When analyzing credit spreads embedded in fixed-income investments, prior structural models have been shown to misrepresent these spreads compared with the actual credit spreads observed in financial markets. This misrepresentation is referred to as the credit spread puzzle. The author attempts to address, quantify, and better understand the key drivers of this phenomenon.
When purchasing fixed-income investments, investors need to evaluate the risk that a company may not be able to make its promised payments. Investors are compensated for this risk through an additional yield known as a credit spread. These credit spreads differ from what prior structural models would suggest, which has given rise to the credit spread puzzle. Although this research acknowledges the existence of the credit spread puzzle, the author notes that the puzzle may be related to an investor who needs compensation that is greater than the expected loss because of some degree of risk aversion.
How Is This Research Useful to Practitioners?
Market participants require a robust understanding of how various financial instruments behave as well as the associated risks involved. This research is very helpful because it detects and highlights potential credit spread mispricings that are found in the fixed-income market. The author determines that A or B rated securities should be purchased dependent on whether leverage is available because investors who purchase A rated securities are relatively more risk averse than, for example, investors who purchase CCC rated securities. In other words, there appears to be an opportunity to exploit the various risk appetite thresholds embedded in the fixed-income universe.
The author then compares actual credit spreads with model spreads to discover that the average mispricing is 283.5 bps. So, the average mispricing is about 72%, meaning that only 28% of an investment’s actual spread can be explained by prior structural models. Moreover, the author discovers that the average mispricing is about 56% for the shortest-maturity bonds compared with 78% for the longest-maturity bonds. This difference presents yet another opportunity to take advantage of potential investment pricing discrepancies. Being able to consistently identify these mispricings and take the appropriate investment action to exploit these opportunities could provide market participants with an advantage over others.
How Did the Author Conduct This Research?
The author uses corporate bond transactions occurring within the Oslo stock exchange (Oslo Børs) from 2008 to 2013. He focuses on bullet senior unsecured bonds that are denominated in the Norwegian kroner and have publicly traded equity, and he excludes all bonds that have embedded options or less than three months to maturity. Also, financial firms are excluded because leverage is typically materially different from nonfinancial institutions. After applying these various criteria, there are 10,595 transactions.
The author defines a credit spread as the spread between a bond’s yield and a risk-free rate, which is estimated as the swap rate minus 0.10%. Furthermore, local sell-side ratings are used to classify the credit quality of the issuer. To accommodate insufficient historical Norwegian default data, European data are substituted. For forward-looking default probabilities, this research relies on the CreditEdge model from Moody’s Analytics. Asset volatility is determined by using five years of monthly observations for the market value of the expected default frequencies of assets.
The default point is then established as an issuer’s current liabilities plus half of its noncurrent liabilities. The author can then determine how many standard deviations the asset is from the default point, which is then translated into a probability of default. These mispricings are observed for periods with both high and low credit spreads to confirm that the results were stable over time.
The author provides a detailed background surrounding the credit spread puzzle. He offers various explanations for this observed phenomenon and identifies tactics to exploit this discrepancy. Ultimately, I would like to see future research expand the scope of this research both in terms of countries analyzed and time periods observed.