Bryan Caplan  

A Controversial Issue Resolved With Econometrics

PRINT
Does Econometrics Resolve Issu... Local Knowledge and Mortgage L...
In answer to David's challenge to name "one controversial issue that had been resolved with econometrics," I nominate the self-interested voter hypothesis - the idea that voters maximize their narrow self-interest.  While many people - scholars included - take this view for granted, the vast majority of people familiar with the data now agree that it's wrong.  In reality, objective self-interest and political views have almost nothing to do with each other.  Any apparent connection usually just reflects the observers' confirmation bias.

Of course, the "people familiar with the data" tend to be political scientists, not economists.  But it was econometrics that convinced them.  (And yes, political scientists do use the word "econometrics" when they apply statistics to political questions.  "Politimetrics" never caught on).


Comments and Sharing





COMMENTS (17 to date)
Gian writes:

"objective self-interest" appears to be a particularly freighted term and loads of philosophical biases are already built into it.

What is my "objective self-interest" and could you know anything about it, without knowing me personally?

Could even I know my "objective self-interest"?

ed writes:

I think there are lots of examples of where econometrics has changed people's minds. David loaded the question a little when he asked for "controversial issues." Issues tend to be controversial precisely where the data are inconclusive, or where no data exist.

I can think of lots of examples in finance, where there is some pricing "anomaly" such as the accrual effect, or momentum, or whatever. Many people have been convinced, but usually only after controlling for other factors using some sort of econometric pricing model (such as the Fama-French 3 factor model.)

I think there are also examples in education research. For example: charter schools on average seem to perform very differently from regular public schools, student demographics are more important than teachers, but teachers are still fairly important, etc. Econometrics has certainly changed my mind about things like that, and I don't think I'm alone.

John writes:
"objective self-interest" appears to be a particularly freighted term and loads of philosophical biases are already built into it.

What is my "objective self-interest" and could you know anything about it, without knowing me personally?

Could even I know my "objective self-interest"?

If "objective self-interest" is meaningless or indefinable, then it clearly doesn't affect our voting behavior in a meaningful or definable way. The self-interested voter theorem, to be interesting, must say something predictable--for example, that my stance on new taxes is largely dependent on whether I will personally be subjected to them.

"Voters self-interestedly vote for whatever they think is ultimately best for themselves" is almost tautologically true, but uninteresting, since it has no explanatory or predictive power.

Andy writes:

I would rephrase it as "Voters self-interestedly vote for whatever they think is ultimately best for themselves to have voted for".

Peter St. Onge writes:

Good point.

@Gian: the studies Bryan cites are fairly clear in definition of self-interest. Eg preferences on flood-control spending depending on whether your house is in a flood plain or are you just taxed for no personal benefit.

Yes there are valid philosophical points, but it's not an impossible task to pretty strongly outline what is and isn't objective self-interest.

neal writes:
And yes, political scientists do use the word "econometrics" when they apply statistics to political questions. "Politimetrics" never caught on

I know it's a lost battle, but I've never really seen the point of calling statistics applied to economics 'econometrics' as if it's fundamentally different (same applies to "biometrics"). Even more so when you start applying "econometrics" to politics! Risks disciplinary silos and reduces cross fertilisation of ideas

Its just stats!

Chris writes:

The debate about the anti-poverty effects of minimum wage v. EITC has been pretty well resolved by econometrics. EITC is the clear winner. Those who continue the debate are now arguing around what is politically feasible (DeLong once argued that politicans will almost always favor min wages because it is easier to rally voter support around min wages than a tax rebate).

Now, whether this is "controversial" enough an example for Prof Henderson....

Pat T writes:

I wonder what econometrics has to say about the phenomenon of humblebragging.

David R. Henderson writes:

Bryan,
I just read through your notes that you linked to above on the Self-Interested Voter Hypothesis. Outstanding! They will be part of what I teach when I next do a segment on Public Choice.

