Symfony Profiler

Request

GET Parameters

Key	Value
page	144

POST Parameters

No POST parameters

Request Attributes

Key	Value
_controller	AppBundle\Controller\UserController::showAction
_h5	false
_permission	[]
_route	user_show
_route_params	[_permission => [], _h5 => false, id => 3]
id	3

Cookies

Key	Value
PHPSESSID	pbu0n5gcie1frc74hi99gbn9n3

Request Headers

Header	Value
accept	/
accept-encoding	gzip, br, zstd, deflate
ali-cdn-forward-client-ip	216.73.216.182
ali-cdn-real-ip	216.73.216.182
ali-swift-log-host	www.cfrisk.org
ali-swift-stat-host	www.cfrisk.org
cookie	PHPSESSID=pbu0n5gcie1frc74hi99gbn9n3
eagleeye-traceid	b4a3921517613349755812564e
host	www.cfrisk.org
user-agent	Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; ClaudeBot/1.0; +claudebot@anthropic.com)
via	cn7174.l1, kunlun6.cn7174, l2cn8734.l2, cache48.l2cn8734
x-client-scheme	https
x-forwarded-for	216.73.216.182
x-php-ob-level	1

Request Content

Request content not available (it was retrieved as a resource).

Server Parameters

Key	Value
CONTEXT_DOCUMENT_ROOT	/www/wwwroot/www.cfrisk.org/web
CONTEXT_PREFIX
DOCUMENT_ROOT	/www/wwwroot/www.cfrisk.org/web
FCGI_ROLE	RESPONDER
GATEWAY_INTERFACE	CGI/1.1
HOME	/home/www
HTTPS	on
HTTP_ACCEPT	/
HTTP_ACCEPT_ENCODING	gzip, br, zstd, deflate
HTTP_ALI_CDN_FORWARD_CLIENT_IP	216.73.216.182
HTTP_ALI_CDN_REAL_IP	216.73.216.182
HTTP_ALI_SWIFT_LOG_HOST	www.cfrisk.org
HTTP_ALI_SWIFT_STAT_HOST	www.cfrisk.org
HTTP_COOKIE	PHPSESSID=pbu0n5gcie1frc74hi99gbn9n3
HTTP_EAGLEEYE_TRACEID	b4a3921517613349755812564e
HTTP_HOST	www.cfrisk.org
HTTP_USER_AGENT	Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; ClaudeBot/1.0; +claudebot@anthropic.com)
HTTP_VIA	cn7174.l1, kunlun6.cn7174, l2cn8734.l2, cache48.l2cn8734
HTTP_X_CLIENT_SCHEME	https
HTTP_X_FORWARDED_FOR	216.73.216.182
LD_LIBRARY_PATH	/www/server/apache/lib
PATH	/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin
PATH_INFO	/user/3
PATH_TRANSLATED	redirect:/app.php/3
PHP_SELF	/app_dev.php/user/3
QUERY_STRING	page=144
REMOTE_ADDR	117.92.155.196
REMOTE_PORT	49896
REQUEST_METHOD	GET
REQUEST_SCHEME	https
REQUEST_TIME	1761334975
REQUEST_TIME_FLOAT	1761334975.6916
REQUEST_URI	/app_dev.php/user/3?page=144
SCRIPT_FILENAME	/www/wwwroot/www.cfrisk.org/web/app_dev.php
SCRIPT_NAME	/app_dev.php
SERVER_ADDR	172.17.183.217
SERVER_ADMIN	webmasterexample.com
SERVER_NAME	www.cfrisk.org
SERVER_PORT	443
SERVER_PROTOCOL	HTTP/1.1
SERVER_SIGNATURE	<address>Apache Server at www.cfrisk.org Port 443</address>
SERVER_SOFTWARE	Apache
USER	www
proxy-nokeepalive	1

Response

Response Headers

Header	Value
cache-control	no-cache
content-type	text/html; charset=UTF-8
x-debug-token	e7919d

Session

Session Metadata

Key	Value
Created	Sat, 25 Oct 25 03:42:36 +0800
Last used	Sat, 25 Oct 25 03:42:50 +0800
Lifetime	0

Session Attributes

Attribute	Value
_csrf/site	lzAR83iWOFHq-L2Z6bACm4mK-XiIbheVSumgWe45Bz4

Flashes

No flash messages were created.

Sub Requests 2

AppBundle:User:headerBlock (token = 46c66f)

Key	Value
_controller	AppBundle:User:headerBlock
_format	html
_locale	zh_CN
user	[ id => 3, email => cpjlikephp@126.com, verifiedMobile => 18510341342, password => q141yZ92zOhdAlPt1SNgaqEgNS/wlJplWZtIcEz0Gjo=, salt => avbo9tcei8wgo44c8sgww4ckokgcowc, payPassword => , payPasswordSalt => , locale => null, uri => , nickname => admin, title => 111, tags => , type => default, point => 0, coin => 0, smallAvatar => public://default/2019/05-23/095559fd639b055006.png, mediumAvatar => public://default/2019/05-23/095559fd47cb521902.png, largeAvatar => public://default/2019/05-23/095559fd1e4f111012.png, emailVerified => 1, setup => 1, roles => [ 0 => ROLE_USER, 1 => ROLE_EXPERT, 2 => ROLE_TEACHER, 3 => ROLE_SUPER_ADMIN ], promoted => 0, promotedSeq => 0, promotedTime => 0, locked => 0, lockDeadline => 0, consecutivePasswordErrorTimes => 0, lastPasswordFailTime => 0, loginTime => 1761301457, loginIp => 120.221.146.37, loginSessionId => sb20acfottmie1oqh47mvjr364, approvalTime => 1561624383, approvalStatus => approved, newMessageNum => 1, newNotificationNum => 0, createdIp => , createdTime => 1552732491, updatedTime => 1761301457, inviteCode => 0N6Z5, orgId => 1, orgCode => 1., registeredWay => , distributorToken => , uuid => a5245103b11ff3d6877f148a4bdf0caf99792ad5, passwordInit => 1, faceRegistered => 0, registerVisitId => , wechatmember => 0, member => 2, virtual_coin => 155, order_sn => 2022051911493929683, chances => 0, remark => , mobileSessionId => evuj8o49l0njcce9oi7kvrst72, course_chance => 1, login_limt => 1, token => null, org_name => , penName => , truename => 常鹏杰, idcard => 41052119920410815X, gender => male, iam => , birthday => null, city => , mobile => , qq => 738576547, signature => 滴答滴答滴答滴答滴答滴答多大, about => 1111 , company => 北京天弈方圆, job => PHP工程师, school => , class => , weibo => , weixin => 技术部, isQQPublic => 1, isWeixinPublic => 1, isWeiboPublic => 1, site => , intField1 => null, intField2 => null, intField3 => null, intField4 => null, intField5 => null, dateField1 => null, dateField2 => null, dateField3 => null, dateField4 => null, dateField5 => null, floatField1 => null, floatField2 => null, floatField3 => null, floatField4 => null, floatField5 => null, varcharField1 => , varcharField2 => , varcharField3 => , varcharField4 => , varcharField5 => , varcharField6 => , varcharField7 => , varcharField8 => , varcharField9 => , varcharField10 => , textField1 => , textField2 => , textField3 => , textField4 => , textField5 => , textField6 => , textField7 => , textField8 => , textField9 => , textField10 => ]

AppBundle:Course/CourseSet:courseSetsBlock (token = d3d88e)

