Educational Production Function

Value-added modeling is, in econometric terms, a particular specification of the educational production function — typically one that controls for a lagged achievement score to proxy the unobserved history of inputs and endowment, then attributes the residual gain to a teacher or school. The production-function literature provides the theoretical justification and the assumptions under which value-added estimates recover causal contributions. See the related Value-Added Modeling entry.

Inputs in education are chosen, not assigned. Families select schools and neighborhoods, schools allocate resources and assign teachers, and students exert effort — all in ways correlated with unmeasured determinants of achievement. A regression coefficient on an input then mixes its true effect with these selection processes. For example, if struggling students are placed in smaller classes, naive estimates can make small classes look harmful. Credible estimates require designs (experiments, instruments, fixed effects) that break this correlation.

This is among the most contested questions the framework addresses. Hanushek's reviews argued the relationship between simple resource inputs and achievement is weak and inconsistent, while other scholars, using quasi-experimental evidence from school-finance reforms, find that well-targeted spending increases can improve outcomes. The disagreement turns largely on methodology — how endogeneity and input history are handled — which is exactly why specification and identification dominate the educational production function literature.