US Treasury Returns: Back To The Future?

By datatrekresearch in Blog US Treasury Returns: Back To The Future?

Surprising factoid: the US 10-year Treasury note actually posted a small gain on a total return basis in 2018: +0.07%. We rely on NYU professor Aswath Damodaran’s (a.k.a. the “Dean of Valuation”) work as the definitive voice on total return calculations, and his team just released that tidbit over the weekend. The late-2018 risk off rally that pushed Treasury prices higher was just enough to keep the 10-year in the black for the year. By a very thin hair, yes…

That got us to thinking: what is the long run track record of 10-year Treasuries in both nominal and real returns as well as compared to simply rolling 3-month T-bills through the year? Damodaran’s long run dataset (link at the end of this section) back to 1928 yields the answers:

#1: Negative total return years for the 10-year Treasury are rare.

Since 1928, the 10-year has only posted total return losses in 16 years (17.6% of this 91 year time period).
Only 4 of these losses have occurred since 1990: 1994 (-8.0%), 1999 (-8.3%), 2009 (-11.1%) and 2013 (-9.1%). For investors with balanced portfolios, it is worth noting that the S&P 500 was higher in all those years.
Back-to-back total return losses for 10-year Treasuries are even rarer, occurring only in twice since 1928: 1955 – 1956 (-1.3% and -2.3%) and 1958 – 1959 (-2.1% and -2.7%). Again, stocks were higher in those years.

#2: However, real negative returns – when the 10-year total return fail to match current year CPI inflation – are much more common. That was the problem with 2018, for example, when CPI inflation rose by 2.2% through November.

In almost half the years since 1928 (40 of 91, or 44%), 10-year Treasury total returns were below same-year CPI inflation.
The worst real returns (negative 10% or lower) cluster during periods of high inflation such as 1974, 1979 and 1980 or 1941, 1946, and 1969.
Unlike nominal returns, negative total returns are “streaky”, often running in sequential years of rising/high inflation. Non-World-War II examples include 1946 – 1948, 1965 – 1969, 1972 – 1975, and 1977 – 1982.

#3: Last year saw 90-day Treasuries returns (+2.4%) soundly beat taking the duration risk of owning 10-years (0.07%), but that is unusual.

Over that last 91 years, rolling 90-day Treasuries has only beaten owning 10-years a total of 31 times. No surprise, but those events are clustered during periods of rising inflation (1965 – 1969, 1972 – 1975, and 1977 – 1981 for example).
On average since 1928, average annual total returns for 10-years have outpaced 90-day bills by 1.7 percentage points.

Finally, our thoughts on how to position in the current environment: with the shorter end of the Treasury yield curve paying almost as much as 10-years, it makes more sense to own shorter term (2-years, for example) paper than going out further on the yield curve. Barring a geopolitical shock large enough to create a US recession, 10-years are unlikely to deliver much in the way of capital appreciation this year. If/when US stocks manage to stabilize, long-term yields may even rise modestly. Conversely, Fed Fund Futures show a 70% chance that short rates will remain at current levels through the end of 2019.

Bottom line: 10-years managed to squeak out a small gain in 2018, but that sets them up for possible disappointment in 2019. As unusual as down years are for this asset, they do happen. Better to play conservatively for now and keep durations short.

Source (click on the link for a spreadsheet that includes 2018): http://pages.stern.nyu.edu/~adamodar/New_Home_Page/datafile/histretSP.html