AIMs: Describe key guidelines for verification and validation of a bank's ORMF and ORMS. Describe key supervisory guidelines for the selection of a reference date for an internal loss. Describe key guidelines for the selection of a bank's Operational Risk Categories (ORCs). Explain key guidelines for modeling the distribution of individual ORCs, including the selection of thresholds, necessary adjustments, and selection of statistical tools and probability distributions.
Questions:
412.1. According to the Committee, each of the following is true about a bank's internal loss data thresholds in the Basel III advanced measurement approach (AMA) EXCEPT which is false?
a. The threshold for internal loss collection processes corresponds to the gross loss amount
b. The appropriate threshold is limited on the upside by the test of materiality: a threshold is too high if it omits material loss event data
c. A bank may use different thresholds for data collection and modeling
d. The Basel III framework has a phase-in period; at the end, all banks will converge on an internal loss threshold of EUR 10,000
412.2. In preparation for your bank's shift from the standardized approach to the advanced measurement approach (AMA) to operational risk, your colleague Peter has drafted an AMA model which includes a large number of operational risk categories (ORC). Peter defends this by referencing "Operational Risk--Supervisory Guidelines for the Advanced Measurement Approaches" on Granularity, specifically "160. In accordance with paragraph 669(c) of the Basel II Framework, an AMA bank’s risk measurement system must be sufficiently granular to capture the major drivers of operational risk affecting the shape of the tail of the loss estimates." On the other hand, you point out that, according to the Committee, EACH of the following can be a problem created by too many (a high number) of operational risk categories (ORCs) EXCEPT which is false?
a. Too many ORCs can lead to increased heterogeneity for the events in each category.
b. A high number of ORCs can cause the number of losses in each category to fall below a model’s data threshold
c. A high number of ORCs may lead to an unrealistically high capital charge when no correlations are modelled and capital charges for all ORCs are summed together
d. A bank modelling correlations that use a high number of ORCs might have difficulty finding statistical means to validate correlation assumptions due to minimal loss data for each ORC
412.3. You are helping to select frequency and severity distributions for your bank's advanced measurement approach (AMA) to operational risk. According to the Committee, each of the following is true EXCEPT which is not?
a. The choice of the severity distribution has a greater impact on the final outcome than the choice of frequency distribution
b. Probability distributions should meet the following criteria: realistic, well specified, flexible and simple; i.e., easy to implement and to generate random numbers for the purpose of loss simulation)
c. It is advisable to use Exploratory Data Analysis (EDA) for each ORC to better understand the statistical profile of the data and select the most appropriate distribution
d. Due to the observable skew and kurtosis of loss severity data, in almost all cases, the extreme tail region of the loss severity distribution should employ internal empirical data
Answers here:
Questions:
412.1. According to the Committee, each of the following is true about a bank's internal loss data thresholds in the Basel III advanced measurement approach (AMA) EXCEPT which is false?
a. The threshold for internal loss collection processes corresponds to the gross loss amount
b. The appropriate threshold is limited on the upside by the test of materiality: a threshold is too high if it omits material loss event data
c. A bank may use different thresholds for data collection and modeling
d. The Basel III framework has a phase-in period; at the end, all banks will converge on an internal loss threshold of EUR 10,000
412.2. In preparation for your bank's shift from the standardized approach to the advanced measurement approach (AMA) to operational risk, your colleague Peter has drafted an AMA model which includes a large number of operational risk categories (ORC). Peter defends this by referencing "Operational Risk--Supervisory Guidelines for the Advanced Measurement Approaches" on Granularity, specifically "160. In accordance with paragraph 669(c) of the Basel II Framework, an AMA bank’s risk measurement system must be sufficiently granular to capture the major drivers of operational risk affecting the shape of the tail of the loss estimates." On the other hand, you point out that, according to the Committee, EACH of the following can be a problem created by too many (a high number) of operational risk categories (ORCs) EXCEPT which is false?
a. Too many ORCs can lead to increased heterogeneity for the events in each category.
b. A high number of ORCs can cause the number of losses in each category to fall below a model’s data threshold
c. A high number of ORCs may lead to an unrealistically high capital charge when no correlations are modelled and capital charges for all ORCs are summed together
d. A bank modelling correlations that use a high number of ORCs might have difficulty finding statistical means to validate correlation assumptions due to minimal loss data for each ORC
412.3. You are helping to select frequency and severity distributions for your bank's advanced measurement approach (AMA) to operational risk. According to the Committee, each of the following is true EXCEPT which is not?
a. The choice of the severity distribution has a greater impact on the final outcome than the choice of frequency distribution
b. Probability distributions should meet the following criteria: realistic, well specified, flexible and simple; i.e., easy to implement and to generate random numbers for the purpose of loss simulation)
c. It is advisable to use Exploratory Data Analysis (EDA) for each ORC to better understand the statistical profile of the data and select the most appropriate distribution
d. Due to the observable skew and kurtosis of loss severity data, in almost all cases, the extreme tail region of the loss severity distribution should employ internal empirical data
Answers here:
Last edited:
I will fix the question numbers here, in the paid section of the forum, and in the pdf. 