No abstract

No abstract

The introduction of green credit policy provides an important idea to solve the contradiction between economic development and environmental protection. Based on fuzzy-set Qualitative Comparative Analysis (fsQCA) method, from the perspective of bank governance structure, this paper selects ownership concentration, independence of the Board, executive incentive, activity of Supervisory Board, degree of market competition and loan quality as antecedent variables to analyze the path of their impact on green credit. It is found that: (1) The main way to achieve high-level green credit is high ownership concentration and good loan quality. (2) The configuration of green credit has causal asymmetry. (3) Ownership structure is the most important factor affecting green credit. (4) There is a substitution between the low independence of the Board and the low executive incentive. The low activity of Supervisory Board and the poor loan quality are also substitutable to a certain extent. The research conclusion of this paper is helpful to improve the green credit level of Chinese banks and win the green reputation for banks.

Agricultural finance is in an embarrassing position in the current financial environment, especially during the process of COVID-19. Based on a small-scale peasant economy, it can no longer meet the rapidly rising demand of farmers for agricultural finance. Moreover, there has been a serious disconnection between the financial system of secondary and tertiary industries, and the quality of development needs to be improved urgently. The agricultural loan risk assessment has always been the main problem that we pay great attention to in the innovation of agricultural finance. Agricultural loans are an indispensable element in supporting agricultural development and promoting rural revitalization strategy. However, financial institutions have certain credit risks in reviewing and issuing agricultural loans. This article studies the speech emotion recognition of farmers in loan review to assess loan risk. As for emotional confusion caused by speech segmentation, a special method of data connection between Convolutional Neural Networks (CNNs) and Bidirectional Long Short-Term Memory (Bi-LSTM) Networks is designed, and a variable-length speech emotion recognition model including CNN and Bi-LSTM is designed. Experimental results show that the proposed algorithm can effectively improve the risk assessment of farmers in loan review.

No abstract