Confidence Intervals for Policy Reforms in Behavioural Tax Microsimulation Modelling

This paper addresses the need for a measure of the uncertainty that is associated with the results calculated through tax policy behavioural microsimulation modelling. Deriving the analytical measure would be extremely complicated, therefore, a simulated approach is proposed which generates a pseudo sampling distribution of aggregate measures based on the sampling distribution of the estimated labour supply parameters. This approach, which is very computer intensive, is compared to a more time-efficient approach where the functional form of the sampling distribution is assumed to be normal. The results show that in many instances the results from the two approaches are quite similar. The exception is when aggregate measures for minor types of payments, involving relatively small groups of the population, are examined.