【作者】 王惠；

【导师】 王静；

【作者基本信息】 西北农林科技大学 ， 农村金融， 2019， 博士

【摘要】 农村金融联结作为新的农贷模式逐渐演变成许多发展中国家重要的农村融资战略。该模式将正规金融机构的资金优势和非正规金融机构的信息优势联结起来,可降低农村金融市场上的信息不对称和交易成本,扩大农户的贷款规模,有效缓解农户的融资问题。而农村金融联结机制运行的成败取决于维系其系统稳定的信用关系。在农村金融联结过程中,农户作为农村金融联结的基础主体,由于组织松散、缺乏可抵押财产及违约机会成本低的特征,其信用风险具有极大的不确定性。随着农村金融联结主体间关联交易的增加,如果具有较大且复杂信用风险的参与者违约,信用风险会随着时间的推移传导至其他参与主体,演化为关联信用风险的传染,进而直接影响农村金融联结系统的稳定性,甚至殃及整个农村金融市场。因此,研究农村金融联结信用风险及其演化,预防和管控农户信用风险的发生,决定着农村金融联结的成败,对保证农村金融联结模式的稳定性及其可持续发展,具有重要的研究意义。本文在分析现有农村金融联结模式及信用风险的基础上,以最普遍的农村金融联结模式“正规金融机构+乡村中介+农户”(Legal Financial institutions+Rural Intermediaries+Farmers,简称LIF)为研究对象,从信用的本质出发,基于信用行为的视角,通过理论建模、动态模拟仿真和实证研究相结合的方法,研究金融联结的信用风险演化及其机理,为推动农村金融联结制度的可持续发展,完善农村金融联结机制提供理论决策依据,促进我国农村金融市场的良性发展。本论文的研究内容和结论主要包括以下几方面:(1)农村金融联结运行特征、信用风险及相互影响机理分析。本章就农村金融联结运行特征和农村金融联结过程中所涉及的风险及其形成和传递机理进行深入分析。然后基于有限理性假设建立了LIF模式的农村金融联结前后贷款演化博弈模型,并对其参与主体的信用行为相互影响机理进行分析。结果表明:(1)农户的决策选择主要取决于抵押资产的金额和贷款所能获得的收益以及信用社的监管力度。(2)银行的决策主要取决于农户选择“还款”策略的比率;(3)合作社使得银行以及农户之间的决策行为发生了积极的变化。(4)通过提高农户归还贷款的初始概率,增加银行的贷款收益来增加银行贷款的可得性,同时提高抵押资产数额,套现比率以及合作社的惩罚来提高农民的还款意愿。(2)基于改进型模糊聚类无权值原理对陕西省阎良区的955户农户信用风险进行微观实证评价研究。结果表明:该地区农户信用风险处于中等偏上水平,信用风险较低,有利于贷款的发放和农户的融资;但农户的家庭结构特征和经营情况评价结果较低,建议提高农户的教育水平,加强农户的技术培训,加快土地规模化程度,进一步提高农户的信用等级。(3)前景理论视角下,农户的偏好对农村金融联结机制下农户的信用风险影响研究。理论分析与实证检验的结果表明:(1)提高农户的收益率和对相对收益的预期,短期内会使得农户的履约意愿降低。但分别超过12%和23%的临界值时,会提高农户的履约意愿。降低农户对相对收益的重视程度也会提高农户的履约意愿。(2)增加农户对相关农户履约比例的预期和对相关农户违约比例的重视程度,将会提高农户的履约意愿。相关农户的履约比例超过约43%时,将使得农户的履约意愿加速上升。降低相关农户的违约比例将会提高农户的履约意愿,当违约比例超过约52%时,农户的履约意愿将加速下降。(3)增加农户对监管和惩罚力度的预期,会使得农户的履约意愿上升,当监管达到一个最大值时,增加监管和惩罚力度的预期反而会使得农户的履约意愿下降;惩罚比例与农户的履约意愿呈近似正相关关系,当惩罚超过收益的8%时,农户的履约意愿将会加速上升。适当的监管将有助于提高农户的履约意愿,监管过度则会引起相反的效果。(4)建立农户的信贷决策经验权重魅力值学习模型,探讨农村金融联结机制中农户的学习和模仿行为对其信贷决策行为的影响。理论分析与实证检验的结果表明:(1)EWA学习模型能够促使农户做出履约决策,提高整体的信用等级,随着博弈次数的增加,农户的还款策略选择概率呈总体上升趋势;(2)随着合作社规模的逐渐增大,农户决策选择概率出现分化,组内差异较小,组间差异较大,农户履约策略选择比率整体呈下降趋势,违约策略选择比率整体呈上升趋势,且在研究中不存在风险控制的最佳合作社规模,应根据自身抗风险能力结合规模效应理论进行合作社规模的选择;(3)在农村金融联结模式下,来自金融机构和合作社内部的惩罚,可以提高履约策略选择比例,且随着惩罚力度的增加而增加,合作社内部惩罚相较金融机构的惩罚,效果更快且更显著。(5)构建农村金融联结小世界复杂系统网络与传染病模型,对其关联信用风险及传染机制进行研究。理论分析与实证检验的结果表明:(1)资产关联有助于分担农村金融联结参与主体之间的关联信用风险,且资产关联越紧密关联信用风险发生的概率越低;(2)及时救助能够降低信用风险爆发后关联信用风险传染的概率;(3)资产关联比的存在能够有效降低LIF网络中关联信用风险爆发后“非健康”主体的密度;(4)“非健康”主体的密度随着传染延迟时间的延长而增大并最终趋于稳定;(5)“非健康”主体的密度值始终小于1,不会使得所有LIF网络中的主体同时感染关联信用风险。(6)农村金融联结信用风险防控优化设计。综合前文研究,建立农村金融联结信用风险防控优化模型,提出了末位淘汰机制,利用实际数据进行分析验证。结果表明:(1)末位淘汰机制对LIF农村金融联结模式下的农户整体信用水平有一定的提高效果,但作用力随着时间推移而减弱,需要其他措施辅助;(2)随着合作社规模的增加,农户整体信用水平增长速度也有所减缓,但稳定性提高;(3)农户整体信用水平随关联成员数的增加而呈现先增大后减小的趋势,当农户的关联人数K=6时,存在一个局部最优关联成员数。(4)利用陕西阎良区某农村金融联结模式下的合作社中甜瓜种植农户的数据,求得最优信贷效益下农户的违约概率为p_i~t≈0.34,最优信贷效益下甜瓜种植农户的规模为N≈59,最优信贷效益下政府的贴息率为r≈3%,最优信贷效益下违约农户的合作社内部惩罚为s≈4,最优信贷效益下银行对违约农户给与的惩罚为K≈8。本研究的创新点主要体现在:第一,创新性的使用改进型模糊聚类无权值原理,建立农户信用风险评价模型。该方法的应用可以分阶段分层次进行评价,无需依靠指标权重,且评价指标体系可以根据地区和情况的不同进行调整,能够更灵活、全面的对农户信用风险进行评价,实用性和可操作性强。第二,引入前景理论和学习模型对农村金融联结中农户的信念和偏好进行研究。研究得出,增加农户对相关农户履约比例的预期和对相关农户违约比例的重视程度,将会提高农户的履约意愿,降低相关农户的违约比例将会提高农户的履约意愿。第三,采用小世界网络与传染病模型相结合的方法研究农村金融联结各主体的关联信用风险传染过程和作用机理。研究得出,资产关联有助于分担农村金融联结参与主体之间的关联信用风险,对受传染的主体给与及时的救助能够降低信用风险进一步传染的概率,且在关联信用风险爆发后始终存在健康主体。更多还原

【Abstract】 As a new agricultural loan model,the rural financial linkage(RFL)has gradually evolved into an important rural financing strategy for many developing countries.The RFL combines the financial advantages of formal financial institutions with the information advantages of informal financial institutions,which can reduce information asymmetry and transaction costs in rural financial markets,expand the scale of farmers’loans,and effectively alleviate farmers’financing problems.However,the success or failure of the RFL depends on the credit relationship