09 August 2007

Forums Comment

Recently Steven Levitt, author of Freakonomics and a very successful blog, moved to the NY Times. One of his first postings there proved to be very controversial. The subject was “If you were a terrorist, how would you act?” Levitt provided his ideas and rationale and asked that readers do the same. Some readers saw the idea as progressive and therefore participated, while others saw the idea as definitely non-economic and culturally subversive. What’s going on here? Let’s take out our economic balance sheet to identify the costs and benefits of such a posting.

(It’s important to note that Levitt is not being cavalier in his question. It is how his mind works as a Micro-economist working in the science of incentives. Thinking this way is a characteristic of intelligence. It is how Einstein came up with relativity. He was able to imagine the properties of physics while traveling at c. It is the same skill that allows humans to empathize.)

First, let’s look at the costs and benefits to Dr. Levitt. He has stated before on his blog that he appreciates all types of commentary. Well, he received ~600 (and counting) comments on this posting. Also, I’m sure the number of hits his site got spiked this week as well and may stay high at least for the near future. On the other hand, he may also have permanently lost some readers and some readers will visit less frequently because of the post. Dr. Levitt probably also cares a little about the NY Times perspective, which depends on their costs and benefits of Levitt’s post.

For the NY Times there are similar benefits as Dr. Levitt’s. Hits and exposure go up. Costs include the possibility that long run hits decrease, but also that there are corporate reputation risks involved. The NY Times does not want to be potentially culpable for publishing methods that terrorists may learn from or even use. Interestingly, there is also a small benefit associated. It might be a positive for the NY Times to be part of a project that identifies some terrorist hatches and therefore become the initiating step in preventing terrorist activities. Econometrically, this is a difficult benefit to quantify. Every time a terrorist activity does not occur does it become a feather in the cap of Levitt and NY Times? (Actually, Dr. Levitt is probably one with the best ability to quantify the benefit.)

There is a glimmer of brilliance in the post. I believe that what Levitt and the NY Times are trying to tap into, wittingly or unwittingly, is a concept outlined in a book called Wisdom of Crowds by James Surowiecki. What the concept states is that if there exists a well-structured public forum, then decisions made with crowd input can always far exceed those without.

The way Surowiecki’s concept would work for Levitt’s blog is that if you believe that Levitt’s blog & commentary format constitutes a well-structured forum, then Levitt’s posting can successfully improve security by preempting plots that bureaucratic experts and homeland security contractors may not have considered.

Thankfully Surowiecki has provided us some guidance into what characteristics are needed for a well-structured forum; diversity, independence, and practicality. If these conditions are met (starting to sound like a theorem) then the existence of the forum guarantees that the decision made will be (weakly) superior to a decision without a format, according to any metric of superiority. We are not saying that the forum should equally weight all input or specify any other aggregation technique (that is a discussion for Social Choice theory). But, what we are saying is that for any definition of superior, having input provided by a well-structured forum will yield superior policy than not having it available, having only the input of a small group of “experts”, or any forum (information market) failure.

Maybe it’s me, but this claim sounds a lot like the Fundamental Theorem of Economic Welfare, which says that under some conditions (if government has intervened to correct the market failures of imperfect competition and externalities) the resulting competitive equilibrium will be Pareto efficient, a generalized notion of optimality. What we’ve just stated in the preceding paragraph about the potential Wisdom of Crowds is that, under some conditions (independence, diversity, practicality; which the government can intervene to guarantee) the resulting policy is superior under a class of metrics.
Perhaps what we’re driving at is a Fundamental Theorem of Public Policy.

The Fundamental Theorem of Welfare Economics is a mathematically proven statement. Its concept (first postulated by Adam Smith) was eventually translated into precise conditions and a conclusion justified by mathematical tools and reason.

Paul Samuelson, in his graduate dissertation advised by an Economist and Physicist (a polymath academic descendent of Gibbs), imported the tools of Thermodynamics into Economics and placed them on the proper shelves. It is with these tools and reasons that The Fundamental Theorem of Welfare Economics was developed and formalized.
Is it time or is it even possible to again import those tools into the field of Public Policy?

Upon consideration, what Surowiecki purports makes sense.
Diversity is critical because it can illuminate the subject matter of the discussion from different perspectives, leaving no facet shadowed and the pull of potential factions negated.
Independence is important so that members do not correlate each other prior to providing their input. It is well understood that in failed forums, those who speak first, those who speak most frequently, those who are verbose, those with better rhetoric, and those who speak with intimidation enjoy disproportionately large bias relative to their input.

