Abstract
Is there an effect of salience of taxes on the income tax return? To address this question we consider the Swedish 2001 reform which introduced a separate line specifying the church taxes on the Swedish income tax returns. Earlier, church taxes had been included as an element of municipal labor income taxes. The tax agency collects about 1 percent of taxable labor income from the majority of Swedes who belong to the Church of Sweden. We use rich individual level data on the full Swedish working population. Using a difference in difference strategy, we estimate that the reform increased the likelihood of having exited among those who entered the labor market as church members and have had a chance to react to one income tax return by 0.4 percentage points. The estimated effect increases with number exposures. We also estimate the reform's effect on membership sensitivity to the church tax rate. In our preferred estimation strategy of this effect we use variation over time in the parish wise specific tax rate that the individual faces. We estimate that an increase in tax rate by 1 percentage point reduces the likelihood of memebership by about 0.5 percentage points more following the reform.