The Union Wage Effect and Ability Bias: Evidence from Ireland

We use longitudinal data from an Irish household survey to measure the union wage premium.  A sub-sample where the worker's payslip was seen by the interviewer is unlikely to have measurement error for the union variable.  The results support the finding that measurement error leads to a large downward bias in fixed effects estimates of the union effect but indicate that ability bias has a small effect on the average union wage premium.