Using Machine Learning- based Linear Regression to Analyze the Impact of M2 on the Stock Market Index from China

Abstract

This paper investigates whether money supply (M2), as published by the People’s Bank of China, can significantly explain or predict fluctuations in China’s stock market, represented by the Shanghai Stock Exchange Composite Index. Using monthly data from June 2015 to June 2025, the research constructs a regression model with Python’s scikit-learn Linear Regression tool. Both variables are log-transformed, and the dataset is divided into training and testing sets to evaluate predictive performance. The model results indicate a positive correlation between log(M2) and log(SSE), but the explanatory power is extremely limited, with R² values close to zero and even negative in the testing set. Error metrics such as MSE, MAE, and RMSE further reveal that prediction accuracy is weak and unstable, with results often distorted by policy interventions and institutional frictions. These findings suggest that while M2 has some influence through the liquidity channel, its role is indirect and fragile in China’s policy-sensitive market. By integrating traditional financial theory with machine learning validation techniques, this study provides transparent visualization, predictive testing, and policy insights, offering quantitative evidence for understanding the complex interaction between monetary policy and the Chinese capital market.