[lbo-talk] Output Falling in Oil-Rich Mexico, and Politics Get the Blame

[Miles Jackson]

Julio Huato wrote:
>
> In the jargon of economists, the mechanism that generates the
> time-series data on extraction or consumption rates is not stationary.
>  With non-stationary systems, models can be consistent with historical
> data but useless as predictive devices.  That's why I'm not impressed
> with peak-oil models, even if they have been adequately back-tested or
> calibrated to historical data.  The future is not a replica of the
> past.
>
>   
This is pretty rich: Hubbert used his model in the 1950s to correctly 
predict the peak of U. S. oil production in the 1970s!  Economists' 
misconceptions notwithstanding, it turns out you can effectively predict 
the future with a well-specified regression model.

Miles