S YSTEMS OF C HURCH F INANCING
3.5 The tax assignment system
3.6.5 Limitations of the model and suggestions for further research
Although the model is quite extensive and helps shed light on some empirical observations, there are still limitations. First, both the church tax and the tax assignment system are represented in a simplified way. In the model, I assume a constant tax rate. In reality, in Germany as well as Italy the church and assignment taxes, respectively, are progressive.
Furthermore, these taxes depend on taxable income, which differs from overall income.
However, since income is completely exogenous in the model, this has no direct implications on the model quality. The model also disregards specific German regulations for heterogamous couples, where under specific circumstances one spouse who is not a church member has to pay church taxes for the other who is a member. Since the model looks only at the choice of a single individual, this is neglected.
For the case of Italy, the model disregards the fact that the individual cannot assign his or her personal tax amount to the institution he or she prefers, but only has a vote in the overall distribution of the taxes. For a rich individual, this would mean that by earmarking his or her taxes for the church, only an amount equal to the average tax payment goes to the church, whereas the remainder of this individual’s taxes is distributed according to the overall partitioning of “otto per mille” revenues.
Second, the model is based on a specific utility function with a standard Cobb-Douglas part and an additive term for utility from church membership. One could particularly question the additivity of the benefits of membership because this assumption implies that membership has no influence on the marginal utility of the other commodities. However, including membership benefits in a Cobb-Douglas type utility function is difficult because this would not allow a discrete choice over the “amount” of membership benefits. The high number of non-members in Germany shows that membership benefits of zero are a common choice whereas in a Cobb-Douglas function the amount of all commodities consumed is strictly
positive. Another possible objection against the Cobb-Douglas utility function is that the marginal utility of income is constant. However, it is often argued that the marginal utility of income is decreasing in reality, which renders the result that church taxes are a heavier burden for the rich than for the poor questionable.52
Third, in the model non-members of the church and even individuals with a disutility from membership make voluntary religious contributions. This results from the Cobb-Douglas utility function, combined with the assumption that the marginal utility from religious giving, captured by , and the benefits of church membership, , are independent. In reality, one probably finds a correlation between the two variables, although not a perfect one. In addition, one might even observe a positive relationship between the preference for secular giving, , and the benefits of church membership. Researchers argue that the correlation of religiosity and charitable giving can be explained either by “conviction” or “community”
(Bekkers and Wiepking, 2011: 944). “Conviction” refers to inherent “values and attitudes” of individuals, e.g. how much they care about others (Bekkers and Wiepking, 2011: 944). Such values can be transmitted during childhood. When someone grows up in a religious community and internalizes its (usually very positive) position on charitable behavior, he or she will at the same time internalize other values and support the religious community itself, in which case the parameters and the membership benefit are closely connected. The
“community” explanation refers to the fact that individuals who are involved in religious networks might be asked to give more often than others or may experience social pressure to give (Bekkers and Wiepking, 2011: 944). Again, being active in a religious network induces a high membership benefit as well as a high preference for giving. Nevertheless, it is also reasonable to assume that the two variables are not perfectly correlated. In Germany, individuals who seek purely economic benefits from church membership, such as finding a job with a religiously affiliated employer or getting easier access to a church-run kindergarten, need not necessarily be very generous. In addition, individuals might disagree fundamentally with some parts of the church doctrine, which makes them feel uncomfortable with being a church member, but still appreciate some very specific charitable projects of the church. Notwithstanding these objections, a possible correlation between membership benefits and preferences for giving might explain differences between predictions derived from the model and empirical observations. In particular, it could explain why charitable giving of church members (exclusive of church taxes) is often higher than that of non-members.
Fourth, in the present model benefits from church membership depend neither on the church tax rate nor on the share of the population with church membership – thereby neglecting the costs of producing these membership benefits as well as possible scale effects in production.
