Association for Public Policy Analysis & Management

One of the key features of skill formation is that early investments raise the productivity of later investments (Cunha and Heckman, 2007). These so called `dynamic complementarities' of skill formation are typically hard to estimate because they require exogenous variation on both skills at one point in time and also in a subsequent investment for the same cohort (Almond and Mazumder, 2013). In this paper we take advantage of exogeneous variations on two dimensions (age and resources) for disadvantaged students in a middle income country. The availability of rich administrative information covering the complete population of students allows us to study whether targeted investments at school benefit more the disadvantaged younger students than their older peers. We examine a host of academic outcomes; not only student achievement scores and GPA, but also other non-test-score outcomes such as attendance, on-time grade progression, schooling attainment, and transitions to post-secondary education. This research supplements previous articles that have found -separately- effects from shocks on age and school funding. Studying their interaction informs us on the nature of the human capital production function. If their interaction generates is significant in early primary grades but then dissipates over time, this phenomenon, while interesting, may not be particularly important. On the other hand, if the effects propagate themselves through the human capital accumulation process into later life, they may have important implications for adult outcomes and productivity.