FRY9CS_20080630_i

Accord Implementation Group
Operational Risk Working Group

2008 Loss Data Collection Exercise
Contents
2008 Loss Data Collection Exercise......................................................................................... 1
Overview and instructions ........................................................................................................ 2
Background .....................................................................................................................2
Participation .................................................................................................................... 2
Scope .............................................................................................................................. 3
Data security and confidentiality ..................................................................................... 4
Results ............................................................................................................................ 4
Timeline .......................................................................................................................... 5
Attachment A: Internal loss data, scenario data and supplemental questions ......................... 6
Internal loss data.............................................................................................................6
Template ......................................................................................................................... 6
Template instructions ...................................................................................................... 6
Scenario data ................................................................................................................ 10
Template .......................................................................................................................10
Template instructions .................................................................................................... 10
Supplemental internal loss data and scenario analysis questionnaire.......................... 16
Process/instructions: ..................................................................................................... 16
Attachment B: Exposure indicators and capital estimates ..................................................... 17
Questionnaire:............................................................................................................... 17
Process/instructions: ..................................................................................................... 17
Attachment C: Supplemental range of practice questionnaire ............................................... 18
Questionnaire:............................................................................................................... 18
Process/instructions: ..................................................................................................... 18

1/18

Overview and instructions
Background
The Operational Risk Working Group of the Accord Implementation Group (AIGOR) is
conducting a voluntary Loss Data Collection Exercise (LDCE). While similar to two previous
international LDCEs, which focused on internal loss data, this LDCE is the first international
effort to collect information on all four data elements – internal data, external data, scenario
analysis, and business environment and internal control factors (BEICFs) – used in an
Advanced Measurement Approach (AMA).
The objective of the exercise is to further the understanding of both supervisors and
participating banking institutions regarding outstanding operational risk implementation
issues, as well as to promote consistency in addressing these issues across jurisdictions.
The exercise will facilitate comparative analysis across jurisdictions by benchmarking losses
at the national/regional and international levels, and will provide data to assess capital levels
relative to internal data and scenario analysis. Collecting data on the four elements of the
AMA framework will provide benefits to participating banking institutions and national
supervisors including:
1.

A greater perspective on the banking industry’s loss exposure;

2.

Insight into how banking institutions are using internal and external loss data,
scenarios and BEICFs for risk measurement and risk management;

3.

Information on the four data elements and their influence on operational risk capital
levels; and

4.

Updated range of practice and new cross-bank comparisons.

Participating institutions will receive customised analysis comparing their data with industry
data at both the international, and when possible, regional/national levels. Results will be
used to benchmark a banking institution’s loss experience and to gain a better understanding
of the completeness of its data. In addition, participating institutions will receive updated
range of practice (ROP) information on scenario analysis, external data and BEICFs. This
ROP information can be used by participating institutions to assess and benchmark their
practices against industry practices.

Participation
Participation is voluntary and the exercise is open to banking organisations at the group-wide
level that are implementing or using one of the three Basel II approaches for calculating
operational risk capital: the Advanced Measurement Approach (AMA), the Standardised
Approach (TSA) or the Basic Indicator Approach (BIA). The exercise has been designed to
minimise resources needed to participate. Banking institutions may choose to participate in
the full exercise or to submit information only for certain parts of the exercise. If your
institution has not done work in an area addressed by a question, please leave the question
blank.

2/18

Banking institutions wishing to participate should inform their national supervisors of their
intent as soon as possible and provide contact information for their LDCE point of contact.
Each national supervisor will provide instructions to their participating institutions regarding
how to submit LDCE data and attachments. Participating institutions are encouraged to
contact their national supervisor with any questions about the exercise prior to submission.
Participating institutions should submit the information requested in this exercise to their
national supervisors by 30 June 2008.

