公司高质量发展水平能否预测其股票市场表现——基于机器学习方法

Abstract

Building on the concept of high-quality development, this study constructed an index system guided by five core development principles: “Innovation, Coordination, Sustainability, Openness, and Inclusivity.” We employed nine machine learning techniques to forecast stock returns, investigating the importance of each variable in the index system, fluctuations in predictive power, and the origins of this forecasting capability. The findings reveal that a company’s level of high-quality development significantly impacts its stock market performance. Notably, indicators related to innovation emerge as the most critical for forecasting stock returns. Except during times of market turbulence, the predictive accuracy stemming from high-quality development levels was fairly stable and even showed signs of improvement. From a component-wise perspective, the predictive strength of innovation and sustainability indicators increased, whereas coordination indicators are losing their forecasting power. The level of high-quality development can indeed forecast a company’s future fundamental shifts; yet, the market appears to underreact to this valuable information. This market mispricing serves as the source of the predictive power of our high-quality development index. We therefore recommend refining the capital market’s pricing mechanisms to better account for a company’s level of high-quality development.