Youth employment, academic performance and labour market outcomes: Production functions and policy effects

Angus Holford
2020 Labour Economics  
We use instrumental variables for teenage employment opportunities to identify the causal effects of part-time work during compulsory education in England on educational performance at age 16 and labour market outcomes to age 25. We identify the total 'policy effect', partly driven by resulting changes in other inputs, and the direct effect or 'production function parameter', which holds these constant. The total effects of an additional hour of part-time work per week at age 15 include
more » ... 15 include reducing educational performance in school-leaving qualifications by males by 2.5% and females by 6.7% of a standard deviation, and increasing duration of unemployment experience before age 25 by two months. Direct effects on long-run outcomes are generally beneficial for women and less so for men. What human capital or signalling benefits there are to teenage part-time work are substantially offset by the effects of reduced educational investments. and van Soest, 2007 ) . Indeed, in the UK there remains a significant direct wage return to obtaining more age-16 academic qualifications, and at higher grades, even holding constant qualifications obtained later ( Dearden et al., 2002; McIntosh, 2006 ) . In this paper we evaluate the effect of part-time work during school term-time at age 15 on educational performance at age 16 and retention in full-time education, unemployment experience, earnings, and occupational attainment up to age 25. Our contribution is to identify both the 'policy effect' effect of part-time work, and its parameter in the 'production function' for educational or labour market performance. The 'policy effect' includes indirect effects of resulting changes to other inputs, while the 'production function parameter' is the direct effect holding these other inputs constant. This distinction, and the mechanisms underlying it, matter for policy implications in education. For example, De Fraja et al., (2010) and Datar and Mason (2008) show that students and parents substitute school inputs for their own, meaning that policy effects of increased school quality are expected to understate such interhttps://doi.
doi:10.1016/j.labeco.2020.101806 fatcat:6e5b47mzgrcz7fhskkxqchw6qm