Key	Value
_controller	AppBundle:Course/CourseSet:courseSetsBlock
_format	html
_locale	zh_CN
courseSets	[ 0 => [ id => 4048, type => message, title => 徐凡：开放式，银行的未来, subtitle => , tags => [], categoryId => 15, cate_child => 0, cate_grand => 0, summary => null, goals => [], audiences => [], isVip => 0, cover => [], status => published, creator => 3, createdTime => 1606521600, updatedTime => 1611832551, serializeMode => none, ratingNum => 0, rating => 0, noteNum => 0, studentNum => 0, hotSeq => 0, recommended => 0, recommendedSeq => 0, recommendedTime => 0, orgId => 1, orgCode => 1., discountId => 0, discount => 10.00, hitNum => 0, maxRate => 100, materialNum => 0, parentId => 0, locked => 0, minCoursePrice => 0.00, maxCoursePrice => 0.00, teacherIds => [ 0 => 3 ], defaultCourseId => 4048, sourceUrl => https://www.yunzhan365.com/73818446.html, body => null, weight => 0, expert_id => null, classroomId => 0, length => , startTime => 1606521600, floorId => 0 ], 1 => [ id => 4049, type => message, title => 江李蕰盈：在COVID期间管理风险, subtitle => , tags => [], categoryId => 15, cate_child => 0, cate_grand => 0, summary => null, goals => [], audiences => [], isVip => 0, cover => [], status => published, creator => 3, createdTime => 1606521600, updatedTime => 1611832571, serializeMode => none, ratingNum => 0, rating => 0, noteNum => 0, studentNum => 0, hotSeq => 0, recommended => 0, recommendedSeq => 0, recommendedTime => 0, orgId => 1, orgCode => 1., discountId => 0, discount => 10.00, hitNum => 0, maxRate => 100, materialNum => 0, parentId => 0, locked => 0, minCoursePrice => 0.00, maxCoursePrice => 0.00, teacherIds => [ 0 => 3 ], defaultCourseId => 4049, sourceUrl => https://www.yunzhan365.com/34449597.html, body => null, weight => 0, expert_id => null, classroomId => 0, length => , startTime => 1606521600, floorId => 0 ], 2 => [ id => 1238, type => message, title => 陈凯丰：美国大选结果对下一阶段的经济分析, subtitle => 陈凯丰,美国汇盛金融公司首席经济学家, tags => [], categoryId => 47, cate_child => 0, cate_grand => 0, summary => null, goals => [], audiences => [], isVip => 0, cover => [], status => published, creator => 3, createdTime => 1606521600, updatedTime => 1614844531, serializeMode => none, ratingNum => 0, rating => 0, noteNum => 0, studentNum => 0, hotSeq => 0, recommended => 0, recommendedSeq => 0, recommendedTime => 0, orgId => 1, orgCode => 1., discountId => 0, discount => 10.00, hitNum => 3, maxRate => 100, materialNum => 0, parentId => 0, locked => 0, minCoursePrice => 0.00, maxCoursePrice => 0.00, teacherIds => [ 0 => 3 ], defaultCourseId => 1238, sourceUrl => https://www.yunzhan365.com/82428158.html, body => null, weight => 0, expert_id => null, classroomId => 0, length => , startTime => 1606521600, floorId => 0 ], 3 => [ id => 1239, type => message, title => 李松民：区域与行业风险分析——基于客户视角, subtitle => , tags => [ 0 => 340 ], categoryId => 15, cate_child => 0, cate_grand => 0, summary => null, goals => [], audiences => [], isVip => 0, cover => [], status => published, creator => 3, createdTime => 1606521600, updatedTime => 1611396617, serializeMode => none, ratingNum => 0, rating => 0, noteNum => 0, studentNum => 0, hotSeq => 0, recommended => 0, recommendedSeq => 0, recommendedTime => 0, orgId => 1, orgCode => 1., discountId => 0, discount => 10.00, hitNum => 1, maxRate => 100, materialNum => 0, parentId => 0, locked => 0, minCoursePrice => 0.00, maxCoursePrice => 0.00, teacherIds => [ 0 => 3 ], defaultCourseId => 1239, sourceUrl => https://www.yunzhan365.com/90543716.html, body => null, weight => 0, expert_id => null, classroomId => 0, length => , startTime => 1606521600, floorId => 0 ], 4 => [ id => 4067, type => information, title => 石智勇：大数据时代信用风险管理变革的思考, subtitle => , tags => [], categoryId => 16, cate_child => 0, cate_grand => 0, summary => null, goals => [], audiences => [], isVip => 0, cover => [], status => published, creator => 3, createdTime => 1606521600, updatedTime => 1617073800, serializeMode => none, ratingNum => 0, rating => 0, noteNum => 0, studentNum => 0, hotSeq => 0, recommended => 0, recommendedSeq => 0, recommendedTime => 0, orgId => 1, orgCode => 1., discountId => 0, discount => 10.00, hitNum => 0, maxRate => 100, materialNum => 0, parentId => 0, locked => 0, minCoursePrice => 0.00, maxCoursePrice => 0.00, teacherIds => [ 0 => 3 ], defaultCourseId => 4067, sourceUrl => null, body => <h1 style="text-align: center;"><strong>大数据时代信用风险管理变革的思考</strong></h1> <p style="text-align: center;">石智勇农业银行总行风险管理部副总经理</p> <p style="text-align: justify;"> </p> <p style="text-align: justify;">      信息是人类活动踪迹的诚实记录，随着互联网飞速发展，我们可以获得的信息量也在飞速增长。大数据引发了一场商业银行经营理念和管理方式的重大变革，开启了一个全新的时代。在大数据时代，风险管理者拥有浩瀚的数据，在拥有更多机会的同时，也面临着很多全新的挑战。</p> <h2 style="text-align: justify;"><strong>一、风险管理工作面临新的特征和挑战</strong></h2> <h4 style="text-align: justify;"><strong>（一）信用风险管理面临新的形势</strong></h4> <p style="text-align: justify;">      大数据时代主要有三方面的特征，一是数据规模十分庞大，其增长速度也非常快，往往呈现几何级的增长；二是处理速度快，相较以往的数据处理速度大数据的处理速度有大幅度提升，数据处理往往结合了一些云计算技术，基本上可以满足实时交互的需要；三是管理的自动化程度高，越来越多的算法、规则、模型运用到风险管理领域，极大地提升了风险管理的效率。随着我们进入大数据时代，风险管理工作面临四个新的形势。</p> <p style="text-align: justify;">      一是客户真实性识别难度增加。