Scope
The LDCE consists of three sections. The first section includes forms to report internal loss
data, scenario analysis data and qualitative information that gives context to the submitted
data (Attachment A). The second section (Attachment B) requests information on exposure
indicators and capital estimates. The final section (Attachment C) contains questions that will
provide range of practice (ROP) information on how operational risk is measured and
managed. These three sections align data access with the degree of confidentiality
associated with the data and allow an option for national supervisors to complete certain
information on behalf of banking institutions within their jurisdictions. These materials are
summarised below:

Attachment A: LDCE templates for internal loss data and scenario analysis data and
supplemental internal loss data and scenario analysis questionnaire:
1.

Internal Data template (Internal Loss Data Template.xls),

2.

Scenario Data template (Scenario Templates.xls, and

3.

Supplemental Internal Data and Scenario Analysis Questionnaire (file name
Attachment A.pdf).

The forms facilitate submission of individual operational losses and scenario analysis data as
well as qualitative information that provides context for these data elements.

Attachment B: Exposure indicators and capital estimates (Attachment B.pdf).
The form requests data on the results of the 31 December 2007 calculation of capital (AMA,
TSA or BIA). The form also requests group-wide and business line exposure indicators,
which will be used to construct national/regional and international benchmarks. These
benchmarks will be integrated into the customised reporting that is provided back to
participating banking institutions. Individual raw data from this questionnaire will not be
shared outside the participating institution’s national supervisory jurisdiction.

Attachment C: Supplemental Range of Practice questionnaire (Attachment C.pdf)
This attachment may be completed either by national supervisors or by participating banking
institutions at the national supervisor’s discretion. These questions address the AMA model
framework, external data, scenario analysis practices and the use of BEICFs.

3/18

Data security and confidentiality
Data security and confidentiality are a paramount concern, and will be handled in a manner
similar to the 2002 LDCE. Each national supervisor will maintain all data and information
submitted by institutions within its jurisdiction in a secure environment. Access to original
submissions with a participant’s name will be restricted to authorised staff of an institution’s
national supervisor(s).
As discussed above under the description of Attachment B, raw data on exposure indicators
and capital received pursuant to Attachment B will be separate from other data submitted
and access to these data will be limited to authorised staff from an institution’s national
supervisor(s). Data requested in Attachment B will remain with the national supervisor(s),
where it will be restricted to authorised staff and maintained in a secure environment.
Upon receipt of submissions from institutions within their jurisdiction, national supervisors will
take steps to anonymise the data and information received pursuant to Attachments A and C.
This will include steps to replace each institution’s name with a numeric code, to convert
country references to geographic regions if necessary to preserve confidentiality, to convert
all internal loss events to a common currency (Euro), and to ensure that scenario
descriptions of losses exclude potential identifiers. Working with the Secretariat of the Basel
Committee on Banking Supervision (BCBS), the national supervisor will add a country code
(or a regional code for jurisdictions with less than three participating banking institutions) to
enhance confidentiality of the analysis.
Each national supervisor will send anonymised LDCE and ROP data (Attachments A and C),
via secure pathway, to the BCBS’s secure “e-BIS” website. The BCBS Secretariat will be the
custodian of the data, except for data provided in Attachment B, which will remain with the
primary national supervisor. In its role as custodian, the Secretariat will secure all submitted
LDCE and ROP data by restricting access and housing the data in a secure environment
with appropriate access controls. The Secretariat will consolidate the data into aggregate
anonymised data sets and forward them to all participating national supervisors and the
Central Processor.
As in the 2002 LDCE, resources from a central processor are necessary to process and
analyse the data in a timely manner. AIGOR has asked the Federal Reserve Bank of Boston
and the Office of the Comptroller of the Currency to be the Central Processor for the 2008
LDCE. Representatives from the two organisations serving as Central Processor will work
under the auspices of the Basel Secretariat throughout the duration of the exercise to limit
access to the LDCE data and maintain the data in a secure environment.

