How About a Regulatory Action Lab?
I have just finished reading Abhijit Banerjee and Esther Duflo’s excellent and very thoughtful book, Poor Economics: A Radical Rethinking of the Way to Fight Global Poverty. MIT economists Banerjee and Duflo reject broad, sweeping arguments concerning either the necessity for infusions of foreign aid or the futility of such efforts. Instead, they advocate detailed studies examining whether and how interventions can actually improve the lives and incomes of poor people.
Such an approach sounds so common-sensical that I was somewhat surprised that no one had thought of it before, but more likely it simply reflects the provincialism and ignorance of the economics
priesthood profession. But no matter: it’s good to see economists finally waking up. In Banerjee and Duflo’s case, field research essentially means the Randomized Controlled Trial (“RCT”), in which one randomly selected group of people receives an intervention (or “treatment”, in the sanitized parlance) and the other is the control group. This varies from what social scientists call the “semi-experimental” method, in which social scientists get a bunch of data and attempt to construct a model to interpret it and control for variables unrelated to the intervention they are examining — usually through techniques such as mutiple regression. Banerjee and Duflo’s Abdul Latif Jameel Poverty Action Lab (“JPAL”) serves as the incubator and funder for much of this research.
It did get me to thinking, though, about the possibilities of using such RCTs in the legal context, particularly when it comes to regulation. Would it be possible to have a Regulatory Action Lab, which would use RCT tests for regulatory actions? Inspectors could enforce rules in different ways, sometimes negotiating things out and sometimes going by the book.
Obviously powerful considerations would seem to weigh against such an idea. Law is supposed to apply to everyone equally, so to exempt some groups or apply it differently to different groups might raise equity problems.
But this overstates the matter, in my view. Agencies and policymakers test pilot enforcement programs all the time, and this hardly threatens the rule of law. The more difficult problem, in my view, is getting agencies to cooperate with researchers in designing RCTs. The great advantage of the JPAL model is that interventions do not necessarily need government approval: you can place chlorine dispensers close to water sources, for example, usually without the health ministry’s approval. With a Regulatory Action Lab, researchers are not so much as testing individuals’ responses to treatments, but testing the interactions of government and regulated parties’ responses to treatments. That is more complicated.
Still, I think it’s an idea worth exploring. I have never seen anyone try to work through JPAL’s model for legal scholars, but it is high time to do so. Empirical legal studies is all the rage now, after decades of law professors just writing about appellate cases. But most empirical legal studies is still using third-party data sets. Legal scholars — and especially environmental scholars should start developing data of their own. JPAL is a good model to work from.