There are many factors involved in corporate bond default loss risk. Among these factors are (1) default rates, (2) recovery rates, (3) the cyclicality of default rates, (4) the cyclicality of recovery rates, (5) the relationship between credit rating and default rates, (6) the relationship between credit rating and recovery rates, (7) credit rating drift (which varies by starting credit rating), (8) the relationship between holding period and default rates (which varies by starting credit rating), and (9) the relationship between how senior a bond is and whether or not it is secured and recovery rates. In this article, we will explore these factors.
To see a list of high yielding CDs go here.
To begin with, we should be clearer about what constitutes a default. From what I have seen, the three major credit rating agencies―Standard & Poor’s (S&P), Moody’s, and Fitch―have similar definitions of what constitutes a default. They all include bankruptcies, distressed exchanges (and the like, at least in the case of Moody’s), and missed payments. S&P also includes any “obligor…under regulatory supervision owing to its financial condition” (i.e., issuers rated R by S&P).
Below is a table from the “Fitch U.S. High Yield Default Insight” report covering thru June ‘13. It gives us an understanding of how often the different types of default occurred in the last 1½ years. We can deduce that a missed payment can evolve to be a Chapter 11 (bankruptcy) filing, a Chapter 7 (bankruptcy) filing, or a distressed exchange. In a Chapter 7 filing, a company’s assets are liquidated (i.e., sold) to pay its debts―to the extent they can be paid. In a Chapter 11 filing, a company is rehabilitated via, among other things, easier debt terms and/or partial debt forgiveness. If a Chapter 11 company cannot be rehabilitated, it may file for Chapter 7.
Per Moody’s “Annual Default Study: Corporate Default and Recovery Rates, 1920-2012”, from 1994-2012, the average dollar-weighted (versus issuer-weighted or issue-weighted) default rate for all rated issuers, on a global basis, was 1.165% per year. Figures like this are of little use to us though. Default rates vary greatly based upon multiple factors.
Default rate is only one half of the default loss risk equation. The other half of the equation is one minus recovery rate. The recovery rate is the amount you will not lose, on average, if your bond defaults. Default loss risk equals the percentage chance that your bond will default times the amount you will lose in an average default for your bond. When a bond defaults, the loss is usually not 100%.
An average recovery rate is a little less than 40%. Figures like this are of little use to us though. Recovery rates vary greatly based upon multiple factors. It is important to understand, however, that, if the percentage chance that your bond will default is 1.165% and the applicable recovery rate is 37.5%, your default loss risk is 0.73%. It is important to understand the principles involved.
1.165% x (1 – 37.5%) = 0.73%
Default rates vary greatly over time. When there is a recession and more companies are struggling, there are more defaults. Below is a chart from S&P. The chart is based on issuer-weighted data. Notice that the percentage of defaults is relatively very high in 1991, 2001, and 2009 and relatively very low in 1981, 1996, and 2007.
The chart above makes investment-grade bonds appear somewhat safer than they actually are, as credit rating drift, which is covered later in this article, is not accounted for. Also, do not conclude that the chart reflects the highest that defaults can be. The chart only reflects the highest that defaults were in the 31-year period covered. The chart does not reflect how high defaults can be.
Recovery rates tend to fall when defaults rise and rise when defaults fall. In bad times, you get hit by a double whammy―high defaults and low recoveries. In good times, you experience a double blessing―low defaults and high recoveries. The chart below was derived from data in the Fitch U.S. high yield report mentioned above. Notice how recovery rates fall when default rates rise and vise versa. The effect is rather drastic, with recovery rates ranging from 22.5% to 66.4% from 2000 to 2012.
The average recovery rate for 2000-2012 shown in the Fitch report was 36.5%. The average recovery rate was this low because the years with a higher percentage of defaults had a greater effect on the overall average recovery rate. Also, recoveries on junk bonds are a little lower on average than they are on investment-grade bonds, as you will see below.
Below is another chart from S&P. In the chart, notice the extremely strong relationship between credit rating and the propensity to default, and that S&P showed no defaults of AAA-or-AA-rated bonds on a global basis from 1981-2012. (A D rating indicates a default.) The chart does not account for credit rating drift, which is covered later in this article. The chart only reflects a bond’s rating immediately prior to default.
Below is a table from the above-mentioned Moody’s report. The table is based on global data. Notice that the average recovery rate for investment-grade bonds appears to be about 4% better than the average recovery rate for junk bonds. When bonds that once had a higher credit rating defaulted, the losses, on average, were not quite as bad. The early year figures for highly rated bonds are deceptive. The sample sizes are much smaller for these data buckets because highly rated bonds rarely default and, when they do, the default usually occurs years later.
