FREEIV: Stata module for instrument-free estimation of a linear structural model with an endogenous regressor

freeiv estimates the coefficient on an endogenous regressor without an external instrument, in the linear triangular model in which a single latent confounder generates the endogeneity. Under scale consistency -- the confounder moves the outcome directly by exactly as much as it moves it through the regressor -- the second moments of the data leave one number free, and the coefficient lies between half the least-squares slope and that slope: an interval that assumes nothing more and is exactly the set of values for which the implied variances are non-negative. The routes of the command impose the same linear structure on the moments of the next orders: at order three the quadratic moment estimator (the default) and a closed form for symmetric noise; at order four a joint and a profiled GMM, each with a Hansen J, and the region where the profiled J stays within 3.84 of its minimum. A second indicator of the same confounder frees the confounder's direct loading and turns scale consistency into a test. The estimators of Lewbel, Schennach and Zhang (2024) and of Lewbel (2012), the Gaussian copula, the rank control function and the coefficient-stability adjustment of Oster (2019) are computed on the same data and judged by the same interval. Standard errors come from the influence functions of the moments, with pweights in sandwich form and Taylor linearization over a survey design declared by svyset (vce(svy)). Companion commands report what the data can carry before any estimation (freeivmenu), what a value of the coefficient implies about the unobservables (freeivdiag), tests of endogeneity and of the agreement between routes (freeivtest), and all of it in one table (freeivreport); a dialog box is included, and the examples of the help run from clickable links (freeiv_examples). The paper that documents the models and the command is included as an ancillary file (freeiv_paper.pdf).