Assuming Predictability

There is a reason that economists are so rarely featured in jokes. They tend to be rather dry people, who can and will estimate anything under magical “assumptions” designed to prove their point. In fact, the only jokes with economists I know involve them making assumptions. Which is really a shame because they make hilarious statements like the following all the time:

Economists are nearly unanimous that Ben Bernanke should be reappointed to another term as Federal Reserve chairman, and they said there is a

71% chancethat President Barack Obama will ask him to stay on, according to a survey.

Wow. Seventy-one percent, eh? That’s an awfully exact number for a prediction isn’t it? It suggests there was some real numbers computed and weighted in order to arrive at a probability slightly less than 5 out of 7 that Bernanke would be reappointed.

But where did those numbers come from? And what exactly would they be? Moreover, who were the “economists” that are so “nearly unanimous” who arrived at this prediction? Inquiring minds want to know.

Of course, per the article, these same economists are all enthusiastic about the economy having bottomed out, and that Bernanke is largely responsible for that, er … “success”. Perhaps they only interviewed family members of the Fed Chairman who also happen to be economists? “*Five out of 7 Bernankes agree!*”

By the way, there is a 26.3% chance that only economists and statisticians will comment on this post, but I’m keeping my top-secret formula for arriving at that conclusion all to my self. I’ll give you a hint, though: assume I know what you’re thinking.

Michael, I think the real problem is the journalist who converted a survey of economists of which 71% predicted Obama would keep Bernanke into “they said there is a 71% chance.”

71% of respondents…

I did consider that, Harun, which is where I came up with the “

Five out of 7 Bernankes agree!”line. But the way the article was written makes the chances of such unlikely. Using my infallible formula I’d say … approximately 52.34% less likely than some economists pretending they can predict reappointment outcomes with any kind of precision.“Try new Whizzo Butter! 9 out of 10 ecconomists can’t tell the difference between Whizzo Butter and a dead crab! “

I think that there’s a 101% chance that Obama will or won’t ask him to return. The extra 1% is due to rounding.

There is a 99.99% chance Erb will show up here and tell you why you are wrong.

Here’s an economist joke: Q: What’s the difference between meteorologists and economists? A: Meteorologists can agree what the weather was like yesterday.

Meteorologist 1: “PARTLY CLOUDY!”

Meteorologist 2: “PARTLY SUNNY!”

71% of all stats are just made up on the spot.

“there is a 71% chance that President Barack Obama will ask him to stay on, according to a survey…”

Wow! Did you hear about the other odds calculated by some other surveys? For instance:

92% say that Keith Dimbulbermann on MSDNC is a complete moron. The other 8% said he a total moron.

82% say that the Christmas shopping season will be terrible this year. The other 18% all said that they think it will be a good season, except that since they lost their jobs they won’t have any money to spend to enjoy it.

22% agree that if you take the word “OBAMA” and turn it around, it spells “AMABO.”

When asked, 100% of the people said that they could not name one member of Obama’s cabinet. When prodded, 99% said that they could care less; the other 1% would like to answer, but they are busy beating up people for the unions at town hall meetings.

3 out of 4 people polled said that “Bernanke” is some kind of fish. The fourth said it was the name of an underarm deodorant.

The problem with equating “71% of respondents” and “71% chance” is that it works on the assumption that all respondents have an equal probability of being correct. It is clearly more accurate to simply state the facts. An aside: for meteorologists, a “71% chance of rain” does not mean they think 71% is the probability of rain. It means they estimate that it

willrain in 71% of the viewing area.Speaking of which, this book,

How to Lie with Statistics, should be required reading for every American.