that maintains its system stability.As the basic participant in RFL,farmers’credit behavior has great uncertainty due to loose organization,lack of collateralized property and low cost of default opportunities.With the increase of connected transactions between RFL connection entities,if participants with large and complex credit risks default,the credit risk will be transmitted to other participating entities over time,and will evolve into the transmission of related credit risks,which will directly affect the stability of the RFL system has even affected the entire rural financial market.Therefore,studying the credit risk and its evolution of rural financial connections,preventing and controlling the occurrence of farmers’credit risks,determines the success or failure of RFL,and has important research significance for ensuring the stability and sustainable development of RFL models.Based on the analysis of existing RFL models and their credit risks,this paper takes the most common RFL model“Legal Financial institutions+Rural Intermediaries+Farmers”(LIF)as the research object.Based on the essence of credit and the perspective of credit behavior,the paper studies the evolution of credit risk and its mechanism,and promotes the realization and sustainability of RFL system through the combination of theoretical modeling,dynamic simulation and empirical research.And the evolution of credit risk and its mechanism in RFL was studied,and provide theoretical basis for promoting the sustainable development of the rural financial linkage system and improving the rural financial linkage mechanism,and promote the sound development of China’s rural financial market.The research conclusions of the thesis are as follows:(1)Analysis of RFL operation characteristics,credit risk and mutual influence mechanism.This chapter conducts an in-depth analysis of the operating characteristics and the risks involved in the process of RFL connections,as well as their formation and transmission mechanisms.Then,based on the assumption of bounded rationality,a game model of loan evolution before and after the rural financial connection of the LIF model is established,and the interaction mechanism of the credit behaviors of its participants is analyzed.The results show that:(1)Farmers’decision-making choices mainly depend on the amount of mortgage assets,the income that loans can obtain,and the supervision of credit cooperatives.(2)The bank’s decision mainly depends on the ratio of farmers’choice of"repayment"strategy.(3)Cooperatives have made positive changes in the decision-making behavior between banks and farmers.(4)Increase the initial probability of farmers returning loans,increase the bank’s loan income to increase the availability of bank loans,and increase the amount of mortgage assets,cash ratio and penalties of cooperatives to increase farmers’willingness to repay.(1)The theoretical basis of the evolution mechanism of RFLcredit risk.This chapter analyzes the operational patterns of typical linkage cases in RFL.At present,the model of“formal financial institutions+rural intermediaries+farmers”is the most common mode of linkage in RFL.The linkage model of agricultural professional cooperatives as a linkage intermediary has a relatively broad development space.Therefore,this paper chooses this model as the research object.