A superior credit exposure: how CDS can help solve performance problems in high yield

High-yield (HY) bonds can deliver an effective combination of attractive spread, carry and income. However, known issues with liquidity, performance and cost tend to be a drag on overall returns. 

In previous articles, we set out the problems with HY investing and explained how an efficient, low-cost solution using high-quality treasury debt and credit derivatives can provide better liquidity. In this article we explain why using credit-default swap (CDS) indices as a more liquid replacement for high-yield bonds in this ‘liquid credit strategy’ can offer a further advantage: higher performance potential.

Read also: Key problems with high-yield bond investing – and steps toward a solution

Better credit exposure

CDS indices significantly outperform the credit exposure of cash bonds, history shows. Unlike treasury bond futures, synthetic credit indices lack an arbitrage mechanism to ensure convergence in the price of the CDS with its reference obligation – except when a credit event occurs or at maturity (which is typically five years). Therefore, CDS index performance can deviate from that of cash bonds in the short term.

Looking at the historical performance of credit returns of HY cash bonds versus CDS, we find that CDS indices significantly outperform – with lower volatility and a higher credit Sharpe ratio (see Figure 1). The improved benefits are retained even if we use only the last 10 years of data, thereby excluding the outperformance generated during the extreme events of the 2008 global financial crisis.

FIG 1. Credit return of cash bond indices vs synthetic indices¹

This is of even greater interest given that, as we explained previously, attempts to outperform the HY benchmark by active mutual funds have a poor track record in general. While the ‘average’ corporate credit fund manager outperforms, in reality this outperformance derives mainly from investment-grade (IG) benchmarked funds². Focusing on active HY mutual funds, most the data shows that most tend to underperform their benchmark.³ Studies show that active mutual funds in HY tend to be underweight credit due to default aversion. Conversely, within IG, the average mutual fund is overweight credit risk, leading to long-term outperformance.  

FIG 2. Underperformance is the norm: average monthly HY credit mutual fund performance⁴ 

Read also: Think high-yield bonds, but with better liquidity – 3 advantages of a more efficient approach

Decomposing CDS outperformance

CDS outperformance is driven by a composition effect and a basis effect. These drivers explain the difference in performance and tracking-error volatility between CDS indices and benchmark credit returns. First, some definitions:

• The composition effect is measured as the difference between the credit return of a CDS index-matched bond portfolio and the relevant bond benchmark⁵
• The basis effect is measured as the difference between the credit return of the CDS index and the CDS index-matched bond portfolio.
   

Next, let’s understand how they impact performance. CDS indices can differ in composition to their bond benchmark in multiple ways. First, CDS indices are equal-notional-weighted, while bond benchmarks are market-weighted. Second, CDS indices may have a sector allocation that differs from the bond benchmark. Third, CDS indices are concentrated on the five-year point, while bond benchmarks are distributed across maturities on the credit curve. Finally, CDS indices are rebalanced biannually, while bond benchmarks are rebalanced monthly.

FIG 3. Historical performance of basis vs composition effects in CDS index returns⁶

The basis effect captures the differences in cash and synthetic exposures. The difference between the credit return of a synthetic CDS index and a CDS index-matched cash bond portfolio creates a ‘basis’ effect. In the short term, this is driven by the large differences in liquidity and the unfunded nature of the derivative.

As shown in Figure 3, the outperformance of HY CDS indices is fully driven by the basis effect, which substantially outweighs the negative impact of the composition effect – both for USD and EUR-denominated bonds. A potential reason for basis outperformance in HY is that credit spread curves are significantly steeper in HY CDS indices when compared to their matched bonds.⁷

Basis also drives tracking error for HY bonds. Decomposition indicates that for HY bonds, the effect of basis on tracking error is two-to-three times higher than the impact of composition. Basis tends to perform much better in a liquidity crisis, as bond prices overreact in a liquidity vacuum. There are usually also lead-lag effects, since CDS indices tend to react faster to economic news than corporate bonds. Outside of liquidity crises, tracking errors are in the range of 3% per year for HY.

