A user’s guide to economic forecasts

  • Share
  • Read Later

As best I can parse it, there are four ways to forecast the economic future. None of them works very well, but each has its uses. And with lots of people right now wondering whether the subprime lending mess and the global stock market jitters presage an economic downturn, I think it might be helpful to trot out my taxonomy (additions, subtractions, and editing suggestions are more than welcome):

1) Assume the future will be like the recent past. This sounds kinda stupid, but it’s how most of us get through our daily lives, and it’s the guiding theme of most economic forecasts generated on Wall Street and elsewhere. Most of the time, it works pretty well: Trends continue, economic growth begets more growth. But this approach of course fails at the most important economic turning points. Which is why, as Dan Gross wrote (available only to TimesSelectoids) in the New York Times a few weeks ago (the subject is a hardy perennial for financial journalists), economists have such a poor record at forecasting recessions.

2) Look for particular indicators that in the past have presaged upturns or downturns. The high priest of this work was the late Geoffrey H. Moore, Alan Greenspan’s statistics professor and the guy who created the Index of Leading Economic Indicators in the 1960s. Moore’s work is carried on today by the Economic Cycle Research Institute, which he founded and which did just about the best job of any forecasting outfit of calling the 2001 recession. But ECRI didn’t so much predict that recession as announce that it had arrived before anyone else did. As soon as you start trying to use the indicator method to look more than a few weeks into the future, you get into trouble, as Moore did in the spring of 1996 when he told Fortune that he had finally tracked down the perfect economic indicator–and it predicted a recession beginning in December of that year. Didn’t happen. Not even close.

3) Assume that what goes up must come down, and vice versa. This was the Leitmotif of Roger Babson, the most famous economic forecaster of the Teens and Twenties of the 20th century. If price-earnings ratios (or the ratio of house prices to rents, or any other such relationship you can come up with) are above the long-term norm, then you should expect them to come down. Yale economist Bob Shiller is the leading exponent of such thinking these days, but lots of other sensible people share this mindset (I certainly do, although I’m not saying I’m sensible). There are two big problems with using it as a forecasting tool: One is that past financial relationships don’t inevitably have to apply in the future–that is, there might be good reasons for price-earnings ratios to be higher now than they were in the early 1900s. But the more significant issue is one of timing. It’s undeniably true that financial markets, and to a lesser extent economic activity, tend to overshoot on the upside and the downside. Eventually, things revert to the mean. But there’s no telling when they’ll do it. Babson is famous for having predicted the 1929 crash, but he started predicting it in 1926. Similarly, Shiller started warning of stock market troubles ahead in the mid-1990s, just in time for several years of spectacular gains.

4) Construct scenarios of how economic forces will shape the future. There was a time, back in the mid-20th century, when some scholars thought complex models of the economy could be used to predict its future path. Nobelist Lawrence Klein was the leading proponent of this in the U.S., and Klein-style models (Klein himself founded the forecasting firm WEFA, which is now part of Global Insight) are still used widely on Wall Street and Washington. But nobody entirely believes in them anymore–the global economy is simply too complex to be captured in any model. Instead, academic economists build much simpler models to explain one phenomenon or another, and a few Wall Street economists (I’m thinking here especially of Morgan Stanley’s irrepressible Steve Roach) spin even less formal if-then scenarios of how different trends might interact. The problem here is that, well, the global economy is way too complex to be captured in any model. Such an approach can be enormously useful in explaining the recent economic past, and every once in a while it will deliver a spot-on prediction. But you can’t rely on it.

There’s one other group that I’ve left out–the stopped-clock types who invariably see either disaster or prosperity ahead. These people can be worth paying attention to if they’re good at identifying weaknesses in the arguments of those who disagree with them (Bill Fleckenstein springs to mind here) but I think it’s a stretch to call them forecasters.

So where does that leave us right now? According to the first approach, we should expect moderate economic growth as far as the eye can see. As for the second, the folks at ECRI haven’t sounded any alarms yet, either–but then they probably won’t until a recession or at least a significant slowdown is already upon us. The third, Babson/Shiller school would see cause for alarm: Stock prices are still high by historical standards, albeit not crazy high like a few years ago, and home prices are too. And finally, those who spin scenarios, like Morgan Stanley’s Roach, are generally spinning worrisome ones at the moment. I’d call this a moderately negative outlook. But then, what do I know? What does anybody know?

Update: This just in from the W$J:

Most economic forecasters in a new WSJ.com survey believe recent turmoil in the subprime mortgage market is likely to spread to the broader mortgage market and they expect a widely followed index of home prices to fall this year. But they still think the U.S. will avoid a recession and even a significant rise in unemployment.

As if they know.