徐同 Tong Xu

Working and Published Paper

An Interpretable High-Dimensional International Factor Model

(Job Market Paper)

Multiple anomalies in different countries form a dataset that is high-dimensional by construction. For this high-dimensional array, also known as tensor, I apply CP decomposition, which is a rank decomposition of tensor, and obtain the factor matrices of time, country, and anomaly dimensions, respectively. The pricing factors from CP decomposition, i.e., CP factors, load differently on different anomalies as well as different countries, suggesting that they can be interpreted as anomaly factor or country factor, which is not attainable by traditional two-dimensional methods like PCA. By treating country and anomaly as two distinct dimensions, CP factors have a Sharpe ratio of 1.51. Besides, CP factors have better pricing power than traditional factors that are constructed locally, thus providing new evidence on global equity market integration.

How Do Implicit Guarantees by Local Governments Affect the Real-Time Credit Risk of Small and Medium-Sized Banks? (In Chinese)

with Zhuo Chen, Zhiguo He, and Xiaoquan Zhu, Journal of Financial Research, 2025 (2) 76-94.

地方政府隐性担保如何影响中小银行实时的信用风险?

合作者: 陈卓, 何治国, 祝小全, 金融研究, 2025年第2期, 76-94.

COVID-19 vaccine: Does It Save the World Stock Markets?

With Kam Fong Chan, Zhuo Chen, and Yuanji Wen, Financial Research Letters 47, 2022.