Results
The Central Processor will cleanse the anonymised aggregate data sets and prepare the
loss data and scenario analysis data for analysis. The Central Processor will prepare
aggregate results for public release and develop comparative tables for distribution back to
participating banking institutions. Aggregate public results will be similar to results from
previous LDCEs, but will also include a high-level analysis of the ROP for quantifying and
managing operational risk. Public results will be presented on a basis that avoids any
disclosure of the identities of participating institutions. Banking institutions will receive
individual reporting templates with comparisons of their loss experience, scenario analysis
data, and ROP against regional/national (where feasible) and international benchmarks,

4/18

derived from the medians of the aggregate data. National supervisors will discuss individual
reports with their participating banking institutions.

Timeline
1 May 2008

LDCE released to banking institutions

30 June 2008

Completed LDCE material due to national supervisors

31 July 2008

National supervisors provide anonymised data to Secretariat

1 September 2008

Secretariat sends anonymised data to Central Processor and national
supervisors

1st Quarter 2009

Public LDCE report issued

2nd Quarter 2009

National supervisors begin
participating institutions

5/18

discussing

individual

reports

with

Attachment A:
Internal loss data, scenario data
and supplemental questions

Internal loss data:

Internal Loss Data
Template.xls

Template instructions
Please submit internal loss data according to the following rules. Deviations from these rules
should be documented in Question 8 of the Supplemental Internal Loss Data and Scenario
Analysis Questionnaire.
Participating institutions should submit a minimum of three years of internal data that they
view to be reasonably complete.1 For participating AMA institutions,2 submitted data should
cover at least the time period used to quantify their operational risk capital charge, which
may be longer than three years. Data should cover loss events through 31 March 2008 or 31
December 2007 if data for March are not yet available.
File type
Data can be submitted using the Excel template provided or via a comma separated value
(*.csv) file, with the same column layout and headings as in this Excel template. To minimise
burden, institutions may also choose to submit data in an alternate layout (in either Excel or
csv format), provided that the data set contains the information requested in this exercise.3
For example, members of a loss data consortium may provide a copy of the data set
submitted to the consortium.

1

When determining what data to submit for the minimum of three years of data, the date should be based on
the date of discovery if possible, or the date of financial impact.

2

For the purposes of these instructions a participant would be considered an AMA, TSA or BIA participant
based on the approach used at the group-wide level.

3

Minor inconsistencies with the information requested in this exercise would be acceptable. For example,
reporting dates in quarter-year format rather than in day-month-year format would be considered a minor
inconsistency. Using a full-text currency name rather than the three character currency codes would be
considered a minor inconsistency. Reporting numeric variables in thousands rather than ones would also be
considered a minor inconsistency, so long as the institution explicitly notes this issue as part of its submission.

6/18

Cell formats
All data should be formatted as either text, numeric or date. Dates should be formatted as
“dd-mm-yyyy.”
Currency
Loss and insurance recovery amounts should be reported in the currency native to the
internal loss database with the currency type reported in a separate column. Participating
institutions should not undertake any currency conversions solely for the purpose of the
LDCE. For example, if an institution converts all losses to a common reporting currency the
institution could provide the losses in the common currency. Alternatively, if an institution
uses the local currency where the losses occurred, the institution could provide the losses in
the local currency.
Units for reporting loss amounts and insurance recovery
All numeric variables should be reported in ones – not thousands. For example, a one million
Euro loss would be reported as 1,000,000.
Threshold
No specific loss threshold is required for this exercise. AMA institutions should submit all of
the internal loss data used to quantify operational risk capital at a minimum; TSA and BIA
institutions should report all data above their loss data collection threshold. If an institution
chooses to submit consortium data, the institution may submit losses based on their own
internal threshold(s) or the consortium threshold.
Credit-related losses
Operational risk-related credit risk losses should generally be excluded. However, as slight
differences may occur across jurisdictions, credit-related operational losses should be
reported in accordance with the rules of the institution’s primary national jurisdiction.

For each loss, please provide the following information.
Column A: reference number
The reference number is the institution’s unique internal identifier for each loss event. Losses
can be submitted individually or aggregated into loss events (groups of associated losses
generated by a single root event). In instances where losses are submitted individually, the
identifier may appear several times in the data set for multiple losses linked to the same
event across dates, business lines and/or event types. If separate identifiers are used for
losses associated with the same event, please provide information such that individual
losses associated with the same event can be aggregated at the event level. If an institution
does not have an internal process that assigns reference number, please number losses as
1, 2, 3, etc.
Columns B – D: date of occurrence, date of financial impact, and date of discovery
For all available dates submit using format “dd-mm-yyyy.”

