A PDF version of this commentary can be downloaded here.
Earlier this week, Ben Carlson, over at his blog A Wealth of Common Sense, authored a post entitled Bonds Are Supposed to be Boring. With people all worked up over bonds lately, this is a refreshing perspective. In the post, he showed a handful of graphs demonstrating:
- The reasonable long-term predictability of nominal bond returns based on their starting yields.
- The real risk to bonds returns is inflation.
- The short-term variations in mark-to-market bond valuations due to interest rate changes.
Ben’s takeaway was simple: short-term return fluctuations are highly volatile, but over the long-run, returns are more-or-less governed by math. And Ben is right. If you can hold a bond to maturity, you pretty much know the nominal return you’re going to get from it.
This analysis raises two questions for us, however:
- Are bonds worth holding right now?
- Can we stomach holding them to maturity?
To steal a graph from Ben’s blog post, we can see that annualized 10-year performance lines up with starting yields.
The present translation: with 10-year U.S. Treasury rates currently at 1.95%, we can forecast about a 1.95% annualized return for 10-year U.S. Treasuries from 2015 to 2025.
For the standard 60/40, that means 40% of our invested wealth will only return about 2% a year over the next decade. Given that annualized inflation over the prior decade (2/2005 to 2/2015) was about 2%, we’ll carry that forward as our estimate and we’re looking at a (simplified) real return estimate of 0%.
So boring does not equal safe. In this case, we’re looking at allocating to an asset class with zero real return potential. For the next decade. As in, until 2025.
In reality, few people hold just 10-year U.S. Treasuries. One thing we often hear is that if interest rates are going up, it’s likely a sign of economic strength, and therefore we should have credit spread compression as a tailwind to our bond portfolio.
So perhaps the Barclay’s Aggregate index is a much better representative of a broad bond portfolio. We can look into the underlying components of the iShares ETF “AGG” to get a view into the index sector, duration, and credit breakdown.
For a broader, more diversified, bond portfolio, there will be two dominant factors that govern returns: interest rate changes and credit spread changes. Using monthly changes in the 10-year U.S. Treasury rate and monthly changes in Moody’s BAA-AAA spread, we can create a multivariate regression against monthly returns of the Barclay’s Aggregate index.
Translated from math-speak to English, we’re more or less saying, “the monthly returns of the bond portfolio is equal to some multiple of rate changes plus some multiple of credit spread changes.” The “some multiple” aspect is known as the “factor loading,” and tells us how sensitive the bond returns are to that factor.
Running the regression from 9/1/2003 to 2/1/2015, we find that our simple model can explain 80% of the variance in returns and the loadings on both factors are statistically significant at a 95% level. For rate changes, we have a loading of -4.19 and for credit spread changes, we have a loading of - 2.64.
The interpretation of this is that when 10-year rates go up by 1%, we estimate the Barclay’s Aggregate to drop by 4.19% (note how well this lines up with the duration profile of the portfolio above). Similarly, if credit spreads increase by 1%, we expect the Barclay’s Aggregate to drop by 2.64%.
It is worth pointing out that not only is the sensitivity to rate changes 1.58x the sensitivity to credit spreads, but the historical volatility of rate changes is 1.47x the historical volatility of credit spreads. Putting it all together, sensitivity to interest rate changes represents over 90% of the portfolio risk.
So yes, credit spreads will play a role – but from a contribution to risk perspective, it is pretty marginal compared to interest rate risk in our bond portfolio.
So right now, we’re staring down a 0% real return prospect and the potential for a whole lot of forward volatility with rising rates. Bonds may be boring, but that does not mean they are not risky.
In Our Models
Momentum, yield, and volatility shifts triggered a rebalance in Multi-Asset Income (“MAI”) this week.
The signals for bank loans and the S&P 500 buy-write strategy turned positive and so we have begun the process of building those positions within the portfolio. To do this, we trimmed several positions including: corporate bonds, emerging market debt, mortgage REITs and U.S. REITs. We also increased our positions in ex-US REITs, which has been exhibiting a more favorable risk- adjusted yield than its domestic cousin.
Looking at the following yield-to-risk tradeoffs, we expect that if the signals for bank loans and S&P 500 buy-write remain consistent, bank loans will become our 3rd largest holding, at just under a 20% weight, behind mortgage REITs and preferreds. These positions will be followed by fairly equally sized positions in U.S. corporates, S&P 500 buy-write, and emerging market debt at around an 8% allocation. Together, we expect these six positions to make up just over 85% of the portfolio.
Due to its sizable position in MAI, Tailwinds Conservative also rebalanced to implement these changes.