(2)Micro-empirical evaluation of the credit risk of 955 farmers in Yanliang District of Shaanxi Province based on the improved fuzzy clustering without weight principle.The results show that the credit risk of farmers in this area is at a moderately high level,and the credit risk is low,which is conducive to the issuance of loans and financing of farmers.Strengthen the technical training of farmers,speed up the scale of land,and further improve the credit rating of farmers.(3)From the perspective of prospect theory,the research on the influence of farmers’preferences on the credit behavior of farmers under the LIF financial linkage mechanism was performed.The results of theoretical analysis and empirical tests show that:(1)Raising the farmer’s rate of return and the expectation of relative income will reduce the farmers’willingness to perform in the short term.However,when the threshold values exceed 12%and 23%respectively,the farmers’willingness to perform will be increased.Reducing farmers’emphasis on relative returns will also increase farmers’willingness to perform.(2)Increasing the farmers’expectations of the relevant farmers’compliance ratio and the importance of the relevant farmers’default ratio will increase the farmers’willingness to perform.When the relevant farmers’compliance ratio exceeds about 43%,the farmers’willingness to fulfill will accelerate.Reducing the proportion of defaults of the relevant farmers will increase the farmers’willingness to perform.When the proportion of default exceeds 52%,the farmers’willingness to perform will accelerate.(3)Increasing farmers’expectations for supervision and punishment will increase the farmers’willingness to perform.When the supervision reaches a maximum,the reference point for increasing supervision and punishment will reduce the farmers’willingness to perform;the punishment ratio and the farmers’performance willingness is approximately positively correlated.When the penalty value exceeds 8%of the income,the farmers’willingness to perform will accelerate.Appropriate regulation will help raise farmers’willingness to perform,and over-regulation will have the opposite effect.(4)The credit value learning model of the credit decision-making experience of farmers was established,and the influence of the learning and imitation behavior of farmers in the LIF financial linkage mechanism on their credit decision-making was discussed.The results of theoretical analysis and empirical test show that:(1)EWA learning model can encourage farmers to make performance decisions and improve the overall credit rating.As the number of games increases,the farmers’repayment strategy selection probability shows an overall upward trend;(2)With the increasing size of cooperatives,the decision-making probability of farmers is different,the difference within the group is small,the difference between groups is large,the ratio of farmers’compliance strategy selection is decreasing,the ratio of default strategy selection is increasing,and there is no risk in the study.The size of the best cooperatives controlled is based on the size of the cooperatives based on their own anti-risk ability combined with the scale effect theory;(3)In the financial linkage mode,internal penalties from financial institutions and cooperatives can increase the proportion of compliance strategies and increase with the increase of punishment.The internal punishment of cooperatives is faster and more significant than that of financial institutions.