7/18

The date of financial impact of a loss event is the date the loss was first recorded in the
general ledger.
In the case of a group of associated losses generated by a single root event that occur or are
discovered over time, institutions should report all the losses with the (same) date of
occurrence of the root event and the pertinent (potentially different) dates of discovery and
dates of financial impact.
In the case of a claim on financial instruments (eg deceptive sales practices), the occurrence
date is the date when the financial instrument has been bought or sold; the date of discovery
is the date the institution received the claim.
Columns E-F: business line
Business line should be reported as one of the Level II Basel business lines, and if such
detail is not available, business lines should be reported by the Level I Basel business line. If
an institution cannot provide losses by Basel business line, internal business line should be
reported in Column F along with a mapping of internal business lines to Basel business lines
on a best efforts basis.
Basel business line (Level II if available, otherwise Level I)
For losses that are not assigned to a Basel business line, the business line should be
reported “UA” for unallocated. See Table 1 (located under a separate tab in the internal loss
data template) for a list of codes that should be used to refer to Basel business lines. While
the use of the listed codes is preferred, internal codes for the Basel business lines also will
be accepted as long as a key is provided.
Losses that affect more than one business line and are apportioned across business lines
should be associated using the same reference number to report the loss in each of the
business lines affected. If such losses are not apportioned across business lines, then the
business line should be reported as “UA” for unallocated.
Data from the insurance business line should be excluded.
Internal business line
If Basel business lines cannot be provided, provide the name of the associated business line
used internally within the participating institution.

Columns G – H: Basel event type (Level I and Level II)
Event types should be reported as one of the Level I Basel event types, and if available,
Level II Basel event types should also be reported. See Table 2 (located under a separate
tab in the internal loss data template) for a list of codes that can be used to refer to event
types. While the use of the listed codes is preferred, internal codes for Basel event types will
also be accepted as long as a key is provided.

8/18

Column I: gross loss amount
Report the total loss amount before any recoveries. The gross loss amount should be
reported in currency native to the internal loss database. The loss should be reported in
currency native to the internal loss database. Institutions should not undertake any currency
conversions solely for the purpose of the LDCE.
Column J: gross loss amount net of all recoveries except insurance
Report the gross loss amount net of all recoveries except insurance on a pre-tax basis,
reported in currency native to the internal loss database. The loss should be reported in
currency native to the internal loss database. Institutions should not undertake any currency
conversions solely for the purpose of the LDCE.
Column K: currency
Enter the three character International Organization for Standardization (ISO) currency code
for the currency in which the gross and net loss amounts are reported.4
Column L: country of origin
Report the country where the loss occurred, if available.
Column M: insurance coverage
Report a “Y” if the loss is currently covered by an insurance contract. Report an “N” if the loss
is not currently covered by an insurance contract. If this information is not available, report
“NA”.
Column N: insurance recovery amount
Report the amount recovered from insurance in the same currency in which the associated
loss was reported.
Column O: used in AMA model
Report a “Y” if the loss is used for AMA capital modelling. Report an “N” if the loss is not
used for AMA capital modelling. If the institution is using the TSA or BIA approach at the
group-wide level, report an “N” even if the subsidiary is using the AMA approach.

4

The standard currency codes are contained in the international standard ISO 4217 and include Euro (EUR),
U.S. Dollar (USD), Japanese Yen (JPY), U.K. Pound Sterling (GBP) and Australian Dollar (AUD). Please refer
to http://www.xe.com/iso4217.php.

9/18

Scenario data:

Scenario
Templates.xls

Scenario data should be submitted using one of the three templates provided. Institutions
should submit scenario data using the template that is most consistent with their scenario
approach. The templates have been designed along the following scenario approaches,
examples of which are contained in the Scenario Templates.
•

The percentile approach, where scenario output includes specified percentiles of
loss severity (50th, 95th, 99th, etc.).

