FOOL'S EYE VIEW
What People Forget About Stock Market Statistics

By Maynard Paton (TMFMayn)
February 13, 2002

Carburton Street, London -- "The are three kinds of lies: lies, damn lies, and statistics" -- Benjamin Disraeli.

Anybody with the slightest interest in investment can't escape statistics. The past performance of the stock market, the past volatility of a particular fund, the past price of a particular share; investors are barraged by such figures.

But do any stock market statistics matter? Some do. But in general, investors would do well to keep Disraeli's scepticism in mind. More often than not, statistics just report the past. They do not foretell the future.

For starters, you may have read something like this at some point:

"Since 1918, the UK stock market has beaten the returns from cash and government bonds (gilts) during:

* at least 75% of all 5-year periods;
* at least 90% of all 10-year periods, and;
* at least 98% of all 20-year periods."

Does this mean the stock market will outperform cash and gilts in the future simply because it statistically has done so in the past? No. To know why the stock market beats cash and gilts, you have to go back to basics.

The stock market is a collection of businesses. And over the long-term, the valuation of each of those businesses correlates to its growth in profits. Thus, a long-term investment rests on the future profitability of the underlying company. 

To cut a long story short, shareholders will only tolerate businesses that, over time, can reinvest their profits at rates greater than those available from risk-free alternatives. Thus corporate profit growth, which drives the stock market, will generally outrun the returns from cash and gilts. (For a full explanation of this phenomenon, read Why Trackers Beat Cash.)

Simple danger

But in the above example, there's a simple explanation for the use of statistics. It's far easier to highlight the long-term benefit of the stock market through past performance rather than go into the nitty-gritty about corporate profitability. Unfortunately, it's the simplicity of statistics that leads many investors astray.

While the above example sensibly concentrates on longer-term horizons, it's straightforward enough for anybody to look at short-term timescales. Ever heard of the "January effect"? Or perhaps you believe in the old "Sell in May and go away" adage? While stock market "historians" can make careers from discovering monthly or yearly patterns in past data, there's no underlying reason whatsoever to suggest those patterns should persist. Why should there be?

In general, too many people put too much emphasis on historic stock market statistics. The advent of spreadsheets and data downloads has allowed people to try and distil the secret of stock market success by analysing various statistics. But it's a pointless task, because the big picture is missed.

What people forget about statistics is that they just a record of what has happened. To succeed on the stock market, you have to judge what will happen. Thus, can you rely on investors to buy during the same "popular" months, year in year out? Can you consistently rely on investors to buy and sell their shares at same prices they did last year, or the year before that? No you can't.

It's companies that count

What really counts on the stock market is judging companies. As touched upon earlier, it's an undisputable fact that over the long-term, a company's share price will generally run in tandem with its profitability. And here's where statistics can come in handy.

Evaluating the past financial performance of a company can help identify suitable long-term investments. For example, does a particular company have a reasonably consistent record of profit and dividend growth? Does the company have a record of stable margins and cash flow?

Such operating statistics will not reveal a sure-fire long-term winner. You have to decide whether the past financial numbers will continue well into the future. Here, the company's business, products, industry, competition and management will all have a significant bearing on this investigation.

That said, there are no statistics that can predict a company's rival will launch a better product, or inform you that the respected management will suddenly up sticks and leave. Nobody ever said finding great long-term investments was easy. But relying on companies whose underlying business is inherently stable and predictable will be far more rewarding than relying on statistics that appear the same way.