(5)Based on the network of LIF financial linkages in the small world complex system and the infectious disease model,the associated credit risk and infection mechanism was studied.The results of theoretical analysis and empirical test show that:(1)Asset association helps to share the associated credit risk between financial participants,and the closer the asset association is,the lower the probability of occurrence of credit risk;(2)Timely rescue can reduce the probability of associated credit risk infection after the outbreak of credit risk;(3)The existence of asset correlation ratio can effectively reduce the density of“non-healthy”subjects after the outbreak of associated credit risk in the LIF network;(4)The density of“non-healthy”subjects increases with the delay of infection delay.Eventually it tends to be stable;(5)“Non-healthy”subjects have a density value of less than 1,which does not cause all subjects in the LIF network to infect the associated credit risk at the same time.(6)Optimized design of credit risk prevention and control in rural financial connection.Based on the previous research,an optimization model for credit risk prevention and control in rural financial linkages was established,and a last elimination mechanism was proposed,which was analyzed and verified using actual data.The results show that:(1)the last elimination mechanism has a certain effect on improving the overall credit level of rural households under the LIF rural financial connection model,but the force weakens with time and requires other measures to assist;The growth rate of the overall credit level has also slowed down,but the stability has improved;(3)The overall credit level of farmers has increased first and then decreased with the increase in the number of associated members.When the number of connected farmers is K=6,there is a local maximum.The number of best associated members.(4)Using data from melon-growing farmers in a cooperative under a rural financial connection model in Yanliang District,Shaanxi Province,the default probability of peasant households with optimal credit benefits is p_i~t≈0.34,and the scale of melon-growing farmers with optimal credit benefits is N≈59.Under the optimal credit benefit,the government’s discount rate is r≈3%.The cooperative penalty of the defaulting farmers under the optimal credit benefit is s≈4.Under the optimal credit benefit,the bank’s penalty for the defaulting farmers is K≈8..The innovations of this research are mainly reflected in:Firstly,the improved fuzzy clustering without weight principle was applied to establish a credit risk evaluation model for farmers innovatively.The application of this method can be evaluated in stages and levels,without relying on index weights,and the evaluation index system can be adjusted according to different regions and situations,which can more flexibly and comprehensively evaluate the credit risk of farmers.Strong.Secondly,the beliefs and preferences of farmers in RFL based on prospect theory and learning model was studied.The results showed that increasing the farmers’expectations of the relevant farmers’compliance ratio and the importance of the relevant farmers’default ratio will increase the farmers’willingness to perform,and reducing the proportion of defaults of the relevant farmers will increase the farmers’willingness to perform.Finally,the small world network and the infectious disease model are combined to study the process and mechanism of the related credit risk infection of RFL.The asset associations help to share the associated credit risk between financial participants and timely assistance to infected entities can reduce the probability of further spread of credit risk,and there is always health after the associated credit risk main body.更多还原