This paper examines under which conditions religious denomination affects public spending on schooling and educational performance. We employ a unique data set which covers, inter alia, information on numerous measures of public school inputs in 169 Swiss districts for the years 1871/72, 1881/82 and 1894/95, marks from pedagogical examinations of conscripts , and results from political referenda to capture conservative or progressive values. Although Catholic districts show on average significantly lower educational performance and spend less on primary schooling than Protestant districts, Catholicism is harmful only in a conservative milieu. We also exploit information on absenteeism of pupils from school to separate provision of schooling from use of schooling. Abstract This paper examines under which conditions religious denomination affects public spending on schooling and educational performance. We employ a unique data set which covers, inter alia, information on numerous measures of public school inputs in 169 Swiss districts for the years 1871/72, 1881/82 and 1894/95, marks from pedagogical examinations of conscripts , and results from political referenda to capture conservative or progressive values. Although Catholic districts show on average significantly lower educational performance and spend less on primary schooling than Protestant districts, Catholicism is harmful only in a conservative milieu. We also exploit information on absenteeism of pupils from school to separate provision of schooling from use of schooling.