| After experiencing the development of financial electronation and internet finance,Fintech has gradually stepped into the mature stage of development,especially with the rapid development of big data,artificial intelligence,and other financial technologies,Fintech has become an emerging industry with high activity in the financial market.With the help of emerging technologies,fintech enterprises continue to expand the boundary of traditional financial businesses and derive many new financial businesses and management models,which has obvious technology spillover and impact effects on commercial banks.One of the main ways for commercial banks to generate credit risk is credit business,while the development of financial technology has greatly affected the level of credit risk management of banks,the credit business developed by financial technology enterprises has also greatly crowded out the similar business market of banks.Therefore,this paper intends to measure the development degree of financial technology from the perspective of macro fintech development and micro-fintech enterprises,measure the credit risk level of banks from the perspective of bank internal control,and further analyze the impact of fintech development on the credit risk of commercial banks on this basis.However,under the background of the development of emerging technologies such as big data,bank risk management and its impact have shown new characteristics,with more diversified influencing variables,more complex measurement models,and more lack of effective measurement methods for the development degree of fintech and credit risk of commercial banks.Based on the research on the connotation,essence and development mechanism of financial technology,this paper conducts research on a number of issues such as the development degree of financial technology,the credit risk measurement of commercial banks and the dual impact relationship,aiming to break through the previous research paradigm,construct a measurement model from the perspective of machine learning,and analyze the impact relationship from the perspective of dual effect,in order to put forward feasible suggestions for the win-win development of commercial banks’ integration of financial technology.The paper attempts to answer the following questions:(1)Fintech is an emerging industry developed in recent years,which is controversial due to its flexible mechanism but lack of service boundary supervision.The basis of fintech research is to answer the questions of what is connotation and essence of Fintech,how is the relationship between fintech and internet finance,and clarify the evolution mechanism and influence ways of fintech to provide theoretical support for subsequent research.(2)How to construct the fintech development index.The existing literatures usually measure the development index based on social financial technology but do not measure the development index of commercial banks’ internal investment and application of Fintech.As the main force of traditional finance,the measurement of the application level of Fintech and the development index of commercial banks is more important.(3)How to measure the development degree of financing and lending fintech enterprises and their impact on bank credit risk when they compete with similar businesses of commercial banks.At present,due to the lack of a priori indicators for the measurement of the development degree of fintech enterprises,it is a problem of unsupervision,especially in the empirical research of large-sample,high-dimensional transaction data,and the general method is powerless.How to design unsupervised machine learning algorithms to measure the development of enterprises,especially to solve the problems of high dimension,poor timeliness and low accuracy,is an urgent problem to be solved in the measurement of the development degree of financial technology enterprises.(4)In the new technology period,the credit risk measurement methods of commercial banks have undergone significant changes.In the context of the development of fintech,bank credit risk is more complex,and it is urgent to adopt machine learning and other artificial intelligence methods to cope with the new changes in credit risk,such as high dimension,complex nonlinear problems.Although some literatures have adopted machine learning methods for measurement,there are still some problems:First,the research objects are usually listed banks or international banks,and lack of small and medium-sized banks,which is not generic and universal.Second,the application of machine learning models is only to improve and compare various models without parameter tuning,as it ignores the parameter tuning and selection of the typicality of the comparison model,so it has a poor effect in dealing with complex nonlinear problems.Third,the early warning signals are not found and identified according to the feature importance,and the effect of early warning supervision is not ideal.(5)Commercial banks have not only been affected by the impact of fintech from external non-financial institutions,but also under pressure from the development of fintech within banks.The use of the Fintech Development Index to measure its dual effect on the credit risk of commercial banks is also a question that this paper strives to answer.At the same time,from a micro perspective,it is also an issue that needs to be explored in this paper to analyze the impact of the development degree of financing and lending fintech enterprises on the credit risk level of similar business of banks.In order to answer the above questions,this paper follows the following ideas to carry out the research:First,it clarifies the connotation,essence and transmission of financial technology and internet finance.Second,the external social fintech development index and the internal development degree of fintech development index of commercial banks are constructed respectively,and the dimensions,processes,methods and results of the development index construction are introduced in detail.Third,based on the idea of parallel computing,the feature selection technology and collaborative particle swarm optimization are used to measure the development degree of typical financial technology enterprises in financing and lending,which solves the problems of poor timeliness and low accuracy of existing models,and also provides a measurement basis for subsequent measurement of the influence relationship between such enterprises and bank credit risk.Fourth,the credit risk level of commercial banks is measured from their own risk control ability of credit business,and a variety of machine learning models are used for comparison and screening,while paying attention to the parameter setting and tuning of the model,and early warning signals are found according to the importance of features in further research,so as to provide important research ideas for pre-loan early warning and the solution of regulatory problems.Finally,on the basis of the research on the development of financial technology and the measurement of credit risk of commercial banks,the dual impact effect of financial technology development on the credit risk of commercial banks is analyzed,and the impact effect on the credit risk of commercial banks is analyzed based on the development degree of typical financial technology enterprises measured above.Based on the above research ideas,the main research conclusions obtained in this paper are as follows:First,clarify the definition of fintech.Fintech refers to all enterprises or departments that specialize in using modern technology to improve the efficiency of financial services and innovate financial businesses.The important carrier of fintech is fintech enterprise.The essence of fintech development is the development brought by technological progress and business innovation of fintech enterprises.Second,it measures the fintech development index within banks.