The healthy operation of the stock market relies on sound liquidity. We utilize the semantic information from disclosure texts of listed companies on the China Science and Technology Innovation Board (STAR Market) to construct a daily corporate network. Through empirical tests and performance analyses of machine learning models, we elucidate the relationship between the similarity of company disclosure text contents and the temporal and spatial correlations of stock liquidity. Our liquidity indicators encompass trading costs, market depth, trading speed, and price impact, recognized across four dimensions. Furthermore, we reveal that the information loss caused by employing Minimum Spanning Tree (MST) topology significantly affects the explanatory power of network topology indicators for stock liquidity, with a more pronounced impact observed at the document level. Subsequently, by establishing a neural network model to predict next-day liquidity indicators, we demonstrate the temporal relationship of stock liquidity. We model a liquidity predicting task and train a daily liquidity prediction model incorporating Graph Convolutional Network (GCN) modules to solve it. Compared to models with the same parameter structure containing only fully connected layers, the GCN prediction model, which leverages company network structure information, exhibits stronger performance and faster convergence. We provide new insights for research on company disclosure and capital market liquidity.

The stock price of a firm is dynamically influenced by its own factors as well as those of its peers. In this study, we introduce a Graph Attention Network (GAT) integrated with WaveNet architecture—termed the GAT-WaveNet model—to capture both time-series and spatial dependencies for forecasting the stock price crash risk of Chinese listed firms from 2012 to 2021. Utilizing node-rolling techniques to prevent overfitting, our results show that the GAT-WaveNet model significantly outperforms traditional machine learning models in prediction accuracy. Moreover, investment portfolios leveraging the GAT-WaveNet model substantially exceed the cumulative returns of those based on other models.

We examine whether large language models (LLMs) can extract contextualized representation of Chinese news articles and predict stock returns. The LLMs we examine include BERT, RoBERTa, FinBERT, Baichuan, ChatGLM and their ensemble model. We find that tones and return forecasts extracted by LLMs from news significantly predict future returns. The equal- and value-weighted long minus short portfolios yield annualized returns of 90% and 69% on average for the ensemble model. Given that these news articles are public information, the predictive power lasts about two days. More interestingly, the signals extracted by LLMs contain information about firm fundamentals, and can predict the aggressiveness of future trades. The predictive power is noticeably stronger for firms with less efficient information environment, such as firms with lower market cap, shorting volume, institutional and state ownership. These results suggest that LLMs are helpful in capturing under-processed information in public news, for firms with less efficient information environment, and thus contribute to overall market efficiency.

This paper examines the impact of asset bubbles on innovation and long-run economic growth within a semi-endogenous growth framework, incorporating idiosyncratic productivity shocks and endogenous credit constraints in the R&D sector. It demonstrates that pure bubbles tied to intrinsically useless assets and equity bubbles linked to intermediate goods firms can coexist, relaxing credit constraints and boosting entrepreneurs’ total factor productivity (TFP), which stimulates R&D and enhances growth along the transitional path. However, these bubbles generally do not influence the long-run economic growth rate. The model’s mechanisms and predictions are supported by aggregate and firm-level evidence, showing a positive correlation between equity bubbles and R&D investment, with stronger effects during periods of tightened financial constraints.

Banks have a significant funding-cost advantage if their liabilities are protected by bailout guarantees. We construct a corporate finance-style model showing that banks can exploit this funding-cost advantage by just intermediating funds between investors and ultimate borrowers, thereby earning the spread between their reduced funding rate and the competitive market rate. This mechanism leads to a crowding-out of direct market finance and real effects for bank borrowers at the intensive margin: banks protected by bailout guarantees induce their borrowers to leverage excessively, to overinvest, and to conduct inferior high-risk projects. We confirm our model predictions using U.S. panel data, exploiting exogenous changes in banks' political connections, which cause variation in bailout expectations. At the bank level, we find that higher bailout probabilities are associated with more wholesale debt funding and lending. Controlling for loan demand, we confirm this effect on bank lending at the bank-firm level and find evidence on loan pricing consistent with a shift towards riskier borrower real investments. Finally, at the firm level, we find that firms linked to banks that experience an expansion in their bailout guarantees show an increase in their leverage, higher investment levels with indications of overinvestment, and lower productivity.

