You’re not Nostradamus

I was looking through the blog for the Portland Mercury and ran across the headline “Millions of New Residents, Not Nearly as Many New Jobs Projected for Portland in 2050.” The projection of a lack of future jobs 40 years into the future bothered me a lot less than the fact that someone was projecting job rates and population levels 40 years into the future.

That someone, as it turns out, is Metro, the regional elected government for the Portland metropolitan area. The numbers the blog post referenced appear to be from the “20- and 50-year regional population and employment range forecasts.” The executive summary for this report states, among other things, that “the forecast indicates a 90 percent chance that there will be between 1.3 and 1.7 million jobs in the statistical area in 2030 and a 90 percent chance that there will be between 1.7 million and 2.4 million jobs in the same area in 2060.”

How 2060 became 2050 for the Mercury is something I choose not to think about too much.

What haunts my thoughts instead is the reminder that the bureaucratic fetish known as forecasting has moved beyond the point of silly to absurd—and slightly insane. To understand my concern, we have to think about what exactly this report is claiming to do: It is claiming to tell a fortune, and not just any fortune, but the fortune of millions of people during the next 50 years.

The problem here is that this fortunetelling is rooted in some very poor assumptions, and my mother taught me that when you assume, you make an ass out of yourself. Just as you do by basing urban planning for a medium-sized metropolitan area on spurious growth projections.

Yes, it is true that we all try to predict the future to some degree, and all those predictions are rooted in assumptions. Like predicting the bus will arrive at 9:15 a.m. based on the assumption that the numbers written on the bus schedule are there for more than just decoration.

The difference is that bus-schedule assumptions are rooted in fairly consistent patterns—the bus usually arrives somewhere in the vicinity of its scheduled time—and the chances of something interrupting that pattern between now and when I need to catch the bus are pretty low. Nothing game-changing is likely to happen by then.

For the same reason, a lot of forecasting is at least somewhat reasonable. Forecasting tax revenue for the state of Oregon for each biennium, for instance, is sensible because things like employment and willingness to pay taxes do not usually change drastically over the course of two years. Game-changing events are not things you can predict.

Another war could break out, we could have a flood or a blizzard, people could stop wearing shoes, causing Nike to go out of business, or a massive earthquake could destroy the entire town. No single one of these events is likely to happen in any given year, but the further in the future you try to predict events the more likely it is something will happen significant enough to render your prediction useless.

And 50 years into the future, that is more or less guaranteed.