How do Levitt’s and Mankiw’s blog and comment forum measure up to these desired characteristics?
Diversity. Given the large amount of people with access to the internet and subsequently their blog, both have the potential to garner sufficient diversity. For Levitt, being now associated with the NY Times site should even further his forum’s diversity. As self-noted on his blog however, his comments are subject to a self-selection bias. Independence. This has become a major concern in Mankiw’s blog and comment forum. It is not so much bias from rhetoric or intimidation, but rather bias generated from mass. With the quantity of comments generated in his forum, there becomes an enormous amount of repetition of the same factors; the equivalent of individual verbosity or a group over-representing a factor. Independence, lack of bias, and lack of pre-correlation are crucial. There is, however, a balance. Should individuals have access to the factors presented already, in case they had failed to consider certain aspects? I think so, minimally. You may argue that this necessarily introduces bias. I claim that it should be the role of a forum governing body to intervene minimally to provide this service in order to correct the potential market failure of information asymmetries. What do you think now? Or did I just present too much bias?
Practicality. This is the most dangerous failure of Levitt’s forum given his post and its greatest detriment in my mind. Since there is no enforcement agency to actively address the results of the forum’s discussion, then we have only illuminated failures without providing for means to correct them. Hence we have a major forum failure.
Indirectly we do have such a means though. It is possible that the post and comments will generate enough interest to actually guide public resources to addressing these failures. It is important to point out that it is not necessary to post the comments in response to Levitt’s post. Levitt could only offer an incentive such as one of his yo-yos to the comments that he deems most insightful in order to compensate the readers for the lost incentive for no longer enjoying seeing their comments published.

Evidently there appears to be a real science to generating efficient forums, just as there is a real science to generating efficient markets (called Economics).

I would like to highlight one of Milton Friedman’s quotations.
“Abraham Lincoln talked about a government of the people, by the people, for the people. Today, we have a government of the people, by the bureaucrats, for the bureaucrats, including in the bureaucrats the elected members of Congress because that has become a bureaucracy too.
And so undoubtedly the most urgent problem today is how to find some mechanism for restructuring our political system so as to limit the extent to which it can control our individual lives. You know, people have the image, have the idea, that somehow ‘we the people’ are speaking through the government. That is nonsense.”

Maybe that is what our discussion is pointing at; working to find some efficient mechanism for our political system to listen to the people.
For the people to be heard we need a forum, and as we’ve seen, it is absolutely critical that that forum needs to satisfy some conditions in order to be efficient.

Just as efficient market operation is important in the regional to the global environment; so is the efficient forum important in local, state, federal and international policy. Efficient forums are needed in academia, private industry, and government.

I believe that structuring efficient forums can in the long run refine an umbrella of policies; from tax to national security to campaign finance (normally a direct enemy to efficient forums and our political system by biasing input with wealth and generating sustainable constitution-endangering factions).

Surowiecki provides a number of examples of forums successfully at work in a variety of situations.
1) The average group guess of the weight of an ox at a fair is consistently significantly closer than any individual. [The reason is that the group is comprised of individuals who will overestimate the weight (a cattle truck driver who may associate the animal with all of its transport gear) and underestimate the weight (a butcher who views the animal as a combination of steaks and forgets about pre-trimmed portions]
2) The significant consistent accuracy of the “ask the audience” in Who Wants to be a Millionaire relative to the performance of “ask a friend/expert.”
3) A navy vessel lost at sea whose discovery was due only to a forum effective at gathering the input of weatherman, coastal current experts, sailors, sociologists familiar with behavior of lost crew, and committees well versed with subtle but critical techniques of rescue operations.
4) The significant accuracy and consistency of bettor’s prediction in sports games relative to the bookies/experts initial lines (remembering that bookies have incentive to accurately predict games to maximize their vigorish)

Beginning to tread in the world of research-needed speculation, I might venture that there exists a preferred location in the political compass dependent on the present position of society relative to the Fundamental Theorems of Economics and (our new) Fundamental Theorem of Public Policy. http://en.wikipedia.org/wiki/Political_compass
Milton Friedman hints at something like this in Capitalism and Freedom.

My claim would state that the optimal position on the fiscal axis of the political compass should depend upon present market situation in the economy.
An economy with perfectly compensated market failures should be optimally situated for no further government intervention and would belong completely at the right end of the fiscal spectrum.
Similarly an economy in desperate need for government intervention to correct gaping market failures should be optimally situated at the left end of the fiscal spectrum.
Further, a policy making forum without any forum failures (including information asymmetries as well as the others outlined earlier) should be optimally situated for the total liberal behavior side of the social spectrum.
Similarly, a policy making forum desperately suffering from lack of diversity, independence, and total information asymmetries would be optimally situated for the power-vested leader representation side of the social spectrum.
Therefore, our (US) optimal position in the political compass depends on our individual perspectives of our current situation; that is, the health of our free markets and public forums. Again, this is not one particular point, but a set characterizing the generalized notion of optimality representing different individuals’ relative weightings between security and freedom.


loyal fan said...

when can we expect your next blog?

Graduate Dissertation said...

Very informative and helpful. I was searching for this information but there are very limited resources. Thank you for providing this information

Graduate Dissertation Examples