Titolo della tesi: Nonpecuniary motivations for tax compliance: evidence and network analysis
In this dissertation, we focus on tax morale, social norms and reputational concerns. The contribution to the existing literature is threefold. First, we provide an updated review of the literature on tax compliance, examining separately the reasons that relate to pecuniary and nonpecuniary motivations. Moreover, we review a new strand of literature, which has investigated the role of social networks and communication on tax compliance. Second, we consider the possibility that being a member of a formal group increases individual’s tax morale, focusing on membership in voluntary and/or volunteering organisations. Based on data provided by European Value Survey 2017, we use an ordered probit model to account for the ranking information of the scaled depend variable (tax morale). According to our estimates, group membership is, in general, not correlated with tax morale because of the heterogeneous effects that each group membership exerts on it. However, our findings suggest that memberships of a certain group are significantly correlated with a lower level of individual’s willingness to pay taxes. Finally, we present an agent-based dynamic model of tax compliance in which taxpayers, who live and interact in a network, decide whether to pay or to evade taxes by considering the expected economic net benefit, the subjective cost (tax morale), and reputation cost (honour and social stigma) of tax evasion. We simulate the model for different values of the parameters and show that the steady-state share of taxpayers is affected by the probability of finding like-minded peers in the reference group, the weight that individuals attribute to reputation, and the share of individuals who update their tax morale. We also present a variation of the previous agent-based model to focus on the role of the reputational costs. We consider two income-groups whose monetary incentive for tax evasion differ. Our results suggest that in an integrated society, in which individuals interact with all the neighbours, no matter their income, all individuals either evade or pay their taxes. By contrast, a segregated society, in which individuals interact only with neighbours who have the same income, generates obstacles to the diffusion process of information around the network and, as a result, tax evasion and tax compliance might coexist. In both models, we perform simulations to consider the possibility for a fiscal authority to increase tax compliance using the knowledge of the network structure to implement moral or reputational policies. Our numerical simulations show that behavioural strategies could be an effective complement to more traditional policies.