“了解你的客户（KYC）”是所有业务风险管控的前提和基础，在大数据时代，我们业务可以突破地域、时空限制，服务对象由“实体的人”变成了“虚拟化的数字”，客户真实性识别难度增加。一些业务失去了“面对面”进行客户身份识别和客户资料收集的基础。在实际工作中除了防范传统的身份证件伪冒风险外，还需要具备防范密码试错及指纹、声纹、虹膜、人脸识别等伪冒欺诈手段的技术，如密码撞库、三维头套等。在“开放银行”模式下，银行经常面临“客户”的“客户”，需要穿透识别客户真实性。</p> <p style="text-align: justify;">      二是增加了“信息被篡改”的风险。交易信息的真实性和一致性是风险控制的关键，线上业务主要通过网络、电话、邮件、短信、微信等渠道办理，虽然客户能够随时随地快捷地享受银行服务，但业务数据在提交、返回和交互的过程中存在被篡改、被替换的风险。传统线下服务主要通过物理网点渠道办理，业务数据在客户和银行之间“点对点”直接传输，银行只需关注客户资料的真实性，无需考虑信息数据在传递过程中被篡改、被替换的风险，而线上业务需要将防信息篡改作为重点。线上业务实时化、智能化主要依托银行跨系统交互，交易信息的一致性校验是关键。我曾观察到某行的线上业务已经发生过类似交易报文信息被恶意篡改的风险事件。</p> <p style="text-align: justify;">      三是风控重点发生了重大变化。风险管理非常依赖经验的积累，在大数据时代，需要将“人”的风险管理经验提炼成规则嵌入系统，并全面梳理“机器”的风险和新的内外部风险因素。在大数据时代，主要基于数据和机器指令自动进行业务处理，业务逻辑和业务规则更为复杂，需要防范规则、模型、系统的不完善的风险。人工智能并不能代替业务人员和风险管理人员的经验判断，大数据背景下，业务更多地采取线上运作，人的实时干预越来越弱。</p> <p style="text-align: justify;">      四是客户信息泄露风险日趋严峻。“大数据”时代客户隐私保护是难点，也是风险点。从内部看，大数据环境下，数据的流转经过前端采集、业务处理、大数据平台入库、集中存储等过程，都有可能导致数据大面积泄露。从外部看，大数据蕴含的巨大价值，吸引了网络黑客的频繁攻击。此外，中国银行业无论是在软硬件设施还是数据服务，都一定程度依赖国外厂商，为潜在第三方数据监听和数据泄露埋下风险隐患。例如最近圆通泄露40万条个人隐私信息，引发全民“安全”危机；2009年，黑客攻击了美国的“信用卡第三方支付付款处理器”的网络系统，导致包括万事达、美国运通、维萨等在内的4000多万信用卡客户数据被窃取。</p> <h4 style="text-align: justify;"><strong>（二）传统风险发生新的变化</strong></h4> <p style="text-align: justify;">      在传统的风险领域也发生了一系列变化。</p> <p style="text-align: justify;">      在信用风险领域，一是客户准入门槛降低。网络融资产品引入更多高风险客户群体，提高异地、长尾客户占比，主要是“轻担保、无抵押”的产品模式，缺少第二还款来源，造成信用风险水平上升，违约概率提高。二是信贷模型风险。网络融资产品利用大数据技术构建信贷模型，但建模数据尚未覆盖一个完整的信贷周期，且面临违约数据不足、数据噪音较大、内外部基础数据缺乏统一规范性、对借款人财务状况和偿债能力等关键变量分析不足等问题，同时模型适用的外部经济、社会环境变化迅速，模型准确性有待时间考验，往往是还没有来得及验证，所适用的客户对象和数据特征已经发生了改变，易产生信贷模型风险。三是过度授信风险。银行、消费金融公司、互金平台、金融科技公司大举进入消费贷市场，消费贷不能及时体现在征信信息中，行业内还缺乏有效、稳定的数据共享机制，多头授信现象较为普遍，增加授信风险敞口。</p> <p style="text-align: justify;">      在操作风险领域，一是交易真实性风险。在远程开户方式下，整形、三维头套、证件伪造等技术的发展提高了有效识别客户的难度，而且无法判断客户是否存在被胁迫等情况，对于开户人身份真实性及开户意愿真实性难以有效识别。二是内部人员操作风险。这个风险发生了转变，从具体的业务操作风险，转变为模型、规则设计的风险，模型、参数的维护风险。三是外部欺诈风险。通过网络虚拟空间发动资金勒索、盗窃、网络欺诈攻击的便利性和可能性日益增大，欺诈犯罪活动逐步形成了制售欺诈工具、盗取客户信息、实施社交工程欺诈、盗取/转移资金等黑色产业链，呈现团伙化、专业化、犯罪线索难以追踪等特点，而欺诈手法一旦被验证可行，将快速复制传播并不断变异。</p> <p style="text-align: justify;">      最后，线上业务的基础要素存在合法合规性问题。在线合同、电子签名等技术方式的法律效力还存在不够明确之处，在产生法律纠纷时可能不利于银行。</p> <h2 style="text-align: justify;"><strong>二、在变革中提升信用风险管理能力</strong></h2> <p style="text-align: justify;">      在新的形势下，对于如何应对挑战，提高风险管理能力，我归纳总结了五个方面的建议。</p> <p style="text-align: justify;">      第一，要从宏观上把握风险，清晰定位目标客群。在大数据时代应用模型和数据进行客户筛选时，不能陷入客户微观数据的“迷宫”，忽视行业等宏观层面的判断。要充分考虑行业等宏观风险因素，重视客户群体选择。在传统的风险识别框架中，行业维度有行业信贷政策、行业限额等，地区维度有信贷差异授权、区域限额等，客户维度有客户评级模型、违约模型、行为模型等。在现行大数据技术下风险识别框架中，我们更加依托微观“指标”评价客户，使用大数据从“近端”和“细节”审视客户风险，没有从“远端”和“总体”按照一定的逻辑和机制对客户进行综合评价和选择。</p> <p style="text-align: justify;">比较典型的就是钢贸行业的信贷危机，也为多家银行带来了深刻的教训。我们需要根据大数据时代下风险多变的特点，建立行业分析和监测机制，通过综合分析宏观经济行业运行的外部因素，以及资产质量变化的因素，前瞻性地判断行业客户的变化趋势。</p> <p style="text-align: justify;">      第二，需及时更新调整风险模型。大数据模式下，风控策略的及时更迭至关重要。传统信贷业务为“模型+专家”模式来管理风险，专家会根据风险形势的变化相机抉择、及时调整；大数据模式下，信贷业务更多依靠模型，需针对最新情况及时调整更新风险防控模型和策略。在大数据时代，商机稍纵即逝，风险管控的时机也是稍纵即逝，动态的风险捕捉能力是风险管理应对大数据时代的核心竞争力。随着大数据技术的发展，银行业的风险管理策略体系应该向动态自动化方向发展，通过整合风险管理各模块的监控维度及指标，提高风险管理策略的系统化、动态化和实效性，及时扫描发现各领域的潜在问题，进而及时调整相应策略，提高风险管理策略的针对性和有效性。</p> <p style="text-align: justify;">      第三，需要建立新型信贷责权对等机制，这是风险管理的基石。2008年金融危机最重要的教训之一就是设计复杂金融产品的人对风险不负责任，在传统信贷管理机制下，我们建立了信贷审查、审批人员责、权、利对等机制。大数据模式下的信贷业务决策由模型和系统自动完成，管理责任出现新的变化。</p> <p style="text-align: justify;">      第四，需要做好模型风险的管控。在提高模型区分能力的同时，还要建立完善的模型风险监控体系，逐步形成精细化的模型分级管理机制，以及高效精准的模型风险预测体系，并建立预警的快速应对机制。要开展压力测试，通过压力测试识别风险、判断风险变化的方向和趋势、变化风险的影响，提前对重大风险进行预警。此外，还要开展周期性的滚动预测，将短中长期的预测相结合，前瞻性地预测风险，进行风险指标的预警监控、差异化分析，提高预警的精度。</p> <p style="text-align: justify;">      第五，需做好大数据与传统信贷的结合。数据分析是风险管理的基础，数据常常面临数据质量和噪声问题，人们也往往更多地关注与自己直觉判断相一致的信息，这些都会造成我们理解数据、运用数据的偏差，而大数据技术的应用有可能无限放大这些偏差。开启数据支持性决策的新时代，无疑是对数据的理解能力、运用能力以及数据和传统的结合能力提出了更高的要求。深耕数据，精细化、实时化、动态化地利用数据构建模型和风控策略，实施匹配的机制，应该成为未来发展的方向。</p> <p style="text-align: justify;">      总而言之，大数据技术的发展和运用将会改变商业银行信息获取、分析和运用的渠道和机制，为信息化的风险管理创造基础条件。大数据时代的风险管理者应该加快创新步伐，全面提升自身的风险管理能力，妥善应对风险管理挑战，提高风险管理的有效性，促进商业银行业务的稳健可持续发展。</p> <p style="text-align: right;">来源：2020（第十六届）中国金融风险经理年度总论坛</p> , weight => 0, expert_id => 0, classroomId => 0, length => , startTime => 1606521600, floorId => 0 ], 5 => [ id => 4068, type => information, title => 江李蕰盈：在COVID期间管理风险, subtitle => 江李蕰盈,瑞士信贷（美国）（CreditSuisseUSAIHC）董事总经理, tags => [], categoryId => 43, cate_child => 0, cate_grand => 0, summary => null, goals => [], audiences => [], isVip => 0, cover => [], status => published, creator => 3, createdTime => 1606521600, updatedTime => 1612142588, serializeMode => none, ratingNum => 0, rating => 0, noteNum => 0, studentNum => 0, hotSeq => 0, recommended => 0, recommendedSeq => 0, recommendedTime => 0, orgId => 1, orgCode => 1., discountId => 0, discount => 10.00, hitNum => 0, maxRate => 100, materialNum => 0, parentId => 0, locked => 0, minCoursePrice => 0.00, maxCoursePrice => 0.00, teacherIds => [ 0 => 3 ], defaultCourseId => 4068, sourceUrl => null, body => <h1 style="text-align: center;"><strong>在COVID期间管理风险</strong></h1> <p style="text-align: center;">江李蕰盈  前瑞士信贷（美国）董事总经理、首席风险官</p> <p style="text-align: center;"> </p> <p style="text-align: justify;">      我今天讲的题目就是在COVID期间银行业是怎么样管理风险的，因为在之前我是在瑞士信贷做数据方面以及交易前台的工作，现在就想跟大家分享一下，在COVID期间，压力测试方面，我们都有哪些额外的挑战，还有怎么样的应对策略。 </p> <h2 style="text-align: justify;"><strong>一、新冠疫情大流行给压力测试带来的额外挑战</strong></h2> <p style="text-align: justify;">      可能大家都知道自从疫情以来，世界各地，包括美国，陆续实行了城市封锁政策，银行业专门做风险管理的人士也一直都在努力把疫情和宏观经济，结合起来，然后考虑有哪些部门受到打击比较大，比如说在美国的食品加工业，受到的打击比较大，那么银行如果对这些行业有信贷关系的话，银行的影响也会比较大，反而其他的行业比如说Amazon，他已经有一个网上订货的平台，反而没有受到重大的负面影响，甚至还使得它有更多盈利和销售的机会。</p> <p style="text-align: justify;">      像我们专门做风险管理的人员，就会把对流行病的测试跟宏观经济结合起来，看如果疫情发生会对哪些部门有比较大的影响，具体是通过什么样的渠道来对GDP产生影响，然后把新的利益回报率以及高收益的一些指数重新加起来，然后对每一个行业、每一个借贷人、每一个借贷公司来说，看他的资产负债表怎么样，看收入怎么样，成本怎么样。一些医疗层面的测试也会加入对宏观经济的预测上。还有货币政策也都加入进来，并且考虑一些特殊的政策，比如说在房地产行业，诸如一些其他的投资人，房贷美、房利美的资产负债表会受什么样的影响，会不会又像2008年的次贷危机期间，再向财政部要钱，所有需要把这些因素叠加起来。大家也看得出，这就会对压力测试带来了很大的挑战。</p> <p style="text-align: justify;">      尽管CCAR提交时间是2020年4月，但2020年初的CCAR情景并未考虑大流行。美联储要求2020年9月重新提交CCAR，以特别考虑大流行的情景。同时，气候风险和贸易战给压力测试和风险管理带来了额外挑战。</p> <h2 style="text-align: justify;"><strong>二、大流行预测与经济预测相结合：情景设计的新元素</strong></h2> <p style="text-align: justify;">      在疫情刚刚开始的时候，有一些学校就跟一些银行临时组合进行疫情和经济形势相关的预测，其实就是预测美国第一季、第二季 GDP的走势，然后把这些数据整理以后，利用这些大家预测的数据，把它放到模型里面，看看他们所关心指标的预测结果。</p> <p style="text-align: justify;">      有的公司开发了大流行导航，并与传染病专家合作，为何时可以实现适当的阈值免疫力提供了可能的方案，并分享了影响时机和成功可能性的关键杠杆——障碍和促进剂。比如流行病专家合作，我也参加了几次流行病专家、律师以及风险管理专家的会议，大家把流行病专家的数据和预测结果拿出来考虑不同地区、不同的行业的具体影响结果，从而想一些对策去应对这些特殊的风险产生的特殊影响。</p> <p style="text-align: justify;">      疫苗的研发作为恢复正常情况的重要指标，是预测工作的核心。基于疫苗研发时间的预估以及有效性的判定，预测的经济恢复时间从2021年4月到2021年12月不等。</p> <h2 style="text-align: justify;"><strong>三、前瞻性风险评估比任何时候都重要</strong></h2> <p style="text-align: justify;">      我归纳一下，在特殊的时期是怎么样把流行病专家、政府健康指导，与传统的银行风险测试或者压力测试，以及行业的特定因素相结合的。然后通过利润、亏损、资本等指标进行预测，即预测中我们应当采取怎么样的具体工作。</p> <p style="text-align: justify;">      银行做的流行病学和政府健康指导的测试，可以分得很细。就封闭地区而言，可以看被封锁的区域、封锁的持续时间以及受影响的特定部门。宏观影响方面，可以看财政政策、货币政策的刺激力度。行业方面，可以看特定行业受到的特定影响。 </p> <p style="text-align: justify;">      针对房地产贷款而言，超过几百万的抵押贷款被允许延期12个月。但是，退出FB的结果是什么？房主还能保住房子吗？银行和GSE的资产负债表上会有不良贷款吗？首先延期12个月偿还主要看12个月以后会有什么样的这种政策，或者有人申请，但是不可能不付房贷。因此金融机构目前的措施就非常重要，因为贷款延期会直接影响银行贷款账户的盈利或亏损。比如当这个政策已经结束，但是债务人还是没有能力偿还，其结果会如何，是否还能保住房子都具有不确定性。但目前拜登刚刚当选，对美国中下低收入的人群是比较保护的。如果是说债务人继续延期偿还房贷，那会对美国的经济有多么大的影响都需要目前的风险管理人进行考量。</p> <h2 style="text-align: justify;"><strong>四、完善的风险管理实践和商机</strong></h2> <p style="text-align: justify;">      在医疗方面的因素加入之后，可能在4-6月份价格可能变成0或者是负值，大家对一些定向的模型都产生了怀疑，就需要运用可以产生零值或负值的调整定价模型，从而使模型能够在外力作用的异常情况下继续正常运行。就压力测试而言，当前的压力测试模型需要更多考虑一个非常严重的场景，包括自然灾害，COVID激增案例，经济封锁，政治不确定性等因素。就银行内控而言，银行内部的风险评估次数愈发频繁，因为大家本来以为不会发生的事情还会经常发生。同时，据说很多民间银行也经常遭受黑客攻击，也就是说网络的安全也要加强。另外，公司试图变得更加灵活，并支持员工在家工作。现在很多银行也都认识到了，在家工作不但可以节省房租，还发现员工的创造力其实并不减，因为其工时有所延长。</p> <p style="text-align: justify;">      面对这样的情况，银行需要识别不断增长的需求和机会，包括高科技（zoom），交付，人口较少地区的需求房屋（单户住宅）等，银行抓住这样的机会，并需要积极在这方面进行投资。</p> <p style="text-align: right;">来源：2020（第十六届）中国金融风险经理年度总论坛</p> , weight => 0, expert_id => 0, classroomId => 0, length => , startTime => 1606521600, floorId => 0 ], 6 => [ id => 1766, type => video, title => 陈公越：智能移动银行业务的技术壁垒与实践, subtitle => , tags => [], categoryId => 10, cate_child => 0, cate_grand => 0, summary => null, goals => [], audiences => [], isVip => 0, cover => [], status => published, creator => 3, createdTime => 1606521600, updatedTime => 1611372567, serializeMode => none, ratingNum => 0, rating => 0, noteNum => 0, studentNum => 0, hotSeq => 0, recommended => 0, recommendedSeq => 0, recommendedTime => 0, orgId => 1, orgCode => 1., discountId => 0, discount => 10.00, hitNum => 0, maxRate => 100, materialNum => 0, parentId => 0, locked => 0, minCoursePrice => 0.00, maxCoursePrice => 0.00, teacherIds => [ 0 => 3 ], defaultCourseId => 1766, sourceUrl => f8d60c6fbcc74e7d9ba6a87950238a23, body => null, weight => 0, expert_id => null, classroomId => 0, length => 615.5306, startTime => 1606521600, floorId => 0 ], 7 => [ id => 1768, type => video, title => 陈公越：智能移动银行业务的服务需求, subtitle => , tags => [], categoryId => 10, cate_child => 0, cate_grand => 0, summary => null, goals => [], audiences => [], isVip => 0, cover => [], status => published, creator => 3, createdTime => 1606521600, updatedTime => 1611372554, serializeMode => none, ratingNum => 0, rating => 0, noteNum => 0, studentNum => 0, hotSeq => 0, recommended => 0, recommendedSeq => 0, recommendedTime => 0, orgId => 1, orgCode => 1., discountId => 0, discount => 10.00, hitNum => 2, maxRate => 100, materialNum => 0, parentId => 0, locked => 0, minCoursePrice => 0.00, maxCoursePrice => 0.00, teacherIds => [ 0 => 3 ], defaultCourseId => 1768, sourceUrl => 0c4c4b749fd746e1a0c95641f7011245, body => null, weight => 0, expert_id => null, classroomId => 0, length => 318.08, startTime => 1606521600, floorId => 0 ], 8 => [ id => 1769, type => video, title => 陈公越：银行普惠金融业务——智能移动银行, subtitle => , tags => [], categoryId => 10, cate_child => 0, cate_grand => 0, summary => null, goals => [], audiences => [], isVip => 0, cover => [ large => public://course/2021/01-29/19153020646f040358.jpg, middle => public://course/2021/01-29/191530207bce985466.jpg, small => public://course/2021/01-29/1915302091ed876486.jpg ], status => published, creator => 3, createdTime => 1606521600, updatedTime => 1619314103, serializeMode => none, ratingNum => 0, rating => 