Historical Performance of Predicted Credit Cycle Regimes
We define the three stages of the cycle in real-time based on the level of our probability indicator and the level of credit spreads versus their long-term average. The rules for our regimes’ definitions (outlined in the chart below) use a probability threshold of 50%. The choice of a 50% threshold is inherently subjective, in this case simply representing even odds, and it is shown for illustrative purposes.
However, we believe our results are robust to a wide range of thresholds, confirming the validity of our methodology.
Next, we back-test the historical performance of the credit risk premium in these predicted regimes and assess whether credit returns map consistently with our expectations. In other words, we determine whether our probability indicator is able to anticipate turning points in the credit cycle and, therefore, in credit markets.
The main highlights are as follows:
- Beginning of the Cycle: Major Spread Compression: This is the least frequent of the three stages, with only 21% of observations since 1989. As expected, this regime delivers the highest excess returns over government bonds, with credit benefiting from both price appreciation and income.
- Mid-Cycle: Stable Spreads: This is the most frequent and persistent state of credit markets, with about 50% of observations. Excess returns are substantially lower than in the first stage, but still positive. For both investment grade and high yield, performance is mostly coming from the additional yield (i.e., the carry) earned over government bonds in a stable market environment, without significant contribution from price appreciation. In other words, despite tight credit spreads and limited upside, credit still offers better returns, and risk-adjusted returns, than government fixed income.
- End of the Cycle: Major Spread Widening: With 29% of occurrence since 1989, this regime tends to deliver outright negative returns. Credit clearly underperforms government bonds, as large spread widening causes price losses in excess of income. Results are directionally consistent between investment grade and high yield.
Overall, this analysis confirms our expectations and provides strong support to the validity of our macro framework. The performance of the credit risk premium maps very intuitively into the predicted stages of the credit cycle, and it is directionally consistent between the two credit universes. In addition, our results are robust to alternative probability thresholds, suggesting that our analysis is not sensitive to a subjective parameter choice.
Follow @OppFunds for more news and commentary.