High yield debt has enjoyed a nice rally during 2012, as have many other risky assets. Articles and press coverage have addressed the asset class’s future as a result; is this the new way to capture high risk but high reward returns? At Redington, we always look at risk-adjusted returns (or *bang for your risk buck*) and it seems that, while this asset class could offer some interesting opportunities for pension funds, the potential downside inherent in it at present could be a showstopper.

Since the credit crisis, the high yield asset class has been traded on an absolute yield basis rather than at spreads to risk free assets (such as treasury bonds), which means that credit risks in the asset class had over weighted the risks from interest rates. However, an investor faced with the challenge to invest in various fixed income asset classes must ascertain the opportunity-cost of foregoing one asset for another. This opportunity-cost is the relative value between the two assets, and the credit spread can be used to compare two similar credit investments of similar maturities.

Credit spreads also reflect the extra risk (over a default-free bond) for which an investor is compensated. This extra risk is usually determined by two factors:

1. The credit worthiness of the borrower. That is, the likelihood of the borrower continuing to honour its debt.

2. The value the investor can recover if the borrower unfortunately defaults on any of the contractual payments. Ie, the recovery value.

Combining the two, investors can form a reasonable idea of the credit risks involved in the investment. It is important to remember, though, that the credit spread itself can move around along with the price of a bond, and this movement can cause a significant volatility (risk) when investing in high yield bonds. Spread volatility is another topic, but one that an interested investor should investigate.

**Probability of default**

As a vital component of credit spread, this deserves further attention. Historically, a tight relationship has been seen between the economic cycle and corporate default rates year on year. The chart below from Moody’s shows the default rate since 1920.

**1. Real world probability of default measure**

Corporate defaults can be observed from history, as shown in Moody’s default and recovery study. The high yield sector had a default rate of 2.6% during 2012, up from the 2011 year-end level of 1.9%. The historical default rate is sometimes referred to as the *real world probability of default,* as it reflects what has actually happened. If we assume an average recovery rate of 40%, then a simple calculation can predict a default loss of 156 bps based only on the two measures.

If we extend the analysis a bit further, taking into account the fact that investors can hold their high yield investment a bit longer, say 5 years, we can look at the cumulative default probabilities over this 5 year period. If I look at the historical annual cohort data from 1970 and take the 95th percentile of the cumulative default probability, high yield bonds would have a cumulative probability of default of 30.6% over a 5 year holding period. This indicates an average of 365 bps default loss every year.

**2. Risk neutral probability of default measure**

On the other hand, we can deduce an equivalent probability of default from the credit spread. Since this spread is derived from market prices, the probability of default under this calculation will indicate investors’ perception on the credit worthiness and uncertainty that they may not receive their investment back.

The spread over US treasuries of Barclays Global High Yield index is 471bps as of 29 Mar 2013. Again, using a 40% recovery value assumption would produce a probability of default of 7.9% per year. If compared to the historical 1 year default rate (i.e. the real world probability of default measure) in figure 1, this implied probability of default *(i.e. the risk neutral probability of default measure) *is clearly too high compared to historical average. So what has caused the difference in the default credit spread *(i.e. credit spreads compensating of the potential risk of default) *calculated from history and a spread of 106bps (471bps-365bps) from the market?

The credit spread can be expressed in the following two components:

Credit Spread = Expected Loss + Risk Premium

If we consider the default loss derived from the real world probability measure, the spread difference stated above can be explained by the risk premium. Risk premium is often affected by investors’ risk aversion, illiquidity premium, etc.

Lastly, we can review the recovery value. 40% recovery is a common assumption used by many market participants. The following analysis from Moody’s has shown that the average recovery value from a high yield bond is slightly lower than 40%. Please note this value can vary according to the actual mix of rating allocations in a high yield investment portfolio. Higher weights toward Ba and B rated bonds would increase the average expected recovery rate in the portfolio.

To summarize the simple calculation provided in this short article, we estimate the expected loss on high yield bonds using the historical measure shows a credit default loss of 365 bps.

This result implies that, given the current spread of 471bps, investors can earn a spread of 106bps net of default. If adjusted by the swap spread, interest insensitive investors can earn a LIBOR plus return of 73 bps. It is important to note that this is based on the assumption that a HY bond would default at a rate 30.6% over 5 years. If you believe there is such a trend following the historical default rate, and you think the global economy is in the process of recovering, the expected default rate for your assumption should somehow be much more optimistic.

High yield indices, it is worth pointing out, have higher weightings toward BB and B rated companies that have much lower historical default rates. If your analysis is based on historical data, you should adjust for the rating composition of your actual investment portfolio or the benchmark the asset manager is trying to follow in order to gain clearer insight into the default risks of your portfolio.

All things being said, the analysis demonstrates there is certainly some value in this asset class given the market rallies in 2012, but the cause for concern may outweigh that value. Net spread over LIBOR after adjusting for default may not be adequate to solve the problems of most troubled pension schemes in the UK. In fact, while I have written this article, the yield on global high yield bonds has reached an historical low. As shown in the figure below, the average yield hit 5.29% as of 30 Apr, and the average credit spread has tightened to 439bps. But don’t forget, this is an asset class which suffered a drawdown of-34% in 2008.

Given where spreads are today, and given the potential for the global economic situation to deteriorate further, the spreads of this asset class can easily revert to the previous high level, which pose a significant downside concern for investors. Furthermore, given current default assumptions, the theoretical spread, which is lower than the traded spread, can be regarded as a floor for further spread tightening (that is, there is limited room for more spread tightening); however, spreads can easily widen.

There is much more on the downside than the potential upside from further spread tightening. On the whole, it seems this asset class is worth monitoring, but not worth jumping into without significant care and attention. Despite the warnings, though, certain specialized managers can still add alpha by picking out names which deliver the return of the principle in the end and capture the full value of the spread.