The composition effect lags over the long term. As shown in Figure 3, the long-term performance of the composition effect in HY is negative for both USD and EUR universes – although it is more than offset by the basis return. Composition effects in HY markets come from the equal weighting of the CDS index. This causes exposure to idiosyncratic shocks and a size bias – if smaller companies default, they are likely to have a bigger impact on CDS indices than on market-weighted bond indices.

Negative composition effects are caused mainly by unrecoverable losses. These negative composition effects in HY are largely a reflection of defaults or realised losses that are not recoverable.⁸  Permanent losses for the CDS index can occur in two ways:

1. Exclusion of a name from the index. A default event may occur in the on-the-run index; this is realised as a cash outflow for protection sellers, with a new version of the index created that excludes the defaulted name⁹
2. Forced sales of a distressed name. A distressed name may become illiquid and thus ineligible for the new on-the-run index when it is created¹⁰.
    

It should be noted that a large realised loss only implies negative composition effects if the default losses in the benchmark are lower. This was very much the case during the Covid-19 crisis, when the US HY CDX index saw realised losses of nearly 10%, three times higher than in the cash bond index (3.3%). However, in all other past crises, the losses incurred by the CDS index were comparable with the cash-bond benchmark.

Time variation

Another factor to consider is the tactical benchmark-relative performance of a CDS-based HY approach. Figure 4 shows its six-month performance relative to the global HY benchmark¹¹, which is roughly a 75/25 mix of USD and EUR-denominated bonds. The performance pattern indicates a strong pattern of mean-reversion, with a period of outperformance followed by underperformance and vice versa. 

FIG 4. Relative performance of a global liquid-credit approach vs the global bond benchmark¹²

Excess spreads and tactical performance

Basis drives excess spread volatility. To further analyse the performance of this CDS-based approach, we can use ‘excess spread’ – defined as the spread difference between this strategy and the bond benchmark. Again, we can split excess spread into basis, defined as the spread difference between the CDS index and the CDS index-matched bond portfolio, and composition, defined as the spread difference (over swaps) between the CDS index-matched bond portfolio and the benchmark.

We focus on the US HY market. This is because most of the underlying single-name CDS in the US HY index can be mapped to a relevant HY bond from inception, and the quality of the CDX HY index has remained relatively stable and comparable with the benchmark. As shown in Figure 5, the excess basis spread for US HY is volatile, with large negative moves in a liquidity shock that reverse after the shock abates. In comparison, the excess composition spread is less volatile

FIG 5. Excess spread by composition and basis – US HY¹³

Read also: Capturing high-yield carry: the benefits of a CDS-based approach

This mean reversion of excess spreads drives relative performance. With excess spread closely linked to performance relative to the HY benchmark, and positively linked to future return potential, the mean reversion of excess spreads is a positive tailwind to relative performance.

Excess spreads are generally flat-to-positive in three cases: 1) in a carry-like environment; 2) after a moderate selloff; and 3) in the initial stages of a liquidity crisis. Our research shows that relative performance (especially from basis effects) increases monotonically (i.e., consistently) with the basis excess spread. Periods of flat-to-positive excess spread (70% of the time) result in outperformance of nearly 2%, while for periods of significant negative basis (the remaining 30% of the time) there is -2% relative return¹⁴. Note that underperformance occurs in a recovery from a liquidity shock (see Figure 6). This pattern of excess spread driving relative performance is not just an outcome of the extremes of the 2008-09 shock but has persisted over the 15 years since the financial crisis.

FIG 6. Excess spread and 12-month average relative performance, 2005-2024¹⁵

Conclusion

As we have shown, a liquid credit strategy has the potential to provide significant potential for outperformance relative to traditional active and high-yield strategies and the benchmark. Next, in the final article in this series, we will discuss the benefits of a liquid credit solution during periods of high volatility in bond markets.