How to move behavioral economics forward with federalism
By Phil Armour
When social and natural scientists run an experiment on say, a college campus, there is always a concern that the results will not extend beyond the lab’s walls. External validity, as researchers call it, is a particularly thorny problem in behavioral economics, and has perhaps been the biggest stumbling block to wider acceptance in the economics community. In particular, economists debate the relevance of laboratory results in a broader, market-based context. Natural and randomized field experiments are often lauded as possessing strong evidence of an effect, but some economists have criticized the ability of policymakers to extrapolate beyond the idiosyncratic circumstances of these often tiny and localized experiments.
To move the discussion forward as to policy implications of behavioral economics, these laboratory results need to be tested on a real world level, out from under the parochial circumstances of college sophomores. Nudges, in other words, need to be implemented and evaluated in a different laboratory. Conveniently, in the United States, there are 50 ready-made laboratories, also known as states. Nearly as important as the aim of the nudge itself is the design of its implementation, which would preferably be in randomized, evaluation-ready experiments. They would be large enough to satisfy concerns over general applicability, while small enough to minimize downside risk from possible failure. Take two examples from Nudge:
- Application of RECAP (for those new to the blog/book, “Record, Evaluate, and Compare Alternative Prices” information provision schemes) to a class of loans, for example credit card loans, with treatment and control groups for evaluation on a nationwide basis.
- Replacement of the current practice of random assignment of “dual eligibles” in Medicare Part D with informed assignment in neighboring Part D plan areas
These small scale implementations allow for a second test: The fallibility of policymakers. One particular criticism of libertarian paternalism is simultaneously suspicious of the findings of human irrationality and doubtful about the ability of policymakers to overcome these findings, if they are indeed true. If well-designed, choice-preserving, government-implemented nudges produce favorable results on this smaller-scale, then these concerns may be largely allayed as wider implementation is considered. Then again, if they fail, then the community of libertarian and asymmetric paternalists, myself included, should take another look at what we thought we knew.
A call for data moves libertarian paternalism out of the abstract, making it easier to evaluate as a proposal or doctrine. Simply put, the economic debate on the use of behavioral economics in policy is ridden with ideological tendencies; bringing policymakers to the process of finding and implementing sound nudges on a somewhat larger scale allows the evidence to catch up to the argument. Psychologists and behavioral scientists helped put behavioral economics on the beach; it may require the politicians to help take it from there.