发帖
雷达卡
DBI咨询：监管趋严，第三方风控行业将面临大洗牌
DBIConsulting: With stricter regulation, the third party risk control industrywill face a big reshuffle
2019-12-09 10:04
[概要] DBI咨询认为，金融及大数据行业政策、银行数字化转型、人工智能等智能化技术等因素将改变第三方风控行业格局，其中单纯的技术型、数据型的第三方风控公司将遇发展瓶颈。
[Abstract]DBI Consulting believes that financial and big data industry policies, digitaltransformation of banks, artificial intelligence and other intelligenttechnologies and other factors will change the pattern of third-party riskcontrol industry, among which the pure technology-based and data-basedthird-party risk control companies will encounter development bottlenecks.
作者：DBI咨询（德信商业智能）
风控管理是金融活动的核心，我国金融产业的发展表现出很强的信贷驱动属性，各类以新技术为支撑的智能化风控产品服务，已经成为信贷业务（包括传统信贷、互联网信贷）的重要支撑，随着金融业的转型升级、监管政策要求，风控也逐步在融合到其他类型业务当中。
自2012年移动互联网爆发以来，科技型金融取得了高速发展，但传统的风控方式和手段已经变得很难适应消费旺盛引发的信贷增长、长久以来被传统金融机构忽视的长尾用户的贷款需求，金融科技极大的促进了信贷智能化风控的发展。在《关于规范整顿“现金贷”业务的通知》出台之前，第三方智能风控企业超过500家。随着互联网消费金融业务的监管要求趋严（未来逐步压缩存量业务），一部分第三方风控企业开始退出市场或向技术服务转型，或进入企业级市场帮助金融机构服务服务企业客户。
据DBI统计，目前第三方智能化风控市场图谱如下（图1）
Accordingto DBI statistics, the current market map of third-party intelligent riskcontrol is as follows (Figure 1)
目前，第三方风控市场格局：银行金融科技子公司趋向于自给自足并有对外能力输出倾向；生态型互联网巨头利用生态优势与大型银行协作发力科技金融；输出型金融科技公司头部（约10家）占据主要市场份额（服务于股份制银行和大部分城商行）；自用型金融科技公司受监管影响业务收缩而开始涉足对外输出风控能力；个人征信市场则由央行牵头、巨头把握；企业征信市场刚刚起步。
At present, the market pattern ofthird-party risk control:
1) bankfintech subsidiaries tend to be self-sufficient and have the tendency to exportexternal capabilities;
2) EcologicalInternet giants make use of ecological advantages to cooperate with large banksto develop technology finance; The head of export fintech companies (about 10)occupies the main market share (serving joint-stock banks and most citycommercial banks);
3) Fintechcompanies for their own use began to set foot in the export risk controlability due to the contraction of their business under the influence ofregulation;
4) Thepersonal credit market is led by the central bank, the giant grasp; Theenterprise credit market has just started.
根据银监会、网信办等监管层意见，金融科技的实质是金融，关系金融业务发展的稳定性和可持续性，正在被纳入到监管范围。与此同时，为了更好的发挥金融科技的应用能力和推动作用，国家层面出台多项政策，来指导金融科技的顶层规划、战略部署和标准制定。
Accordingto the opinions of the CBRC, the Cyberspace Administration of China and otherregulators, the essence of fintech is finance, which is related to thestability and sustainability of the development of financial business, and is being included in the scope ofsupervision. At the same time, in order to better play the application abilityand promoting role of fintech, the national level has issued a number ofpolicies to guide the top-level planning, strategic deployment and standardformulation of fintech.
DBI认为，行业政策在主导金融科技行业的发展。在大力倡导普惠金融的今天，未来2-3年风控行业的发展将主要看与银行业的融合情况。目前存在三大因素在影响第三方风控行业未来格局，分别是金融及大数据行业政策、银行数字化转型、人工智能等智能化技术。
DBIbelieves that industry policy is driving the development of the fintechindustry. Today, the development of the risk control industry in the next 2-3 years will mainly depend on theintegration with the banking industry. Atpresent, there are three major factors affecting the future pattern of thethird-party risk control industry, namely, financial and big data industrypolicies, digital transformation of banks, and intelligent technologies such asartificial intelligence.
（I）Financial andbig data industry policies
政策监管主要体现在：其一，对金融业务不合规、不符合发展要求的监管；其二，对金融行业大数据，尤其是涉及个人隐私的信用类数据的监管。
Policysupervision is mainly reflected in the following aspects: First, supervision offinancial business that is not in compliance with the requirements ofdevelopment; Second, the supervision of big data in the financial industry,especially credit data involving personal privacy.
政策促进主要体现在：其一，促进金融科技与传统金融机构业务融合；其二，促进金融科技与传统监管体系融合；其三，促进金融科技的顶层规划、行业标准制定、体系建设等，如《金融科技（FinTech）发展规划（2019-2021年）》。
Policypromotion is mainly reflected in the following aspects: First, to promote thebusiness integration of financial technology and traditional financialinstitutions; Second, to promote the integration of fintech and traditionalregulatory systems; Third, promote the top-level planning, industry standardsand system construction of FinTech, such as the Fintech Development Plan(2019-2021).
在此类因素影响下，与不合规金融业务发生深度绑定的、涉及违规采集和使用隐私数据的风控技术方将首选被淘汰，符合监管要求、符合金融科技规划标准、与金融机构协同良性发展的风控技术方将得到长足发展机遇。
Under the influence of such factors, risk control technologists thatare deeply bound to non-compliant financial businesses and involve illegalcollection and use of privacy data will be the first choice to be eliminated, while risk control technologists that meet regulatoryrequirements and fintech planning standards and develop healthily incoordination with financial institutions will get considerable developmentopportunities.