Seth writes:

I've read through your notes as well and I am still unconvinced.

What the notes tell me is that voters are bad at determining the actual consequences of their votes or that they make trade-offs when they vote. That doesn't mean that their vote isn't self-interested.

It seems the definition for self-interest does not specifically address the direct benefits of voting for a lot of people. The closest it comes to those direct benefits is group-interested and party identification, but even those are too aggregated to capture the individual direct benefits of voting, which are many and varied.

For example, I know folks who view their vote as a signal that they care about the intentions of some program or ideology. Even if it impacts them negatively in some other way, it serves their self-interest because they think it allows them to claim the moral high ground in political debates. They also say they sleep better at night for voting with their conscience.

It may all come down to is trade-offs and the econometrics is missing the other side of the trade. Folks make trade-offs when they vote with dollars and in the voting booth.

I doubt many economists would say that someone wasn't serving their self-interest because they gave up $2 for a cup of coffee. It seems the econometrics here is saying "a person gave up $2, why would they do that? How does giving up $2 serve their self-interest?" Of course, we all know it serves their self-interest because they valued the cup of coffee more than the $2.

When people vote at the booth they make similar trade-offs. I may vote against a politician that supports a program that directly benefits me, because I can claim the moral high-ground. That trade-off may be worth it.

Or, I may just be a sucker and I assume all the moral high-ground benefits of my vote, without understanding what I traded off.

Ezadarque writes:

I would point the economic effects of migration restriction - perhaps trillions of dollars

http://www.cgdev.org/files/1425376_file_Clemens_Economics_and_Emigration_FINAL.pdf

David R. Henderson writes:

@Seth,
Based on his book, I’m positive that Bryan wouldn’t disagree with you. The key is the adjective “narrow” in front of self-interest.

Trespassers W writes:

@neal:

I claim, with tongue only partly in cheek, that econometrics is just linear regression done very carefully on natural (rather than experimental) data.

Yes, it's "just" a subset of statistics, but econometricians face somewhat different issues than, say, biostatisticians, so there's some justification for calling it something different.

Arthur_500 writes:

the typical voter votes with emotion rather than objective or any other kind of self-interest.

Taking care of the poor is not in my objective self interest. But, it makes me feel good - especially if someone else does it like the gubment.

Ultimately there are too many issues to narrow a vote so the voter goes with habit or some specific issue to which they have an emotional attachmnet.

Mike Rulle writes:

Most of us commenters, myself included, immediately jumped on the term and phrase "objective self interest" and "maximize narrow self interest" with good reason. These are value laden phrases which already makes us suspicious as to who has the confirmation bias.

I prefer actual more "objective" standards to measure voting preferences against. One example might be "people tend to vote against their short term monetary interest"; or, people will often vote for a person who supported a law because they believe the law did "X", when it actually does "Y".

Using phrases like "objective self interest" creates confusion as it has to be defined in such a way that many will disagree, relative to everyday usage.

A racoon writes:

An experiment where people are asked to vote on discrete question is not analogous to the votes we actually cast because people only rarely vote on discrete issues (I can't think of any Federal referendums).

We mainly vote to pick one person among 2 to do a job. Those 2 persons have made lots of promises, will vote on many bills not yet printed, and may face crises not yet anticipated. Every single candidate will certainly act contrary to your desires at least some of the time and have loyalties divided among voting constituents, contributors, party politics and self-regard.

I know no better way to vote than trusting your intuition about a candidate's common sense, humility and integrity. But most people (~2/3) use political party as proxy for whatever they care about.

(Experiments also don't replicate the emotion or effort most people take into the booth; in fact the entire ritual of community voting imbues seriousness and focus.)

Kendall writes:
student demographics are more important than teachers, but teachers are still fairly important

I thought I heard on an econ Talk episode that the difference in performance between teachers in the same school was greater than the difference between schools? If this is correct does that fit with student demographics being more important?

Comments for this entry have been closed
Return to top