What's new

P2.T7.502. External operational loss data sources

Nicole Seaman

Chief Admin Officer
Staff member
Learning outcomes: Explain the motivations for using external operational loss data and common sources of external data. Compare the characteristics of external operational loss data from different sources.


502.1. Which of the following statements is true about the role of external data in Basel III?

a. External data is a requirement in all three approaches (BA, SA and AMA) to operational risk under Basel III
b. External data is technically not required under any of the three approaches (BA, SA, and AMA) to operational risk under Basel III; it is merely a recommendation in the Principles for the Sound Management of Operational Risk
c. In general banks have resisted collected external data due to lack of perceived relevance; for example, among systemically important banks, the majority do not use external data elements
d. External data is a requirement neither of the basic indicator (BIA) nor standardized approach (SA) to operational risk under Basel III, but external data is required in the advanced measurement approach (AMA)

502.2. Among the seven Level I operational risk event types, which type is both (i) easily the most common in the IBM Algo FIRST database producing 46% of records and 48% of losses and (ii) also carries the majority of losses (54%) in the Operational Riskdata eXchange Association (ORX) database?

a. Business disruption and system failures
b. Clients, products, and business practices
c. Damage to physical assets
d. Employment practices and workplace safety

502.3. Sources of external loss data include subscription and consortium databases. About these sources each of the following is true EXCEPT which is false?

a. IBM Algo FIRST is a subscription-based database that relies on public data sources (ie, legal, regulatory, and press reports of events) and therefore suffers "availability bias"
b. The Operational Riskdata eXchange Association (ORX) is a consortium-based database (ie, collects data from members) that does not suffer availability bias because all operational risk events are provided anonymously into the database; however, as only 67 member banks participate, the data relate to only a subset of financial services
c. With respect to operational losses by event type categories, the ORX database shows a significantly higher percentage of Execution, Delivery, and Process Management (EDPM) losses than is indicated in the Algo FIRST database
d. With respect to operational losses by business line, there is no material difference between the ORX and Algo FIRST databases; for example, in both databases the number and size of losses is largest for the Payment and Settlement (aka, Clearing) business line

Answers here: