Regularizing double machine learning in partially linear endogenous models
From MaRDI portal
Abstract: The linear coefficient in a partially linear model with confounding variables can be estimated using double machine learning (DML). However, this DML estimator has a two-stage least squares (TSLS) interpretation and may produce overly wide confidence intervals. To address this issue, we propose a regularization and selection scheme, regsDML, which leads to narrower confidence intervals. It selects either the TSLS DML estimator or a regularization-only estimator depending on whose estimated variance is smaller. The regularization-only estimator is tailored to have a low mean squared error. The regsDML estimator is fully data driven. The regsDML estimator converges at the parametric rate, is asymptotically Gaussian distributed, and asymptotically equivalent to the TSLS DML estimator, but regsDML exhibits substantially better finite sample properties. The regsDML estimator uses the idea of k-class estimators, and we show how DML and k-class estimation can be combined to estimate the linear coefficient in a partially linear endogenous model. Empirical examples demonstrate our methodological and theoretical developments. Software code for our regsDML method is available in the R-package dmlalg.
Recommendations
- Regularization statistical inferences for partially linear models with high dimensional endogenous covariates
- Instrumental regression in partially linear models
- Sparse linear models and \(l_1\)-regularized 2SLS with high-dimensional endogenous regressors and instruments
- A simple estimator for partial linear regression with endogenous nonparametric variables
- Regularized LIML for many instruments
Cites work
- scientific article; zbMATH DE number 3827056 (Why is no real title available?)
- scientific article; zbMATH DE number 4098524 (Why is no real title available?)
- scientific article; zbMATH DE number 52749 (Why is no real title available?)
- scientific article; zbMATH DE number 1134987 (Why is no real title available?)
- scientific article; zbMATH DE number 1533566 (Why is no real title available?)
- A Generalized Classical Method of Linear Estimation of Coefficients in a Structural Equation
- A Monte Carlo Study of Alternative Simultaneous Equation Estimators
- A Monte Carlo Study of Estimates of Simultaneous Linear Structural Equations
- A method for calculating bounds on the asymptotic covariance matrices of generalized method of moments estimators
- Analysis of testing-based forward model selection
- Analytical Small-Sample Distribution Theory in Econometrics: The Simultaneous-Equations Case
- Anchor Regression: Heterogeneous Data Meet Causality
- Applied nonparametric instrumental variables estimation
- Asymptotic theory of statistics and probability
- Bandit algorithms
- Causality. Models, reasoning, and inference
- Data-adaptive doubly robust instrumental variable methods for treatment effect heterogeneity
- Double/debiased machine learning for logistic partially linear model
- Double/debiased machine learning for treatment and structural parameters
- Doubly Robust Estimation in Missing Data and Causal Inference Models
- Doubly robust instrumental variable regression
- Efficient Estimation of Models with Conditional Moment Restrictions Containing Unknown Functions
- Elements of causal inference. Foundations and learning algorithms
- Estimation of the Parameters of a Single Equation in a Complete System of Stochastic Equations
- Estimation with weak instruments: Accuracy of higher‐order bias and MSE approximations
- Evaluation of the Distribution Function of the Limited Information Maximum Likelihood Estimator
- Evaluation of the Distribution Function of the Two-Stage Least Squares Estimate
- GMM estimation in partial linear models with endogenous covariates causing an over-identified problem
- Generalized random forests
- Handbook of Graphical Models
- High-dimensional additive modeling
- Identification of Causal Effects Using Instrumental Variables
- Instrumental Variables Regression with Weak Instruments
- Instrumental regression in partially linear models
- Invariance, causality and robustness
- Large Sample Properties of Generalized Method of Moments Estimators
- Least squares after model selection in high-dimensional sparse models
- Locally Efficient Estimators for Semiparametric Models With Measurement Error
- Machine Intelligence for Individualized Decision Making Under a Counterfactual World: A Rejoinder
- Minimax optimal rates of estimation in high dimensional additive models
- Nonparametric and semiparametric models.
- On Deriving the Inverse of a Sum of Matrices
- On adaptive estimation
- On the asymptotic optimality of the LIML estimator with possibly many instruments
- Origins of the limited information maximum likelihood and two-stage least squares estimators
- Penalized quasi-likelihood estimation in partial linear models
- Probability. Theory and examples.
- Root-N-Consistent Semiparametric Regression
- Semiparametric Regression
- Simultaneous analysis of Lasso and Dantzig selector
- Some Properties of a Modification of the Limited Information Estimator
- Sparsity in multiple kernel learning
- Statistics for big data: a perspective
- Statistics for high-dimensional data. Methods, theory and applications.
- The Bias and Moment Matrix of the General k-Class Estimators of the Parameters in Simultaneous Equations
- The Dantzig selector: statistical estimation when \(p\) is much larger than \(n\). (With discussions and rejoinder).
- The Exact Distribution of LIML: I
- The Exact Distribution of LIML: II
- The Existence of Moments of the Ordinary Least Squares and Two-Stage Least Squares Estimators
- The nonlinear two-stage least-squares estimator
- Third-Order Efficiency of the Extended Maximum Likelihood Estimators in a Simultaneous Equation System
- Unified methods for censored longitudinal data and causality
Cited in
(3)
This page was built for publication: Regularizing double machine learning in partially linear endogenous models
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q115460)
