Shoiuld Humans Lie to Machines: The Incentive Compatibility of Lasso and General Weighted Lasso

2021/01/04 by Mehmet Caner, Caner, Mehmet, Kfir Eliaz +1
Decision Sciences · Mathematics · #Advanced Causal Inference Techniques #Auction Theory and Applications #Benford’s Law and Fraud Detection #Econometrics (econ.EM) #FOS: Economics and business

paper · pdf · doi:10.48550/arxiv.2101.01144

openalex publication_date 2021/01/04 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

We consider situations where a user feeds her attributes to a machine learning method that tries to predict her best option based on a random sample of other users. The predictor is incentive-compatible if the user has no incentive to misreport her covariates. Focusing on the popular Lasso estimation technique, we borrow tools from high-dimensional statistics to characterize sufficient conditions that ensure that Lasso is incentive compatible in large samples. We extend our results to the Conservative Lasso estimator and provide new moment bounds for this generalized weighted version of Lasso. Our results show that incentive compatibility is achieved if the tuning parameter is kept above some threshold. We present simulations that illustrate how this can be done in practice.

Citations

• Statistical Theory for High-Dimensional Models

Related

• The Mismatch Principle: The Generalized Lasso Under Large Model Uncertainties
• Risk-consistency of cross-validation with lasso-type procedures
• False Discoveries Occur Early on the Lasso Path
• On cross-validated Lasso in high dimensions
• De-Biasing The Lasso With Degrees-of-Freedom Adjustment