When an analyst is "shocked", is that usually a good sign?

[tpfkaw]

The economy cannot be predicted, pretty much by definition. There is nothing wrong with economists who predict things wrongly.

If, by definition, something cannot be predicted, why would somebody try to predict it?

Perhaps I should have qualified my statement. Some things can of course be predicted. Like, I can predict that China's GDP per capita relative to that of the US will be higher in 20 years than it is now. Or I can predict that the US debt will be higher next year because there will still be a deficit this year. Stuff like that.

But things that one makes money out of can't be predicted. Which is why few real economists do that. Of course, if someone pays you to make up predictions there is every incentive to do so. But thoe guys who make a living out of doing that on tv typically don't know as much as they pretend to.

That's actually the Austrian position you just took, y'know...

[tpfkaw]

Eh...how so?

That it's impossible to create an accurate economic model because the economy is far too complex and has too many variables to model with any degree of certainty.

(This is apparently an outrageous proposition to some folks, even though it's essentially chaos theory).

[tpfkaw]

No one claims they have an economic model that can produce absolute accuracy.

But plenty of people claim they have economic models that produce high levels of accuracy for years or even decades on end, which would have you laughed out of the room at a meteorological conference, for example.

I'd also point out I never said anything about "absolute accuracy," but whatever.

[tpfkaw]

Meteorologists have access to highly-accurate measurements of thousands of different variables, worldwide, in real time.  With this information, they are able to predict certain the values of certain variables (high/low temperature, humidity, dewpoint, etc.) in a given location for up to about 7 days.  After 7 days, the prediction of even the most sophisticated meteorological model loses any statistically-significant relation with reality.  Economists have access to questionable estimates of a few hundred different variables, in certain jurisdictions, months after the fact.  With this they claim they can predict the workings of the global economy for years or even decades.  This is blatant pseudoscience of the highest order.  It is mathematically impossible for any economic model to have any real predictive power (any data is too old to have predictive power for future data, and in any case even real-time data could only be useful for a matter of days).  Any accurate "modelling" would simply be the statement of a pseudo-tautology:  "When it's humid, it's likely to rain," would be a meteorological equivalent.