(1)The fintech development index of joint-stock commercial banks is the highest and the growth rate of the development index is the fastest.The fintech development index of state-owned banks has developed rapidly since 2017 and has reached a similar level to joint-stock banks after increasing investment in recent years.Compared with state-owned banks and joint-stock banks,the overall fintech development index of city banks is lower and the growth rate is slower.(2)From the bank’s internal financial input level,degree of combination of science and technology,business innovation ability,cognition,four dimensions of the segment,significantly better than that of state-owned Banks and joint-stock banks city bank,including the integration of state-owned banks level higher than the city bank,investment level,business innovation ability,and cognition is lower than the city bank.Third,the new method is used to measure the development degree of typical fintech enterprises and the credit risk of commercial banks,which can solve the problems of high dimension,large sample and non linearity,and improve the accuracy and operation efficiency of the model.(1)Based on the multi-source and complexity of lending transaction data,the development degree of fintech enterprises is designed by using feature selection and collaborative particle swarm optimization,and it is found that the accuracy of the improved model is 93.1%,which is 7.1%higher than the traditional CPSO algorithm and the convergence speed is increased by 61.1%.It is found that China’s listed financial technology enterprises have the highest development degree,the best asset quality and capital liquidity,and have strong comprehensive strength.The transaction characteristics that have a greater impact on the development of financing and lending fintech enterprises are,in order,transaction volume,proportion of the top ten borrowers’collection amount,operating time,and average borrowing period.(2)The credit risk measurement of commercial banks is based on the comparison of multiple models,and it is found that the accuracy of random forest after parameter adjustment reaches 93.5%,which is 8.6%higher than that of the traditional model.The measurement analysis shows that the credit risk control level of commercial banks in China shows an overall upward trend,among which the credit risk control level of joint-stock banks and urban commercial banks with more flexible mechanisms increases the fastest.Using the optimized model to calculate the relative importance,it is concluded that the synthetic factors of six indicators,such as the bank’s return on net assets,total asset growth rate,and deposit-loan ratio,constitute the early warning signal of credit risk,and achieve a better actual early warning effect.Fourth,the dual effect from the development of the internal and external Fintech of commercial banks have a significant negative correlation with the level of credit risk control of commercial banks.The dual effect generally aggravates the credit risk level of commercial banks and reduces the level of bank credit risk control.The dual effect generally increase the risk level of commercial banks and reduce the business performance of banks.Commercial bank limited "boost" the effect of internal financial technology,external financial technology "yoke" more obvious,the bank internal financial application advantages of science and technology is not obvious,short-term development efforts have not been fully coped with external non-financial institutions of the impact of the financial technology enterprises,the disincentive effect of external fintech still plays an important effect.At the same time,the technology spillovers generated by financing and lending fintech enterprises generally have a positive promoting effect on the credit risk control level of commercial banks,but this effect is weak,while it has a significant inhibiting effect on the credit risk control level of city banks.Finally,in terms of heterogeneity effects.(1)Due to the flexible mechanism,joint-stock financial technology enterprises have strong adaptability to financial technology business and better absorption capacity of financial technology,showing greater enterprise development potential.The development of state-owned enterprises is relatively stable,but they have the characteristics of slow adjustment and transformation,and the improvement of product and service efficiency is not obvious.Private enterprises are more affected by the impact of similar business and technological development of external enterprises,and their development capacity is reduced.(2)The use of financial technology by commercial banks can significantly reduce the level of risk and improve the ability of banks to control credit risk,but this role has little effect in small and medium-sized banks such as city banks.State-owned and joint-stock banks have a significant crowding out effect on city banks.Empirical studies have shown that fintech can help large banks upgrade traditional lending technologies and improve the credit supply capacity of small and micro enterprises,which may have an impact on the advantageous business of small and medium-sized banks,and put more pressure on small and medium-sized banks such as city banks.Compared with the existing literature,the possible innovations in this paper are as follows:First,the index system of the fintech development Index has been enriched.Most of the existing literature measures the external social fintech development index,and lacks research on the internal fintech application and development index of commercial banks.This paper comprehensively analyzes the application of financial technology in commercial banks and the development of financial technology in external society,and refines the index system of the two levels,considering a more novel perspective,which provides a strong basis for subsequent research on dual effects.Second,this paper extracts early warning signals for the credit risk of commercial banks.This paper innovatively constructs the index combination of early warning signals by calculating the characteristic importance and verifies the correlation between financial indicators,early warning signals,and risk levels,which plays an important role in early warning supervision for commercial banks to manage credit risks before the loan and avoid risks in time.Third,This paper systematically studies the dual effects of the development of financial science and technology on the credit risk of commercial banks.Most of the existing literature analyzes the impact of financial technology development on commercial banks from the external society.This paper analyzes the comprehensive impact of commercial banks on the credit risk of commercial banks from the two levels of internal fintech application and external social fintech development,which is more comprehensive than the previous literature that studies the impact relationship from a single perspective.Fourthly,the measurement model is designed innovatively based on big data.The complexity of financial markets in the new technology period makes the traditional measurement methods face new changes.This paper designs a new measurement model under the thinking of big data breaks through the hypothesis testing and statistical limitations of traditional econometric models,and comprehensively considers the characteristics of a large sample,high dimension,and complex nonlinearity of data.Machine learning(including decision tree,artificial neural network,random forest,Adaboost ensemble learning)and artificial intelligence(collaborative particle swarm optimization algorithm)methods are applied for empirical analysis in the part of fintech development and bank credit risk model measurement,and feature selection techniques in data mining technology is used for dimensionality reduction.The measurement effect is better than that of the traditional model. |
