Institutional Constraints and Random Heterogeneity in Structural Discrete Choice Models of Household Labour Supply

Despite a growing literature in the application of Conditional or Mixed Logit methods to the estimation of labour supply models, relatively little attention has been paid to the role of institutional constraints in employment choices. We argue that the failure adequately to control for optimising errors in models of this form lead to substantial biases in structural preference parameters. We show how an alternative class of DOGEV estimator can be used to control for general patterns of optimising error in models of this form, and provide empirical examples of the effects such errors have on modelled preferences