【作者】 李丹；

【导师】 李妍峰；

【作者基本信息】 西南交通大学 ， 管理科学与工程， 2023， 博士

【摘要】 商业银行效率评价是银行管理的一项重要活动,而识别投入产出指标角色是准确测度银行效率的关键环节。在现有银行效率评价研究中,关于如何确定存款的投入产出角色尚未达成共识。“生产法”将银行视作利用人、财、物等资源为顾客提供存款和贷款等服务的“生产者”,此时银行存款可以视为产出指标;而“中介法”将银行视为吸纳存款并发放贷款以实现盈利目标的金融“中介者”,此时银行存款可以视为投入指标。现有银行效率评价研究中,不同视角下所确定的银行存款角色不尽相同,从而导致“存款角色困境”问题。这种类似存款等需要识别其投入或产出角色的指标,通常称为“灵活指标(Flexible Measure)”。由于灵活指标的特性,如何在银行效率测度过程中合理地刻画灵活指标,成为公平有效地测算银行效率的重要前提。数据包络分析法(Data Envelopment Analysis,DEA)作为一种广泛应用于银行领域的数据驱动方法,不仅能有效处理“存款角色困境”问题,也能为改善决策单元效率以提升其相对竞争力提供决策支持。然而,现有研究并未考虑“不良贷款指标的影响”“所有制类型不同导致的技术异质性”“不良贷款跨期结转的动态特性”“银行内部两阶段运营结构”等现实情形下的银行效率评价以及对应的存款角色识别。存在灵活指标时,如何考虑非期望产出指标的影响、技术水平异质性、动态视角以及两阶段视角,分别构建公平合理的效率评价方法,并应用于上述不同情形下的银行效率评价以及存款角色识别是本文的主要研究问题。为此,本文分别提出了新的DEA模型,并应用于中国上市商业银行效率评价。本文主要的研究内容可概括如下:(1)基于银行“存款角色困境”问题,研究不良贷款指标对银行效率的影响。不良贷款作为银行系统中的非期望产出理应负向影响效率值。然而,在应用DEA方法过程中,由于待评价银行有更大机会选取能使自己处在有效前沿面的权重组合,导致部分银行的效率值出现不降反升的反直觉结果。基于此,本文第3章提出了考虑非期望产出指标影响的FMDEA方法,并将所提方法应用于中国上市商业银行效率评价。当考虑不良贷款指标影响时,大型银行比中小型银行受新冠疫情影响更大,且效率下降更为明显。最后,将提出的模型与未考虑非期望产出指标影响的模型以及固定存款角色为投入或产出的DEA模型进行比较分析,研究表明第3章提出的模型可以得到帕累托改进的效率结果,进而提升评价结果的有效性和公平性。(2)基于银行“存款角色困境”问题,考虑不同所有制类型的银行所面临的外部运营环境不同而导致的技术水平异质性。本文第4章从内部指标角色异质性视角扩展到外部运营环境异质性视角,提出了考虑技术异质性的FMDEA方法。提出的方法可以深入探究银行相对于群组前沿的效率和相对于共同前沿的效率及对应的存款角色,并将无效率情况分解为管理无效率和技术无效率,探究无效率来源。最后,将提出的模型与固定角色的模型进行比较分析,验证所提模型的合理性和有效性。(3)基于银行“存款角色困境”问题,考虑不良贷款跨期结转特性,本文第5章从银行的静态效率研究扩展到动态效率研究,提出了考虑跨期要素的动态FMDEA方法。随后,将提出的模型应用于中国银行业动态效率评价,并以待评价银行改进潜力最大化为目标确定存款角色。最后,将提出的方法与考虑灵活指标的静态效率评价模型、固定角色的动态效率评价模型进行比较分析,验证了该方法的有效性和公平性。(4)基于银行“存款角色困境”问题,本文第6章从银行“黑箱”系统视角扩展到两阶段网络结构视角。针对银行系统中包含增值子阶段以及盈利子阶段的序列两阶段结构,本文提出两阶段视角下的FMDEA方法。提出的模型从多个视角考虑银行“存款角色困境”问题,分析银行的整体效率以及子阶段分解效率,探究银行系统的无效率来源。随后,将所提模型应用于中国银行业效率评价。最后,将所提模型与固定存款角色的模型进行比较分析,并对模型中的参数进行了灵敏度分析。从理论方面,针对灵活指标及非期望产出指标、技术异质性问题以及两阶段结构,提出了四种不同的FMDEA方法,研究银行效率及存款角色情况,以提升效率评价的有效性和公平性。从实践方面,考虑银行不良贷款指标影响、所有制类型异质性、跨期要素以及银行内部网络结构,对中国银行业进行效率评价研究。针对存在灵活指标的其他生产和服务系统(如考虑科研经费为灵活指标的高校运营系统),当系统中存在非期望产出、运营环境异质性、跨期结转项以及两阶段运营结构时,可应用本文提出的方法分析其效率以及灵活指标的角色情况,为管理者提供决策参考。更多还原

