Default options and tax avoidance: Evidence from the church tax

Abstract

This article investigates how the design of the tax system affects tax revenue when avoidance is legal. Between 1952 and 1995, the Swedish church tax was constructed as an opt-out system: Swedish citizens were automatically enrolled in the church at birth but were free to opt out. I compare children born shortly before and after the system’s discontinuation and find that birth memberships significantly affect church tax payments later in life. The baseline estimates imply that changing from an opt-out to an opt-in system reduced church tax revenue by 8.2 percent. The default effects are significant on the full population but are especially strong among individuals born in low-income households, implying the opt-out system made the church tax more regressive. I find no evidence that the default option affected religious socialization: reform-affected individuals are equally likely to baptize their children and equally likely to join a different congregation.