Bryan Caplan  

Wanted: Counterexamples to Klein's "Forsaken Liberty Syndrome"

Next Malcolm X Book Club 9/14... Cristina Saralegui, The Dems, ...
When my favorite economists change, I usually think they change for the worse.  Dan Klein is a glowing exception.  I've always liked his work.  But in recent years (see e.g. here, here, and here) Dan's metamorphosized from very good to truly great.  The latest example: his "The Forsaken-Liberty Syndrome: Looking at Published Judgments to Say Whether Economists Reach a Conclusion." 

At first blush, this paper is just a cool meta-analysis of Econ Journal Watch's series, "Do Economists Reach a Conclusion?"  Klein argues that when economists publicly "go on the record," they offer better answers than they would in an anonymous survey:
[O]n-record scholarly judgments remain a special way to target knowledge and responsibility. When an economist steps forward to publish a judgment on a policy issue, he does so because he thinks he has something to add, and he attaches his name to the statement. He opens himself up to challenge. Such circumstances usually impel greater care. Moreover, the self-sorting means that the on-record economists are not a random sample of economists at large.  All this is not meant to suggest that stepping forward and going on the record means getting it right, only that judgment emerging from that combination is generally better than judgment solicited by others and rendered in seclusion and anonymity.

In this paper, the question that really ends up being the one that matters is this: How strong are the forces that impel an economist to step forward and publish a judgment that is more pro-liberalization than he really believes? I do not claim that such forces are entirely nonexistent, but I think they are small enough in relation to the healthy incentives to justify the approach pursued here.
After carefully reviewing the fascinating evidence, Klein concludes that on-record judgments are indeed noticeably much more pro-liberalization.  Then he bends over backwards to make his argument persuasive to skeptics - and to advance the cause of human knowledge:
It would be natural for a skeptic to object as follows: "We know that the researchers involved in DERAC are staunch believers in the free market. Even if we grant the findings of the studies provided, what about the studies not provided? What about cases in which it would be found that on-record economists are less pro-liberalization than at-large economists? We cannot count on you to turn those up for us."

The ad hominem is sound and proper, and the logic is flawless. There may well be cases in which on-record economists show a consensus in favor of some intervention, while the at large economists are less favorable to the intervention.

My own suspicion is that there are no such cases. It would be very interesting to query economists of various ideological stripes about whether they might postulate any such cases, or whether they agree that such cases are probably an empty set. Would any economist dispute my suspicion? (emphasis mine)
We should take up Dan's challenge.  Would you dispute his suspicion?  If so, how?  Please show your work.  If we're lucky, Dan will join the discussion in the comments.

COMMENTS (5 to date)
Daniel Klein writes:

Your post is very kind, Bryan. I hope you receive speculations of policy issues on which at-large economist judgments (as would be recorded in an anonymous survey) are more pro-liberalization than on-record economist judgments, even if hard evidence is lacking.

I would be glad even to hear people's guesses as to policy issues on which that would be most likely, even if the guesser thought that even the most likely cases were still unlikely.

Pace Mark Twain, the reports of disagreement among economists are greatly exaggerated. But they are about things other than they appear to be on the surface (the politics surrounding the issue).

Maniel writes:


In my work as an engineer, we generally use three categories of argument to support decisions: technical, to describe measurable attributes of performance (speed, reliability, capacity, etc.) related to the system under discussion; programmatic, to describe related constraints (cost, benefits, schedule, available resources, etc); and political (less quantifiable factors such as the agendas of decision makers). Our best decisions, as you might expect, are supported by factors within the first two categories.

With the disclaimer that I retain my amateur status as an economist, I would offer that the best public policy decisions will be in harmony with the "laws" of economics. Not unlike in physics, these "laws" (I refer to them as "axioms" in an attempt to describe the process) allow us to predict outcomes. As an example, the law of supply and demand allows us to predict that price controls (capping the price of a service or commodity) will lead to shortages. I show my work here.

If we interpret the laws of economics as the technical category, there is also a programmatic category related to public policy. For example, it is easy to argue technically that Social Security is inefficient because it violates various laws (taxing work, subsidizing retirement, increasing hiring costs, establishing a monopoly (FICA), etc). However, the existence of such a policy over time creates interest groups. To move away from an inefficient policy such as this, the programmatic case is that we need a transition strategy; many have been proposed for SS (see Ron Paul), but the point is that the public cannot (and will not) come to a full stop even on a technically unsound policy.

I believe that economists, pros and amateurs alike, contribute the most when they expose the technical and programmatic aspects of proposed and existing policies. If we do our work with integrity, our contribution to the evaluation of a policy is more likely to be constructive than political.

I suspect that Patrick Sullivan is correct.

KevinH writes:

It would be pretty easy for this type of situation to create group-think as well, as the costs associated with going on record are significantly different based upon which side you publicly supported.

Another way to say this is there is some cost associated with uncertainty which should work in Klein's favor if the certainty distribution is skewed. However, if the cost-of-making-the-statement-distribution is skewed, this could create an identical observed behavior.

Cole Terlesky writes:
There may well be cases in which on-record economists show a consensus in favor of some intervention, while the at large economists are less favorable to the intervention.

If I had to make a guess of where I would find such a case I'd look for a couple of features:

1. An area of research with outcomes that are hard to distinguish. They can't easily be called out for potentially being wrong, thus there is a lower cost to holding the interventionist beliefs they had prior to their economics training.

2. Gaining expertise in the particular area tends to select for people that hold interventionist beliefs going in. Something like environmentalism, an area of research that has traditionally appealed to liberals.

3. There are political side benefits to openly declaring interventionist beliefs. These benefits could range from more respect at their institution, additional funding for their projects, or just not facing the disincentives of holding an unpopular non-interventionist belief.

I haven't seen any of the survey, but I will make two predictions, so that I can be wrong if the issues have already been addressed:

Issues of foriegn policy intervention:
1. Its an area fraught with counter-factuals. It is incredibly complex, making it difficult to control for variables with the few solid examples we do have.
2. Getting interested in foreign affairs often goes along with an interest in getting involved in foreign affairs.
3. Politicians and Military leaders will be happy to find an outspoken supporter of their beliefs. However, it is at times an unpopular belief to hold in universities.

Certain extreme measures to protect the environment:
1. Same problems as above.
2. Same problem as above.
3. Its a form of signalling for many liberals, and non-interventionist views on the subject are not popular at most universities.

While I could see these two examples being the kind of corner cases that Klein is looking for. I would still be willing to guess that the positions of outspoken economists in these areas is still more non-interventionist than non-economists on the same issues.

Comments for this entry have been closed
Return to top