Generalized Trust and Taxation -A cross-country study

Abstract

This paper investigates whether generalized trust can help solve the large N social dilemma that financing public goods by taxes entails. Using data for generalized trust from the World Values Survey we examine if the level of generalized trust can explain cross-country variations in tax revenues and in marginal tax rates. Besides controlling for demographics and institutional factors, we also divide our sample by the share of the population belonging to hierarchical religions. Our OLS regressions show that generalized trust has a positive effect on tax revenues in countries where less than 60 % of the population is affiliated with Catholicism or Islam and in countries with already high tax revenues. To circumvent endogeneity and causality issues we use an IV regression with sport organization membership as the instrument. Our IV regression show that generalized trust has a positive effect on the marginal income tax rate in the complete sample. We also conclude that it is plausible that trust does not have an effect on either tax revenues or marginal tax rate in developing countries as they may be stuck in a low trust equilibrium.