Re: future of statistics and econometrics

On Mon, 10 Aug 2009 16:46:51 -0700, Richard Startz
<richardstartz@xxxxxxxxxxx> wrote:

Let me disagree a touch with Rich with regard to the econometrics part
of his responses.

Thanks, I'm pleased to hear more information.
I've got a couple of casual questions, below.


On Mon, 10 Aug 2009 16:17:55 -0400, Rich Ulrich
<rich.ulrich@xxxxxxxxxxx> wrote:

Nobody has posted any comment in a couple of days, so I will
take a shot.

On Sat, 8 Aug 2009 07:57:24 -0700 (PDT), oercim <oercim@xxxxxxxxx>
wrote:

Hello. As being a fan of statistics and econometrics , I
wonder the road of statistics, recent developmets and future of
statistics and econometrics (especially in methodogical areas and
mathematical statistics) .

Are you interested in "econometrics" or "statistics"?

I've only brushed against econometrics, and it seems to
be engaged in building enormous models, which use a lot
of mathematical sophisitication. Testing is difficult or
not attempted. The problem may be ... finding or shaping
the raw data to fit the terms of the model.

Econometricians usually estimate single equations. "Enormous models"
were in vogue 30 years ago. And yes, getting useful data is tough.


I was thinking of huge, matrix models, and I never heard what
happened to them. - didn't seem like that long ago.

Maybe, once the computers were big enough to 'fit' them,
they didn't add that much? Or, what?



snip

I don't associate MCMC methods or bootstrap with econometrics,
so here is where I think you are referring to other general areas
of statistics.

Econometricians certainly use MCMC methods and the bootstrap. In fact,
we're happy to steal anything that statisticians invent and forget to
hide.

Doesn't a lot of the work in econometrics have to deal with time
series, and parallel time series? I guess that one reason I
didn't expect MCMC to apply so much is that time series data are
especially hard to model that way.


The central issue in econometrics is almost always pushing out the
limits of how to control for causality in the absence of controlled
experiments.


Thanks,

--
Rich Ulrich
.

Relevant Pages

  • Re: future of statistics and econometrics
    ... statistics and econometrics (especially in methodogical areas and ... I've only brushed against econometrics, ... Maybe, once the computers were big enough to 'fit' them, ... Pagan, Three econometric methodologies: ...
    (sci.stat.math)
  • Re: future of statistics and econometrics
    ... statistics and econometrics (especially in methodogical areas and ... I've only brushed against econometrics, ... Maybe, once the computers were big enough to 'fit' them, ... Pagan, Three econometric methodologies: ...
    (sci.stat.math)
  • Re: future of statistics and econometrics
    ... Let me disagree a touch with Rich with regard to the econometrics part ... statistics and econometrics (especially in methodogical areas and ... be engaged in building enormous models, ... Econometricians certainly use MCMC methods and the bootstrap. ...
    (sci.stat.math)
  • Re: Econometrics in Panel data?
    ... econometrics stuff? ... languages, along with Python. ... There are many statistics packages and programming languages used in ... Enough programmers still use Fortran that major hardware vendors such ...
    (comp.lang.python)
  • ANNOUNCE: Hidden Markov Models short course April 2005, Sydney
    ... Institute of Statistics and Econometrics, ... Hidden Markov models have been used for about three decades in signal ... general-purpose models for time series, ...
    (sci.stat.edu)
Loading