Inference on varying coefficients in spatial autoregressions

• DOI: 10.48550/arxiv.2502.03084
• Publication Date: 2025-02-05
• Authors: Abhimanyu Gupta; Xi Qu; Sorawoot Srisuma; Jiajun Zhang
• Keywords: FOS: Economics and business; Econometrics (econ.EM); Economics - Econometrics

We present simple to implement Wald-type statistics that deliver a general nonparametric inference theory for linear restrictions on varying coefficients in range of spatial autoregressive models. Our covers error dependence form, allows degree misspecification robustness via weights and permits both regression coefficients. One application our method finds evidence constant returns scale the production function Chinese nonmetal mineral industry, while another nonlinear impact distance employment center housing prices Boston. A simulation study confirms tests perform well finite-samples.