By Mark Hulbert
Give the benefit of the doubt to the classic 60% stock/40% bond portfolio. That’s not easy, given its awful performance so far this year. On Jun. 13, the day the S&P 500 /zigman2/quotes/210599714/realtime SPX +0.27% officially entered bear-market territory by closing more than 20% below its Jan. 3 high, long-term U.S. Treasurys were sitting on an even bigger loss — down 22.1%, as judged by the Vanguard Long-Term Treasury ETF /zigman2/quotes/201457689/composite VGLT +0.20% .
Bad as this performance is, it isn’t significantly different than what the 60/40 portfolio has experienced at other times. Armed with Morningstar data back to 1926 on the U.S. stock market and long-term U.S. Treasurys, I calculated rolling five-month returns of the 60/40 portfolio. I found 17 other occasions over the past 96 years in which the portfolio performed even worse than it has this year.
The last such occasion came at the end of the Global Financial Crisis in 2009. Prior to then, the most recent such occasion came near the end of the 1973-1974 bear market.
What this means: If the 60/40 portfolio’s recent loss was in fact sufficient to disqualify it from future consideration, then you shouldn’t have been following it in the first place.
Such an observation is not the final word, of course, since there are more rigorous tests to check whether there’s been a significant deterioration in a portfolio’s recent performance. One, known as the Mann-Whitney test, divides a portfolio’s entire track record into two segments and measures whether the more recent one has significantly different statistical properties than the first.
(I mention this specific test only to illustrate the care you should take when deciding whether you’re justified in switching advisers or strategies. While you may find it too daunting to conduct the test yourself, the internet offers many video tutorials on how to apply the test using Excel or Google Sheets.)
At standard levels of statistical significance, the Mann-Whitney test does not find that the 60/40 portfolio’s recent performance is different than in prior decades. This is yet more evidence that, from a statistical point of view, there’s just as much reason to be following the approach today as before.
What about interest rates?
The Mann-Whitney test focuses only on raw performance numbers and not on the underlying market conditions that might cause a deterioration in the future. One such factor that worries many investors is interest rates. Even though rates have risen over the last year, they remain low by historical standards. Doesn’t that mean the bond portion of the 60/40 portfolio will face strong headwinds in coming years?
Once again this is something that can be tested empirically. One test I conducted involved segregating all months since 1926 into four quartiles according to their 10-year yields. The table below shows, for each quartile, the average subsequent-five-month return of the 60/40 portfolio:
|Average return of 60/40 portfolio over subsequent five months|
|The 25% of months with the lowest 10-year yields||4.1%|
|The next highest 25% of months||3.9%|
|The next highest 25% of months||2.5%|
|The 25% of months with the highest 10-year yields||5.2%|
Notice the absence of any obvious pattern. I confirmed this absence by calculating the correlation between the 10-year yield and the subsequent performance of the 60/40 portfolio. The correlation coefficient was statistically indistinguishable from zero.
To be sure, you would want to avoid the bond portion of your 60/40 portfolio if you were certain that rates will rise in coming years. But confidence that you can time the bond market represents a triumph of hope over experience.
Take a look at the data in the table below, which reflects the results of the Hulbert Financial Digest’s tracking of bond timers’ timing-only performance. (By “timing only,” I mean that the calculations reflect solely the timers’ decisions to get into or out of bonds and not their abilities, or lack thereof, to pick individual bonds or bond funds.)
|% of monitored bond timers who made more money than a bond-market index fund on a timing-only basis|
The investment implication of this discussion is that you should not be too quick to give up on the 60/40 portfolio.
There’s a broader investment implication too: Given that the 60/40 portfolio has performed so poorly this year and still needs to be given the benefit of the doubt, a strategy or adviser’s performance has to be really awful before you have statistical justification to stop following it or him.
David Aronson is a statistician who has authored several books on how to base your investment decisions on a sound statistical foundation, including Evidence-Based Technical Analysis and Statistically Sound Machine Learning for Algorithmic Trading of Financial Instruments (co-authored with Timothy Masters). Aronson uses a wonderful analogy to respond to investors who object to how long they must be willing to stick with an underperforming adviser: They have no choice if they want “to scientifically approach the question of picking advisers and strategies,” he says. “The alternative is to flit like a moth from flame to flame, an approach that is destined to fail over the long term.”