•

The interval approach, where scenario output is the frequency of losses estimated
within a series of distinct loss ranges.

•

The individual scenario approach, where scenario output consists of individual
data points linked with a probability of occurrence.

Each institution should submit only one template for scenario data. Any inconsistencies
between the scenario data requested and the scenario template submitted should be
documented in Question 7 of the Scenario Analysis section of the Supplemental Internal
Loss Data and Scenario Analysis Questionnaire.

Template instructions
A. Percentile approach
Submit data for all scenarios used in the measurement and management of operational risk.
Reported scenario data should be the direct output of a workshop or other method used to
generate scenario data and should reflect raw data prior to use in modelling. If reporting the
full scenario data set is infeasible, report information on the following scenarios:
(i)

At least the 20 highest-severity scenarios where at least 20 of these scenarios are
expected to happen once or more every 1,000 years (ie with a mean annual
frequency greater or equal to 0.001). The highest-severity scenarios should be
determined according to the loss amounts reported for the highest percentile that
have been provided.

and
(ii)

The five highest-severity scenarios for each of the seven Level I Basel II event types
(if available).

Scenarios that meet the criteria of both (i) and (ii) need not be reported twice.

10/18

Section 1: scenario reference number
Provide the institution’s unique internal identifier for each scenario. If an institution does not
have an internal process that assigns reference number, please number scenarios as 1, 2, 3,
etc.
Section 2: frequency
Percentile 1 – On line 10, please indicate how frequency is estimated in the scenario
process. This will be either via the mean (enter “mean”) or via a specific percentile of the
frequency distribution (eg enter “50th percentile”). If another percentile of the frequency
distribution is estimated, please indicate the relevant percentile (eg “90th percentile”) on line
10.
On line 11 and below, provide the corresponding frequency for each scenario. Frequency
should be reported as the number of events per year in decimal form (i.e. a 1 in 100 year
event would be reported as .01).
Section 3: severity
Percentile 1 – Many institutions estimate scenario severity via a series of increasing
percentiles of the severity distribution. Such institutions would indicate the relevant
percentiles on line 10. For example, an institution that estimates the 75th, 90th and 99th
percentiles would enter “75th percentile”, “90th percentile” and “99th percentile” on line 10.
Some institutions estimate scenario severity via the mean of the severity distribution,
supplemented with a series of increasing percentiles of the severity distribution. Such
institutions would enter “mean” and the relevant percentiles on line 10. For example, an
institution that estimates the mean and the 90th and 99th percentiles would enter “mean”, “90th
percentile” and “99th percentile” on line 10.
On line 11 and below, provide the corresponding severity data for each scenario. Please add
columns to the worksheet for any additional percentiles as necessary.
Section 4: currency
Enter the three character ISO standard currency code for the currency in which the severity
data are reported.5
Section 5: Basel business line
If a scenario is associated with exactly one Basel business line, report the Basel Level I
business line. See Table 1 for a list of codes that should be used to refer to Basel business
lines. While the use of the listed codes is preferred, internal codes for the Basel business
lines also will be accepted as long as a key is provided. Report “UA” for unallocated, for
scenarios that are associated with more than one business line, including group-wide
scenarios.

5

The standard currency codes are contained in the international standard ISO 4217 and include Euro (EUR),
U.S. Dollar (USD), Japanese Yen (JPY), U.K. Pound Sterling (GBP) and Australian Dollar (AUD). Please
refer to http://www.xe.com/iso4217.php.

