In first-to-default trades investors earn, or pay, premium to sell, or purchase, protection on a basket of credit default swaps, which terminates after the first credit event or fixed term of typically five years if there are no credit events. These are leveraged trades because first-to-default protection is more sensitive to changes in the spreads of the underlying pool of names than the outright sale of protection on any one of the names.
These types of trade may be useful in the current tight spread environment to help investors earn attractive returns, or to hedge existing positions. The default correlation between the reference credits drives the basket's future behavior and spread.
Structure
If one of the names in a basket experiences a credit event, the trade terminates and the protection buyer receives the difference between par and the recovery of the credit (cash settlement), or par in exchange for the credit instrument, typically a bond or a loan (physical settlement) in the same way as any other International Swaps and Derivatives Association documented credit default swap.
For example, an investor could sell USD10 million first-to-default protection on five names and earn an attractive premium. The trade terminates after the default of the first name and the investor pays (1 recovery) x USD10 million. At that point, the investor stops receiving the premium and there's no further exposure to the remaining four names.
The notional size of these trades typically range from USD5-50 million and they reference two to 10 credits. Most of the deals have a five-year maturity, but they can vary from one to 10 years.
Pricing: Correlation & Spreads
It's uncertain at the outset of a first-to-default trade which name, if any, will default first and trigger a payout from and a loss to the protection seller. A key input to help determine the value of a first-to-default trade is default correlation, which is the tendency of the credits to default simultaneously. This correlation helps quantify the collective default behavior of the credits, implying the likelihood of the first default in the basket, and therefore how much premium should be paid.
All things being equal, low correlation implies a higher premium. To understand this intuitively, consider two idealized scenarios: independence (correlation = 0%) and perfect dependence (correlation = 100%).
Independence
If there is absolutely no relationship among the credits in a basket, then selling first-to-default protection is the same as selling protection on all of the credits individually. The event of one credit defaulting is completely independent to the event of any other credit in the basket defaulting, and the likelihood of a default event depends on the likelihood of default of each of the credits in the basket separately. Therefore, if desiring a hedge, a seller of first-to-default protection would have to effectively write protection on each individual credit and would expect to earn the sum of the spreads of each individual credit in compensation.
Perfect Dependence
If the credits are completely related, then selling protection on a first-to-default basis is the same as selling protection on the poorest-quality credit. With perfect dependence, if the poor-quality credit defaults, all credits in the basket are assumed to default as well. Therefore, the likelihood of a default event depends simply on the riskiest of the credits in the basket, and a seller of first-to-default protection would expect to receive the spread of the worst credit.
Reality
Of course, independence and perfect dependence are two theoretical extremes and correlation is somewhere in between. Keeping our simple and intuitive framework in mind, consider an investor seeking to earn premium by selling first-to-default protection. Assuming a set of names trading at the same spread, an investor desiring the highest premium should sell protection on a basket of names with low correlation, such as across several unrelated industries. The investor on that basket receives a high spread because the low correlation increases the basket's default risk. Conversely, the same investor should avoid selling protection on a highly correlated basket, such as one composed of names from the same sector because this reduces premium.
Table 1 lists two five-name first-to-default baskets with similar credit quality. The first basket is diversified across five sectors and has a relatively low correlation (about 25%), while the second basket has several auto names and a relatively high correlation (about 55%). With similar underlying asset credit quality, the low-correlated diversified basket has a 45 basis point spread pick-up over the highly-correlated auto-skewed basket.
On the other hand, consider an investor seeking a cost-effective hedge, such as for a portfolio with a high telecom concentration. The fund should buy protection on a telecom basket, picking names to maximize correlation, as this would minimize the premium paid for credit protection. The fund would profit from a credit event in any of the telecoms, thereby hedging its credit risk more cheaply than by buying protection individually on the same credits. Of course, once there's a default, the fund no longer has protection on any of the remaining names.
Spread Distribution
Given correlation, the value of first-to-default protection is further influenced by the spread distribution of the assets in the basket. Generally, the first-to-default premium is most enhanced for baskets containing names trading at similar levels.
Consider an investor that has already selected an acceptable correlation and narrows down a five-name basket to four single A rated credits each at 100bps and a fifth CCC-rated credit at 1,000bps. In this basket, the pick-up above the widest spread isn't compelling, as the outlying CCC component dominates the others. Since the trade will likely terminate as a result of its poor quality outlying component, the value of selling protection on the other clustered components is degraded. Therefore little premium will be generated by the first-to-default feature of the trade and the principal source of value will be the premium generated by selling protection on the outlying component.
In order to maximize premium generated, investors should select components with spreads that are clustered. As an example, Table 2 provides two baskets with similar correlation (34-37%), credit quality (Aa1 - Aa3), and recovery rates (around 50%). The first basket has five names with spreads clustered near 25bps, while the second has four names at 25 bps and a fifth outlier at 75 bps.
The second basket generates a higher absolute spread (120bps versus 80bps) but a smaller pick-up relative to its outlier (160% versus 320%). In other words, selling protection on the second basket is skewed toward selling protection on the fifth outlying name, but the extra income generated from this outlying name may not compensate for its additional risk.
Putting It All Together
First-to-default trades allow for spread pick-up, specific name and sector exposures and cost-effective credit hedging. They are often able to accomplish all of the above by allowing investors to take a leveraged view on correlation. Since a first-to-default trade is a leveraged transaction, the basket spread is sensitive to movements in the underlying spreads of the components. Specifically, CDS bid/offer liquidity is important in realizing value if an investor chooses to unwind a first-to-default trade. Finally, different relative movements in credit quality among the names will affect the first-to-default probability.
This week's Learning Curve was written by Lee McGinty, head of credit derivatives strategy, and Rishad Ahluwalia, a credit derivative strategist at JPMorgan in London.
For a more detailed review of the product see JPMorgan's First-to-default baskets--A Primer.
