Description:
This paper analyzes the relationship between cash and tax evasion by studying the effects of two measures to fight evasion: accessing taxpayers’ banking data to check the consistency with reported income and discouraging the use of cash with thresholds on payments. We show that variation in cash holdings triggered by these policies can be substantial, about 1.5 per cent of the GDP. Exploiting regional heterogeneity in tax evasion propensity and using a simple model of tax evasion and payment choice, we find that both policies do not significantly raise tax revenues. Accessing banking data pushes evaders to convert deposits into cash and keep avoiding taxes for about 2 per cent of total annual tax revenues. The difficulty of tracing cash transactions makes thresholds effective mostly for tax compliers