Robin Hanson argued some time ago that politics isn’t about policy. This was his theory for why we have so many excessive regulations that “make everyone act like high status folks act, regardless of how appropriate that is for their situation”. Here are some of his examples:
“Consider one-size-fits-all building codes, food and drug regulations, safety rules, professional licensing, and medical insurance regulations. Such rules tend to make sure that a typical rich person wouldn’t accidentally buy a product or service of a much lower quality than they would desire.”
His favored theory for why these types of laws are pervasive is that politics isn’t about policy, in that:
“We (unconsciously) don’t care much about the consequences of such policies – we instead support policies to make ourselves look good. If our support for regulations pushing high status actions is taken as a signal of our personal status, then we can want to support such regulation regardless of what results when such regulations are implemented.”
How would we know if Robin was wrong? I think that no matter what your policy priors are, there are some obvious things policy should incorporate if in fact we did care about policy outcomes. The lack of these policy features suggests to me that Robin is correct.
For one thing policies should be designed so that we can tell if they worked. One way to do this would be to incorporate randomization into their design. State and county differences in laws are already used as instruments for economic studies, but endogeneity of policy is always a concern. The experimentalist school of economics is growing more and more prominant, and government’s could hire staffs of these economists to design very subtle and smart ways to insert randomization into policies and tease out causality.
A state that wants to raise the minimum wage, for instance, could raise it in randomely selected counties. Or consider what could be learned about payday lending if some state or group of states hired Dean Karlan and Jonathan Zinman to design a study oriented policy. It certainly would have been a good use of money to randomize some of the first round stimulus package. If it worked as proponents claim then it would make it easier to get a second round of it, and if it didn’t as critics claim then it would be much less likely we would waste money on future stimulus.
Of course this would not work with all policies. Sometimes the effects are significantly long-run, indirect, or they impact hard to measure outcomes, such that it would be difficult or impossible to empirically determine whether they’ve worked. Here we must be guided by theory. But surely there are many laws where this is not the case.
Another thing you would observe if politics were about policy is sunset provisions in laws where the efficacy is examined some predetermined number of years out and they are automatically repealed unless they are shown to have had a demonstratable effect. This would go hand in hand with policies designed so that effectiveness can be measured.
Policies that are designed so that we know if they worked as intended, and are be automatically repealed if they don’t: why is such a common sense idea so foreign, and what does that tell you about politics and policy?
(Thanks to the always insightful Sister Y for inspiring this post)