(MarketWatch) The 2019 investment performance scoreboards—which will soon be published, if they haven’t been already—are undeniably tantalizing.
If only we had been so lucky to follow those advisers at the top of those rankings.
As a guide to which advisers will do well in 2020 and beyond, however, these scoreboards are next to useless, if not worse.
That’s because regression to the mean is one of the most powerful forces at work in the investment arena: An outstanding return in one calendar year is more likely than not to be followed by a subpar return in the next, and vice versa. It’s a far better bet to focus on advisers with good long-term performance.
To illustrate how powerful a force is regression to the mean, I am focusing in this column on investment newsletters. Though they are distinct from mutual funds, hedge funds and other advisory offerings, they are quite similar in their tendencies to regress to the mean.
Specifically, I created a database containing all calendar-year returns since 1980 produced by the hundreds of model portfolios recommended by the stock-market-focused investment newsletters monitored by my Hulbert Financial Digest — nearly 12,000 entries in all. I then measured how much each portfolio’s return in the subsequent year was better or worse.
Consider first the 20% of calendar-year returns since 1980 that were the highest. On average, the returns of the portfolios that produced these gains were 33.4 percentage points worse in the immediately subsequent year. Now consider the 20% of calendar-year returns with the worst returns since 1980: On average in the immediately succeeding year, the portfolios that produced those losses did 32.9 percentage points better. (See accompanying chart.)
Here’s another way of making my same point: Consider a hypothetical portfolio that, on Jan. 1 of each year, started following the investment newsletter portfolio with the best return in the immediately preceding calendar year. Over the last three decades, this hypothetical portfolio lost 16.7% annualized, which means that it produced a total loss.
This is why I so highly recommend that you follow advisers with good long-term performance. How long is long?
The data in the table below provides the basis for an answer. The difference in the three strategies reported in the table is solely the amount of past performance on which it bases each Jan. 1 choice of adviser to start following.
| Each Jan. 1 start following the adviser with the best return over… | Annualized since 1991 |
| The previous calendar year | -16.7% |
| The trailing 5 years | 4.3% |
| The trailing 10 years | 9.3% |
These results tell a powerful story. After contemplating these results, why would any of you pick an adviser solely on the basis of his or her 2019 returns?
To be sure, not every newsletter at the top of the one-year rankings did poorly the next, and vice versa. But to bet that they won’t regress to the mean you are definitely playing against the odds.
