We examine whether large language models (LLMs) can extract contextualized representation of Chinese public news articles to predict stock returns. Based on representativeness and influences, we consider seven LLMs: BERT, RoBERTa, FinBERT, Baichuan, ChatGLM, InternLM, and their ensemble model. We show that news tones and return forecasts extracted by LLMs from Chinese news significantly predict future returns. The value-weighted long-minus-short portfolios yield annualized returns between 35% and 67%, depending on the model. Building on the return predictive power of LLM signals, we further investigate its implications for information efficiency. The LLM signals contain firm fundamental information, and it takes two days for LLM signals to be incorporated into stock prices. The predictive power of the LLM signals is stronger for firms with more information frictions, more retail holdings and for more complex news. Interestingly, many investors trade in opposite directions of LLM signals upon news releases, and can benefit from the LLM signals. These findings suggest LLMs can be helpful in processing public news, and thus contribute to overall market efficiency.

The Chinese government has actively traded in the stock market through governmentsponsored institutions, the National Team, since the 2015 market crash. I adopt Koijen and Yogo’s (2019) demand system approach in China’s stock market to understand the impact of government participation. Estimation results indicate the government tilts towards large, less risky, and SOE stocks. During the crash, government participation indeed stabilized the market: the large-scale purchases reduced the cross-sectional market volatility of annual return by 1.8% and raised the market price by 11%. When the market ffuctuation returns to normal, the government acts more like an active investor; its price impact remains high but does not contribute to the cross-sectional volatility. Based on the theoretical framework of Brunnermeier et al. (2020), I investigate the interaction between the Nation Team and retail investors to reveal the government trading strategy. No evidence shows that government participation signiffcantly distorts market information efficiency.

With China being a large developing economy, the trading in China’s stock market is dominated by retail investors, and its government actively participates in this market. These features are quite different from those of typical developed markets, and This review focuses on two important questions: how do retail and institutional investors trade in China and why? We have three main findings after reviewing 100+ previous studies. First, small retail investors have low financial literacy, exhibit behavioral biases, and not surprisingly, negatively predict future returns; whereas large retail investors and institutions are capable of process information, and they positively predict future returns. Second, the macro- and firm-level information environment in China is slowly but gradually improving. Finally, the Chinese government actively adjusts their regulations of the stock market to serve the dual goals of growth and stability, with many of them being effective, while some may not generate intended consequences.

In order to reveal the impact of securities investment fund behavior on market quality, this paper starts from the perspective of microstructure of the securities market and utilized the transactional accounts of Shanghai Stock Exchange(SSE) to analyze the effect of impact on market quality (including liquidity, volatility and information efficiency) by securities investment funds by applying the cross-sectional model. The empirical result showed that institutionalization of the structure of domestic investors hasn’t improved market quality significantly. The increase (decrease) of positions by funds has significant impact on immediate liquidity and possesses permanent shocking characteristics. Net changes of positions by funds have led to higher hetero-volatility, whereas funds，functioning as institutional investors, do stabilize market to some degree in the adjustment phase of bull market, especially during the market turbulence of “2.27” and “5.30” in 2007; during the rising phases of stock market, the changes of positions by funds will improve market liquidity and enhance informational efficiency of securities market.

The B-share markets in China, originally for foreign investors only, were opened to local investors in 2001. This reform was expected to improve the information efficiency in B-share markets, since local investors were supposed to be better informed than foreign investors. Meanwhile, we find that, after opening to local investors, B-share price synchronicity increases, and firm-specific return variation (idiosyncratic risk) decreases. Opening B-share markets to local investors fails to improve or even deteriorates the information capitalization of B-share prices. The findings may help us understand Chinese government’s policy making. For instance, in August 2007, Chinese government announced that Chinese citizens would be allowed in public to buy and sell Hong Kong stocks through special accounts with domestic commercial banks. But after hearing opinions from different entities, Chinese government decides to infinitely postpone this policy.