0, noteNum => 0, studentNum => 0, hotSeq => 0, recommended => 0, recommendedSeq => 0, recommendedTime => 0, orgId => 1, orgCode => 1., discountId => 0, discount => 10.00, hitNum => 0, maxRate => 100, materialNum => 0, parentId => 0, locked => 0, minCoursePrice => 0.00, maxCoursePrice => 0.00, teacherIds => [ 0 => 3 ], defaultCourseId => 1769, sourceUrl => 07dbab46cdbd402aa715632d104a313d, body => null, weight => 0, expert_id => null, classroomId => 0, length => 498, startTime => 1606521600, floorId => 0 ], 9 => [ id => 8169, type => video, title => 叶大清：数字时代的零售金融风险管理的几点思考（完整版视频）, subtitle => , tags => [], categoryId => 10, cate_child => 0, cate_grand => 0, summary => null, goals => [], audiences => [], isVip => 0, cover => [], status => published, creator => 3, createdTime => 1606521600, updatedTime => 1734328153, serializeMode => none, ratingNum => 0, rating => 0, noteNum => 0, studentNum => 0, hotSeq => 0, recommended => 0, recommendedSeq => 0, recommendedTime => 0, orgId => 1, orgCode => 1., discountId => 0, discount => 10.00, hitNum => 0, maxRate => 100, materialNum => 0, parentId => 0, locked => 0, minCoursePrice => 0.00, maxCoursePrice => 0.00, teacherIds => [ 0 => 3 ], defaultCourseId => 8169, sourceUrl => a4fdf487839e4e48853f2f489cb2ae28, body => null, weight => 0, expert_id => null, classroomId => 0, length => 1827, startTime => 1606521600, floorId => 0 ], 10 => [ id => 1770, type => video, title => 陈公越：移动科技与智能获客的重点——普惠金融, subtitle => , tags => [], categoryId => 10, cate_child => 0, cate_grand => 0, summary => null, goals => [], audiences => [], isVip => 0, cover => [], status => published, creator => 3, createdTime => 1606521600, updatedTime => 1611372537, serializeMode => none, ratingNum => 0, rating => 0, noteNum => 0, studentNum => 0, hotSeq => 0, recommended => 0, recommendedSeq => 0, recommendedTime => 0, orgId => 1, orgCode => 1., discountId => 0, discount => 10.00, hitNum => 1, maxRate => 100, materialNum => 0, parentId => 0, locked => 0, minCoursePrice => 0.00, maxCoursePrice => 0.00, teacherIds => [ 0 => 3 ], defaultCourseId => 1770, sourceUrl => 62680ecf6eb5471f945c4230fff90192, body => null, weight => 0, expert_id => null, classroomId => 0, length => 255.3173, startTime => 1606521600, floorId => 0 ], 11 => [ id => 1771, type => message, title => 陈公越：银行智能移动技术与信贷业务发展, subtitle => , tags => [], categoryId => 15, cate_child => 0, cate_grand => 0, summary => null, goals => [], audiences => [], isVip => 0, cover => [], status => published, creator => 3, createdTime => 1606521600, updatedTime => 1620977203, serializeMode => none, ratingNum => 0, rating => 0, noteNum => 0, studentNum => 0, hotSeq => 0, recommended => 0, recommendedSeq => 0, recommendedTime => 0, orgId => 1, orgCode => 1., discountId => 0, discount => 10.00, hitNum => 1, maxRate => 100, materialNum => 0, parentId => 0, locked => 0, minCoursePrice => 0.00, maxCoursePrice => 0.00, teacherIds => [ 0 => 3 ], defaultCourseId => 1771, sourceUrl => https://www.yunzhan365.com/80968089.html, body => null, weight => 0, expert_id => null, classroomId => 0, length => , startTime => 1606521600, floorId => 0 ], 12 => [ id => 4345, type => video, title => 袁先智：cafe全息风险评估体系背景及框架, subtitle => , tags => [], categoryId => 10, cate_child => 0, cate_grand => 0, summary => null, goals => [], audiences => [], isVip => 0, cover => [], status => published, creator => 3, createdTime => 1606521600, updatedTime => 1617073785, serializeMode => none, ratingNum => 0, rating => 0, noteNum => 0, studentNum => 0, hotSeq => 0, recommended => 0, recommendedSeq => 0, recommendedTime => 0, orgId => 1, orgCode => 1., discountId => 0, discount => 10.00, hitNum => 0, maxRate => 100, materialNum => 0, parentId => 0, locked => 0, minCoursePrice => 0.00, maxCoursePrice => 0.00, teacherIds => [ 0 => 3 ], defaultCourseId => 4345, sourceUrl => c9a0bfcdc00c49afae53d96b60d27ee5, body => null, weight => 0, expert_id => null, classroomId => 0, length => 451.7333, startTime => 1606521600, floorId => 0 ], 13 => [ id => 4346, type => video, title => 袁先智：案例分享, subtitle => , tags => [], categoryId => 10, cate_child => 0, cate_grand => 0, summary => null, goals => [], audiences => [], isVip => 0, cover => [], status => published, creator => 3, createdTime => 1606521600, updatedTime => 1617073785, serializeMode => none, ratingNum => 0, rating => 0, noteNum => 0, studentNum => 0, hotSeq => 0, recommended => 0, recommendedSeq => 0, recommendedTime => 0, orgId => 1, orgCode => 1., discountId => 0, discount => 10.00, hitNum => 0, maxRate => 100, materialNum => 0, parentId => 0, locked => 0, minCoursePrice => 0.00, maxCoursePrice => 0.00, teacherIds => [ 0 => 3 ], defaultCourseId => 4346, sourceUrl => 00ca065f11704fe5a793d62f7c319530, body => null, weight => 0, expert_id => null, classroomId => 0, length => 332.608, startTime => 1606521600, floorId => 0 ], 14 => [ id => 4347, type => video, title => 袁先智：欺诈特征提取, subtitle => , tags => [], categoryId => 10, cate_child => 0, cate_grand => 0, summary => null, goals => [], audiences => [], isVip => 0, cover => [], status => published, creator => 3, createdTime => 