Stocks

Confidence in Buy-and-Hold Is Rooted in the Small-Sample-Size Fallacy of Statistical Analysis

 Valuation-Informed Indexing #385

By Rob Bennett

It happens every Spring. The Phillies invite some fellow that hardly anyone has ever heard of to Spring training and, perhaps because he faces no expectations, he tears the cover off the ball. Articles are written about how, if things keep going the way they have been going for this fellow, he will be competing for the batting title. People who have been around awhile point out that the fallacy of statistical analysis called “Small Sample Size” applies. But some still believe. The fellow is hitting now. It’s real.

This is what is going on with Buy-and-Hold, in my assessment. People believe in it. People are betting their financial lives on it. And these are not credulous people. The sort of people who believe in Buy-and-Hold are the type of people who laugh at Spring phenoms. People who believe in Buy-and-Hold are no-nonsense types. They understand the Small-Sample-Size fallacy and they generally don’t fall for that sort of thing. Yet they are falling for it in this case, I believe.

What would it take for you to believe that a second baseman was the real thing and not just a Spring training wonder? If a guy can hit .300 for an entire season, he’s the real thing. Every now and again, there’s a baseball player who hits well for a single season and then can never make it happen again. And it is of course common to see a baseball player hit .300 one season and then hit only .290 or .280 or .270 in his following seasons. But a second baseman who can hit .300 for an entire season has what it takes to hit a baseball.

The odds are strong that someone who can do that once can do it again, or at least some close to doing it again. While that is not true for someone who hits .300 only for a single month or perhaps a few months. Hitting .300 for a full season takes the encouraging performance out of the Small-Sample-Size phenomenon and transforms it into something meaningful.

If you talk to Buy-and-Holders, they will tell you that their strategy has been working for a long time. The annualized real return on stocks from 1982 through today is 8.9 percent. That’s an amazing return. And 36 years is not a short stretch  of time. It’s certainly not the sort of time-period that anyone would ordinarily describe as a “Small-Sample-Size.”

I believe that in the stock investing realm it may be just that. I think it may be this odd phenomenon — the way in which a time-period that stretches out for over three decades is in an important sense not a long time-period at all — that causes the subject of stock investing to be so confusing for so many.

Stock prices play out in a hill-and-valley pattern that usually extends for about 35 years (the bull/bear cycle that is nearing an end today is the longest yet seen in the historical record). Valuations move gradually upward during the early years of the cycle. It can take 20 years before they hit the dangerous levels that cause the upward movement to break. And then they can take years either moving gradually downward or remaining at low levels.

Say that the pattern that has always applied in the past is playing out once again. Does that change our assessment of how Buy-and-Hold has performed over the past 36 years? If we are nearing the end of a long bull/bear cycle, we will be seeing a price drop of 50 percent within the next year or two or three (stocks are today priced at two times fair value and the cycle cannot end until prices have dropped to fair-value levels or lower). If prices were to drop by 50 percent, that would reduce that long-term return considerably. And of course it would be the return that applied after the price crash that mattered, not the one that applied before it.

Does Buy-and-Hold work? It depends on how you assess things. If you get caught up in the excitement of high-return periods, it more than works — Buy-and-Hold appears to have provided flat-out astounding results during the heat of bull markets. But bull markets inevitably produce bear markets and the most exuberantly irrational bulls produce the most depressingly irrational bears. At the end, the long-term return is always something in the neighborhood of 6.5 percent real. So, no, Buy-and-Hold never works. It encourages investors to get caught up in the emotionalism of temporary bull-market prices, making effective financial planning an impossibility. Investors would be better off investing more heavily in stocks when prices are low and less heavily when prices are high.

But Buy-and-Hold sure seems to be working when the bull/bear cycle has not yet resolved itself. Most of us invest for a time-period of perhaps 60 years. So it is hard for us to notice a cycle that takes 35 years to play out. Most of us only get to see two of them in an entire lifetime. We don’t anticipate how the cycle that we are living through is going to end because we have never lived through an entire cycle before.

This is the benefit of using the historical-return data as a guide for how to invest in stocks. The data goes back to 1870 and covers nearly four full bull/bear cycles. We can see things in the historical data that we could never see with our own eyes because we just don’t live long enough for more than one or two bull/bear cycles to pass before our vision. 36 years of stock-market history is not enough to go by when you consider how long it takes for a bull-bear cycle to come to completion.

Rob’s bio is here.