BudgetIV: Optimal Partial Identification of Causal Effects with Mostly Invalid Instruments

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Penn, Jordan, Gunderson, Lee M., Bravo-Hermsdorff, Gecia, Silva, Ricardo, Watson, David S.
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929761012416512
author Penn, Jordan
Gunderson, Lee M.
Bravo-Hermsdorff, Gecia
Silva, Ricardo
Watson, David S.
author_facet Penn, Jordan
Gunderson, Lee M.
Bravo-Hermsdorff, Gecia
Silva, Ricardo
Watson, David S.
contents Instrumental variables (IVs) are widely used to estimate causal effects in the presence of unobserved confounding between exposure and outcome. An IV must affect the outcome exclusively through the exposure and be unconfounded with the outcome. We present a framework for relaxing either or both of these strong assumptions with tuneable and interpretable budget constraints. Our algorithm returns a feasible set of causal effects that can be identified exactly given relevant covariance parameters. The feasible set may be disconnected but is a finite union of convex subsets. We discuss conditions under which this set is sharp, i.e., contains all and only effects consistent with the background assumptions and the joint distribution of observable variables. Our method applies to a wide class of semiparametric models, and we demonstrate how its ability to select specific subsets of instruments confers an advantage over convex relaxations in both linear and nonlinear settings. We also adapt our algorithm to form confidence sets that are asymptotically valid under a common statistical assumption from the Mendelian randomization literature.
format Preprint
id arxiv_https___arxiv_org_abs_2411_06913
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle BudgetIV: Optimal Partial Identification of Causal Effects with Mostly Invalid Instruments
Penn, Jordan
Gunderson, Lee M.
Bravo-Hermsdorff, Gecia
Silva, Ricardo
Watson, David S.
Methodology
Statistics Theory
Quantitative Methods
Instrumental variables (IVs) are widely used to estimate causal effects in the presence of unobserved confounding between exposure and outcome. An IV must affect the outcome exclusively through the exposure and be unconfounded with the outcome. We present a framework for relaxing either or both of these strong assumptions with tuneable and interpretable budget constraints. Our algorithm returns a feasible set of causal effects that can be identified exactly given relevant covariance parameters. The feasible set may be disconnected but is a finite union of convex subsets. We discuss conditions under which this set is sharp, i.e., contains all and only effects consistent with the background assumptions and the joint distribution of observable variables. Our method applies to a wide class of semiparametric models, and we demonstrate how its ability to select specific subsets of instruments confers an advantage over convex relaxations in both linear and nonlinear settings. We also adapt our algorithm to form confidence sets that are asymptotically valid under a common statistical assumption from the Mendelian randomization literature.
title BudgetIV: Optimal Partial Identification of Causal Effects with Mostly Invalid Instruments
topic Methodology
Statistics Theory
Quantitative Methods
url https://arxiv.org/abs/2411.06913