Bunching Estimates of the ETI

This chapter turns to an estimation method which relies on the `bunching' of individuals at income thresholds, where marginal income tax rates change. The method avoids the need for longitudinal information about income changes of individuals over time, separating `treatment' and `control' groups, and/or finding suitable `instruments' to deal with endogeneity. It is shown that a direct relationship exists between the elasticity of taxable income and the extent of observed bunching: the ETI is directly proportional to the extent of bunching. Cross-sectional income distribution data for NZ are used, and estimates are obtained for a variety of tax kinks and years. Furthermore, elasticity estimates obtained from persistent bunching at kink points when the tax regime is unchanged are compared with the more transitory effects on bunching associated with tax reforms. Adjustment costs and/or inattention biases associated with a shift in a tax threshold are found to be important.