[Paleopsych] NYT Mag: Bush Landslide (in Theory)!
Werbos, Dr. Paul J.
paul.werbos at verizon.net
Thu Aug 19 20:27:46 UTC 2004
Fair's model (below) is interesting. Fair is a highly respected educator.
I vaguely remember thirty years ago... that one of my students tested
something like Fair's simplified econometric model of the US economy, and
showed how it could
predict more accurately if retuned using certain algorithms I developed.
But... there are caveats.
The "2.5 percent (mean?) error"... is for a limited set of observations.
A limited number of elections.
It's a nice simple model, it sounds like. Two inputs, one binary output.
Now consider another. One input: the last digit of the year of the election.
The output: whether the elected President gets shot.
Fewer variables, more cases (since the data goes back further.)
Not even 2.5 percent mean error; zero error. Even more reliable.
So if we believe Fair, we should have even more confidence that Bush will
President Cheney. How about that?
What do we make of this?
We shouldn't totally discount either model -- but we should remember that
additional information, as the Bayesian would say. Information about
complex cause and effect aspects. Information about equally simple models
that may point
in other directions.
In fact -- predicting terrorism and predicting war and predicting the stock
market involve similar issues.
The best experience from stock market predicting is all mucked up by
proprietary barriers to
giving away information. But I gather there are periods where simple models
by phase shifts... and then by other simple models... and so on.
For terrorism and war and such, I would look for a totally different kind
of model, making
better use of the richness of the data. Too bad no one seems to know how
who does it for
a living -- but really, it has subtle aspects.
And right at the moment I tend to feel that such a model would be a tad
more favorable to
Kerry than to Bush right now, though the world is full of free will
could change it either way.
Fair's model also reminds me a little of some amusing econometric models I
For example, how people's expectations of oil price twenty years' hence is
by stuff like the one-year average trend and the present price. How energy
prices predict the
DOE budget much better than they predict energy consumption (in the past...
People's expectations of what happens to the economy under Bush versus
undoubtedly be one causal factor here... among others...
At 09:15 PM 8/18/2004 -0400, Premise Checker wrote:
>Bush Landslide (in Theory)!
>QUESTIONS FOR RAY C. FAIR
>New York Times Magazine, 4.8.15
>Interview by DEBORAH SOLOMON
>Q As a professor of economics at Yale, you are known for
>creating an econometric equation that has predicted
>presidential elections with relative accuracy.
>My latest prediction shows that Bush will receive 57.5
>percent of the two-party votes.
>The polls are suggesting a much closer race.
>Polls are notoriously flaky this far ahead of the election, and there
>is a limit to how much you want to trust polls.
>Why should we trust your equation, which seems unusually
>It has done well historically. The average mistake of the
>equation is about 2.5 percentage points.
>In your book ''Predicting Presidential Elections and Other
>Things,'' you claim that economic growth and inflation are
>the only variables that matter in a presidential race. Are
>you saying that the war in Iraq will have no influence on
>Historically, issues like war haven't swamped the
>economics. If the equation is correctly specified, then the
>chances that Bush loses are very small.
>But the country hasn't been this polarized since the 60's,
>and voters seem genuinely engaged by social issues like gay
>marriage and the overall question of a more just society.
>We throw all those into what we call the error term. In the
>past, all that stuff that you think should count averages
>about 2.5 percent, and that is pretty small.
>It saddens me that you teach this to students at Yale, who
>could be thinking about society in complex and meaningful
>I will be teaching econometrics next year to
>undergraduates. Econometrics is a huge deal, because it is
>applied to all kinds of things.
>Yes, I know one of your studies used the econometric method
>to predict who is most likely to have an extramarital
>In that case, the key economic question was whether
>high-wage people are more or less likely to engage in an
>affair. They are slightly more likely to have an affair.
>But the economic theory is ambiguous because if your wage
>is really high, that tends to make you work more, and that
>would cut down on how much time you want to spend in an
>Are you a Republican?
>I can't credibly answer that question. Using game theory in
>economics, you are not going to believe me when I tell you
>my political affiliation because I know that you know that
>I could be behaving strategically. If I tell you I am a
>Kerry supporter, how do you know that I am not lying or
>behaving strategically to try to put more weight on the
>predictions and help the Republicans?
>I don't want to do game theory. I just want to know if you
>are a Kerry supporter.
>Backing away from game theory, which is kind of cute, I am
>a Kerry supporter.
>I believe you entirely, although I'm a little surprised,
>because your predictions implicitly lend support to Bush.
>I am not attempting to be an advocate for one party or
>another. I am attempting to be a social scientist trying to
>explain voting behavior.
>But in the process you are shaping opinion. Predictions can
>be self-confirming, because wishy-washy voters might go
>with the candidate who is perceived to be more successful.
>It could work the other way. If Kerry supporters see that
>I have made this big prediction for Bush, more of them
>could turn out just to prove an economist wrong.
>Perhaps you could create an equation that would calculate
>how important the forecasts of economists are.
>There are so many polls and predictions, and I am not sure
>the net effect of any one of them is much.
>Yes, everyone in America is a forecaster. We all think we
>know how things will turn out.
>So in that case, no one has much influence, including me.
>paleopsych mailing list
>paleopsych at paleopsych.org
More information about the paleopsych