1606521600, updatedTime => 1617073785, serializeMode => none, ratingNum => 0, rating => 0, noteNum => 0, studentNum => 0, hotSeq => 0, recommended => 0, recommendedSeq => 0, recommendedTime => 0, orgId => 1, orgCode => 1., discountId => 0, discount => 10.00, hitNum => 0, maxRate => 100, materialNum => 0, parentId => 0, locked => 0, minCoursePrice => 0.00, maxCoursePrice => 0.00, teacherIds => [ 0 => 3 ], defaultCourseId => 4347, sourceUrl => 73638fd1cdb4493ebfaa377d51bb9e76, body => null, weight => 0, expert_id => null, classroomId => 0, length => 332.2667, startTime => 1606521600, floorId => 0 ], 15 => [ id => 3580, type => information, title => 徐凡：开放式，银行的未来, subtitle => , tags => [ 0 => 441, 1 => 434 ], categoryId => 16, cate_child => 0, cate_grand => 0, summary => null, goals => [], audiences => [], isVip => 0, cover => [], status => published, creator => 3, createdTime => 1606521600, updatedTime => 1611825038, serializeMode => none, ratingNum => 0, rating => 0, noteNum => 0, studentNum => 0, hotSeq => 0, recommended => 0, recommendedSeq => 0, recommendedTime => 0, orgId => 1, orgCode => 1., discountId => 0, discount => 10.00, hitNum => 2, maxRate => 100, materialNum => 0, parentId => 0, locked => 0, minCoursePrice => 0.00, maxCoursePrice => 0.00, teacherIds => [ 0 => 3 ], defaultCourseId => 3580, sourceUrl => null, body => <h1 style="text-align: center;"><strong>开放式，银行的未来</strong></h1> <p style="text-align: center;">徐凡盛宝银行大中华区首席执行官、全球执委会委员</p> <p style="text-align: justify;"> </p> <p style="text-align: justify;">      盛宝银行做开放式银行的时间较长，海外目前大概有一百多家银行使用我们的平台来做业务。各个公司做金融转型都面临很大的困难以及激烈的竞争，我们公司为如何进行转型提供了一个不同的思路。我认为无论哪家公司都应该考虑是使用完全自己投入开发的模式，还是尽可能地使用一些可以借鉴的模式。</p> <h2 style="text-align: justify;"><strong>一、开放式银行的核心</strong></h2> <p style="text-align: justify;">      开放式银行的核心实际上是合作。一家金融机构做数字化转型需要的技术投入是非常昂贵的，而技术升级换代几乎是不间断的，所以投入也是持续的。此外，虽然客户的需求不断变化，但是各大证券公司以及银行等机构在业务方面很多技术是通用的，没有本质区别。从工业化的角度来看，专业化的技术协作是现代社会化的方向，比如汽车、飞机、手机等，几乎没有一家公司能够做到从头到尾进行整个产业链的生产，依然很有生命力，这是金融业应当借鉴的一件事情。</p> <p style="text-align: justify;">      开放是为了合作，合作动力来自于什么？聚焦价值的创造。这里有一个比较常用的例子，就是咖啡的价值链。从种植咖啡豆到收割、销售、烘培、浸泡，咖啡的价值链很长。耗费时间最长的阶段是种植和收割，但咖啡豆本身是非常便宜的，加工过程中会有所涨价，但最后价值中真正最大的部分是在最终咖啡的销售部分，销售出的咖啡一杯就有5块钱，但咖啡豆可能只值两分钱。价值链中升值最高的部分在最后环节，原因在于客户体验。每家店的咖啡豆种类其实都很相似，咖啡机也差不多，咖啡作为饮料（解渴、提神）也并非没有替代品，但价格却可以相差很多，这主要就是靠体验的区别，客户体验逐渐成为真正的价值来源。咖啡店如果做得好，它就会提供一些比较独特的感受，无论是环境还是宣传，让客户觉得值得付出。所以说价值主要来自体验的创造，而不是产品或功能本身。</p> <p style="text-align: justify;">      这在金融服务中也是类似的。基础设施投入是目前所有的银行在考虑转型时的核心，这一部分的投入十分大，无论是数据中心、连接还是云计算，都主要集中在这一块，但是UI/UX客户体验部分的投入则相对较弱，这与把客户体验作为价值来源的重心有所冲突，因此我们认为需要进行改造。开放式银行某种程度上在基础设施投入上可能更大，使得云计算更强，以至于可以服务更多客户，但是因为它是一个金融行业的共享经济，因此每一家真正使用这个架构的机构在这方面的平均投入就变得很少，这样就可以在体验方面有较高的投入。以一个非常极端的例子来讲，工农中建四大银行的核心系统虽然有所不同，但是可以把它们共有的部分抽取出来放在一起，这时投入可能增加了一倍，但是在四家共享后，实际上每个银行投入只有原来的一半。</p> <h2 style="text-align: justify;"><strong>二、云计算与全功能开放式银行</strong></h2> <p style="text-align: justify;">      我们现在做的就是云计算与全功能开放式银行，云计算是强大功能的基础，基础架构分成很多不同的模块，在云计算的基础上分布于各个地方，人们所需要的服务可以从里面去获取，这样就形成了一个共享的基础设施。各方面数据的存储各个机构可以各自进行，但是提供的客户体验就要完全根据自己的设置来做，客户体验包括产品设计、产品运营，以及客户服务等各个方面，这时才算是真正回归本源。 </p> <p style="text-align: justify;">      金融机构的核心并不在于有多先进的技术，也不在于工作人员有多少，而是在于能否给客户带来最大的价值，如何来服务客户。金融机构的本质一是记账，二是服务客户，它本身属于一个服务型行业。金融行业都是高科技行业是因为只有这样才能更好服务客户，但是不能过于专注于技术投入，投入的本质是要提供好的服务，因此开放式银行实际上会让金融机构更集中精力回到自己应该做的事情，就是如何服务好客户。</p> <p style="text-align: justify;">      微服务就是提供不同种类服务的能力模块，开放式平台整合了这样一整套微服务的能力，在该平台中有各种各样的模块，然后各个机构通过个性化，根据不同的场景给客户提供十分贴心而有特色的服务。开放式银行成本较低，也十分有效，这是因为使用开放式银行后，金融机构的精力可以集中在服务客户，而不是如何提高自己的技术能力。</p> <p style="text-align: justify;">      在我们公司中，我们可以在同一套技术架构的基础上提供一个非常简单的用户界面，也可以提供一个非常复杂的用户界面，它背后的技术架构一模一样，没有任何区别，只不过提供的客户体验完全不同。对于一个专业的投资者，我们就可以提供一个不限制屏幕的、最高端的交易平台；对于一个普通客户，则可以让他们使用手机来进行交易。这背后的基础设施完全一样，包括云计算、人工智能、客户化、定制内部风控等，都在后台进行。我们的伙伴银行使用同样的技术平台来提供不同的客户体验。事实上，客户并不关心我们使用了什么技术，而只关心他能不能得到想要的服务，开放式银行就可以满足这种需求。</p> <h2 style="text-align: justify;"><strong>三、未来方向</strong></h2> <p style="text-align: justify;">      更安全、在云端、智能化、个性化是未来金融的大方向。大数据、云计算、人工智能对于金融科技发展以及对金融机构竞争力来说是核心要素，但是这些新科技的应用对于最终用户是隐形的，最终用户只享受更好的服务。例如苹果手机所使用的具体技术对于其用户来讲就不那么重要，最终用户只享受其成果。</p> <p style="text-align: justify;">      我认为就像其他高度成熟的工业一样，开放式专业分工合作是金融行业未来的方向。在过去，中国每一家银行都是要大而全，公司设置高度类似，做的事情也高度类似，但这会造成巨大的资源浪费。未来的方向是开放式银行，通过分工合作，使得机构更加重视客户体验等更加重要的方面。</p> <p style="text-align: right;">（责任编辑：许泰琦）</p> <p style="text-align: right;">来源：2020（第十六届）中国金融风险经理年度总论坛</p> , weight => 0, expert_id => 0, classroomId => 0, length => , startTime => 1606521600, floorId => 0 ], 16 => [ id => 1021, type => information, title => 程建：关于信用风险内评法的几点思考, subtitle => , tags => [ 0 => 86, 1 => 308, 2 => 312, 3 => 262, 4 => 322, 5 => 339 ], categoryId => 16, cate_child => 0, cate_grand => 0, summary => null, goals => [], audiences => [], isVip => 0, cover => [], status => published, creator => 3, createdTime => 1606521600, updatedTime => 1611818423, serializeMode => none, ratingNum => 0, rating => 0, noteNum => 0, studentNum => 0, hotSeq => 0, recommended => 0, recommendedSeq => 0, recommendedTime => 0, orgId => 1, orgCode => 1., discountId => 0, discount => 10.00, hitNum => 5, maxRate => 100, materialNum => 0, parentId => 0, locked => 0, minCoursePrice => 0.00, maxCoursePrice => 0.00, teacherIds => [ 0 => 3 ], defaultCourseId => 1021, sourceUrl => null, body => <h1 style="text-align: center;"><strong>关于信用风险内评法的几点思考</strong></h1> <p style="text-align: center;">程建中国建设银行总行风险管理部系统重要性银行管理处处长</p> <p style="text-align: center;"> </p> <p style="text-align: justify;">      建设银行是国内第一批批准使用资本计量高级方法——内评法的银行，做内评法到现在也有15年左右的时间了。