3.4 Empirical strategy 3.4.1 Main model Our empirical strategy builds on the model of Cunha and Heckman (2007). This model expresses the stock of skills as a function of the stock of initial skills, parental characteristics, and investments. This aligns with other literature, which highlights the cumulative nature of skill development and emphasizes the essential roles of family and school inputs in the skill development process (e.g., Todd & Wolpin, 2003). While Cunha and Heckman (2007) present a multi-stage model, we focus on one stage of this technology of skill formation, namely the primary school period of students. We treat the initial skill level of the student at primary school entry as given, and examine its relationship with the student’s skill level at the end of primary education. We model the skill level of a student (SKILLGt+p) as a function of the student’s initial skill level (SKILLGt) and the student’s skill production since the last period (∆SKILLGt): SKILLGt+p =SKILLGt +∆SKILLGt (3.1) where t,p ∈ N0. A positive skill production component represents learning advancement, whereas a negative component implies skill depreciation. As we want to capture the relationship between the stock of skills in a subsequent grade and the initial stock of skills, we break the production component into smaller components. Cunha and Heckman (2007) mention that the production component consists of initial 55
RkJQdWJsaXNoZXIy MjY0ODMw