The Economist: The art of forecasting

Tucker Hart Adams //October 1, 2011//

The Economist: The art of forecasting

Tucker Hart Adams //October 1, 2011//

Economists have been taking it on the chin recently. After the failure of most of my colleagues to warn of the Great Recession – some arrogantly boasted that their computer models had eliminated risk and the possibility of recession – the profession’s ability to foretell the future has pretty much been written off.

I’ll let you in on a secret: Neither I nor any of my colleagues has ever had any special ability to foretell what lies ahead, no matter how many newt eyes and frog toes we stir into the black cauldron that produces our forecasts. If we did, we’d all be enormously wealthy.

If we can’t foretell the future, what can we do? Or, is forecasting just a giant scam? 
Fortunately, there are some things that economists do better than other disciplines, things that add value to a business or investment decision.

First, we can analyze where the economy is today. That’s not as easy as it sounds, since data are released with a substantial time lag. And, there are frequent revisions, as we’ve seen with output growth for the first two quarters of 2011. I learned that lesson back in the 1980s, when customers were telling us times were terrible, but the Department of Labor was telling us jobs were growing rapidly. Months later data were revised to show that not only was there no job growth, jobs had been disappearing for months.

Second, we can analyze how we got to where we are. Why are interest rates low? Why are housing prices still falling and foreclosures rising? Why are so few jobs being created? Why is Facebook successful, while MySpace is not? If you have a good understanding of where we are and how we got here, you are probably way ahead of the competition.

A third thing we attempt to do is to identify which of the forces already in place will be important determinants of the path that lies ahead. And, we can speculate about the surprises – we call them exogenous events – that will change the expected course of the economy. This is where what we do becomes more art than science.

There’s a final component of what we do that adds value, and it is the reason we dare to attach figures to the future, why we are paid to say that output will grow by x, the mortgage rate will increase to y and housing permits will decline by z. This is useful not because we are going to be correct, but because there is internal consistency to our forecasts – we wouldn’t say consumer spending will plunge, causing industrial output to grow, or that mortgage rates will rise, causing foreclosures to decline.

And, when planning is done based on an explicit, internally consistent forecast, it means everyone is operating from the same set of assumptions, whether he agrees with them or not. You then know what to monitor as the year unfolds and where to adjust your plans if inflation or interest rates or exports differs significantly from what was assumed when the plan was conceived. This is the part of an economist’s job that requires a tad of arrogance and a generous portion of fairy dust.

That said about why a forecast may be useful, what went so wrong in the forecasting community in 2007? The short answer is that there was too much dependence on computer models. A few years after I emerged from graduate school, shiny new Ph.D. in hand and a concentration with honors in econometrics (statistical economics and computer modeling), I had an epiphany. I wasn’t going to build the perfect model of the economy and neither was anyone else. The economy is too complex, too many simplifying assumptions are necessary and the data are too inaccurate. There had to be a better way to spend my time.

If you want to know my secret, don’t miss the November issue of ColoradoBiz.
{pagebreak:Page 1}