Article Tools
The forecast season is upon us. This year, as every year, economists will line up in December to predict what will happen in the 12 months ahead. These superbly educated experts will consider all the data and draw on years of experience to tell us where the economy, employment, oil prices and the stock market are headed next.
Oh, but there is one hitch — you shouldn’t count on those forecasts to be all that accurate. In particular, don’t count on finding out what matters most to Canadians this year — when the economy will turn around.
The unfortunate truth about economic forecasting is that it is fundamentally unreliable. So if you’ve been reading the experts’ predictions, and thinking about making major changes to your portfolio as a result of what you’ve read, think again. The failure of most forecasters to foresee the recent downturn is nothing unusual. “Economists are simply hopeless when it comes to forecasting recessions,” writes James Montier, an economist at Société Générale in London who has written extensively about forecasting. “Actually, I could have stopped that sentence before the word recessions.”
Smart investors don’t forecast; they anti-forecast. They realize how uncertain the future is and put little faith in anyone’s predictions. Rather than using forecasts to shape their portfolios, they use them to ensure that their portfolios aren’t blindsided by unseen risks. Here’s how to make anti-forecasting work for you.
Anti-forecasting begins with the proposition that nobody can predict the future. At first this seems hard to believe. Why would banks and brokerages employ so many staff to produce forecasts that don’t pan out? It’s an excellent question — and the answer may be that if you’re running a bank or brokerage you have to act as if you can predict the future, because clients expect it. Fortunately for financial institutions, most clients don’t keep tabs on how forecasts pan out. If they did, clients might be dismayed.
Flash back to December of 2007. Most economists in Canada and the U.S. predicted the economy would stride ahead in 2008, growing by 2% to 3%. They saw equally good or better prospects in store for 2009. The forecasters believed home prices in Canada would either stabilize or make small gains. While sliding home prices in the U.S. were recognized as grounds for concern, the forecasters thought other factors, such as growing U.S. exports because of a cheaper greenback, would provide enough boost to offset the real estate downturn. Typical of the measured optimism of the time were comments by Jeff Rubin, chief economist and chief strategist at CIBC World Markets: “A two-quarter slowdown, contained by more Fed rate cuts, should set the stage for a gradual re-acceleration of economic growth over the course of 2008, and a subsequent rally in North American stocks to new cyclical highs.”
Look what actually happened. The Canadian and U.S. economies slowed in early 2008. U.S. investment banks crumbled. The downturn in U.S. home prices gained speed and turned into outright collapse. Global stock markets crashed. By late fall, it was clear that the U.S. was in recession and the world was in the grip of the biggest financial crisis since the Great Depression. While it had not been confirmed that Canada had fallen into recession as we went to press, home prices had begun to fall, the Toronto Stock Exchange was in freefall, and Stephen Harper was vowing to fight the downturn.
The most candid forecast during this dreadful period came in October from the University of Toronto Institute for Policy Analysis. As markets crashed, it issued a report with the headline “We Don’t Have a Clue and We’re Not Going to Pretend that We Do.” Forecasters Peter Dungan and Steve Murphy wrote in the report: “The short-term outlook at this point depends on a huge number of either/ors, some of which we do not even pretend to understand, never mind try to forecast.”
Forecasters are rarely so honest about their predicting abilities. Rather than confess to ignorance, most will continue to make forecasts, adjusting them as necessary. After all, that’s what economists are paid to do, despite a marked lack of success.
Forecasters have a particularly bad record when it comes to calling turning points in the economy. Most of them failed to predict the extent of the dotcom crash at the turn of the century. Most failed to identify the dire repercussions of the U.S. housing bubble. And most also failed to predict any of the last three U.S. recessions, based on evidence from the quarterly survey of economic forecasters by the Federal Reserve Bank of Philadelphia. “I don’t think we, as a profession, ever had an ability to forecast recessions,” Jeffrey A. Frankel, professor of economics at Harvard University told The New York Times in 2007. “It’s hard enough to know when a recession has started, looking at it with hindsight.”
Even when the economy is growing, forecasts are blurry. Steve Murphy of the University of Toronto Institute for Policy Analysis and Bryan Campbell of Concordia University reviewed Canadian economic forecasts from nine banks and other organizations over the period from 1984 to 2003. They found that the average of the annual forecasts for growth in gross domestic product missed the actual result by an average of 1.3 percentage points. That is, if the economy grew by, say, 3%, the average forecast would typically have been 1.7% or 4.3% — not much help at all if you’re trying to plan your investing strategy based upon those forecasts. Why is the economy so hard to predict? At any one time, there are multiple forces pulling it in different directions and it’s hard to know which forces will prevail. To make the challenge even more difficult, collecting numbers takes time and forecasters are often working with data that is months old.
But the problem extends beyond data or economic models. The future seems to have an endless capacity to surprise us, and it’s not just economists who get their forecasts wrong. Political observers and social commentators also mess up. One of the most comprehensive studies of forecasting accuracy was performed by Philip Tetlock, a professor of business and political science at the University of California at Berkeley. Beginning more than 20 years ago, he selected 284 people who make their living offering advice on political trends and asked them to assess the probability that various events would occur: would Canada split up?; would the U.S. go to war in the Persian Gulf?; would apartheid end peacefully?; and so on. All told, the experts made more than 82,000 forecasts by the time the study ended in 2003.
When Tetlock analyzed the results, he came to a surprising conclusion. Experts can’t predict broad trends any better than a random guess would. And specialists in an area can’t forecast more accurately than non-specialists. Yes, subjects who know a little about an area can forecast better than someone who knows nothing, but knowing a lot doesn’t seem to improve forecasting accuracy. Forecasters seem to rapidly reach a point where adding more knowledge doesn’t make predictions any more reliable.
Rate this article
