Working Paper: CEPR ID: DP57
Authors: Richard Blundell; Costas Meghir
Abstract: In this paper we develop a set of selection criteria for the specification of a female labour supply model for microdata. We look at both economic and statistical criteria in assessing various models, paying particular attention to the flexibility of the economic model and the stochastic specification of unobservable factors. Using these criteria, a model is selected from a sample of United Kingdom Family Expenditure Survey data.
Keywords: model selection; female labour supply; microdata
JEL Codes: 211; 212; 229; 824
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| appropriate stochastic specifications (C15) | maximum likelihood estimator's properties (C51) |
| model's flexibility (C52) | accommodate various demographic characteristics (J79) |
| selection of a stochastic specification (C52) | integration back to an underlying indirect utility function (D11) |
| presence of young children (J13) | marginal value of time for women (J29) |
| educational attainment (I21) | likelihood of working longer hours (J29) |