March 2011

IZA DP No. 5609: Tax Evasion, Welfare Fraud, and "The Broken Windows" Effect: An Experiment in Belgium, France and the Netherlands

revised version published as 'Tax Evasion and Social Information: An Experiment in Belgium, France and the Netherlands' in: International Tax and Public Finance, 2014, 22, 401-425.

In a series of experiments conducted in Belgium (Wallonia and Flanders), France and the Netherlands, we compare behavior regarding tax evasion and welfare dodging, with and without information about others' behavior. Subjects have to decide between a 'registered' income, the realization of which will be known to the tax authority for sure, and an 'unregistered' income that will only be known with some probability. This unregistered income comes from self-employment in the Tax treatment and from black labor supplementing some unemployment compensation in the Welfare treatment. Subjects have then to decide on whether reporting their income or not, knowing the risk of detection. The results show that (i) individuals evade more in the Welfare treatment than in the Tax treatment; (ii) many subjects choose an option that allows for tax evasion or welfare fraud but report their income honestly anyway; (iii) examples of low compliance tend to increase tax evasion while examples of high compliance exert no influence; (iv) tax evasion is more frequent in France and the Netherlands; Walloons evade taxes less than the Flemish. There is no cross-country difference in welfare dodging.