(2)Digital transformation of banks
银行数字化转型
为适应金融业新阶段的发展，银行数字化转型在银行业已经成为共识。银行的风控能力与其数字化水平息息相关，其智能化风控发展可分为三个阶段：初级阶段、过渡阶段、高级阶段。
Inorder to adapt to the development of the new stage of financial industry, the digital transformation of banks has become aconsensus in the banking industry. The risk control ability of a bank isclosely related to its digitalization level, and the development of itsintelligent risk control can be divided into three stages: the primary stage,the transition stage and the advanced stage.
初级阶段：银行处于数字化转型初期，业务和管理体系尚未完全数字化，此时以传统风控为主（通过传统评分卡模型和规则引擎等强特征进行风险评分），以智能风控为辅，第三方风控技术方以帮扶改良为主。目前，银行业正处在此阶段。
Primary stage: Banks are inthe early stage of digital transformation, and their business and managementsystems have not yet been fully digitized. At thistime, traditional risk control is the priority (risk scoring is conductedthrough strong features such as traditional scorecard model and rule engine),supplemented by intelligent risk control, and thethird-party risk control technology is mainly to help and improve. The bankingindustry is now at this stage.
过渡阶段：银行的数字化转型进入实质性阶段，基本数字化体系已经建立，在部分业务层面已经实现了智能化并有了一定量的数据积累，智能风控模型得以初步完善，但精度不足、覆盖面不广，此时传统风控和智能风控处于交叉使用阶段。
Transition: the digital transformation of Banks has comeinto a substantial stage, basic digital system has been established, on thepart of the business level has achieved intelligent and had accumulated acertain amount of data, the intelligent risk control model to preliminaryimprovement, but lack of accuracy, coverage is not wide, the traditionalcontrol and intelligent risk control in the stage of cross use.
高级阶段：银行的数字化转型已较为成功，基本实现全面数字化和大部分智能化，无论运营机制、组织架构、数据体系，均较为完善，此时风控流程基本依赖智能风控体系，且智能风控体系已经融入到全业务全流程当中，在出现异常时才需要人力介入。
Advanced stage: the digital transformation of bank has been moresuccessful, basic realize comprehensive digitization and the most intelligent,regardless of the operating mechanism, the organizational structure, the datasystem, and each is perfect, the risk control process basic rely on intelligentrisk control system, risk control and intelligent system has been integratedinto the whole business of the whole process, in the abnormal need humanintervention.
在此类因素影响下，单纯以数据交易或技术交易为业务模式的第三方风控企业，很难与银行形成协同化发展。尤其是在银行开展数字化转型的初期，相比风控能力，银行更青睐能带来业绩和用户增长的生态型伙伴。
Underthe influence of such factors, it is difficult for third-party risk controlenterprises with data trading or technology trading as their business model todevelop synergistically with banks. Especially inthe early stage of digital transformation, banks prefer ecological partnersthat can bring performance and user growth compared with risk control ability.
(3)Artificial intelligence and other intelligent technologies
人工智能等智能化技术
目前，由于数据规模和质量问题，通过人工智能手段训练出的风控模型存在诸多瓶颈：
Atpresent, due to data scale and quality problems, risk control models trained byartificial intelligence methods have many bottlenecks:
其一，大多数数据源来自场景金融，主要针对个人消费信贷领域，且数据维度较为单一、风险相关性较弱、数据可移植性差，缺乏中小企业数据，由此类数据训练的风控模型在实际意义上并不可靠。
First,most of the data sources are from scenario finance, mainly aimed at the fieldof personal consumer credit, and the data dimension is relatively single, therisk correlation is weak, the data portability is poor, the lack of small andmedium-sized enterprise data, the risk control model trained by such data isnot reliable in the practical sense.
其二，数据源的数据质量不可控，尤其在造假严重的互联网电商领域，低质量数据、数据噪音对风控模型的训练存在极大干扰，极易造成模型过拟甚至错误。
Secondly,the data quality of data sources is uncontrollable, especially in the field ofInternet e-commerce where fraud is serious. Low-quality data and data noisehave great interference to the training of risk control model, which is easy tocause over-fitting or even errors of the model.
在此类因素影响下，缺乏稳定可靠的数据源则无法训练出好的模型，缺乏核心技术则无法形成有竞争力的产品，缺此两样的风控技术企业很难有立足之本——风险识别能力，数据中介、纯技术供应商均存在被边缘化的可能。
Under the influence of such factors, without stable and reliabledata sources, a good model cannot be trained; without core technology, acompetitive product cannot be formed; without these two kinds of risk controltechnology enterprises can hardly have a foothold -- risk identificationability. Data intermediaries and pure technology suppliers are likely to bemarginalized.
BDM_10.zip
(42.49 MB, 需要: 99 个论坛币)
BDM_11.zip
(41.95 MB, 需要: 99 个论坛币)
BDM_12.zip
(32.82 MB, 需要: 99 个论坛币)
BDM_13.zip
(37.73 MB, 需要: 99 个论坛币)
BDM_14.zip
(42.92 MB, 需要: 99 个论坛币)
BDM_15.zip
(35.15 MB, 需要: 99 个论坛币)
BDM_16.zip
(23.5 MB, 需要: 99 个论坛币)
BDM_17.zip
(50.66 MB, 需要: 99 个论坛币)
提升卡
置顶卡
沉默卡
变色卡
抢沙发
千斤顶
显身卡
京公网安备 11010802022788号
