After the delivery of Enterprise Wide Stress Testing (EWST) on August 12, 2015, CCFA Risk Study group held the fifth study session on December 7, 2015. The topic is on Operational Risk. It was an honour to invite two experienced practitioners in the field of operational risk to lead the round table discussion. They are: Baifang Xing, Manager of Risk Methodologies at RBC, and Shunbai Li, Manager of Internal Rating Management at Scotiabank. Both speakers have extensive experience in the operational risk area. Yicent Chen, Model Risk Specialist from BMO was the facilitator. Other practitioners from leading Canadian financial institutions also joined the discussion.
The session covered Advanced Measurement Approach (AMA) methodology specifications, model challenges and limitations from model development perspective. A high-level overview of the operational risk capital model methodology was provided by the speakers. At first the speakers explained the definition of operational risk from Basel Guideline and in practical terms. Then they moved onto explaining the main factors that are associated with operational risk and operational risk capital calculation. In the end, they shared some of the on-going challenges industry faces from regulators and banks perspective.
Operational risk is defined as the risk of loss resulting from inadequate or failed internal process, people and systems or from external events. In practice, it contains all risk types that can’t be classified under market risk, credit risk, or any of the major risk categories. Basel uses two dimensions to define the characteristics of operational risk: event types and business lines. For operational risk capital modelling, a two dimensional table is created by setting event types as column and business lines as row. Each cell of this table contains a unit of measure which is used for loss severity and loss frequency estimation. There are two fundamental assumptions for the Advanced Measurement Approach: (1) it assumes loss events happen with the frequency they happened before, and (2) the magnitude of the loss, while random, is adequately captured by the past. This approach is being developed on the top of basic indicator approach and the standardised approach which are simpler and more straightforward. Modeller uses four different data inputs: internal data, external data, scenario data and Business Environment and Internal Control Factors (BEICF). Even though the model has been studied for more than ten years globally, significant challenges still exist as followings: internal and external data scarcity, capital volatility from low frequency and high severity loss event, truncated data due to reporting bias, inflation adjustment, foreign exchange adjustment, and common root loss event.Due to the nature of the operational risk definition and the data used, it is a challenge to compare the operational risk Advanced Measurement Approach among different banks. Each bank used different approach and dataset to solve the problem.
就企业整体压力测试（EWST）研讨后，CCFA于2015年12月7日迎来了第五次风险管理研讨会，课题是操作风险。我们荣幸地邀请到了在这一领域的两位资深从业者领导为我们主导这次的圆桌讨论，他们是加拿大皇家银行（RBC）的风险方法经理Baifang Xing，以及丰业银行（Scotia bank）的内部评级管理经理Shunbai Li。两位演讲者在操作风险领域都有丰富的经验。担任本次研讨会主持的是满地可银行（BMO）的模型风险专家Yicent Chen。参加讨论的还有来自加拿大各大金融机构的专业人士。