Abstract

Improving productivity in education has become a central policy concern, particularly in public education systems facing fiscal constraints and rising accountability pressures. Education Production Function (EPF) studies provide the dominant analytical framework for examining how educational inputs are transformed into learning outcomes. However, despite decades of research, there is no clear consensus on how to measure productivity and efficiency in education, largely due to methodological fragmentation across studies. The paper provides a critical and integrative review of the education production function literature, synthesizing theoretical foundations, empirical model specifications, and estimation strategies employed in productivity measurement, including contemporaneous, value-added, and growth EPF models. The review evaluates regression-based approaches and frontier methods such as Stochastic Frontier Analysis and Data Envelopment Analysis, with particular attention to methodological assumptions, data requirements, and sources of bias. The study contributes by linking production function models with efficiency analysis and identifying major sources of empirical inconsistency in EPF studies. It emphasises the importance of multi-output, joint production, and hybrid approaches for improving productivity measurement. The review shows that conflicting findings mainly arise from methodological problems such as endogeneity, omitted variable bias, misspecification, restrictive functional forms, and poor treatment of multiple outputs. Although value-added and frontier models improve on simple regressions, their reliability depends on data quality and assumptions. The study highlights the need for integrated, dynamic, and multidimensional approaches, especially in higher education and developing-country contexts, to support evidence-based education policies.

Keywords: Education Production Function, Educational outcomes, Efficiency analysis, Frontier methods, Productivity measurement