery rate on the reference obligation in a risk-neutral world;

s = premium per dollar of notational protection from time zero to T;

u(t) = the present value (PV ) of payments at the rate of $1 per year on payment dates between

zero and t

e(t) = PV on an accrual payment at time t, with payment equal to t-tl,

(where tl is last payment date)

v(t) = PV of $1 received in time t

A (t) = accrued interest on the reference obligation at time t

II. AN ALTERNATE METHODOLOGY: THE CHEAPEST-TO-DELIVER BONDS FOR

ARGENTINA AND BRAZIL

A CDS contract usually refers to a few bonds in a ÐŽobasketÐŽ± of bonds at the time of the contract,

allowing the protection seller a valuable option to deliver the cheapest bond upon a default (or, a

credit event). These cheapest-to-deliver (CTD) bondsЎЄor reference bondsЎЄare generally priced at a

discount to other similar bonds due to their illiquidity, denomination, legal jurisdiction, and other

special features (interest-capitalization). Figure 2, illustrates the case of Argentina where, at default,

bond prices varied considerably between 24 cents and 38 cents, with the euro denominated issues

priced higher owing to large retail ownership in Europe.

Figure 2. Argentine Bonds at Default

Argentina Recovery Values

Average Dirty Mid-price in Jan 2002 vs. Remaining Average life, years

20%

22%

24%

26%

28%

30%

32%

34%

36%

38%