This study investigates the impact of two types of FinTech developed and utilized by banks and non-bank financial institutions on fraudulent behavior in China's A-share listed companies. Based on panel data from 2011 to 2020, the research findings suggest that both types of FinTech can suppress corporate fraud by enhancing internal control levels and external monitoring levels. Heterogeneity analysis indicates that the inhibitory effects of both FinTech types are more pronounced in companies with higher levels of digital transformation and lower levels of information disclosure. Additionally, due to differences in operating conditions, strategies, and objectives of FinTech developers, the inhibitory effect of bank FinTech is significant across all firms, whereas the effect of non-bank FinTech is only significant in high-risk firms. When distinguishing types of corporate fraud, both FinTech types significantly inhibit fraudulent activities related to information disclosure, fund utilization, and other categories. Further analysis reveals a complex interaction between the application effects of bank FinTech and non-bank FinTech. Specifically, the inhibitory effect of bank (non-bank) FinTech is significant when the development of other FinTech is high (low). By simultaneously incorporating both types of FinTech and their interaction terms, significant synergistic inhibitory effects are observed in fund misuse and other types of fraud. Finally, the results indicate that the synergistic development of both types of FinTech may introduce potential risks. In summary, this paper, by identifying the impact of FinTech development on corporate fraudulent behaviors, highlights the common characteristics and individual differences of different types of FinTech, emphasizes potential future cooperation opportunities between bank and non-bank FinTech, and points out potential risks in the development of FinTech.
Entrepreneurship is one of the key means to ease the pressure on social employment, and it is also a long-term driving force to ensure medium-high economic growth. The in-depth development of digital inclusive finance has stimulated the vitality of entrepreneurship, but whether it can effectively improve the quality of entrepreneurship is still a topic worth exploring. Using data from the China Household Finance Survey (CHFS), we find that digital inclusive finance has a positive impact on improving the performance of household entrepreneurship. Further analysis shows that optimizing the external entrepreneurial environment such as regional credit environment, regional innovation level, and market integration, is an important way for households to improve their entrepreneurial performance. In addition, based on the differences in the characteristics of entrepreneurial subjects and regional characteristics, the paper finds that the impact of digital inclusive finance on entrepreneurial performance is more significant in groups with medium and high financial literacy, long-distance groups, and groups in more developed areas. This paper has certain reference significance for further promoting the development of digital inclusive finance and better improving the quality of entrepreneurial development.
In the context of the development of financial technology, we start with the complex characteristics of financial big data and elaborate on the importance of transfer learning of using multi-source data information to assist target tasks. We explain the significance of transfer learning technology in dealing with data heterogeneity from the perspective of multi-source data, and summarize the relevant concepts and methods of transfer learning technology, including data-driven and model-based transfer learning methods. In addition, this paper proposes the unified framework of transfer learning method based on generalized moment estimation (GMM), gives the effective algorithm of transfer learning, and applies the proposed method to the application of transfer learning in risk value (VaR) and risk measure based on expected quantile (expectile) under multi-source data. Then, we simulate two scenarios where samples are of insufficient or imbalanced sample sizes, respectively, in the application to personal bank credit evaluation, with tests of the prediction accuracy of three transfer learning methods, and analysis of the importance of filtering resource domain information. Finally, we described more application scenarios and development prospects of transfer learning in the financial field.
Promoting mass entrepreneurship and innovation is of great significance for advancing economic structural adjustment, creating new engines of development, enhancing new driving forces for development, and pursuing a path of innovative-driven development, as well as promoting social upward mobility and fairness and justice. This study utilizes data from three rounds of the China Household Finance Survey (CHFS) conducted from 2013 to 2017 to measure the degree of opportunity inequality in China and empirically investigates its effects on resident entrepreneurship and the underlying mechanisms. The results indicate that the rise in opportunity inequality significantly stimulates the entrepreneurial motivation of resident households. For each standard deviation increase in opportunity inequality, the probability of household entrepreneurship increases by 1.14%. Mechanism analysis shows that opportunity inequality stimulates residents' pursuit of status, thereby promoting entrepreneurship among residents. Furthermore, expanded research findings reveal that digital finance strengthens the positive driving effect of opportunity inequality on entrepreneurship. This enhancement effect is only present in urban areas and is achieved by improving loan accessibility. Additionally, this study finds that livelihood-oriented fiscal expenditures also amplify the promotion effect of opportunity inequality on resident household entrepreneurship. In heterogeneous analysis, households with lower educational levels, lower total assets, and no unemployment insurance show stronger entrepreneurial motivations. Finally, this study finds that opportunity inequality suppresses the entrepreneurial performance of entrepreneurial households, indicating that government policies should focus on strengthening equal opportunities and supporting resident entrepreneurship.
With the rapid development of digital finance, digital transformation has become a strategic imperative for commercial banks. This paper employs a more comprehensive data set from China Banking Database (CBD), and examines the relationship between bank digital transformation and systemic vulnerability risks. Empirical results indicate that there is a significant inverted U-shaped relationship between the level of bank digital transformation and its own systemic vulnerability risks, and the results are robust, remaining valid even after addressing endogeneity issues. Mechanism analyses reveal that bank digital transformation indirectly affects its own systemic vulnerability risks by altering income acquisition efficiency and active risk-taking. The empirical findings of this paper have important practical implications for preventing systemic financial risks in the banking industry and understanding the balance between digital innovation and risk management.
