Quantile Regression with Multiple Proxy Variables

TL;DR

This paper introduces a novel quantile regression method that effectively integrates multiple proxy variables across datasets, accommodating nonlinear relationships and unobserved covariates, demonstrated through simulations and real data application.

Contribution

It proposes a unified estimation approach for quantile regression with multiple proxies, handling nonlinearities and unobserved covariates without requiring linearity assumptions.

Findings

Successfully integrates multiple proxies in simulations

Reveals quantile relationships in nonlinear data

Applied to household finance data to analyze assets and income

Abstract

Data integration has become increasingly popular owing to the availability of multiple data sources. This study considered quantile regression estimation when a key covariate had multiple proxies across several datasets. In a unified estimation procedure, the proposed method incorporates multiple proxies that have various relationships with the unobserved covariates. The proposed approach allows the inference of both the quantile function and unobserved covariates. Moreover, it does not require the quantile function's linearity and, simultaneously, accommodates both the linear and nonlinear proxies. Simulation studies have demonstrated that this methodology successfully integrates multiple proxies and revealed quantile relationships for a wide range of nonlinear data. The proposed method is applied to administrative data obtained from the Survey of Household Finances and Living Conditions provided by Statistics Korea, to specify the relationship between assets and salary income in the presence of multiple income records.