It is reasonable to assume that some of the benefits are indeed independent from overall revenues of churches, in particular afterlife benefits that are tied to membership itself.
Furthermore, churches currently generate high enough revenues to provide religious festivities like weddings, baptisms or funerals without problems. It is therefore unlikely that individuals take into account that leaving the church or assigning the tax to the state could reduce the availability or quality of these services. Nevertheless, some membership benefits like the availability of clerics do depend on the revenue situation of churches.
Fifth, one could challenge the model on the fact that in the voluntary giving system membership does not have any cost. In reality even in these systems (like the US) membership benefits do at least partly depend on the amount contributed. However, even if this aspect was integrated in the model, the voluntary giving system would still differ significantly from the church tax system because the former allows individuals to choose their individually optimal level of contributions where the marginal cost of an increase in contributions equals the marginal benefit of this increase, while the tax system prescribes the amount of taxes without taking into account the benefits.
Sixth, a comparison of the church disaffiliation decision in Germany and the decision for whom to earmark the assignment tax in Italy is more complicated in reality than it is in the model. The main reason is that the decision that has to be taken in the Italian context can be revised every year, whereas the decision to disaffiliate is usually a long-lasting one which the majority of individuals never revise, or if they do, only after several years. These differences in planning horizons cannot be adequately represented in my simple one period model. The result could be that in Germany people remain church members even if in the current period costs exceed benefits, as long as they expect that in the future benefits will dominate costs.
The model also neglects that in Germany changing the church membership status entails transaction costs, both monetary and in terms of time, while this is not the case in Italy.
Transaction costs can induce individuals to remain in a status even though the costs exceed the benefits. A discussion of transaction costs in the context of the church membership decision in Germany can be found in Section 4.4.1.
Seventh, the model postulates a clear distinction between religious and secular non-profit organizations, whereas it is sometimes difficult to draw a line between these two in reality. On
the one hand, religious communities are expected to use part of their revenues for their charitable activities, which often lack a strong relation to religious teachings (see Feldhoff, 1996: 38). On the other hand, in Italy a large share of the assignment tax earmarked for the state is used for financing the conservation of cultural heritage, of which the majority belongs to religious communities, in particular the Catholic Church (Pistolesi, 2006: 178). In this vein, one can principally question the implicit assumptions made about the substitutability of charitable goods in the model. My assumption is that church taxes and religious giving are net substitutes – eventually, they buy the same good for the individual. Religious and secular giving do not buy the same good. In contrast, Borgloh (2012) assumes that the church tax is a perfect substitute for all types of charitable giving. In her model the church tax has a much larger effect on secular voluntary giving than in my model.
Eighth, the model does not include the possibility that the financing of a particular activity through the government share of the assignment tax might crowd out funding through regular government revenues. In reality, crowding-out is likely to happen (see Pistolesi, 2006: 173-174). What is more, the Italian government even uses the revenues for non-charitable purposes (Pistolesi, 2006: 175; see also Servizio studi del Senato, 2012: 3-4 and Governo italiano, 2012). In the terms of the model, such behavior would further reduce the fitting parameter of the state and make the church a more attractive alternative for the assignment of the tax.
Finally, the model disregards the fact that different types of voluntary contributions can have different prices. As far as donations are tax deductible, the price is defined as one minus the marginal tax rate. In Germany, church taxes paid are tax-deductible, thereby reducing the amount of income tax that needs to be paid. If voluntary contributions were not tax-deductible, it would be less costly for the individual to pay church taxes instead of giving the same amount of money as a voluntary contribution. However, donations are tax-deductible in Germany, therefore price differences between religious and secular donations do not exist.53 In the Italian system, a similar logic applies. Assume that the tax code favored the deduction of secular donations, but not of religious giving. If this was true, individuals would have an incentive to attribute their tax to the church and make additional contributions to secular causes.
53 Borgloh (2012) includes a uniform price for church taxes and voluntary giving in her model. Nevertheless, it