Daniel Carpenter: evaluation

Transcript context

…If your least-captured agency is a very new agency, doesn’t that make you out to be a lot more sympathetic to the capture theory than your published work might make it sound? Well, you have to remember my published work in the book with David Moss suggests that the problem with capture is it’s too sexy a term. David and I wrote during the midst of the financial crisis when every time somebody didn’t like the decision of a regulatory agency — a pundit, an academic — the term capture would be thrown around. You can think that an agency is doing a good job, a bad job, efficient, nonefficient, and capture can be potentially orthogonal to that. That’s the first thing I would say. The second, we really got hit, I’d arguably say, more from the left than from the right on this as we actually argued it’s hard to prove capture. You have to have a counterfactual of what an agency that wouldn’t be captured would — in the same situation ceteris paribus — be doing. A lot of the existing empirical studies — and this goes back generations, and I’m indicting a set of things in political science as well as economics — suffer from the usual omitted variable bias problems. Some of them, I think, were making the wrong inferences from the evidence that exists, such as noticing, for instance, that in certain kinds of patterns of regulatory decision-making, larger and older firms did better, enjoyed, say, quicker or more favorable decisions than did smaller and newer firms. As I wrote in a — it was a mathematical essay published almost 20 years ago, that can be true even when the underlying regulator or the regulator in question has zero implicit or explicit preference over the kind of firm it’s regulating. It has everything to do with, not even risk avoidance, but uncertainty avoidance. How can we improve the process for public comments on regulations or pending regulations?…

Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.