We leverage machine learning to introduce belief dispersion measures to distinguish different opinion (DO) and information asymmetry (IA). Our measures align with the human-based measure and relate to economic outcomes in a manner consistent with theoretical prediction: DO positively relates to trading volume and negatively linked to bid-ask spread, whereas IA shows the opposite effects. Moreover, IA negatively predicts the cross-section of stock returns, while DO positively predicts returns for underpriced stocks and negatively for overpriced ones. Our findings reconcile conflicting disagree-return relations in the literature and are consistent with Atmaz and Basak (2018)’s model. We also show that the return predictability of DO and IA stems from their unique economic rationales, underscoring that components of disagreement can influence market equilibrium via distinct mechanisms.

This study shows how price limit rules, intended to stabilize markets, inadvertently distort price dynamics by fostering strategic speculation. Through a dynamic rational expectations model, we demonstrate that price limits induce post limit-up price jumps by impeding full information incorporation, enabling speculators to artificially push prices to upper bounds and exploit uninformed traders. The model predicts two distinct patterns: (1) stocks closing at price limits exhibit positive overnight returns followed by long-term reversals, and (2) stocks retreating from upper bounds suffer sharp reversals with partial recovery. Empirical analysis confirms these predictions. A natural experiment from China’s 2020 GEM reform —- which widened the price limit -— further provides causal evidence that relaxed limits mitigate speculative distortions.

I conduct an exploratory study about the feasibility of factor timing in the Chinese stock market, covering 24 representative and well-identiffed risk factors in ten categories from the literature. The long-short portfolio of short-term reversal exhibits strong and statistically signiffcant out-of-sample predictability, which is robust across various models and all types of predictors. However, such results are not evident in the prediction of all other factors’ long-short portfolios, as well as all factors’ long-wing and short-wing portfolios. The high exposure to the market beta, together with the unpredictability of the market return, explains these failures to some degree. On the other hand, a simple investment strategy based on predicted returns of the reversal factor’s long-short portfolio obtains a signiffcant return three times higher than the simple buy-and-hold strategy in the sample period, with a signiffcant annualized 20.4% CH-3 alpha.

We conceptually identify and empirically verify the features distinguishing FinTech platforms from non-financial platforms using marketplace lending data. Specifically, we highlight three key features: (i) Long-term contracts introducing default risk at both the individual and platform levels; (ii) Lenders’ investment diversification to mitigate individual default risk; (iii) Platform-level default risk leading to greater asymmetric user stickiness and rendering platform-level cross-side network effects (p-CNEs), a novel metric we introduce, crucial for adoption and market dynamics. We incorporate these features into a model of two-sided FinTech platform with potential failures and endogenous participation and fee structures. Our model predicts lenders’ single-homing, occasional lower fees for borrowers, asymmetric p-CNEs, and the predictive power of lenders’ p-CNEs in forecasting platform failures. Empirical evidence from China’s marketplace lending industry, characterized by frequent market entries, exits, and strong network externalities, corroborates our theoretical predictions. We find that lenders’ p-CNEs are systematically lower on declining or well-established platforms compared to those on emerging or rapidly growing platforms. Furthermore, lenders’ p-CNEs serve as an early indicator of platform survival likelihood, even at the initial stages of market development. Our findings provide novel economic insights into the functioning of multi-sided FinTech platforms, offering valuable implications for both industry practitioners and financial regulators.

In the face of severe climate change, researchers have looked for assistance from financial instruments. They have examined how to hedge the risks of these instruments created by market fluctuations through various green financial derivatives, including green bonds (i.e., fixed-income financial instruments designed to support an environmental goal). In this study, we designed a green bond index option contract. First, we combined an autoregressive moving-average model (AMRA) with a generalized autoregressive conditional heteroskedasticity model (GARCH) to predict the green bond index. Next, we established a fractional Brownian motion option pricing model with temporally variable volatility. We used this approach to predict the closing price of the China Bond–Green Bond Index from 3 January 2017 to 30 December 2021 as an empirical analysis. The trend of the index predicted by the ARMA–GARCH model was consistent with the actual trend and predictions of actual prices were highly accurate. The modified fractional Brownian motion option pricing model improved the pricing accuracy. Our results provide a policy reference for the development of a green financial derivatives market, and can accelerate the transformation of markets towards a more sustainable economic development model.