TL;DR
This study investigates how economic policy uncertainty (EPU) and investor sentiment (IS) influence the returns of China's Growth Enterprise Market (GEM). Using a novel approach to construct an IS index from online forums and a time-varying parameter model, the research reveals that EPU generally dampens market sentiment and returns, while IS exhibits a short-term boost followed by a long-term decline.
What Problem Does This Paper Address?
The Chinese stock market, particularly its Growth Enterprise Market (GEM), is a dynamic and crucial component of the national economy, supporting innovative startups. However, understanding the complex interplay of factors influencing its performance is challenging. Specifically, the impact of economic policy uncertainty (EPU) and investor sentiment (IS) on GEM returns, especially their dynamic and asymmetric effects, has not been fully explored. Existing research often uses broader market indices or less granular sentiment data. This paper aims to fill this gap by providing a detailed, time-varying analysis of how EPU and IS, derived from specific online investor communities, affect GEM returns and its various industry sectors.
Methods
The researchers employed a sophisticated econometric model, the Time-Varying Parameter Stochastic Volatility Vector Autoregression (TVP-SV-VAR) model. This model is crucial because it allows for parameters to change over time, capturing the evolving nature of the relationships between EPU, IS, and market returns. To construct the investor sentiment (IS) index, they utilized the ERNIE (Enhanced Representation through Knowledge Integration) deep learning model. ERNIE is a powerful natural language processing model adept at understanding nuanced language, which was applied to analyze textual data from online investor forums to gauge sentiment. The TVP-SV-VAR model then integrated these sentiment indices, along with an economic policy uncertainty index, to analyze their dynamic impacts on GEM returns.
Datasets
The study utilized monthly data from January 2016 to December 2023.
- Economic Policy Uncertainty (EPU) Index: Sourced from a specialized index for China, building on previous work by Huang and Luk (2020).
- Investor Sentiment (IS) Index: This was a novel contribution. The researchers crawled 760,000 posts from the Guba forum (a major Chinese stock forum) and used the ERNIE model to classify sentiment, constructing a monthly IS index for the GEM.
- Growth Enterprise Market (GEM) Returns: This included the GEM Composite Index return and average monthly stock returns for various industry sectors within the GEM (e.g., Manufacturing, Information Technology, Scientific Research, Forestry). Data for these were sourced from the China Stock Market & Accounting Research (CSMAR) database.
Key Insights and Ideas
- EPU is a Drag on Sentiment and Returns: Economic policy uncertainty consistently has a negative effect on both investor sentiment and the overall returns of the GEM. This suggests that when policymakers are unclear or policies are volatile, investors become more cautious, impacting market performance.
- Investor Sentiment's Double-Edged Sword: Improved investor sentiment can provide a short-term boost to stock prices. However, this effect appears to be temporary, leading to a subsequent decline over the longer term – a "reversal effect." This highlights the potential for speculative bubbles driven by sentiment to eventually burst.
- Time-Varying Dynamics are Crucial: The relationships between EPU, IS, and GEM returns are not static. They change significantly over time, particularly showing heightened sensitivity during bull markets. This underscores the need for dynamic modeling approaches.
- Asymmetric Industry Impacts: The effects of EPU and IS are not uniform across all industries within the GEM. Different sectors exhibit distinct responses, suggesting that some industries are more vulnerable or resilient to these macroeconomic and sentiment-driven shocks.
Key Results
- EPU's Negative Impact: EPU negatively influences both investor sentiment and GEM returns. While it might cause short-term price dips, the empirical evidence suggests these impacts are generally short-lived.
- IS Reversal Effect: Investor sentiment shows a reversal effect on GEM returns. Initial positive sentiment can boost prices, but this is followed by a decline over the long term.
- Time-Varying Sensitivity: The relationship between EPU, IS, and GEM returns is time-variant, with increased sensitivity observed during bull markets.
- Asymmetric Industry Responses: The impact of EPU and IS on the returns of different GEM industries is asymmetric, meaning sectors react differently to these shocks.
Limitations and Future Work
The authors acknowledge that their analysis is limited to the EPU index from Huang and Luk (2020), IS derived from Guba forum posts, and the GEM composite index. They suggest that future research could explore the impact of EPU and IS on stock price volatility and stock price crashes within the GEM. Additionally, expanding the analysis to include international evidence or geographical differences would provide a broader perspective. The study also notes that other political, economic, and financial factors might influence these relationships, which were not explicitly included in this model.
Summary updated