Abstract
This article reviews recent developments in the econometrics of early childhood human capital and investments. We start with a discussion about the lack of cardinality in test scores, the reasons it matters for empirical research on human capital, and the approaches researchers have used to address this problem. Next, we discuss how the literature has accounted for the errors in human capital measurements and investments. Then, we focus on the estimation of production functions of human capital. We present two different specifications of the production function and discuss when to use one versus the other. We describe how researchers have addressed cardinality, measurement errors, and endogeneity of inputs to estimate the technology of skill formation. Finally, we take stock of the work to date, and we identify opportunities for new research directions in this field.