11/18

Section 6: internal business line
If an institution cannot provide losses by Basel business line, internal business line should be
provided along with a mapping of internal business lines to Basel business lines on a best
efforts basis.
Section 7: Basel event type
Event types should be reported as one of the Level I Basel event types. See Table 2 for a list
of codes that can be used to refer to event types. While the use of the listed codes is
preferred, internal codes for Basel event types will also be accepted as long as a key is
provided.
Section 8: used in AMA model
Enter “Y” if the scenario is used for AMA capital modelling. Enter “N” if the scenario is used
only for risk management purposes.
Section 9: brief description of loss
Provide a brief description of the scenario loss for at least the 20 highest severity scenarios
that are expected to happen once or more every 1,000 years. To preserve confidentiality,
scenario descriptions should exclude any potential identifiers (such as names of
companies/subsidiaries and locations). Scenario descriptions may be provided in the
language that is most convenient to the respondent.

B. Interval approach
Submit data for all scenarios used in the measurement and management of operational risk.
Reported scenario data should be the direct output of a workshop or other method used to
generate scenario data and should reflect raw data prior to use in modelling. If reporting the
full scenario data set is infeasible, report information on the following scenarios:
(i)

At least the 20 highest-severity scenarios where at least 20 of these scenarios are
expected to happen once or more every 1,000 years (ie with a mean annual
frequency greater or equal to 0.001). The highest-severity scenarios should be
determined according to the values reported for the upper bound of the highest
interval reported.

and
(ii)

The five highest-severity scenarios for each of the seven Level I Basel II event types
(if available).

Scenarios that meet the criteria of both (i) and (ii) need not be reported twice.
Section 1: scenario reference number
Provide the institution’s unique internal identifier for each scenario. If an institution does not
have an internal process that assigns reference number, please number scenarios as 1, 2, 3,
etc.

12/18

Section 2: intervals
Severity – In the column labelled “Low” enter the lower bound of the lowest interval range. In
the column labelled “High”, enter the upper bound of the lowest interval range.
Frequency – If frequency intervals are estimated in addition to severity intervals, in the
column labelled “Low” enter the lower frequency bound and in the column labelled “High”
enter the upper frequency bound. If frequency is estimated as a single number rather than an
interval, enter the single frequency estimate in both columns under “Low” and “High.”
Additional Intervals (including Interval 2 and Interval 3) - For Interval 2 and subsequent
intervals enter the next highest severity and/or frequency ranges in ascending order. Add
additional columns as necessary to report all intervals used by your institution.
Section 3: currency
Enter the three character ISO standard currency code for the currency in which the severity
data are reported.6
Section 4: Basel business line
If a scenario is associated with exactly one Basel business line, report the Basel Level I
business line. See Table 1 for a list of codes that should be used to refer to Basel business
lines. While the use of the listed codes is preferred, internal codes for the Basel business
lines also will be accepted as long as a key is provided. Report “UA” for unallocated, for
scenarios that are associated with more than one business line, including group-wide
scenarios.
Section 5: internal business line
If an institution cannot provide losses by Basel business line, internal business line should be
provided along with a mapping of internal business lines to Basel business lines on a best
efforts basis.
Section 6: Basel event type
Event types should be reported as one of the Level I Basel event types. See Table 2 for a list
of codes that can be used to refer to event types. While the use of the listed codes is
preferred, internal codes for Basel event types will also be accepted as long as a key is
provided.
Section 7: used in AMA model
Enter “Y” if the scenario is used for AMA capital modelling. Enter “N” if the scenario is used
only for risk management purposes.

6

The standard currency codes are contained in the international standard ISO 4217 and include Euro (EUR),
U.S. Dollar (USD), Japanese Yen (JPY), U.K. Pound Sterling (GBP) and Australian Dollar (AUD).

13/18

C. Individual scenario approach
Submit data for all scenarios used in the measurement and management of operational risk.
Reported scenario data should be the direct output of a workshop or other method used to
generate scenario data and should reflect raw data prior to use in modelling. If reporting the
full scenario data set is infeasible, report information on the following scenarios:
(i)

At least the 20 highest-severity scenarios where at least 20 of these scenarios are
expected to happen once or more every 1,000 years (ie with a mean annual
frequency greater or equal to 0.001). The highest-severity scenarios should be
determined according to the reported loss amounts.

and
(ii)

The five highest-severity scenarios for each of the seven Level I Basel II event types
(if available).

