Oz Blog News Commentary

Expertise and punditry

July 6, 2017 - 06:42 -- Admin

I concluded my post “Against Epistocracy” with the question “Who gets to decide who is well-informed? And who gets to decide who gets to decide?”. This is, I think, a fatal flaw in any system proposing to replacing democracy with rule by a well-informed elite, or any kind of putative aristocracy. But even in a democratic system, we have to make decisions about who should decide things. In many cases, we would like to call on expert advice, and that brings us back to the question “who, if anybody, is an expert on a given topic”. I don’t have a complete answer, but I think it’s helpful to distinguish between experts and pundits or, better, between expertise and punditry.

An easily accessible example is that of forecasting election outcomes. For a long time, this was the domain of pundits, the archetypal example being David Broder, the last (AFAICT) holder of the office of Dean of the Washington Press Corps*. Pundits like Broder and his Australian equivalents (the closest would be Laurie Oakes) drew on their deep connection with the American (or Australian) people to make pronouncements about the way their people were likely to vote.

Political pundits were pretty much put out of the business of electoral forecasting when experts (Nate Silver in the US and psephbloggers like Peter Brent and William Bowe in Australia) arrived on the scene. The experts applied statistical tools to poll results, ignored the temptation to analyse shifts within the margin of error or to extrapolate trends and generally did a much better job.

As a first approximation, expertise involves command of a body of knowledge on a given topic that is based on a coherent (if sometimes implicit) theory and validated by evidence. It can be as concrete and specific as a plumber’s knowledge of how to fix a leaky pipe or as abstract as a mathematicians ability to check the correctness of a proof. Expertise is not infallible, but it’s almost always better than the alternative.

By contrast, successful punditry requires knowledge of specific facts relevant to the topic in question and, ideally, a capacity for intuitive insight, combined with the rhetorical gifts necessary to present these as a source of authoritative knowledge. The most important of these is a capacity to make statements that look like testable predictions but in fact cover all possibilities. Failing this, a social environment in which established pundits are never called to account for their errors will do the trick.

I’ll leave it to readers to discuss which areas of public policy discussions are dominated by expertise, which by punditry and which by conflict between the two.

* As a working academic, I’ve been trained to be wary of anyone with the title “Dean”, but others seem to think it’s an honorific.