Credit rating downgrades are more common than credit rating upgrades. Per S&P’s “2012 Annual Global Corporate Default Study and Rating Transitions” report, from 1981-2012, over the course of a calendar year, upgrades occurred 8.48% of the time, downgrades occurred 11.85% of the time, defaults occurred 1.55% of the time, and withdrawn ratings occurred 7.23% of the time. Counting defaults, upgrades were only 63% as prevalent as downgrades. Not counting defaults, upgrades were only 72% as prevalent as downgrades. Also, per data from the above-mentioned Moody’s report, a junk bond is significantly more likely than an investment-grade bond to experience a rating change, with or without defaults considered.
Below is a simplified version of a table in the above-mentioned S&P report. Notice that, after 10 years, the effect of ratings drift and its downward tendency is rather large. Also notice that, if a bond with a rating at or below about BBB does not default, it is more likely to end up with a higher credit rating than it had than it is to end up with a lower credit rating.
Below is yet another chart from S&P. In the chart, notice that credit ratings tend to fall significantly prior to a defaults occurring.
You may think that, the longer you hold a bond, the more likely it is to default. This is not always true. This is true for highly rated bonds only. Lowly rated bonds are more prone to default in the early years. Below is a table I created based on data in the Moody’s report I mentioned above. There are some data anomalies in the table that you should ignore. For instance, AAA bonds are not less likely to default after year 13; and Caa-C bonds are not totally immune from default in years 18-20. Anomalies like these occur because of limited sample sizes.
In the table, notice that investment-grade bonds are more likely to default as time progresses. This is because the bonds have had more time to deteriorate in credit quality due to credit rating drift. Notice that junk bonds are increasingly less prone to default after year 3. This is because, as stated above, “if a bond with a rating at or below about BBB does not default, it is more likely to end up with a higher credit rating than it had than it is to end up with a lower credit rating”. Notice that the peak year for Caa-C bond defaults is year 1, the peak year for B bond defaults is year 3, the peak year for Ba bond defaults is year 4―although Ba bonds have a fairly equal risk of defaulting in years 2-19, the peak year for Baa bond defaults is year 16, and the peak year for Aaa-A bond defaults is year 20. This makes sense because the Caa-C bonds are on the verge of defaulting or relatively close to it. The other bonds need to migrate downward in credit quality to be highly susceptible to defaulting. The more highly rated a bond is, the more it needs to migrate downward in credit quality to be highly susceptible to defaulting.
The All (Rated) figures are somewhat deceptive because junk bonds tend to be much shorter in years-to-maturity in comparison to investment-grade bonds. The later years have a lower percentage of junk bonds and a higher percentage of investment-grade bonds.
The table above covers the 1970-2012 time period and is based on Moody’s global data. If you study a different time period and/or use a different data set you will get results that differ somewhat―even if the data sets used have sufficient integrity, which data sets do not always have. One reason for this is the default rate cyclicality discussed further above.
Average recovery rates vary depending upon how senior the bond is and whether the bond is secured or unsecured. Below is a reduced version of a table in the above-mentioned Moody’s report. In the table, notice that the average senior secured bond recovery rate is much higher than the average senior unsecured bond recovery rate. Also notice that the average recovery rate for subordinated bonds is much lower than the average recovery rate for unsubordinated bonds. Also notice that there is little difference between the average recovery rates for senior subordinated bonds and (non-senior) subordinated bonds, but junior subordinated bonds have a much lower average recovery rate than other subordinated bonds.
The data in the table above is based on, with some exceptions, 30-days-post-default trading prices. If you hold a defaulted bond until ultimate recovery (which typically occurs a year or two after default in a bankruptcy), your recovery will probably be a little greater―even after adjusting the ultimate recovery amount for the time value of money.
This article provides good quantification of default loss risk elements, but it does not specifically address the most important question. The most important question is: What default loss risk percentage should I apply in evaluating my specific bond? Your bond has a yield-to-maturity (YTM) or yield-to-worst (YTW). In evaluating the bond, there is a certain percentage you should subtract from the YTM or YTW that represents the average default loss you will experience. Above, using Moody’s 1994-2012 global data, we derived an average default loss of about 0.73%; but this figure is for an average bond. It is not applicable to your specific bond. In the next article in this series, we will use the data in this article or similar data to address the most important question.
(Comments are desired, but I do not guarantee their publication. If your comment[s] indicates that you did not read and genuinely consider the all of the article’s contents, I may not publish it; or I may remove it. You do not need to agree; but do not be derogatory. The comment section is meant for legitimate questions or concerns and well-intended discussion.)
About Kurt Shrout
Kurt has a BA and MA in Communication and over 20 years of business experience, almost always serving as a project or program manager, director, or consultant or as an analyst. He lived and worked in many different locations in the U.S., London, England, and Hong Kong. He has experience in at least 18 different industries and 31 different enterprises. Although he was only 49, he essentially retired in 2008 and began spending a lot more time studying investing. His articles are largely written as a public service. They provide investors with a rare totally unbiased view of the investing landscape and often include unique analysis. You can read more of his articles by clicking here.
Want to learn how to generate more income from your portfolio so you can live better? Get our free guide to income investing here.