Sunday, January 18, 2009

Statistics and rare events

People with a fondness for statistical foolishness should check out Statistics.org for a regular dose of the follies perpetrated in the media. Sometimes ESPN serves the same purpose. A writer there commented that in a match between two pro teams, one of them was coming up against the "law of averages" because they'd already won the previous two games.

But in addition to the general misuse of statistics, there is also the problem of a general disregard for the wings of a normal distribution. Unlikely events happen, or to put it another way, the law of averages says that sooner or later, things that seem unlikely do in fact happen. Maybe not often, maybe not soon, but they will happen.

The problem with public fiscal policy is that it's impossible to impress that fact on the voting public. They will always favor the solution that discounts what they deem to be unlikely if there are short-term rewards. Such as having happy public employees, who are promised secure pensions at a rate that presumes no evil economic periods in the future, without enough taxpayer contributions to ensure the result.

All across America, there are public entities who made big promises that they now can't keep. Not the feds, of course, who simply print whatever they need, but at the state, county, and local levels, disaster looms. It's one thing to hope for an 8% ROI on pension assets, quite another to bet the taxpayers' future on it.

There's going to be a lot of resistance from public unions, with a lot of "but you promised ..." They are right that we made the promises and hired politicians on the basis of easy answers. On the other hand, we simply can't fulfill those promises and compromises are going to be made or bankrupcies will occur.

No comments: