Estimating intergenerational income mobility on two samples: Sensitivity to model selection
|When:||Wednesday, 17 April 2019 - Wednesday, 17 April 2019|
|Where:|| Braamfontein Campus West
Room 247, New Commerce Building
The School for Economic and Business Sciences will host Patrizio Piraino from the University of Cape Town.
Much of the international evidence on intergenerational income mobility is based on sub-optimal data. This is especially true for low-income and emerging countries. In particular, a considerable number of studies make use of two-sample techniques to impute parental incomes. Piraino will propose a machine learning method that may improve the reliability and comparability of such estimates. Our approach minimises the out-of-sample prediction error in the parental income imputation. This provides an objective criterion for choosing across different specifications of the prediction equation.Add event to calendar