In their highly influential book describing behavioral economics, Nudge, Richard H. Thaler and Cass R. Sustein devote 2 pages to the notion of “bad nudges.” They describe a “nudge” as any aspect of the choice architecture that alters people’s behavior in a predictable way without forbidding any options or significantly changing their economic incentives. The classic example of a nudge is the decision of an employer to “opt-in” or “opt-out” employees from a 401(k) plan while allowing the employee to reverse that choice; the empirical evidence strongly suggests that opting employees into such plans dramatically raises 401(k) participation. Many parts of the book advocate for more deliberate choice architecture on the part of the government in order to “nudge” individuals in the social planner’s preferred direction.

Thaler and Sunstein provide short discussion and uncompelling examples of bad nudges. They correctly note “In offering supposedly helpful nudges, choice architects may have their own agendas. Those who favor one default rule over another may do so because their own economic interests are at stake.” (p. 239) With respect to nudges by the government, their view is “One question is whether we should worry even more about public choice architects than private choice architects. Maybe so, but we worry about both. On the face of it, it is odd to say that the public architects are always more dangerous than the private ones. After all, managers in the public sector have to answer to voters, and managers in the private sector have as their mandate the job of maximizing profits and share prices, not consumer welfare.”

In my recent work (with Jim Marton and Jeff Talbert), we show how bad nudges by public officials can work in practice through a compelling example from Kentucky. In 2012, Kentucky implemented Medicaid managed care statewide, auto-assigned enrollees to three plans, and allowed switching. This fits in with the “choice architecture” and “nudge” design described by Thaler and Sunstein. One of the three plans – called KY Spirit – was decidedly lower quality than the other two plans, especially in eastern Kentucky. For example, KY Spirit was not able to contract with the dominant health care provider in eastern Kentucky due to unsuccessful rate negotiations. KY Spirit’s difficulties in eastern Kentucky were widely reported in the press, so we would expect there to be greater awareness of differences in MCO provider network quality in that region.

Given the virtually identical and non-existent financial differences across the three Medicaid plans (they were essentially free to Medicaid clients), the standard economic framework with rational consumers and trivial transaction costs would predict all enrollees would switch out of lower quality plans. In this case, it would suggest mass defections from KY Spirit. In contrast, the “nudge” framework suggests enrollees would be for more likely to remain in inferior plans. The nudge – in this case a bad nudge – worked. In each of the other two plans – both of higher quality – approximately 95% of those assigned to those plans stayed in them. For KY Spirit, the percentage was lower, but very far from the prediction of full-scale exit. Specifically, 57% of those assigned to KY Spirit remained enrolled in the plan in 2012, despite its well-documented problems. For sicker individuals, 44% remained in KY Spirit, despite the serious problems in accessing healthcare providers. Very few individuals who opted out of their assigned health plan made the active choice to enroll in KY Spirit, consistent with the notion of its low quality. Of more than 37,000 individuals in eastern Kentucky assigned to the other two health plans, slightly more than 100 actively moved into KY Spirit.

Why would public officials assign Medicaid enrollees to a low quality health care plan? After all, virtually all examples of government nudges in the Thaler and Sunstein book portray officials as steering clients in the right direction. In the Kentucky context, the underlying motivation appears to be program costs. The state paid different reimbursement rates to each of the three health plans, and most of the time, KY Spirit was the “low cost, low quality” plan. In reality, this “bad nudge” – from the Medicaid enrollee’s perspective – was a cost saving from the taxpayer’s point of view. Compare to an objective of maximizing the quality of plans for Medicaid enrollees, the actual plan assignment which included some “bad nudges” reduced program costs by approximately 5%.

Although policymakers might be applauded in this case for reigning in program costs through behavioral economics, it is far from the optimistic framework portrayed in Thaler and Sunstein about maximizing client interest.