Extracting gold risk premium via dimension reduction tools: implication on the gold-inflation relationship

被引：0

作者：

Roh, Tai-Yong ^{[1
]}

Lee, Byung Yoon ^{[2
]}

Xu, Yahua ^{[3
]}

机构：

[1] Liaoning Univ, Li Anmin Inst Econ Res, Shenyang, Peoples R China

[2] Seoul Natl Univ, Coll Social Sci, Seoul 08826, South Korea

[3] Cent Univ Finance & Econ, China Econ & Management Acad, Beijing, Peoples R China

来源：

APPLIED ECONOMICS LETTERS | 2024年

关键词：

Gold risk premium; dimension reduction approach; supervised learning techniques; return predictability; inflation hedge; TIME-SERIES;

D O I：

10.1080/13504851.2024.2364002

中图分类号：

F [经济];

学科分类号：

02 ;

摘要：

We forecast future gold returns with a long sample starting from 1980, using various dimension reduction tools. We demonstrate that the models based on dimension reduction tools can generate sizable fit (approximately 20% of in-sample R-2) and strong out-of-sample performance (approximately 15% of out-of-sample R-2). By utilizing the extracted gold risk premiums, we test the gold-inflation relationship on an ex-ante basis. We find a strong positive relationship between the expected inflation and the gold risk premium when forecasting with feature extraction methods (PLS, PCA, scaled-PCA, target-PCA).

引用

页数：10