The blind box market has been a booming business among young population groups in the past few years largely due to its aspect of ambiguity and limited availability. Various factors affect the consumer decisions. These are psychological, social and environmental factors. Based on the behavioral economics theory and using Pop Mart as an example case study this paper discusses how three major cognitive forces such as loss aversion, perceived scarcity, and influence buying behavior. Variable-ratio reinforcement- influences the buying choices in the blind box consumption. Research has shown that the products make consumers buy the merchandise again by heightening their feeling of missing stuff concealed in them. Also, the fact that they are being widely disseminated via social media platforms enhances this act of consumption, redirecting the motivation of buying blind boxes towards more personalized values of possession to those of belonging and identification with a social group. In general, it seems that the results indicate that, as much as the use of blind box marketing leads to a substantial increase in brand value and user loyalty, it also brings up possible negative aspects to it, such as the tendency to make purchases on impulse and relying on unstable consumption habits. The study can be regarded to facilitate the understanding of the functioning of the blind box economy and provide empirical research that may be used to explain the formulation of a more effective consumer protection strategy.

Show more

Under the background of strengthening sustainable development, companies' ESG performance is also becoming an important part in the capital market. In this paper, it is studied that the influence of corporate ESG performance on COD with the samples which are Chinese A-shares listed corporations from 2018-2023. And also we study on the different role of governance dimensions and the moderating role of externals information environment. Use the 2-way Fixed effect method to control firm size, financial leverage, profit level, growth potential, etc., and do an empirical study of our hypothesis. After controling for the firm and year fixed effects, it's clear that overall ESG is in line with what we'd expect from theory regarding its influence on COD, but not stat sig. It is clear from this that its probably mostly there because of longer run risk and such. Of all the dimensions, only governance's performance shows weak explanation, but this does not change its overall effect on ESG composite, so governance is still important. As for the moderation result, it's observed that external information environment did not have significant effect on ESG-COD relation but its interaction term has sign that matched with our theoretical expectations which implies that the value of ESG information might be higher when surrounded by opaque information. We look at the dynamics through a creditor's view, prices of risk, it gives us empirical takeaways from China's capital market on ESG implies. And it helps us know more about the way of ESG information, corporate governance and COD.

Show more

In the current era, corporate bond default incidents occur frequently, and the financial sector has been focusing on how to accurately identify potential default risks of enterprises. Most traditional default prediction methods rely on financial indicator data, which have certain explanatory power, but they are still subject to factors such as information lag and susceptibility to manipulation, leading to limitations in prediction effectiveness. This paper introduces the Management's Discussion and Analysis (MD&A) textual information and integrates it with financial data to construct a default prediction model. The study selects annual report data from 2021 to 2024 of 67 A-share listed companies in our country, generating panel data with 251 observations. This article chooses Light GBM as main algorithm, while financial indicators such as the debt-to-asset ratio and net profit rate, are selected to depict the financial condition of enterprises. From textual data, features such as sentiment orientation and forward-looking descriptions are extracted, by using Bert semantics. This features are then formed into a feature system through dimensionality reduction. The results shows that the model possesses strong discriminatory capability and good identification of high-risk enterprises and a low misjudgment ratio. At the same time it is discovered that forward-looking information contributes to enhancing the accuracy and forward-looking nature of default risk identification, which proves the supplementary value of the text information.

Show more

Stock market prediction has become a difficult problem in the field of financial engineering due to high data noise, non-stationarity and nonlinear characteristics, which traditional methods can hardly handle effectively. This paper systematically combs the research achievements of artificial intelligence technology in the field of stock forecasting from 2015 to 2026. Starting from traditional machine learning to deep learning, from single modality to multimodal fusion, and from general architectures to financial-specific pre-training, it adopts a five-dimensional evaluation framework to quantify model performance. Studies have found that convolutional neural networks enable the automatic recognition of financial patterns, recurrent neural networks address the challenge of modeling temporal dependencies, Transformer and state-space models break through the efficiency bottleneck of long sequences, and multimodal fusion and financial pre-trained models significantly improve domain adaptability. Current research still faces challenges such as insufficient out-of-distribution generalization and lack of interpretability. Lightweight architectures and causal inference represent the main future directions. This study provides systematic technical references for the design of stock prediction models.

Show more