The term “Bond Bubble” has intermittently captured the headline spotlight since August of 2010 when USA Today asked a rather poignant question:, “Is a bubble brewing in the […] U.S. government bond market?” A short glance at the 10 Year U.S. Treasury rate’s downward march since its peak in 1981 — while remembering that decreases in rates necessarily imply increases in fixed income prices upon which those rates are derived — should evoke a resounding “yes” from the reader.
If bond prices were a concern on the 23rd of August, 2010 (the date the USA Today article was published) they most certainly should be a grave concern given the current day. Consider the following table which compares today’s values to the values when USA Today published the article, for:
- 10 year U.S. Treasury rate
- 2 year U.S. Treasury rate
- Moody’s Aaa Bond Yield
- Moody’s Baa Bond Yield
- Relative value changes
|Rate||Aaa||Baa||10 Year||2 Year||Baa - Aaa|
If bond prices were debatably inflated in 2010 because of abnormally low yields, there should be little disagreement that the situation has only gotten worse. Unfortunately, for the many articles that cite our current debt debacle (a short Google search using the term “Bond Bubble” [quotations intact] using a time horizon of “one month” yields nearly 7,000 results), very few give guidance regarding how to face this impending devaluation of an asset class whose usual mandate is to reduce a portfolio’s risk.
Sure there’s a consistent mantra of “keep duration short,” but that’s like telling your agoraphobic friend (fear of public places) to “stay inside” to allay his fear — it’s accurate because it’s tautological. Greater duration implies greater sensitivity to interest rates. So if interest rates rise, bonds with higher duration will depreciate more — ceteris paribus. Following the guidance of “keep duration short” will likely keep you out of harms way when the bond bubble bursts; however, during your potentially protracted interim, expect a near zero real yield with little opportunity for upside participation.
Fortunately, there is a little known phenomenon the more adventurous reader should further explore: lower credit quality — like short duration — falls less when interest rates rise.
Rates and Credit
Consider monthly rates from April of 1953 to February of 2013 for the following three levels of credit quality:
- 20 Year U.S. Treasury Yield: Risk Free
- Moody’s Aaa Bond Yield: Investment Grade
- Moody’s Baa Bond Yield: High Yield
The Aaa and Baa rates provided by Moody’s “include bonds with remaining maturities as close as possible to 30 years. Moody’s drops bonds if the remaining life falls below 20 years,” and the 20 Year Treasury Yield approximates the effective yield of a Treasury coupon bond maturing in 20 years. So the three series should be fairly close in duration. Using, the time series we should be able to effectively isolate the impact on fixed income prices that are specifically due to credit quality.
Because the “Bond Bubble” concern is the impacts of a rising rate environment, we focus on periods where the 10 Year U.S. Treasury Yield increased at least 4 times consecutively. The plot below indicates those consecutively increasing periods with a shaded grey region (NOTE: the rate data for 20 Year U.S. Treasury Yield, which serves as our “risk free” credit quality, is not available from January of 1987 to September of 1993, therefore no periods are used [i.e. shaded grey] during that time frame).
Armed with several different time periods characterized by rising interest rates, the cumulative loss was then calculated for each of the three bonds of differing credit quality.1 If lower credit quality bonds depreciate less during rising interest rates, there should be a noticeable relationship between cumulative change in rates and the cumulative loss incurred by bonds of a differing credit quality. That information is plotted on the following three scatter plots, along with the equation of the estimation line — the intercept has been anchored at zero to match the intuition that, no change in rates should likewise reflect no change in bond prices. If our posit is correct, we should expect to see less sensitivity (i.e. a lower slope) for lower credit quality bonds.
As expected, the highest credit quality (i.e. risk free) has the highest sensitivity to interest rate increases, whereas the lowest credit quality (i.e. High Yield) has the lowest sensitivity to interest rate increases. Another way to phrase the phonemenon is: as bond credit quality decreases from Risk-Free, to Investment Grade, to High Yield, the per unit expected loss decreases given a per unit increase in interest rates.
Application to Investable Assets
The above summary shows that when interest rates increase, the investor’s fixed income holdings would theoretically experience a varying degree of loss depending on credit quality, all else equal. “Theoretically” is emphasized because bond losses were estimated using available rate data during those time periods. Although this may roughly equate to losses and gains investors “might have” achieved during those time periods, the calculations do not represent actual holdings or investable assets. The savvy investor should further inquire: does the phenomenon likewise exist for tradable assets such as ETFs?
Three deeply liquid, tradable assets that closely mirror the credit quality comparison made above are the following iShares ETFs: IEI (3-7 Year Treasuries), LQD (Investment Grade Corporates), and HYG (High Yield Corporates). With historical data that begins in 2007, this analysis was run based on daily 10 Year U.S. Treasury rates (instead of monthly), where an increase of at least 3 consecutive days was considered a rising interest rate environment. During those time periods, the cumulative loss for each of the three ETFs was calculated, along with the cumulative rise in the 10 Year Rate. Scatter plots are again provided, illustrating the cumulative loss for a given cumulative rise in interest rates from April 2007 to March 2013.
The scatter plots show the same result as was previously derived: lower credit quality reduces the loss exposure of debt portfolios when interest rates increase. Unfortunately, in this last example, because of limited data only going back to 2007, we were unable to sample from different economic environments and isolate credit quality as the descriptive factor; however, the visual results show a very clear relationship that holds, even during the shortened time frame using instruments whose duration changed over time.
The reader may wonder why this phenomenon exists. As a good friend and distressed debt PM succinctly put it:
Interest rate risk and credit risk tend to move inversely to one another […] The more economically sensitive portfolio will outperform in a rising rate environment because it’s levered to an improving economy. As the cycle matures (lower credit underperforms) as corporate defaults accelerate.
Similar to the way investors rationalize holding different lengths of interest rate duration based on the business cycle and rate expectations, credit quality too goes through periods of favor and disfavor based on the changes in borrowing costs driven by the business cycle and interest rates. The good news is that the economy will likely continue to mend, which bodes well for lower credit debt as borrowing costs continue to improve. This can provide investors an opportunity to participate on the upside while simultaneously dampening the negative impacts to fixed income prices should interest rates rise. So the next article you read about the impending “Bond Bubble”, just remember that B’s might be your best bet.