Long Term
I once chaired a "Blue Ribbon" citizen's committee for the City of Los Angeles. Our job was to evaluate pension plan proposals and suggest alternatives. I was the actuary on the committee. We had investment people, labor people, pension administrators, and one economist.

At one point the economist offered to make a long term forecast -- five years. The rest of us just looked at him. In a pension plan you invest money to pay for retirements decades away. You may be putting aside money for a 25 year old to retire on someday. That's money that will be paid out perhaps 60 or more years in the future.

One job of the pension actuary is to model that payout. For me it was a humbling experience; I found that I couldn't predict interest rates one year in advance, let alone five or seventy. For that matter neither could or can anyone else.

Pension actuaries use models that are self correcting. We know that any one prediction is wrong, but over time, our models provide a relatively smooth way to fund those pensions in advance.

Most forecasters in other professions blithely assume their future will be accurate. Some even announce their results knowing they are wrong. A few go to the extent of producing alternatives (a worst case, best case and average scenario), but even these disguise their unrevealed assumptions. Who in their right mind would have predicted the fall of the Soviet Union until just before it happened?

I don't have a large point to make here; just that I am wary of any forecasts over a few days. From the Club of Rome to the Rand Corporation to the local weatherman, they don't have a way of being terribly accurate.

We never did find out what was in our economist's forecast. On the day he was to deliver it, he got stuck in rain stalled traffic and never got to our committee meeting.

The day before, the weatherman had predicted rain.