Scenarios that meet the criteria of both (i) and (ii) need not be reported twice.
Section 1: scenario reference number
Provide the institution’s unique internal identifier for each scenario. If an institution does not
have an internal process that assigns reference number, please number scenarios as 1, 2, 3,
etc. Scenarios that have several possible loss amounts with associated frequencies should
be reported in separate rows for each severity and use the same scenario reference number.
Section 2: frequency
Enter the frequency associated with the scenario. Frequency should be reported as the
number of events per year in decimal form (i.e. a 1 in 100 year event would be reported as
.01).
Section 3: loss amount
Enter the estimated loss amount associated with the scenario.
Section 4: currency
Enter the three character ISO standard currency code for the currency in which the severity
data are reported.7
Section 5: Basel business line
If a scenario is associated with exactly one Basel business line, report the Basel Level I
business line. See Table 1 for a list of codes that should be used to refer to Basel business
lines. While the use of the listed codes is preferred, internal codes for the Basel business
lines also will be accepted as long as a key is provided. Report “UA” for unallocated, for

7

The standard currency codes are contained in the international standard ISO 4217 and include Euro (EUR),
U.S. Dollar (USD), Japanese Yen (JPY), U.K. Pound Sterling (GBP) and Australian Dollar (AUD).

14/18

scenarios that are associated with more than one business line, including group-wide
scenarios.
Section 6: internal business line
If an institution cannot provide losses by Basel business line, internal business line should be
provided along with a mapping of internal business lines to Basel business lines on a best
efforts basis.
Section 7: Basel event type
Event types should be reported as one of the Level I Basel event types. See Table 2 for a list
of codes that can be used to refer to event types. While the use of the listed codes is
preferred, internal codes for Basel event types will also be accepted as long as a key is
provided.
Section 8: used in AMA model
Enter “Y” if the scenario is used for AMA capital modelling. Enter “N” if the scenario is used
only for risk management purposes.
Section 9: brief description of loss
Provide a brief description of the scenario loss for at least the 20 highest severity scenarios
that are expected to happen once or more every 1,000 years. To preserve confidentiality,
scenario descriptions should exclude any potential identifiers (such as names of
companies/subsidiaries and locations). Scenario descriptions may be provided in whichever
language is most convenient to the respondent.

15/18

Attachment A: Instructions
Supplemental internal loss data and scenario analysis questionnaire

Questionnaire:

Attachment A

Process/instructions:
The questionnaire includes supplemental questions that give context to the internal loss and
scenario analysis data submitted.
In some questions, please select one response as requested. In other questions, please
select all responses that apply. If your institution has not done work in an area addressed by
a question, please leave the question blank.

16/18

Attachment B: Instructions
Exposure indicators and capital estimates

Questionnaire:

Attachment B

Process/instructions:
Collection of exposure indicators and capital estimates will facilitate cross-bank comparisons
and benchmarking of the LDCE results.
Please note: The data requested in this attachment are solely for use by the participating
institution’s national supervisor and will not be shared outside of an institution’s national
supervisor(s). The national supervisor will use the data to compute various ratios that will be
submitted to the central processor for aggregate analysis. The central processor will create
national/regional (where feasible, based on confidentiality concerns) and international
benchmarks that will be presented on a basis that avoids any disclosure of the
identities of participating institutions.
In some questions, please select one response as requested. In other questions, please
select all responses that apply. If your institution has not done work in an area addressed by
a question, please leave the question blank.

17/18

Attachment C: Instructions
Supplemental range of practice questionnaire

Questionnaire:

Attachment C

Process/instructions:
This is a range of practice questionnaire designed to give context to the other elements of
this exercise by providing information regarding participating institutions’ operational risk
measurement and management practices.
This questionnaire will be completed by either the banking institution or the national
supervisor, at the national supervisor’s discretion. Please contact your national supervisor to
determine the approach selected by your home jurisdiction.
In some questions, please select one response as requested. In other questions, please
select all responses that apply. If your institution has not done work in an area addressed by
a question, please leave the question blank.

18/18