While the Foolish Four Portfolio is underperforming the S&P 500 this year, its sister strategy, Beating the S&P, is far ahead of the same index. The returns for stock strategies in any particular year can be extremely misleading, especially when there are small numbers of stocks in a portfolio. Although these strategies are likely winners in the long run, in any particular year the returns are often the result of random chance, as today's experiment shows.
By
--Bonnie Raitt
It's no secret there are a lot of chagrined Foolish Four investors out there. Our official F4 Portfolio for the year is down 7.2% as of yesterday. Fools who have owned Sears <% if gsSubBrand = "aolsnapshot" then Response.Write("(NYSE: S)") else Response.Write("(NYSE: S)") end if %> and Goodyear Tire <% if gsSubBrand = "aolsnapshot" then Response.Write("(NYSE: GT)") else Response.Write("(NYSE: GT)") end if %> in their portfolios likely have done even worse; not only has the price of these two stocks taken a beating, but they've also been unceremoniously kicked out of the Dow clubhouse.
But the model Beating the S&P (BSP) Portfolio is up 8.1% this year as of Tuesday, way ahead of the S&P 500 (up 0.3%) and also ahead of four of the five Foolish portfolios we follow in the strategies area. Recall that the BSP strategy is intended to be a sister strategy to the Foolish Four: It uses a different set of large-cap stocks from which to choose a smaller subset of stocks based on yield and price criteria.
If F4 and BSP choose similar types of stocks, what can explain the gaping difference in the returns of the two strategies this year? Why should one portfolio be pushing up daisies while the other, based on similar principles, is coming up roses?
It's in the luck of the draw, baby.
Before searching for answers to explain this apparent discrepancy, let's check the returns of the F4 and BSP portfolios for the last few years:
Returns(%)
F4 BSP Dow
1998 +16 +61 +16
1999 +22 + 2 +21
2000(YTD)* - 7 + 8 - 5
*as of August 9, 2000
The returns for the F4 and BSP portfolios are all over the map. One year (1998) it looks like BSP is the killer strategy, while another year (1999) the very same strategy woefully underperforms both the F4 and the Dow. Then, just when we've given up on BSP, we find it whipping the market again this year, easily outdistancing the F4 and the Dow. How can one strategy look so great one year, then flop the next -- just as its sister strategy performs in the opposite manner?
It's in the luck of the draw, baby.
It's time now for a little experiment that demonstrates just how much the fortunes of individual portfolios depend on this luck of the draw. Below, I've listed the nine F4 and BSP stocks at the start of 1992, along with their subsequent returns that year. The first four represent the Foolish Four stocks and the last five represent the BSP stocks.
American Express +24.1%
Union Carbide +63.1%
General Motors +10.6%
Sears +22.0%
American Brands -6.0%
KMart +6.4%
Amoco +3.7%
Dow Chemical +11.3%
Dun&Bradstreet +4.3%
F4 +29.9%
BSP +3.9%
Dow 30 Average +11.0%
Dow '61-'99 +12.4%
I chose 1992 as the year for the experiment because it represents the most recent year when the Dow's return approximated its historical average, and the average of the combined Foolish Four and BSP returns was about six percentage points higher than the Dow -- slightly under historic averages. The experiment is applicable to any year, but the results are easier to conceptualize if we use a year that approximates the general returns for the Dow and the historic Foolish Four/BSP results.
Looking at the individual returns for each of these stocks, we notice tremendous variation. Even though these stocks were chosen based on characteristics that are similar, their returns vary from -6% for American Brands to +63% for Union Carbide. There's relatively little difference among them in regard to market cap, yield, and price parameters when considered in the larger total universe of stocks, but still the differences in their returns are huge.
Let's put these nine stocks into a hat, then choose four at random to represent our portfolio stocks for any given year. At the end of the first year, we'll place our four randomly chosen stocks back into the hat and again randomly choose a different set of four stocks. In this manner, year after year, we'll keep tabs on three separate portfolios that have been randomly drawn this way.
Here's the results of our experiment after the first year, showing the returns of our three randomly chosen four-stock portfolios (rounded to the nearest percent for simplicity):
Portfolio 1 +10%
Portfolio 2 +28%
Portfolio 3 +18%
Dow30 +12%
After the first year of our experiment, Portfolio 2 seems the clear winner. But it's only the winner because of the sheer luck of the draw. The large disparities in the returns for the portfolio for any given year, or any portfolio that contains only a small number of stocks like the Foolish Four or BSP, are highly dependent on random fluctuations. Since there's only a handful of stocks in each portfolio, in any given year the inclusion (or exclusion) of any one stock can turn a Superman portfolio into a Pee Wee Herman portfolio -- or vice versa.