Enhancing the welfare of the people is one of the core goals of high-quality development in China's new era. Digital financial inclusion plays a crucial role in improving the subjective well-being of Chinese residents. Utilizing the data from the China Household Finance Survey from 2013 to 2019, and integrating city tiers with municipal digital financial inclusion indices, this paper empirically investigates the impact of digital financial inclusion development on residents' subjective well-being using the ordered Probit model. The findings indicate that the development of digital financial inclusion significantly enhances the subjective well-being of residents. In terms of dimensions, its breadth of coverage and depth of use have a positive impact on residents' well-being, while the degree of digitalization has a negative effect. Moreover, the impact of digital financial inclusion development on subjective well-being varies significantly across different relative income and educational levels. Mechanism analysis shows that the development of digital financial inclusion enhances subjective well-being through three pathways: Improving residents' financial literacy, improving economic conditions, and enhancing social security levels.
Utilizing the 2015, 2017, and 2019 China Household Finance Survey (CHFS) data, combined with the income transition matrix analysis method and empirical analysis method, this study systematically investigates the impact of digital finance on income mobility and income inequality among rural households. The income transition matrix analysis reveals that rural households using digital finance have a higher probability of upward income mobility compared to those not using digital finance. Empirical research has found that digital finance significantly promotes upward income mobility and significantly reduces income inequality among rural households. The mechanism of action indicates that digital finance enhances households' income mobility by improving financial accessibility, facilitating the accumulation of development factors, and promoting off-farm employment opportunities. Furthermore, compared to middle and high-income rural households, digital finance has a greater impact on financial accessibility, development factor accumulation, and off-farm employment for low-income rural households. This consequently reduces income inequality, showcasing the inclusive growth characteristic of digital finance. Further analysis reveals that digital finance primarily impacts rural households' property income and wage income through these three pathways, ultimately promoting overall income mobility and reducing income inequality among households. Both digital payments and digital wealth management significantly contribute to upward income mobility and the reduction of income inequality among rural households, while digital lending has a negligible impact. This study provides empirical evidence to support the enhancement of policies aimed at fostering sustained income growth for rural households and optimizing the rural income distribution pattern through digital finance.
How can digital government construction empower real economy high-quality development through a sound data base system? This paper takes the local government SME financing service platform as an exogenous shock of digital government construction to study the impact of such digital government construction on SME investment. This paper finds that the establishment of the financing service platform significantly increases the investment level of SMEs. Heterogeneity analysis shows that the impact of the financing service platforms on corporate investment is more significant among firms with lower collateral value, underinvestment and among firms in areas with higher legal protection. Mechanism analysis further indicates that financing service platforms mainly improve the investment level of small and medium-sized enterprises by alleviating financing constraints and strengthening supervision and governance. This study expands the relevant literature on digital government construction and enterprise investment, and has enlightening significance for the government to guide financial institutions to provide high-quality financial services for enterprises, especially small and medium-sized enterprises.
The end-to-end portfolio selection strategy based on deep learning exhibits high decision-making performance, but its black-box nature hinders interpretability of the decision mechanism. In this paper, we propose a comprehensive end-to-end portfolio selection strategy that combines decision-making capability with interpretability using deep learning, reinforcement learning, and knowledge distillation method. Firstly, by leveraging an improved Transformer to alleviate its quadratic complexity issue, a long sequence representations extractor is proposed. Then, through the employment of a cross-assets attention network and reinforcement learning algorithm, a non-linear "black-box" model is constructed to facilitate dynamic allocation in financial assets. Next, by calculating the gradients of model's outputs with respect to the asset features, we compute significance vectors in the feature space to identify key influential features. Finally, a linear regression model is applied to the identified key features, resulting in a straightforward and economically interpretable end-to-end portfolio selection strategy. Empirical results demonstrate that this interpretable end-to-end portfolio selection strategy based on Transformer and key features achieves favorable return and risk performance, with both decision-making power of deep learning and interpretability. This study provides a portfolio selection strategy that combines efficient decision-making capability and interpretability, contributing to the application of deep learning in the financial domain.
In order-driven markets, limit order book serves as a crucial information carrier that centrally reflects traders' intentions and market liquidity. Quantitatively assessing the information content of the micro-characteristics of the limit order book lays the foundation for in-depth exploration of market dynamics and is significant for precise stock price prediction and effective evaluation of market efficiency. This paper rebuilds limit order book using tick-by-tick order flow data. Based on deep learning model, it incorporates incremental information of order flow measured from the horizontal time dimension and stock information of order flow measured from the vertical space dimension, in addition to multi-level price and transaction information. The information content of limit order book characteristics is comprehensively explored through predicting multiple microstructure variables. Empirical results indicate that compared to transaction data, unfilled order flow data contains richer and more effective information, demonstrating significant advantages in predicting micro indicators. As the prediction window extends, this information advantage becomes more prominent, highlighting the important role of order flow data in market analysis and prediction. However, transaction data has higher information transmission efficiency and can provide complementary information for predicting market indicators. The combined use of these two features can effectively improve prediction performance. This conclusion remains robust in the analysis of heterogeneity regarding stocks' own characteristics. Additionally, price levels and cross-asset effects significantly impact the information content of characteristics. Therefore, careful selection of market depth and thorough consideration of market environmental factors are necessary during the process of order flow data mining.
