14

Microeconometrics: Current Methods

and Some Recent Developments

A. Colin Cameron

Abstract

This chapter surveys microeconometrics methods with the emphasis being on recent develop-
ments in these methods. The survey presumes the basic theory for the standard estimation methods
(LS, ML and IV). Estimation methods surveyed include GMM, empirical likelihood, simulation-
based estimation, Bayesian methods, quantile regression and semiparametric estimation. Inference
methods include robust inference and bootstrap methods. The chapter addresses the recent lit-
erature on estimation of marginal effects that can be given a causative interpretation, notably
treatment effects. The common data complications of nonrandom sampling, missing data and
mismeasured data are also discussed.

14.1 Introduction 730
14.2 Identification 731
14.2.1 Point identification 731
14.2.2 Partial identification 732
14.3 Estimation 733
14.3.1 Generalized method of moments 733
14.3.2 Empirical likelihood 735
14.3.3 Simulation-based ML and MM estimation 737
14.3.4 Simulation-based Bayesian analysis 739
14.3.5 Quantile regression 741
14.3.6 Nonparametric and semiparametric methods 742
14.4 Statistical inference 744
14.4.1 Robust inference for Wald tests 744
14.4.2 Hypothesis tests 746
14.4.3 Model specification tests 747
14.4.4 Bootstrap 748
14.5 Causation 750
14.5.1 Treatment effects 751
14.5.2 Instrumental variables methods 757
14.5.3 Panel data 758
14.5.4 Structural models 760
14.6 Heterogeneity 760

729