A large body of research has investigated the effects of religion on individual behavior and, more recently, the collective performance of societies. Religion is predominantly credited with favorable outcomes, such as pro-social behavior, better health and higher life satisfaction. Religious and non-religious individuals also differ in their values and preferences. Moreover, religious institutions such as churches also have a large direct effect on society by being an employer or a social welfare provider. Against this background, the constant decline in church membership rates in Germany since the late 1960s is an economically relevant phenomenon. The aim of this dissertation is to investigate the economic causes and consequences of church disaffiliation, from both a theoretical and an empirical point of view. Based on an extensive literature review of models of religious consumption choices I conclude that most of the previous work does not pay sufficient attention to the institutional framework of church membership and church financing in Europe. I develop a theoretical model in which I analyze an individual’s decisions on church membership and on voluntary giving to religious and secular organizations against the institutional backgrounds of the German church tax system and the Italian tax assignment system. The model predicts that in a church tax system individuals with a higher income and those who have to pay a higher church tax rate are more likely to disaffiliate from the church. In contrast, in a tax assignment system cost-benefit considerations of church membership should not take place. Furthermore, the model shows that church and assignment taxes can crowd out voluntary giving to religious and non-religious purposes. In the empirical part of the dissertation I focus on the case of Germany, testing the hypotheses derived from the theoretical discussion with the help of the German Taxpayer Panel for the years 2001 to 2006. The main research question in the first empirical chapter is whether the institutional framework in the form of the existing church tax regulations has a statistically significant effect on the decision to leave the church. The hypothesis is that ceteris paribus an increase in the price of church membership increases the probability that an individual disaffiliates from the church. The estimation results show that both the price of church membership in the first year of the observation period and the change in price experienced by the individual have a significant positive, but moderate effect on the probability of church disaffiliation. In the second empirical part of the dissertation I ask if church members and non-members differ in their voluntary giving and if the giving behavior changes from before to after disaffiliation. I distinguish between the decision whether to make a contribution at all and the decision how much to give. The results imply that church members are not less, but rather more likely to make a charitable contribution than non-members. However, I do find that the average amount given by church members is below the amount given by non-members. This finding suggests that church taxes and additional voluntary donations might be substitutes at the intensive margin. With respect to church disaffiliation, I find evidence that giving is moderately higher after individuals have left the church than before. However, results are inconsistent in whether the increase is due to a higher inclination to give, a higher amount given by those who make a contribution, or both.