A PDF version of this commentary can be downloaded here.
Mid- and long-dated treasuries (as measured by the ETFs IEF and TLT) continue to make fresh all-time highs after five consecutive positive return weeks. Equities, on the other hand, have been largely range-bound since the beginning of December, failing to make any material advances or declines. It appears that they are following the usual post-QE playbook: more or less throw a temper tantrum until the Federal Reserve agrees to another round of easing.
But easing may not be on the horizon. Forward rates implied by futures contracts tell us that the market is generally betting for a Q3 federal funds rate increase. While the market’s track-record of predicting the date of increase has been astonishingly bad since 2008, it is worth considering whether this broken clock will finally tell the right time.
So with that in mind, the question that sits on the forefront of our minds is, “how should we be thinking about our bonds?” As long-dated U.S. Treasuries continue to rally, should we be thinking about shoring up our duration further?
Duration is a measure of a bond’s sensitivity to a change in its relevant interest rate. For example, a bond with a duration of 5 is expected to lose 5% of its value if its relevant interest rate increases 1% and gain 5% if its relevant interest rate decrease 1%. The reason we keep repeating “relevant interest rate” is because this point is often lost in the broad use of duration when people talk about “rising rates”. If we are looking at a 3-year bond, we should be looking at 3-year rates and for a 10-year bond, 10-year rates.
Most yield curve movements can be summed up in three ways: level change, slope change, and curvature change. Level is when the entire yield curve moves up or down in unison. Slope is when the near and far ends of the curve move in opposite directions, making the curve “steeper” or “flatter.” Curvature explains the bowing of the curve, pushing the middle of the curve in the opposite direction of the near and far ends.
When people talk about “interest rates rising,” they generally mean the federal funds rate, but translate it as a level shift in the entire yield curve. If that is the case, then duration is a perfectly fine measure for risk management. The problem is, empirically, that’s not what happens. Let’s consider some previous changes in the federal funds rate.
|Start Date||End Date||Start Rate||End Rate||Difference|
The issue here is that while the federal funds rate shifted considerably, the actual change in the yield curves themselves were varied. For example, the 1993 to 1995 change appears to be largely level driven while the 1998 to 2000 change contains a significant change in curvature.
Perhaps most extreme is the 2004 to 2006 change, which is a large flattening move: a large level change and a negative slope change combined. In that instance, the 20 Year U.S. Treasuries barely moved while short-term yields took quite a ride.
While the general market advice for safety still seems to be towards “low duration,” we think it is important to point out that two portfolios with the same duration can behave very differently during a rising rate environment. Consider a portfolio that holds only mid-duration bonds and a portfolio that has a barbell position in both short- and long-duration bonds. While these portfolios may have the same average duration, they are subject to very different risks based on how the yield curve ultimately moves.
Moving entirely to low duration bonds is akin to moving entirely to the Utilities sector before a recession because it has low beta: just because you’re trading market risk for sector specific risk doesn’t make your position any less risky.
Ultimately, the evolution in the dynamics of the yield curve will dictate the returns for fixed-income. Whether rising rates mean a change in level, slope, or curvature will have a dramatic effect on whether “short duration” is ultimately the safe haven.
So what are our models telling us? At the moment, our models are pointing towards a diversified approach. Currently, our Newfound 3% Target Excess Yield portfolio has an approximate yield of 2.3 years and an average credit quality of BBB – but it gets there through a more barbelled approach:
- Approximately 25% of the portfolio is in 0-1 Year A-BBB bonds
- Approximately 20% of the portfolio is in 0-3 Year BB-CCC bonds
- Approximately 30% of the portfolio is in 3-5 Year BB-CCC bonds
- Approximately 20% of the portfolio is in 0-20 Year AAA bonds (spread fairly evenly)
(The remaining 5% is fairly scattered in 0-10 year AA-BBB positions). As we’ve said in recent commentaries, this positioning is largely due to recent changes, where our models have shored up positions in short-term junk, due to recent volatility, in favor on enhanced short duration bonds and core fixed-income, whose diversification benefits against junk have been increasing as of late. We believe this sort of reactive approach will help us navigate not just a rise in rates, but the evolution of the yield curve – which is ultimately more important.
In Our Models
Our Dynamic Alternatives portfolio rebalanced this week on its end-of-month schedule. Broadly, the long-only sleeve was slightly reduced in size. Currently, the alternatives sleeve comprises 55% of the portfolio. The biggest change within the portfolio was the near 50% reduction in size of the multi-strategy alternatives ETF LALT, with the capital being spread nearly pro-rata across the other alternatives holdings. This reduction was largely driven by LALT’s increased realized volatility profile, which was in-turn due to its underlying currency exposure. While we expect that the volatility incurred due to Swiss franc exposure is instantaneous in nature and not symptomatic of a broadly increased long-term volatility profile of the position, it serves as a reminder of the jump risk that these alternative positions can be susceptible to. Until trailing realized volatility of this position settles down, we expect a reduced position size.
The rest of our portfolios remain unchanged (with the exception of Total Return, which holds sleeve exposure to our Dynamic Alternatives portfolio). Broad market losses bring our models much closer to de-risking: