Income Effects and the ETI

The majority of elasticity of taxable income studies make the explicit assumption that there are no income effects, using a specification in which there is a constant elasticity of taxable income. This chapter examines a possible role for income effects. Derivations of alternative specifications are critically examined. An alternative and preferred rationale for a specification involving the average tax rate is given, involving a simple extension of the quasi-linear utility function consistent with the standard model where only the marginal tax rate is relevant. The empirical approach makes use of the `expected tax rate' proxy for each individual, derived in the previous chapter, and based on each individual's projected conditional distribution of income in the post-reform year, given incomes in previous years. The use of this proxy avoids the need to use instrumental variable methods. Empirical results show that income effects are, at most, relatively small relative to estimated substitution effects.