Consultation · HKMA · 22 Dec 2021
This consultation sets out the HKMA's supervisory approach to operational resilience for all authorized institutions (AIs). AIs must develop an operational resilience framework covering identification of critical operations, tolerance for disruption, severe but plausible scenarios, mapping, risk management, testing and incident management. The Board is ultimately responsible for approving the framework and setting tolerance for disruption, with reviews at least annually or after major operational changes. By one year after the final module is issued, AIs must have developed their framework and determined an implementation timeline.
This module should be read in conjunction with the Introduction and with the Glossary, which contains an explanation of abbreviations and other terms used in this Manual. If reading on-line, click on blue underlined headings to activate hyperlinks to the relevant module. ————————— Purpose To set out the HKMA’s supervisory approach to operational resilience and provide AIs with guidance on the general principles which they are expected to consider when developing their operational resilience framework. Classification A non-statutory guideline issued by the MA as a guidance note. Previous guidelines superseded This is a new guideline. Application To all AIs. Structure
1. Definition of operational resilience
2. Operational resilience framework
3. Role of the Board and senior management
4. Determining operational resilience parameters
4.1 Identifying critical operations
4.2 Setting tolerance for disruption
4.3 Identifying severe but plausible scenarios
5. Mapping interconnections and interdependencies underlying critical
operations
6. Preparing for and managing risks to critical operations delivery
7. Testing ability to deliver critical operations under severe but
plausible scenarios
8. Responding to and recovering from incidents
9. Implementation of operational resilience requirements
9.1 Application
9.2 Timeline for implementation
9.3 Supervisory approach
1. Definition of operational resilience
1.1 Operational disruptions (including those due to pandemics, cyber
incidents, technology failures and natural disasters) can affect the viability of individual financial institutions, and in turn, the stability of the wider financial system. This underscores the significance of operational resilience as a supervisory focus and has motivated many regulators around the world and standard setting bodies to issue guidance that aims to improve the operational resilience of financial institutions.
1.2 The Principles for Operational Resilience (POR) issued by the Basel
Committee on Banking Supervision (BCBS) in March 2021 defines operational resilience as the ability of a bank to deliver critical operations through disruption. This ability enables a bank to identify and protect itself from threats and potential failures, respond and adapt to, as well as recover and learn from disruptive events in order to minimise their impact on the delivery of critical operations through disruption.
1.3 The HKMA expects all AIs in Hong Kong to be operationally resilient.
The HKMA will consider an AI to be operationally resilient if it is able to satisfy the following requirements:
Identify and mitigate risks that may threaten delivery of critical operations. In relation to an AI, “critical operations” refers to: (i) activities, processes and services performed by the AI, as well as (ii) the supporting assets (including people, technology, information and facilities) necessary for the delivery of such activities and services, which if disrupted, could pose material risks to the viability of the AI itself or impact the AI’s role within the Hong Kong financial system 1 .
1 These should include any “critical financial functions”, as defined in the Code of Practice “CI-1 Resolution Planning – Core Information Requirements”, that may be performed by the AI.
Continue to deliver critical operations when disruptions occur, including under severe but plausible scenarios. For this purpose, disruptions to an AI’s critical operations must not exceed its “tolerance for disruption”, which is defined as the maximum level of disruption to a critical operation that an AI can accept, and is in practice the point after which further disruption would pose risks to the viability of the AI or impact its role within the Hong Kong financial system. “Severe but plausible scenarios” refers to situations that would result in significant disruptions, and while unlikely to occur, remain probable. Resume normal operations in a timely manner after disruptions occur; and Absorb learnings from disruptions or near-misses to continually improve its ability to prevent, adapt to and recover from risks and disruptions to critical operations delivery.
2. Operational resilience framework
2.1 An AI should develop an operational resilience framework which
enables it to satisfy the requirements detailed in Section 1.3.
2.2 Given the importance of operational resilience for an AI to operate
smoothly and remain viable under extreme scenarios, an AI’s Board of Directors (Board) and senior management are expected to actively participate in establishing, implementing and overseeing the operational resilience framework.
