US bonds as a hedge: it’s complicated
F. Scott Fitzgerald noted that the test of a first rate intellect was being able to hold two opposing ideas, yet still function. The same could be said of any investor who aspires to follow a rules-based portfolio management strategy. Inevitably, such rigor occasionally requires messy compromises with reality as is now the case in the US bond market. Such is my dilemma as I try to broadly follow three rules:
This has to be the starting point and relies on the “golden rule” that in the long run, long-dated government bond yields always converge toward the nominal structural growth rate of the US economy; the proxy for this measure is the 10-year moving average of the annual growth rate. My valuation model includes as a second variable the one-year Fed funds rate. When actual yields are at least one standard deviation above the level implied by the model, then investors should prepare for yields to decline and extend portfolio duration. Conversely, when yields are one standard deviation or more below the level suggested by the model, then duration should be reduced. Presently, yields are well below the model-indicated target, although not yet into outright “overvalued” territory.
When the reading on my US recession indicator declines below -5, there is a fairly high chance that the US economy will formally move into contractionary territory (currently the indicator is at -7). And since government bond yields tend to fall sharply in recessions, preparation for such a scenario would necessarily involve a lengthening of duration and a shift to higher quality bonds as recessions typically are accompanied by a huge blow-out in spreads. In the chart such periods are shaded pink.
In the event that yields on long-dated bonds are up 22% YoY then the market is deemed to be “oversold” and historically this has been a good time to extend duration; such periods are shown with green shading in the in the chart. In the event that yields have fallen -22%, then the reverse is true as denoted by blue shading. In late June US treasuries entered such overbought territory.
From 2010 to the summer of 2015 my US bond recommendations were based almost exclusively on valuations and overbought/oversold measures. These produced advice to sell treasuries in late 2012, buy them in late 2013 and then sell them again in early 2015. What changed from mid-2015, however, was that my US recession
indicator started to flash red. This situation called extension of duration, especially to protect the component of a balanced portfolio.
So with US long bonds being overbought and close to overvalued territory, I face a dilemma. Long-dated treasuries have hugely outperformed equities since the emergence of the US recession signal last year, yet such an economic outcome would point to bonds being held as a hedge.
Being a coward, my approach over the last year to more challenging valuations has been for investors to shorten duration to about seven years. The problem with such a cautious strategy is that it lessens the offsetting power of the hedge. Hence, to be consistent in my advice, logic dictates that shortened duration should be accompanied by a shift in the recommended balanced portfolio allocation from a 50:50 position to 60:40 in favour of bonds.
I will confess that this last suggestion is based more on instinct than particular rules. Yet, when even the best rules contradict each other, difficult choices need to be made. Fitzgerald had a point. for an equity