这个工作也是一个很强大的事情，其中在这里面也涉及到很多如何将监管的规则和银行的管理现状结合起来的一些问题。因为从巴塞尔协议和内评法的规则来看，监管包括银行业也是想解决资本、风险和收益三者之间的衔接和协调的问题。因为资本是表，它主要是看得见，很多时候监管去看一家银行的实力或者状况，就是想通过资本来进行约束。资本的下面就是风险，监管也是希望能够通过资本的变化跟银行风险的水平挂上钩，这个也是银行内部需要解决的问题。因为风险背后还有一个收益的问题，这也是银行包括企业逐利的目的。那么追逐收益无可厚非，但是在这个过程中，是不是能够做到与风险的偏好、自身风险管理的能力相匹配，这就是一个大问题。我们现在看到一些金融机构出现问题，包括2008年金融危机大部分原因就是这两者之间，或者这三者之间，出现了一些失衡的问题。过度逐利实际上就有可能引发了一些金融机构破产，甚至整个金融体系崩溃的问题。</p> <h2 style="text-align: justify;"><strong>一、巴塞尔协议的演变和资本充足率内涵的扩展</strong></h2> <p style="text-align: justify;">      巴塞尔协议的框架从1988年开始到现在也将近有32年的历史，也是一个正值“青壮年”的规则体系。巴塞尔体系的内容也经历过好几轮的演变，目前是第三版。这三个版本从内容包括方法来看，都经历了比较大的提升和变化——风险的覆盖面越来越全，方面也是越来越精细、多样化。实际上也就考虑了国际银行业的实践差异，大家可以根据自己的情况去选择。基础比较好的银行、能力和资源比较充裕的银行可以选择相对比较复杂的内部评级法或者高级方法来计量监管资本的参数。</p> <p style="text-align: justify;">      巴塞尔协议Ⅲ对资本充足率的范围、计算的规则和内涵、比率都进行了扩充。包括比率对资本的要求、风险加权资产的计量规则都引入了很多监管的考虑，或者说是对风险管控的要求。比如巴塞尔协议Ⅲ引入了宏观审慎监管的要求，包括逆周期的资本缓冲、留存资本缓冲，都是一些最新的变化。这背后也是业界的一些实践和经验教训，也是在损失惨重或者导致整个金融体系出问题了之后，监管也是在不停地调整这些资本约束的指标。</p> <h2 style="text-align: justify;"><strong>二、内评法下风险加权资产的计算</strong></h2> <p style="text-align: justify;">      我们国内银行业主要还是管控信用风险，监管资本上信用风险加权资本占到了90%。内评法下信用风险加权资本的计算，第一步是分类，基本上是对客户或产品按照风险暴露的水平作一个大致的分类。第二步就是对每一类都会有一个监管的参数和规则的设定，在这里银行更多的是做客户评级和债券评级两大参数。算完参数之后在计算的时候我们还要考虑风险缓释的处理，主要是抵质押。比如说我们有一些用保证金和国债抵押，它的权重可能需要被适当地替换。经过上述处理之后我们最终通过公式得到风险加权资产，EAD相对来说简单一些，主要就是通过PD和LGD来计算K。</p> <p style="text-align: justify;">      实际上巴塞尔的信用风险计量规则背后是有一个理论模型，这个模型就叫渐近性单风险因子模型。这个模型是Gordy这个专家在2003年的一篇文章中专门介绍的，是监管资本计量的一个基石。在这个模型里面，他提出了一个假设——组合不变性，意思就是组合要充分、要分散，不能集中度过高。其实巴塞尔协议Ⅱ是建议国际活跃性银行来实施，因为这些国际化的银行它相对能够做到在全球或者行业中资产组合的一个均匀分布。当然这只是一个理想化的想法，但实际上来看这个很难做到，我们看到很多银行出问题也是这方面出问题了。这套公式的核心是通过资产的相关性来推断出违约相关性，也就是图1中所表示的ρ，是监管资本计量的一个核心参数，巴塞尔协议不同类别的客户算自己的资本都是在算这个ρ值。</p> <p style="text-align:center"><img alt="" src="data:image/png;base64,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" style="width: 511px; height: 203px;" /></p> <p style="text-align: center;"><span style="line-height: 150%;"><font color="#000000"><span style="font-size: 12pt;"><span style="line-height: 150%;"><span style="font-family: 宋体;"><font face="宋体">图</font></span></span></span><span style="font-size: 12pt;"><span style="line-height: 150%;"><span roman="" style="font-family:;">1</span></span></span><span style="font-size: 12pt;"><span style="line-height: 150%;"><span style="font-family: 宋体;">  <font face="宋体">渐近性单风险因子模型</font></span></span></span></font></span></p> <p style="text-align: justify;">      从银行账户信用风险暴露的分类来看，里面每一个风险暴露它都有自己的计量规则。这些规则里面比较重要的一个参数相关系数，相关系数的一个重要的输入参数就是PD，计算资本还需要加入LGD。这里面可以看到每一类风险暴露，其相关系数都有不同的算法。其中可以看到，零售风险暴露有两类——包括个人住房抵押贷款和合格循环零售风险暴露，它们的相关系数就是一个常数。但随着经济金融的发展，我们习惯的改变等等，这可能也会出现一些问题，这些监管规则在未来也可能会进行校准。</p> <h2 style="text-align: justify;"><strong>三、关于内评法应用的思考</strong></h2> <p style="text-align: justify;">      第一，内评法计量规则存在差异性。非零售信用风险暴露，包括主权、金融机构和公司，是从客户维度去分类的；零售信用风险暴露，主要是从产品维度进行分类的，因为它主要包括了住房、合格循环和其他零售。所以这两者在内评法规则的内部有一定的差异，这个差异的原因就在于，对于零售来说，在当时要实现对客户的风险计量和评级，可能技术和成本上都不支持。那么大数据的出现就会对此有一定的影响，可能就会统一这个规则的计量。现在的零售小微企业是全国上下包括银行业都在重点关注的一类业务，单从计量规则来看的话，它对银行是一个对公类的业务；但监管为了鼓励大家来做这类业务，在资本上如果满足一定的标准就可以按照零售去计量。所以在业务管理上是公司部，资本计量上又按零售，那两者之间的协商、协调，在真正落地上涉及到很多数据、管控流程上的调整问题。</p> <p style="text-align: justify;">      第二，相关性系数的校准。巴塞尔ASFR模型的假设就是分散化、流动性和相关性。分散化刚才已经提到了，其实汇丰在2008年能够躲过一劫也是跟它区域的分散化战略是有一定关系的，但现在它也遇到了问题，恰恰也有区域分散化战略有关系。因为它大部分的业务都是集中在亚洲，尤其是香港，但是现在欧洲和北美的情况也不怎么样，这个时候它也遇到了一些问题。所以分散化和集中度并不是表面看上去“三分天下”就可以了，还有一个动态的结果的影响的问题。流动性在第三版巴塞尔协议中专门提出了一个框架。但是相关性其实在巴塞尔协议Ⅲ里也没有做太多的调整，主要的调整就是针对金融机构，将其相关系数扩大了1.25倍，其他的还是沿用巴Ⅱ的框架。原来零售的逻辑是充分分散的，所以它的相关系数就比较简单，可以设置成一个常数；但现在由于大数据的发展，生活方式、交通和技术的改变，人与人之间的相关性可能会被低估。就比如前不久的公积金的骗贷案，这6000多人并不一定是有一个交集，但现在他们通过这种技术连接在一起，对银行的风险管理、资本管理就会形成一定的挑战。</p> <p style="text-align: justify;">      第三，亲周期的问题。银行主要是PIT（Point In Time，即点违约率），这是银行经营管理上所需要的；但是监管是TTC（Through-The-Cycle，跨周期违约率）。这两者之间在资本高级方法实施的过程中，实际上就是两者之间的一个权衡。我们现在看到监管对于解决亲周期问题其实是采取了一些措施，包括PD校准要求银行将违约率跟整个银行的评级在长期的方向进行校准。另外在计算等级PD的时候，监管可能就会要求银行用均值、众数或者上限值来计算监管资本。这实际上是一个审慎的调整，尽量避免当前时点数据的亲周期影响。此外，还有对downturn LGD，即经济下行时期的违约损失率的要求。最后监管有一个对风险加权资产（RWA）的校准，包括其惩罚因子、放大因子、并行期和底线。当初在批准大行实施内评法的时候，我们一些银行因为监管的校准有一些风险加权资产解决不掉。所以监管的也是考虑到由于周期性等原因可能会低估风险因素，对银行施加了一些约束。但是从结果来看，是不是能够起到相应的防控作用，可能还需要根据情况再去验证。</p> <p style="text-align: justify;">      第四，金融科技的影响。金融科技的影响是比较大的，一个是流程。原来银行的评级管理流程要求要有不同的岗位去制约，现在新的一些金融科技技术的应用，可能就是一个程序就把这些事情全都做完了。当前的管控流程和大数据的模型是有一些相互作用的，也需要银行去对二者进行协调。另外金融科技既会对既有的流程进行冲击，也会产生一些新的风险——因为系统或者算法的应用所产生的模型风险，这可能也是未来防控的重点。最后还有一个能源问题。尽管大量的代码和计算机算法的应用很高级、很先进，但如果系统没电了，一切都归零，而现在一些网络攻击、黑客的手法是可以做到的，这也是科技发展所催生的另外一个问题。这个问题可能对于银行而言稍微有些远，但是实际上这种技术的进步很快，可能相应的问题到来的比我们想象的早。</p> <p style="text-align: justify;"> </p> <p style="text-align: justify;"><strong>提问1：内评体系主要是用于资本监管，但是在内部使用上还有很多地方不尽人意，您在这方面是否也有同感？</strong></p> <p style="text-align: justify;">      回答1：是这样。实际上巴塞尔协议在落地的过程中银行就要解决一个问题，就是资本、风险和收益三者之间衔接的问题。资本就是从监管导向去要求银行要根据风险的计量、风险的状况来挂钩资本。而且在这个过程中，监管也希望银行能将内部评级模型尽量用到经营中的方方面面。