Extending output produced by MITTS

(a) Empirical confidence intervals on predicted values in MITTS

MITTS-B incorporates a behavioural component, which means that individuals are allowed to adapt their level of hours worked in response to policy changes. Results and output produced by MITTS-B, such as changes in the participation and hours probability, transition matrices and changes in revenue and expenditure, are all based on the individuals' behavioural responses.

Given the complexity of the behavioural models involved, it is not clear how precise the results and output are, which are produced by MITTS-B. The aim of this project is to add a measure of reliability of the results predicted to MITTS when using the behavioural component. We propose to compute confidence intervals for predicted values to give an idea on the accuracy of the results.

Behavioural changes are predicted through the use of labour supply models. The point estimates of the parameters in the labour supply models together with their variance/covariance estimates can be used to create empirical confidence intervals around, for example, the expected change in hours of labour supply, the change in participation probabilities or changes in expenditure.

(b) MITTS-B Simulation for individual households

In MITTS-A, net income before and after a reform (broken down in components) can be calculated for individual households. These households can be selected from the SIHC sample or hypothetical households can be created. This can be useful in the analysis of effects. At the moment such an analysis is not possible in MITTS-B, because in the behavioural component of MITTS there is not just one labour supply outcome before and after the reform. Instead MITTS-B takes a probabilistic approach, where the probabilities of being at each of the available discrete labour supply points are computed. As a result, after a reform a distribution of net incomes is computed rather than just one value for net income. The tables in MITTS-B report the expected value of expenditure and transitions.

Although we cannot follow exactly the same approach as in MITTS-A, it would be useful to have a similar tool included in MITTS-B, where the effects of a reform can be examined separately for individual households.

More information about extending output produced by MITTS (pdf 91KB)

The Melbourne Institute contact for this project is Dr Guyonne Kalb.

The final report has been submitted.