2.3 At a minimum, an AI should include the following components within
its operational resilience framework. Further guidance on how AIs may approach each of these components is provided in the subsequent sections of this module. Mechanism for determining the operational resilience parameters, namely critical operations, tolerance for disruption and severe but plausible scenarios. (Section 4) Mapping exercises which enable an AI to develop a detailed understanding of the interconnections and interdependencies that underlie critical operations delivery, and in turn, identify
what risks or events may affect or disrupt critical operations delivery. (Section 5) Risk management policies and frameworks that help an AI prepare for and manage the various risks to critical operations delivery in an integrated and holistic way. (Section 6) Scenario testing which enables an AI to regularly assess whether it is able to continue delivering critical operations through disruption, including under severe but plausible scenarios. (Section 7) An incident management programme which allows an AI to effectively respond to and manage disruptions to critical operations delivery. (Section 8)
2.4 An AI may determine the most appropriate approach to developing
its operational resilience framework, taking into account its particular circumstances. AIs may refer to Diagram 1 for an illustration of how the different components can be brought together to create a holistic operational resilience framework. It is important to note that developing operational resilience is an iterative process. The process will not always be linear. An AI should actively apply learnings from its implementation of the framework and the management of actual incidents to continually improve on the effectiveness of the framework.
Diagram 1: Step-by-step approach to developing a holistic operational resilience framework
3. Role of the Board and senior management
3.1 The Board should be ultimately responsible for approving an AI’s
operational resilience framework and for overseeing its implementation. When formulating the framework, the Board should take into consideration the AI’s risk appetite. For overseas incorporated AIs, this role should rest with the management team at the head office or the regional headquarters overseeing the Hong Kong operations of the AI.
3.2 Senior management should implement the operational resilience
framework and ensure that sufficient resources (including financial, technological and otherwise) are allocated to this purpose. To facilitate the Board’s oversight, senior management should provide regular and timely reports to the Board on the ongoing operational resilience of the AI’s business units, particularly when significant deficiencies could affect the delivery of the AI’s critical operations.
3.3 The Board and senior management should actively participate in the
setting and review of an AI’s operational resilience parameters. Specifically:
The Board should approve and regularly review: (i) the criteria for determining an AI’s critical operations; and (ii) the actual list of critical operations. The reviews should be conducted no less than annually or when major operational changes occur. The Board is responsible for setting the tolerance for disruption. Assisted by senior management, it should also review the tolerance for disruption at least on an annual basis or when major operational changes occur. Senior management should identify and the Board should approve the severe but plausible scenarios which will be used to review whether an AI is operationally resilient. Both the Board and senior management should regularly review the continued relevance of the scenarios identified.
3.4 The Board bears ultimate responsibility for ensuring that an AI
remains operationally resilient. This would require the Board to take appropriate action to address any deficiencies identified in an AI’s ability to remain within its tolerance for disruption. In the event that there is more than one source of deficiency, the Board should suitably prioritise the remedial actions. As a general principle, the Board should place its focus on making improvements to those areas that would result in larger disruptions, higher risks or are facing more significant deficiencies. For instance, an AI should prioritise a critical operation that would more sooner breach its tolerance for disruption over one that is less time sensitive, or a critical operation that is further away from remaining within its tolerance for disruption over one that is largely within its tolerance for disruption.
3.5 The Board and senior management should regularly review the
suitability and effectiveness of the AI’s operational resilience framework. These reviews are particularly important following operational changes and during the transitory period after an
operational change comes into effect.
3.6 The Board should play an active role in establishing a broad
understanding of the AI’s operational resilience framework. It should clearly communicate the objectives of the framework to all relevant parties, including staff, intragroup entities and third parties. Regular training on the AI’s operational resilience framework should be provided to these parties to reinforce their understanding.
4. Determining operational resilience parameters
4.1 Identifying critical operations
4.1.1 As a first step to developing a sound operational resilience
framework, an AI should identify its critical operations. The number of critical operations identified should be commensurate with the size, nature and complexity of the AI’s operations.
4.1.2 When identifying its critical operations, an AI should take into
consideration a set of defined criteria. These criteria should allow an AI to critically assess whether an operation, if disrupted, would affect:
(a) The AI’s viability. Possible factors to consider include the impact on customers and personnel, and financial, reputational, legal and regulatory implications. (b) The AI’s role in the Hong Kong financial system. Possible factors to consider include how disruptions may affect specific market roles played by the AI (e.g. note issuance or clearing) as well as relationships with counterparties in the interbank market. For the avoidance of doubt, while the set of criteria defined by AIs for identifying critical operations should encompass elements of both (a) and (b) above, a given operation need not impact both (a) and (b) in order for it to be classified as a critical operation.
4.1.3 In the process of identifying its critical operations, an AI may,
where appropriate, leverage on relevant concepts covered within its recovery and resolution plans.
