Learn how to estimate coefficients up to scale when the true model is unknown.
This book (a research paper) explores how average derivatives connect to covariances and how to estimate key parameters without fully specifying the distribution of unobserved terms.
This work develops a general single-index framework that covers many limited dependent variable models, including discrete choice, censoring, and selection. It introduces estimators that rely on the marginal distribution of X and shows how to use score vectors and instrumental variables to obtain consistent estimates up to scale.
- How average derivatives relate to sample covariances and what this implies for estimation.
- Ways to estimate coefficients up to scale using different data moments and instrumental variables.
- Situations where ordinary least squares can bias the estimate and when the proposed methods remain robust.
- Practical guidance on applying the approach to models with a variety of dependent variables.
Ideal for readers of econometrics and applied economics who want a rigorous, model-agnostic path to consistent estimation when the exact functional form is uncertain.