Despite widespread predictions to the contrary last year, interest rates this year have plunged and bonds soared.
The yield on the 10-year Treasury TNX, -1.61% has been nearly halved in just nine months, for example, from a high of 3.19% late last year to its current 1.70%. You shouldn’t have been surprised: As I wrote in a February 2018 column for Retirement Weekly, interest rates at that time were not below average—when properly judged on an after-inflation and after-tax basis. So, from the perspective of their historical range between low and high extremes, they had just as much a chance of falling as rising.
For this column I update that analysis, since it’s not obvious that the situation today is any different than in February 2018. While nominal interest rates have plunged, so has expected inflation. Very recently the Labor Department reported that wholesale inflation was dead in the water.
In updating this analysis, I am following the lead of a study that was circulated a couple of years ago by the National Bureau of Economic Research (NBER). Its authors were Daniel Feenberg, a research associate at NBER, Ivo Welch, a finance professor at UCLA, and Clinton Tepper, a Ph.D. student in finance at UCLA. The key contribution of their study is the construction of an historical database of Treasury yields after inflation and taxes are deducted.
To illustrate their approach, consider the 1-year Treasury, which currently yields 1.88%. Assuming a marginal tax rate of 25% and an expected inflation at 1.67% (according to the Cleveland Federal Reserve) this yield drops to minus 0.26%. Though that certainly seems low, it’s actually higher than the 15-year average of comparable yields (which stands at minus 0.91%)—as you can see from the accompanying chart.
Furthermore, believe it or not, the current net-net yield of the 1-year Treasury is higher than where it stood in February 2018. So, based on where rates stands today relative to their historical range, you should be no less surprised by falling rates over the next 18 months than you were over the last 18.
To be sure, the 1-year Treasury’s net-net yield, while higher than the historical average, is not significantly higher according to traditional statistical criteria. Those criteria typically require a reading to be at least two standard deviations higher or lower than the historical average before being classified as meaningfully different, and that’s not the case here. Still, given the steady drumbeat of pronouncements that rates are low, it’s remarkable that—when judged properly—they are even at average levels.
Similar conclusions, by the way, are reached when adjusting the yields on intermediate- and longer-term Treasurys.
What this all means: If you want to argue that rates will rise from current levels, you can’t base your argument on the notion that rates currently are low by any historical standard. If anything, as we’ve seen, they’re above average.
Retirees and soon-to-be retirees might object to this analysis because they hold bonds in tax-deferred accounts. So adjusting bond yields by taxes isn’t appropriate for them.
But retirees and soon-to-be retirees are not the only people who invest in bonds, and many of those who do must pay taxes on their interest income. So taxes undeniably play a powerful role in the bond markets. When judging whether rates or higher or lower than the historical average, the NBER researchers therefore argue, taxes shouldn’t be overlooked—regardless of whether you or I individually hold bonds in a tax-deferred account.
The researchers came up with a complex methodology for determining just how big a role taxes play in setting bond market yields. The 25% rate I used in my illustration above is close to the effective rate they found.
Note carefully that this analysis doesn’t mean that rates will fall. Interest rates that are not significantly different than average presumably have just as much historically-based probability of rising as falling. The broader point here is that, as you determine the bond allocation of your retirement portfolio, it’s important that you base your analysis on an objective reading of history.