Why is price momentum absent in China? Since momentum is commonly considered arising from investors’ under-reaction to fundamental news, we decompose monthly stock returns into news- and non-news-driven components and document a news day return continuation along with an offsetting non-news day reversal in China. The non-news day reversal is particularly strong for stocks with high retail ownership, relatively less recent positive news articles, and limits to arbitrage. Evidence on order imbalance suggests that stock returns overshoot on news days due to retail investors' excessive attention-driven buying demands, and mispricing gets corrected by institutional investors on subsequent non-news days. To avoid this tug-of-war in stock price, we use a signal that directly captures the recent news performance and re-document a momentum-like underreaction to fundamental news in China.

We construct a daily Chinese Housing Market Sentiment Index by applying GPT-4o to Chinese news articles. Our method outperforms traditional models in several validation tests, including a test based on a suite of machine learning models. Applying this index to household-level data, we find that after monetary easing, an important group of homebuyers (who have a college degree and are aged between 30 and 50) in cities with more optimistic housing sentiment have lower responses in non-housing consumption, whereas for homebuyers in other age-education groups, such a pattern does not exist. This suggests that current monetary easing might be more effective in boosting non-housing consumption than in the past for China due to weaker crowding-out effects from pessimistic housing sentiment. The paper also highlights the need for complementary structural reforms to enhance monetary policy transmission in China, a lesson relevant for other similar countries. Methodologically, it offers a tool for monitoring housing sentiment and lays out some principles for applying generative AI models, adaptable to other studies globally.

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 “Internet celebrity economy” is a business model aimed at capitalizing on online traffic based on the purchasing power of users on social media in which “influencers”—highly influential individuals—exercise their marketing power to create a fandom. China has witnessed an abrupt outbreak in its “Wanghong” (internet celebrity) economy since 2016, eventually leading to consecutive high closes for related stocks from around 2020. The empirical findings are as follows: First, investors’ attention to Wanghong stocks and cumulative abnormal returns (CARs) are significantly positively associated. However, operational results and CARs are weakly linked, implying that the economic impact of intense influencer marketing is short-lived, and abnormal returns constitute an anomaly. Second, the positive abnormal returns of Wanghong stocks last approximately six months, which overlaps with the boom period of the Wanghong index based on influencer news articles.

In this paper, we introduce the first dataset that records the news co-mention relationships in the Chinese A-share market. In total, we collected 1,138,247 pieces of news articles that at least mentioned one listed firm in the A market from major Chinese media and financial websites from September 1999 to December 2022. The development of this dataset could enable data scientists and financial economists to investigate the network of stocks through news co-mention in the Chinese stock market. The dataset could also help to construct novel portfolio strategies like the cross-firm momentum strategy with news-implied links as in Ge et al. (2023).

We use supervised machine learning to develop a Chinese language financial sentiment dictionary from 3.1 million financial news articles. Our dictionary maps semantically similar words to a subset of human-expert generated financial sentiment words. In article-level validation tests, our dictionary scores the sentiment of articles consistently with a human reading of full articles. In return validation tests, our dictionary outperforms and subsumes previous Chinese financial sentiment dictionaries such as direct translations of Loughran and McDonald’s (2011) financial words. We also generate a list of politically-related positive words that is unique to China; this list has a weaker association with returns than does the list of otherwise positive words. We demonstrate that state media exhibits a sentiment bias by using more politically-related positive and fewer negative words, and this bias renders state media’s sentiment less return-informative. Our findings demonstrate that dictionary-based sentiment analysis exhibits strong language and domain specificity.

This paper examines a board's decision to hire a star CEO and analyzes the consequences of this decision for firm performance. We propose a new methodology to identify star CEOs by analyzing the texts contained in 18,240 Wall Street Journal news articles. Unlike previous measures, our new metric accounts for the time series variations of executives’ visibility as well as how favorably these executives are portrayed in the business press. The proposed measure indicates that boards with short industry tenure or busy boards are more likely to select a star CEO. Consistent with previous evidence, firms that hire star CEOs perform subsequently worse than firms that hire non-star CEOs. However, in contrast to previous work, we show that this underperformance is attributable to boards with short industry tenure or busy boards, rather than the inabilities of star CEOs. Furthermore, our event studies of stock market reactions to hiring news imply that investors prefer star CEOs selected by boards with long industry tenure. Our work contributes to the literature by offering insights into how board composition affects firm performance.

Based on a sample of financial holding companies listed on the Taiwan Stock Exchange, we examine the impact of media coverage of corporate social performance on corporate financial performance. Our findings are as follows. First, information about a firm’s social actions provided by the media is more relevant than provided by the financial holding company (FHC) itself, and the quantity of news articles about positive social activities of FHCs is positively correlated with financial performance; however, strikingly, that of news articles about FHCs’ negative social activities is also positively correlated with financial performance. In addition, we find that news articles about FHCs’ positive social activities for shareholders will trigger a positive evaluation by shareholders; however, rather interestingly, news articles about FHCs’ positive (negative) social activities for employees will trigger a negative (positive) evaluation by shareholders. But if the news articles about FHCs’ positive social activities for employees are initiated by the media, rather than by the company itself, they will trigger a positive evaluation by shareholders. Therefore, the evidence suggests that “doing good” can be expected to be “doing well” if the positive CSP information is provided by the media, rather than by the company itself.