Want a winning investment? Avoid forecasts

Most investors believe you should invest in countries with high economic growth. Most also believe that forecasters can predict when that growth is going to occur.

Both notions sound eminently reasonable - until you examine the evidence.

Economic growth does not equal stock market returns, particularly over short periods of time. Given investors' penchant for overreaction, areas predicted to have better growth can sometimes get overvalued (see: the U.S., circa 2000), while those predicted to have lower growth become undervalued (see: the U.S., circa 2009).

In fact, there's evidence that gross domestic product growth has little to do with stock market returns. In a 2009 paper, Rajiv Jain and Daniel Kranson of Vontobel Asset Management looked at data in 16 developed countries from 1900 to 2002. "The data clearly shows that, over long periods and when adjusted for inflation, stock market returns and GDP per capita growth are negatively correlated," they wrote.

Mr. Jain and Mr. Kranson cited several reasons for this. First, of course, high expectations often lead to high valuations for a country's stocks. In addition, private and government-owned companies (and some newly formed firms) generate sales that contribute to GDP, but generally aren't a part of the stock market.

Then, there is the question of profitability. GDP measures output and is analogous to sales, while stocks tend to be reflections of profits, Mr. Jain and Mr. Kranson say. A company with huge sales but very low profit margins can give a big boost to GDP while its stock sputters.

Growth is not only a dubious lure for investors; it's also very hard for even professionals to predict. For proof, look at the Survey of Professional Forecasters historical data, available on the Federal Reserve Bank of Philadelphia's web site.

In a May, 2008, report, the group said the average predictions for U.S. gross domestic product growth in the third and fourth quarters of 2008 were 1.7 per cent and 1.8 per cent, respectively. In reality, they turned out to be minus 4 per cent and minus 6.8 per cent. At that point, the average forecast for the 2009 unemployment rate was 5.6 per cent; in reality, unemployment ranged from 7.8 per cent to 10.1 per cent throughout the year.

Forecasters also often miss the mark when it comes to predicting stock market returns. In a February, 2000, report, for example, the mean forecast from the Survey of Professional Forecasters for annual stock returns over the next decade was 9.1 per cent. Over the next 10 years, however, the S&P 500 ended up well in the red.

Canadian-born fund manager David Dreman has researched analysts' earnings forecasts. In his book Contrarian Investment Strategies, he wrote that "there is only a 1 in 130 chance that the analysts' consensus forecast will be within 5 per cent for any four consecutive quarters. … To put this in perspective, your odds are 10 times greater of being the big winner of the New York State Lottery than of pinpointing earnings five years ahead."

Some evidence shows that analysts' forecasts may actually be more of an extrapolation of the recent past than a prediction of what's to come.

Looking at data for the past three decades or so, Hussman Funds' Bill Hester noted in a 2010 paper: "The correlation between year-ahead earnings growth expectations and the actual growth in earnings over the same period is 0.28 (statistically, this means that Wall Street's forecasts explain less than one-tenth of the variation in actual earnings growth over the following year)." The correlation between earnings growth from one year and earnings growth projections for the next year was a much-higher 0.75, however.

What's more, while strong actual earnings growth often leads to higher prices for individual stocks, there's evidence that high aggregate earnings growth for companies in a particular market leads to lower overall stock returns for that market.

In a 2003 paper for the Journal of Financial Economics, S.P. Kothari, Jonathan Lewellen, and Jerold B. Warner found that "over the last 30 years … stock prices increased 5.7 per cent in quarters with negative earnings growth and only 2.1 per cent otherwise." Part of the rationale: When earnings growth is particularly high, interest rates tend to rise, the future implications of which usually override the fact that earnings are growing at a nice pace.

Now, all of this isn't to say that forecasters are always wrong, and none of this is to say that Canada's strong economic performance in recent years means bad times are ahead. But what it does mean is that there is enough unpredictability in the world's economies and markets that putting all of your faith in your ability (or even in that of professional forecasters) to predict what will happen over the next 10 years is a very risky proposition - just think of what befell all those investors who loaded up on U.S. stocks in the late 1990s when the consensus was that the U.S. economy was an unstoppable growth machine.

If you stick to a well diversified portfolio, however, you allow yourself to get exposure to a variety of areas and sources of both growth and value. And in doing so, you can limit your risk without sacrificing long-term returns. In fact, you just might improve them.