实际上资本办法也提出了五项基本用途和五项高级用途，比如说信贷政策、授信审批、经济资本等。这样使用之后确实能把风险传导到后端的资本，但是这个过程中，监管资本计量公式里面本身也有一些缺陷，所以监管在计量资本中对参数的设定上，也是有一些考虑的，比如计算的过程中对PD的校准。银行内部算经济资本，更多的用的是算的每一个客户的PD；但是在算监管资本的时候，它可能要求等级客户统一地用一个PD，比如均值或者中位数。所以在银行自己算出来的风险水平跟监管资本之间，就可能产生差异。监管资本它很现实的就是需要一部分钱，银行必须达到监管所要求的持有数量，这部分钱是有成本的。比如是股权就需要按照ROE的回报去设置监管资本，债券就需要按照不同债的回报去设置。所以必然会传导到前台，对前台会有压力，这种压力可能就是跟实际计量之间的差异要怎么去传导，也就是监管校准这部分资本怎么去分配的问题。这可能也是和经验管理中不一致的地方。</p> <p style="text-align: justify;"> </p> <p style="text-align: justify;"><strong>提问2：因为受疫情影响，今年一季度都已经是负增长了，那么在做压力测试的时候，最严重的情景也能够把负增长考虑进去吗？</strong></p> <p style="text-align: justify;">      回答2：压力测试的核心的确是情景。像是GDP负增长这样的情况，我觉得将来压力测试肯定要放进去。但现在问题就是说，从疫情的发生到传导，到最终带来严重的影响可能它还需要一个过程。相关的数据、结果观察出来之后，你可能才能去设定它的情景和传导机制，这方面是肯定要考虑的。现在对于其严重性和中间过程的传导，还需要一定的时间再看一看。</p> <p style="text-align: justify;"> </p> <p style="text-align: justify;"><strong>提问3：现在有的银行做的模型主要利用的是财务数据，相对来说数据比较单一；再一个数据反映比较慢。您有没有考虑通过大数据等各类数据的外部引入，来调整传统评级模型的架构，对模型本身进行一个升级和调整？</strong></p> <p style="text-align: justify;">      回答3：这个的确是实践中遇到的一个问题，就是大数据怎么用到评级模型里面。但这里面不光是有银行的意愿问题，可能还涉及监管的要求。比如现在监管要求内评法的数据，PD要有至少五年，LGD至少要有七年。但是我理解大数据其实更多的是一个宽表的概念，就是类型的数据更多，但是它有的时候可能不太关注历史，而是更多地关注当前财务和非财务的数据。监管资本计量是要考虑跨周期的要求；而银行希望用得更准，可能银行就是需要借款人现在或者未来一年的状况，大数据就可以看到现在、明天是一个什么状况。因此两者之间就存在一个衔接的问题，所以巴塞尔委员会也在探讨大数据在监管上的运用，可能也要涉及到一些监管规则的调整。</p> <p style="text-align: justify;"><strong>提问3补充：如果不是只用大数据，还是以原来的评级模型为主，基本和内核上的东西不变，再加入大数据去反映财报以外的因素，您对此有何看法？</strong></p> <p style="text-align: justify;">      回答：这个我觉得可行。原来一些定性的指标，比如管理层的一些信息，以前就靠专家去打分，现在就可以引入外部数据，让打分更加客观化。</p> <p style="text-align: right;">（责任编辑：唐寅灏）</p> <p style="text-align: right;">来源：2020（第十六届）中国金融风险经理年度总论坛</p> <p style="text-align: justify;"> </p> , weight => 0, expert_id => 0, classroomId => 0, length => , startTime => 1606521600, floorId => 0 ], 17 => [ id => 1022, type => message, title => 卜永强：信用风险建模中的量化问题, subtitle => , tags => [ 0 => 86, 1 => 308, 2 => 262, 3 => 270, 4 => 313, 5 => 325, 6 => 426, 7 => 447, 8 => 444 ], categoryId => 15, cate_child => 0, cate_grand => 0, summary => null, goals => [], audiences => [], isVip => 0, cover => [], status => published, creator => 3, createdTime => 1606521600, updatedTime => 1611396767, serializeMode => none, ratingNum => 0, rating => 0, noteNum => 0, studentNum => 0, hotSeq => 0, recommended => 0, recommendedSeq => 0, recommendedTime => 0, orgId => 1, orgCode => 1., discountId => 0, discount => 10.00, hitNum => 3, maxRate => 100, materialNum => 0, parentId => 0, locked => 0, minCoursePrice => 0.00, maxCoursePrice => 0.00, teacherIds => [ 0 => 3 ], defaultCourseId => 1022, sourceUrl => https://www.yunzhan365.com/11511321.html, body => null, weight => 0, expert_id => null, classroomId => 0, length => , startTime => 1606521600, floorId => 0 ], 18 => [ id => 1023, type => message, title => 程建：关于巴塞尔协议Ⅲ最终规则实施的一些思考, subtitle => 程建,建设银行总行风险管理部系统重要性银行管理处处长, tags => [ 0 => 308, 1 => 312, 2 => 262 ], categoryId => 47, cate_child => 0, cate_grand => 0, summary => null, goals => [], audiences => [], isVip => 0, cover => [], status => published, creator => 3, createdTime => 1606521600, updatedTime => 1611396765, serializeMode => none, ratingNum => 0, rating => 0, noteNum => 0, studentNum => 0, hotSeq => 0, recommended => 0, recommendedSeq => 0, recommendedTime => 0, orgId => 1, orgCode => 1., discountId => 0, discount => 10.00, hitNum => 1, maxRate => 100, materialNum => 0, parentId => 0, locked => 0, minCoursePrice => 0.00, maxCoursePrice => 0.00, teacherIds => [ 0 => 3 ], defaultCourseId => 1023, sourceUrl => https://www.yunzhan365.com/91346292.html, body => null, weight => 0, expert_id => null, classroomId => 0, length => , startTime => 1606521600, floorId => 0 ], 19 => [ id => 457, type => message, title => 风险管理 2013年第2期, subtitle => , tags => [], categoryId => 4, cate_child => 0, cate_grand => 0, summary => null, goals => [], audiences => [], isVip => 0, cover => [ large => public://course/2020/11-25/1847368ee497475119.png, middle => public://course/2020/11-25/1847368ef8b0547517.png, small => public://course/2020/11-25/1847368f09ed347463.png ], status => published, creator => 3, createdTime => 1606299882, updatedTime => 1441592447, serializeMode => serialized, ratingNum => 0, rating => 0, noteNum => 0, studentNum => 0, hotSeq => 0, recommended => 0, recommendedSeq => 0, recommendedTime => 0, orgId => 1, orgCode => 1., discountId => 0, discount => 10.00, hitNum => 0, maxRate => 100, materialNum => 0, parentId => 0, locked => 0, minCoursePrice => 0.00, maxCoursePrice => 0.00, teacherIds => [ 0 => 3 ], defaultCourseId => 457, sourceUrl => https://book.yunzhan365.com/qpwi/isvs/mobile/index.html, body => null, weight => 0, expert_id => null, classroomId => 0, length => , startTime => 0, floorId => 0 ] ]
view	list

Symfony Profiler

https://www.cfrisk.org/app_dev.php/user/3?page=144

Request

GET Parameters

POST Parameters

Request Attributes

Cookies

Request Headers

Request Content

Server Parameters

Response

Response Headers

Session

Session Metadata

Session Attributes

Flashes

Flashes

Sub Requests 2

AppBundle:User:headerBlock (token = 46c66f)

AppBundle:Course/CourseSet:courseSetsBlock (token = d3d88e)