Try it before you buy it
Over the past 15 years, the economics profession has become increasingly enthusiastic about using randomized control trials (RCT) to conduct social policy experiments. This isn't to say that we don't believe in cost-benefit analysis -- but it's pretty easy to rig if you know how to cheat and are inclined to do so. There are other ways of uncovering causal inference as well, but they tend to involve statistical techniques that are completely inaccessible to non-statisticians (unfortunately, this would be a good description of most of my academic work). But, control trials are pretty easy to design and access, though they can't be used for everything.
Control trials have not yet won a following in politics. I won't speculate as to why, but I think being a first adopter could be a great signature theme for the DCGOP, and not only because you'd win the hearts of the 100 or so economists in Durham County.
Rather than making unfounded, speculative claims about merits or drawbacks of a proposal, the RCT approach is to design an experiment -- and put a lot of the onus on the proposers to do so, and then to get both advocates and opponents to agree to a set of measurable outcomes on which to judge the project. Better Angels could be a model for the discussion, and a non-partisan arbiter would be needed. One also needs to identify people or regions who would be "treated" and others that would be "controls."
One could do this for everything from low-income housing support mechanisms to voucher systems to Medicaid expansion.
---
Charles M. Becker
Research Professor of Economics - and -
Research Professor of Economics - and -
Director of Graduate Studies, MS program in Economics & Computation
Duke University
Duke University
Showing 1 comment
Sign in with