【Abstract】 Measuring bank efficiency is one of the important activities in banking operations,and the prior process is to clarify the input and output measures.The input/output role of other measures can be normally defined in banking industry,except for the deposit.In the production approach,banks are the“producers”that utilize the resources(like human resources,capital,and fixed assets)to provide the deposit and loan services for customers,and the deposit acts as an output.While in the intermediation approach,banks are the“intermediaries”that absorb deposits and make loans for profit,and the deposit plays as an input.From different perspectives,bank deposit can be treated as an input or an output,which is referred to the“deposit dilemma”in bank efficiency measurement.The variables that can act as either inputs or outputs are recognized as“flexible measures”.Because of this characteristic of flexible measure,it is paramount important to deal with flexible measures in efficiency estimation,so that individual banks can get more effective and equitable evaluation results.Data envelopment analysis(DEA)is a widely applied data-driven approach in the banking industry,as it can not only cope with“deposit dilemma”,but also improve the efficiencies of decision-making units,so as to improve their competitiveness.However,the prior studies neglect the impact of non-performing loans,the technological heterogeneity due to different bank ownerships,inter-temporal non-performing loans,and network operation system when measuring the bank efficiencies and classifying the role of deposits.Given above,the research question is:how to develop fair and effective evaluation methods to tackle with the issues of the impact of undesirable outputs,the technological heterogeneity,intertemporal carryovers and two-stage operation system respectively,so as to solve the“deposit dilemma”in the bank efficiency evaluation process practically.Therefore,this study proposes new DEA models to measure the bank efficiency while classifying the role of deposits,and then applies these approaches to Chinese banking industry.The contributions can be summarized as follows:Firstly,this study focuses on the impact of non-performing loans on bank efficiencies while coping with the“deposit dilemma”.Theoretically,incorporating the undesirable output is assumed to decrease the operation efficiency.However,the additional measure improves the possibilities for each bank to approach the efficient production frontier.Therefore,some of the banks may get the counter-intuitively increased efficiency scores after including the non-performing loans.To solve this problem,this study proposes a new DEA model considering flexible measures and the impact of undesirable outputs.The proposed model is applied to Chinese banking industry,and it shows that the larger banks,with more significant decrease in efficiencies,suffer more than the medium-and-small banks when the impact of the undesirable loans is considered.Finally,the proposed model is compared with those with fixed deposit role strategies,which indicates that the proposed model is superior in providing the Pareto-improved results,therefore,it improves the effectiveness and fairness.Secondly,extending from bank system internal factors(flexible measure)to the outside factors(operation environment)to ensure the fairness,this study addresses the technological heterogeneity that originates from differences in bank ownerships while coping with the“deposit dilemma”problem.This study proposes a meta-frontier DEA model with flexible measures to investigate the meta-frontier efficiency and group-frontier efficiency and the corresponding deposit roles of each bank.Meanwhile,the inefficiency score is decomposed into managerial inefficiency and technological inefficiency,which helps to clarify the inefficiency sources.Finally,the proposed model is compared with fixed deposit role models to verify the rationality and effectiveness.Thirdly,coping with the“deposit dilemma”problem,this study extends the static bank efficiency evaluation case to its corresponding dynamic one.As the inter-temporal non-performing loans can be regarded as a carryover from one period to the next,this study proposes a dynamic DEA model with flexible measures.The proposed model is empirically applied to Chinese banking industry,and each bank under examination classifies the role of deposits based on the potential for improvement.Finally,the proposed approach is validated by comparison analysis with the static model considering flexible measure and the dynamic model with fixed deposit role.Fourthly,the research on bank efficiency is extended from the“black box”system to a clarified two-stage serial operation system concerning the“deposit dilemma”.Since the bank system incorporates the value-added subprocess and the profit-earning subprocess,this study proposes a two-stage DEA model with flexible measures to solve the“deposit dilemma”.From multiple perspectives,the proposed model provides an alternative solution to“deposit dilemma”,while measuring the overall efficiency and sub-stage efficiencies of each bank,which enables the managers to find the sources of inefficiency.Then,this approach is applied to Chinese banking industry.Finally,the comparison analysis and sensitive analysis validate the effectiveness and fairness of the proposed model.Theoretically,this study proposes four new DEA approach to address the flexible measures,while considering undesirable outputs,technical heterogeneity,and two-stage network structure in efficiency evaluation.The effectiveness and fairness of efficiency evaluations is enhanced by classifying the role of deposits endogeneously for each bank.Practically,this study measures the performance of the Chinese banking industry,considering the impact of non-performing loans,ownership heterogeneity,inter-temporal carryovers,and internal two-stage structure.The proposed methods can be applied to other production and service systems with flexible measures,such as universities that consider“research funding”as a flexible measure.These methods can analyze system efficiency and the role of flexible measures in the presence of undesirable outputs,technological heterogeneity,inter-temporal carryovers,and two-stage operational structures,providing management insights for decision makers.更多还原