Let's turn to real-world examples. Without the outsized returns of PepsiCo <% if gsSubBrand = "aolsnapshot" then Response.Write("(NYSE: PEP)") else Response.Write("(NYSE: PEP)") end if %> in the BSP portfolio, BSP would have gained only 3.0% so far in 2000, instead of 8.1%. Or, without last year's anemic 5% return for Caterpillar <% if gsSubBrand = "aolsnapshot" then Response.Write("(NYSE: CAT)") else Response.Write("(NYSE: CAT)") end if %>, the Foolish Four would have crushed the Dow, adding about five percentage points to its overall return of 22%. Without Caterpillar in the portfolio, we'd all be cheering last year's performance and trumpeting another victory for F4 investing. With Caterpillar included, we had a mediocre year.
Check out the scoreboard for the Fool strategies page, and you'll see the Drip Portfolio returns are way ahead of all the portfolios this year, up 29% as of yesterday. But without the 54% return for Intel <% if gsSubBrand = "aolsnapshot" then Response.Write("(Nasdaq: INTC)") else Response.Write("(Nasdaq: INTC)") end if %>, the Drip Portfolio would have returned a more earthbound 7% this year. What a difference a stock makes!
We've seen how randomness cam greatly influence portfolio returns in any given year. But over longer periods of time, the wide variation in returns for our randomly drawn portfolios lessens considerably. Let's see how each of our random portfolios performs over the long haul.
The following table represents the compound annual growth rates (CAGR) for our three portfolios over various time frames, using a different set of randomly drawn four stocks each year for each portfolio:
CAGR (%)
Years Port1 Port2 Port3 Dow
1 10 28 19 12
3 12 15 18 12
5 17 13 19 12
10 14 14 18 12
20 14 14 17 12
The longer the portfolio time frame, the less variation that exists among the three portfolios. If you follow the portfolios long enough, they'll all tend to gravitate to the mean return for the portfolio average, in our case about 15% for our original group of nine stocks. Reversion to the mean strikes again!
Blackjack card counters inherently understand the implications of our little experiment. Although counting cards at blackjack gives the skillful player a small advantage over the casino for every hand played, in any given session there are wide variations from this average. For any single card session in which small numbers of hands are dealt, a card counter is almost as likely to lose money as win. It's only over the long haul that the large variations in returns cancel themselves out, giving the player an advantage.
Foolish investors, especially when choosing mechanical strategies, should think like blackjack card counters. We've chosen strategies that over the long run should pay off. We think we've stacked the deck in our favor. But Fools should realize that with only a limited number of stocks in our portfolio each year, the variation of returns for any strategy in the short run will be enormous. In the short run, it's basically just the luck of the draw, baby.
Beating the S&P year-to-date returns
(as of 08-09-00):
Bank One <% if gsSubBrand = "aolsnapshot" then Response.Write("(NYSE: ONE)") else Response.Write("(NYSE: ONE)") end if %> +12.0%
PepsiCo <% if gsSubBrand = "aolsnapshot" then Response.Write("(NYSE: PEP)") else Response.Write("(NYSE: PEP)") end if %> +25.6%
Ford Motor Co. <% if gsSubBrand = "aolsnapshot" then Response.Write("(NYSE: F)") else Response.Write("(NYSE: F)") end if %> +3.5%*
Bank of America <% if gsSubBrand = "aolsnapshot" then Response.Write("(NYSE: BAC)") else Response.Write("(NYSE: BAC)") end if %> +7.7%
Fannie Mae <% if gsSubBrand = "aolsnapshot" then Response.Write("(NYSE: FNM)") else Response.Write("(NYSE: FNM)") end if %> -8.4%<
Beating the S&P +8.1%
Standard & Poor's 500 Index +0.3%
*Includes Visteon <% if gsSubBrand = "aolsnapshot" then Response.Write("(NYSE: VC)") else Response.Write("(NYSE: VC)") end if %> spin-off
Compound Annual Growth Rate from 1-2-87:
Beating the S&P +23.8%
S&P 500 +17.1%
$10,000 invested on 1-2-87 now equals:
Beating the S&P $181,500
S&P 500 $85,800