NEWS

---

ivreg2r 0.1.0 (2026-07-21)

Initial CRAN release. ivreg2r provides extended instrumental variables and GMM estimation with automatic diagnostics, inspired by Stata's ivreg2 (Baum, Schaffer & Stillman).

• Estimators: 2SLS, LIML, Fuller, k-class, two-step efficient GMM, and the continuously-updated estimator (CUE).
• Variance estimators: classical, heteroskedasticity-robust, one- and two-way cluster-robust, HAC/AC with eight kernels and automatic bandwidth selection, Kiefer, and Driscoll-Kraay.
• Diagnostics reported at estimation time: weak identification (Kleibergen-Paap, Cragg-Donald, Stock-Yogo critical values), underidentification, overidentification (Sargan, Hansen J, Stock-Wright S), first-stage tests (Sanderson-Windmeijer, Angrist-Pischke), endogeneity, and orthogonality.
• Tidyverse integration: a three-part formula interface and tidy(), glance(), and augment() methods, plus diagnostics() and first_stage() accessors that return the specification tests and first-stage results as tidy objects.
• Bundled datasets used throughout the help-file examples and vignettes: card, mroz, wagepan, griliches, klein, grunfeld, abdata, nlswork, phillips, stockwatson, and cigar.
• All outputs verified against Stata ivreg2 within tight numerical tolerances.