An application of cart algorithm to administrative data: analysis of youth initial employment trajectories

This work presents an application of the classification and regression tree (CART) algorithm to administrative data on employment in the Veneto region, an area in northern Italy with a strong economy and a dynamic labour market. This data, derived from the national stream of declarations due by employers to notify each activation, termination, extension, or transformation of employment relationships, allows to investigate the occupational condition of young people who enter for the first time in the regional labour market: we classified their initial work trajectories and then we analysed the individual and working features characterising each identified class. In the current socio-economic context, an insight on youth working conditions is crucial to facilitate the successful design of labour and education policies.