Bryan Caplan  

Why Academic Economists Aren't Bayesians

What Is Money Velocity?... An Apology...
I don't think I'll find a better answer than Robin's:
[T]he main social function of academia is to let students, patrons, readers, etc. affiliate with credentialed-as-impressive minds.  If so, academic beliefs are secondary - the important thing is to clearly show respect to those who make impressive displays like theorems or difficult data analysis. And the obvious way for academics to use their beliefs to show respect for impressive folks is to have academic beliefs track the most impressive recent academic work.

So it won't do to have beliefs bounce around with every little common sense thing anyone says, however informative those may be.  That would give too much respect to not-very-impressive sources of common sense. Instead, beliefs must stay fixed until an impressive enough theorem or data analysis comes along where beliefs should change out of respect for it.  Academics also avoid keeping beliefs pretty much the same when each new study hardly adds much evidence - that wouldn't offer enough respect to the new display.

Relative to the Bayesians that academic economic theorists typically assume populate the world, real academics over-react or under-react to evidence, as needed to show respect for impressive academic displays.
Still, I should point out that Robin has only explained, rather than justified, academic economists' deviations from Bayesianism.  Anyone care to try to do that?

Comments and Sharing

COMMENTS (18 to date)
dWj writes:
Foobarista writes:

Why bother? Something is a human trait or it isn't.

I would think that economists, in particular, should be wary of normative arguments.

Doc Merlin writes:

Um, but humans aren't Bayesians... at all. Bayesian thinking is just the add on to our models which work on the assumption of perfect knowledge to allow for imperfection in knowledge. Its all bunk, because humans don't intuitively understand probability or how probability works.

Johan writes:

Is being Bayesian computational feasible?

ao writes:

The truth content of a study is nonlinear in statistical significance and experiment rigor. The vast majority of studies are too weak to reject the null, which is predermined by the slim minority of studies that are good enough to matter. This is why beliefs formed by the Rand HIE, especially Robins, have not moved in 25 years: nothing has come close to it in terms of empirical validity. Outlier studies should drive 95% of our testable beliefs.

Michael writes:

I reread J. K. Galbraith's Affluent Society last night. He suggests "conventional wisdom" in his own delightful descriptive analysis.

Virginia Postrel suggests that people are dynamists or enemies of the future. I will note that these two arguments share some principle assumptions.

Now we can think of the academy and its role as a filter. There are so many hoops to jump through to prove that you are worthy of the title of professor. Do dynamists prosper anywhere besides subversive places like GMU?

David Zetland writes:

It's easy to justify: status quo is easy; change takes work. They don't want to work, and tenure (as well as "profession decorum") allows them to ignore the reality outside the ivory tower.

That's why I am quitting the academic world (even as I enjoy the thinking that *sometimes* occurs within it...)

Lee Kelly writes:

I don't think people should be Bayesians, especially of the subjective variety. That doesn't mean I approve of how economists make decisions, but merely that I do not care for Bayesianism as an alternative.

William Dunn writes:

re Lee Kelly (12:42):

"I don't think people should be Bayesians, especially of the subjective variety."


Bo Zimmerman writes:

Foobarista - They *should* be wary of normative arguments? Should you? Should I? What should we do with normative statements?

Sorry, I couldn't resist an apparant demonstrative contradiction...

Dirtyrottenvarmint writes:

I will justify. Here goes. (Actually, skip the next 3 paragraphs for the short and sweet answer.)

Bryan seems to want economists, and probably everyone else, to behave as if they were rational. Note that unless we are God, we cannot say definitively that we are behaving rationally. However, we can behave as we believe, given the inherent imperfectness of our information, we should behave if we are to be rational.

There is nothing irrational about an economist (or anyone else) creating models of human behavior in interaction with each other and scarce resources in which the assumption is that human beings are Bayesian, and at the same time behaving in his or her interactions with other humans and scarce resources.

This is because when modeling, we necessarily make assumptions in order to simplify the conditions and thus assist our necessarily limited intelligences in drawing some useful data from the model. This is true of all models, including the ones we carry around in our heads and through which we process our experiences. There is no rule that says that assumptions between models must be consistent. In fact, most of us think that it would be incredibly foolish to hold all assumptions constant amongst all models. This would negate all the benefits of being able to create and consider multiple models. Thus there is nothing sinister about creating academic models with the assumption that humans are Bayesian, yet following a different model or models in our daily interactions with others. Sapience means being able to think about multiple models. This is why the computer overlords have yet to take over the world.

In a nutshell: Your Model Is Not Reality.

As a side note, other comments mentioned Postrel, Galbraith, and the idea that the status quo is necessarily opposed to change. You are missing the obvious: it is possible for the status quo to be change. This is sometimes called Progress.

Jesse writes:

So Bryan and Robin think that academia all comes down to signalling. Am I really supposed to believe that they are actually hitting on the "truth," rather than just trying to signal their own intelligence by placing extreme weight on a particular aspect of human existence?

Daublin writes:

Foobista and others ask why bother.

The reason is that academics are supposed to be better. Universities put them through stringent review processes, and far more people apply than are admitted to the club. If the current process is not putting good knowledge seekers into these positions, then those processes should be amended.

To make an analogy, not all judges make fair and impartial judgements, and not every policemen saves the use of force for when it's really necessary. It's wrong in all of these cases.

Michael writes:

I would be very interested in the statement:

"You are missing the obvious: it is possible for the status quo to be change. This is sometimes called Progress."

This is not obvious -- for it to be obvious it would have to be first distinct and second clear for all that wanted to see it.

Progress and change are not a starting point. If they were then it wouldn't be definable. Could you give an example of a starting point that is in flux? Maybe we should take one of the clear examples just for those of us that don't see what you are saying.

axg writes:

[Comment removed pending confirmation of email address and for rudeness. Email the to request restoring your comment privileges. A valid email address is required to post comments on EconLog.--Econlib Ed.]

Johan writes:

"Academics also avoid keeping beliefs pretty much the same when each new study hardly adds much evidence - that wouldn't offer enough respect to the new display."

The obvious counterexample to this is mathematics.
In math people frequently reprove old theorems. They get respect for that, yet nobody updates their beliefs. (Since we already had a proof.)

The law of quadratic reciprocity has hundreds of proofs. Presumably the one who published proof number 100 was hoping to get respect from more than just the people who failed to be persuaded by the previous 99.

Dirtyrottenvarmint writes:


I think you may have misread. I said that "it is possible for the status quo to be change." You ask about a "starting point". The status quo is simply "the current state of affairs". No starting point is necessary in this formula.

The "current state of affairs" has been Progress - so we are told - for a very long time now. Read the NYTimes and they will tell you about the progress we are making.

As an aside to the aside, (!) though as mentioned my prior comment didn't have anything to do with starting points, for a (theoretical) example of how a starting point can be in flux I direct you to quantum theories of universe structure and formation. You may well ask, but what would be the starting point for that, but then we would be talking about religion, and this is an econ blog.

Dirtyrottenvarmint writes:


Of course they are self-interested! You have the opportunity to take what they say about their model(s), run that as data through your own models, and hope that you can draw some conclusions which lead you closer to the truth! What you conclude from your own model regarding the veracity of their statements is more information for your model. This is a good thing!

What, you thought someone was going to just hand you all the answers?

Comments for this entry have been closed
Return to top