Underhill, Kristen, and Ian Ayres.

Abstract

Some have suggested that including sunset clauses in legislation makes laws easier to pass, in part because sunsets may facilitate bipartisan compromise. We use a randomized experiment to assess whether sunset clauses actually change people's support for legislation, along with other compromise beliefs such as the perception of sponsors' good faith and the likelihood that the law will be effective. We randomly assigned 1,639 U.S. adults to read laws with one of three sunset conditions (none, standard sunset, or conditional sunset that would be contingent on an evaluation of the law), one of two topic areas (drug overdoses and Medicaid beneficiary health), and one of three political valences (neutral, liberal, or conservative). Participants estimated their support for the law to which they were assigned, and they identified their own political leanings and party affiliations. Sunsets did not increase overall support for laws, contrary to prior suggestions of an overall compromise effect. But in an interaction between sunset and political valence, we found that sunsets increased support for conservative but not liberal legislation. Subgroup findings confirmed liberals' tendency to increase their support for conservative laws that contained a sunset clause (either standard or conditional). Conservatives, however, did not increase their support for liberal policies in the presence of sunset clauses. This asymmetry gives rise to what we characterize as a "sucker" effect--a willingness to compromise that is not reciprocated, even if these effects are unconscious. We explore possible explanations for this finding and consider whether debiasing strategies may be needed when sunset clauses are used.