Self-productivity and Cross-productivity in the Process of Skill Formation

Given the insight that individual skills crucially impact various life outcomes, questions about the process of skill formation are increasingly being researched. Evidence about path dependency and about substantial and lasting effects of early childhood events emphasizes the importance of the dynamic component in the skill formation process. This dynamic component has been incorporated in skill formation models featuring self- and cross-productivity, while empirical evidence is scarce. Filling this gap I estimate an instrumental variable model, using as instrument a randomized controlled working memory training intervention, to investigate the question of whether skills are self- and cross-productive, i.e., whether skills boost skills over time. My results show that, first, an exogenous shock to one specific skill (working memory capacity) at an initial stage leads to that same skill being improved in a later stage, but only to the extent of the initial skill shift without any extra effect on the production of this skill (self-productivity in the broader sense but not in the narrower sense). Second, I find the exogenously shifted skill, while having no immediate effect on other skills, boosting the production of a number of other skills over time. Hence, I provide evidence about skills being dynamically cross-productive. My findings imply that early disadvantages can be the reason for skill gaps opening up over the life cycle and they explain why early interventions can have significant long-term effects for individual human capital accumulation. My results have implications for the design of policies intended to foster human capital and to augment equality of opportunity.