Essays on Optimal Mix of Taxes, Spatiality and Persistence under Tax Evasion

This dissertation analyzes the optimal mix of direct and indirect taxes in an economy with multiple tax collecting authorities when both the taxes are subject to evasion and to what extent the tax compliance behavior of individuals in the United States are persistent and spatially dependent. Essay I derives and provides an intuitive interpretation of: (i) impact of the changes in the government instruments on tax evasion by firms, the expected prices they charge, and the expected tax rates they face; (ii) a generalized version of Ramsey rule for optimal commodity taxation which accounts for income tax evasion from either or both the tax authorities; (iii) generalized formulae for the optimal income tax rate for each of the tax authorities; and (iv) the tradeoff between optimal tax rates and audit probabilities for each of the tax authorities. It also re-examines controversies surrounding the uniform income taxes and the differentiated commodity taxes, and investigates how income tax evasion affects the progressivity of the income tax rates. It concludes that whether or not tax evasion calls for reductions in the optimal income tax rates hinges on how tax evasion and the associated concealment costs vary across individual taxpayers. Essay II introduces the twin issues of spatiality and persistence in the individual income tax evasion. While the issue of persistence arises through accumulated learning over time, spatiality arises for several reasons. Some these include the exchange of information between taxpayers; the social norm of tax compliance: an individual would comply if everybody in the society complies and vice versa; individuals faced with dynamic stochastic decision problems that pose immense computational challenges may simply look to others to infer satisfactory policies and interpersonal dependence works through learning by imitating rather than learning by doing. State-level annual per return evasion of individual income tax and related data were used to examine the above hypotheses and found supports for both of them in the individual income tax evasion in the United States.