Qiyang Yu

Research

High-dimensional econometrics, financial econometrics, and factor models.

Job Market Paper

Spillovers Under Common Shocks in Panel Data: Estimation and Inference

Abstract. I propose a latent-network model that recovers spillover links while controlling for common shocks. The model estimates interaction structures in a data-driven way rather than imposing them ex ante. I establish inferential theory in large panels.

Published Papers

Estimation and Inference for CP Tensor Factor Models

with Bin Chen and Yuefeng Han

Journal of Econometrics, 2026

Journal Preprint

Diffusion Index Forecasting with Tensor Data

with Bin Chen and Yuefeng Han

Journal of Econometrics, 2026

Journal Preprint

Working Papers

Useful Factors Are Fewer Than You Think

with Bin Chen and Guofu Zhou

Paper

Work in Progress

Semiparametric Inference in Panel Data with Interactive and High-dimensional Confounding

with Bin Chen and Yukun Ma

We develop Factor-Augmented Double/Debiased Machine Learning (FA-DML) for treatment-effect inference in panels with high-dimensional covariates and latent interactive confounders. A factor-orthogonal score combines machine learning, factor estimation, and block cross-fitting to deliver valid inference, illustrated through bank branching deregulation and income inequality.

Missing Imputation of CP Factor Model

with Bin Chen and Yuefeng Han

We are developing a CP tensor method built on the CC-ISO algorithm by Chen et al. (2026) for completing tensor-valued time series with missing observations.

Asset Pricing with a Latent Network

with Qixiang Gao

We develop a model and estimation algorithm to recover characteristic networks and their risk premia from large asset panels. We apply the method to U.S. stocks to study the cross-sectional pricing of characteristic networks.