4.2 Setting tolerance for disruption
4.2.1 A tolerance for disruption should be set for each critical
operation. It should include at least a time-based metric, but may also include a combination of other quantitative (e.g. volume or value of transactions) and qualitative metrics (e.g. reputational or legal implications).
4.2.2 In setting the tolerance for disruption, consideration should
be given to an AI’s operational capabilities given a broad range of severe but plausible scenarios that would affect its critical operations. AIs should be aware that their operational capabilities may vary during different business cycles or as a result of seasonal factors. For instance, during the periods of time when more initial public offerings are launched, an AI’s trading systems are more likely to come under stress, which could weaken the AI’s ability to respond under severe but plausible scenarios.
4.3 Identifying severe but plausible scenarios
4.3.1 AIs should identify a range of scenarios of different nature,
severity and duration relevant to its business and risk profile. Examples of scenarios that AIs may consider include, but are not limited to, pandemics, natural disasters, and failures or disruptions at a third party or within the third party’s supply chain.
4.3.2 When identifying the scenarios, AIs should make reference
to previous incidents or near misses within the institution or across financial sectors, as well as in other sectors or jurisdictions, or any situations that could result in significant disruptions given the changing operational landscape.
5. Mapping interconnections and interdependencies
underlying critical operations
5.1 The appropriate functions within an AI should identify and
document: (i) the people, processes, technology, information, facilities; and (ii) the interconnections and interdependencies among these factors that are necessary for the AI to deliver its critical
operations. When considering (ii), an AI should also include those interconnections and interdependencies that depend on third parties and intragroup arrangements.
5.2 The approach and level of granularity of mapping should be
sufficient to enable the AI to identify vulnerabilities and facilitate the testing of the AI’s ability to deliver critical operations through disruptions. AIs should also consider whether the approach adopted for mapping under its operational resilience framework is appropriately harmonised with that adopted for recovery and resolution planning purposes.
5.3 The mapping documentation should be prepared in a way that is
proportionate to the AI’s size, scale and complexity. It should also be usable by all relevant parties in the event of disruptions.
5.4 AIs are expected to update their mapping documentation on a
regular basis, but no less than annually or following any material changes to their operations.
6. Preparing for and managing risks to critical operations
delivery
6.1 AIs should be prepared to manage all risks with potential to affect
critical operations delivery. As a given critical operation may face a number of risks, AIs should leverage different risk management frameworks, as appropriate, to offer holistic and comprehensive support to the critical operation.
6.2 The HKMA expects that AIs should, at a minimum, take into
consideration the following risk management components with respect to operational resilience:- Operational risk management: As operational risk management focuses on preventing and minimising operational losses, it contributes to an AI’s efforts to maintain operational resilience. Operational risk management should therefore be considered as a crucial element of an effective operational resilience framework. Business continuity planning and testing: Business continuity planning and testing supports an AI’s ability to prepare for and
recover from emergencies or disasters, and therefore contributes to an AI’s ability to continue delivering its critical operations through disruptions. Accordingly, AIs should ensure that their critical operations are subject to appropriate business continuity planning and testing arrangements. Third-party dependency management: As AIs increasingly engage third parties or intragroup entities for the provision of services or delivery of functions, they must ensure that disruptions at these entities will not affect critical operations delivery. To ensure potential risks to critical operations are minimised, AIs should manage their dependencies on third parties and intragroup entities as they would with outsourcing arrangements. Prior to entering into arrangements that support the delivery of critical operations, an AI should verify whether the relevant third parties or intragroup entities have at least equivalent level of operational resilience to that of the AI. During the course of engagement, an AI should have adequate arrangements in place to continually satisfy itself that the third party or intragroup entity has maintained an acceptable level of operational resilience. In addition, an AI should develop appropriate business continuity and contingency planning procedures and exit strategies to maintain its operational resilience in the event of a failure or disruption at a third party or intragroup entity which may impact its delivery of critical operations. An AI should not enter into, or continue, any third party or intragroup arrangements that may weaken the operational resilience of the AI’s critical operations. Information and Communication Technology (ICT) including cyber security: Growing technology adoption raises the likelihood that an AI’s critical operations may depend or may be affected by lapses in ICT risk management. To minimise risks in this regard, AIs should have in place an ICT policy which covers cyber security, as well as arrangements for ensuring the confidentiality, integrity and availability of critical
information assets.
6.3 AIs should note that most of the risk management considerations
associated with operational resilience are not new, and are already covered by existing HKMA guidance. These include but are not limited to: Supervisory Policy Manual (SPM) modules “TM-G-1 General Principles for Technology Risk Management”, “TM-G-2 Business Continuity Planning”, “OR-1 Operational Risk Management”, “SA-2 Outsourcing”, as well as “Cyber Resilience Assessment Framework 2.0”. AIs should refer to and ensure that they are compliant with the supervisory requirements contained therein.
7. Testing ability to deliver critical operations under severe
but plausible scenarios
7.1 AIs should conduct regular testing of their operational resilience
framework to ensure that they are able to continue delivering their critical operations through disruptions, including under severe but plausible scenarios.
7.2 When considering the testing requirement, AIs should take into
account the following:
The testing exercises should include realistic assumptions, and should encompass the AI’s interconnections and interdependencies, including those through relationships with intragroup entities and third parties. The frequency of testing should be determined based on a variety of factors, including the potential impact of a disruption, how many critical operations an AI has, and whether the operating environment has materially changed. Different types of testing (e.g. paper-based, simulations or live-systems testing) serve different purposes and AIs should deploy the most appropriate type of testing based on the nature or needs of the specific testing exercise. An AI should also consider and carefully manage the risks that may be introduced by the testing itself. AIs should deploy staff with appropriate expertise to conduct
the testing. The testing approach should dictate the type of staff involved, including their seniority, qualifications as well as the function (e.g. first, second or third line of defence) from which they are sourced. AIs should consider how they may leverage the testing exercises to enhance their staff’s operational resilience awareness and readiness to operate during disruptions, thereby improving their ability to effectively adapt and respond to different types of disruptive events.
7.3 Where practicable, AIs may leverage on existing testing
arrangements, including those devised for business continuity planning purposes, to fulfill the testing requirement relating to operational resilience. An AI should be able to demonstrate how an existing testing exercise enables it to achieve the specific objectives of scenario testing for operational resilience purposes.
7.4 After each testing exercise, an AI should prepare a formal testing
report to record any gaps or weaknesses identified, as well as document the remedial actions planned. The reports should be reviewed by the AI’s senior management.
8. Responding to and recovering from incidents
8.1 While an AI should dedicate adequate efforts to preventing
disruptions, it should recognise that disruptions will occur no matter how robust its operational resilience framework is. An AI should therefore be prepared to manage and recover from incidents.
8.2 Specifically, an AI should establish an effective incident
management programme to manage all incidents, especially those that may impact its critical operations. The programme should cover those incidents that may arise due to dependencies, including those on third parties and intragroup entities.
8.3 The incident management programme should capture the full lifecycle of any incidents and involve:
Classification of an incident’s severity based on predefined criteria. This should enable the AI to prioritise and allocate resources to respond to an incident.
Incident response and recovery procedures. These should be reviewed, tested and updated on a regular basis. Their connection to the AI’s business continuity, disaster recovery and other associated management plans and procedures should also be clearly documented. Communication plans for reporting incidents to both internal and external stakeholders. Communication should take place during the incident (e.g. to provide performance metrics), and after, including to convey analysis of lessons learned. Root cause analysis of incidents to help with the prevention or minimisation of recurrence.
8.4 The incident management programme should be supported by an
inventory of internal and third party resources to enable prompt incident response and recovery. It should also reflect the lessons learned from previous incidents, including those experienced by others.
8.5 AIs should note that the above requirements complement existing
HKMA guidance on incident management. These include but are not limited to SPM modules “TM-G-2 Business Continuity Planning” and “TM-G-1 General Principles for Technology Risk Management”, and the HKMA’s circular on “Incident Response and Management Procedures” issued in June 2010. AIs should review relevant materials and ensure that they are compliant with the supervisory requirements contained therein.
9. Implementation of operational resilience requirements
9.1 Application
9.1.1 The requirements contained in this module apply to all AIs.
Locally incorporated AIs should endeavour to implement the guidance of this module with respect to their subsidiaries and overseas operations, and for overseas incorporated AIs with respect to their operations in Hong Kong.
9.1.2 In line with the HKMA’s risk-based approach to supervision,
AIs are expected to implement the requirements in a
proportionate manner and develop an operational resilience framework that is “fit for purpose”, i.e. commensurate with its nature, size, complexity and risk profile.
9.2 Timeline for implementation
9.2.1 By [1 year after the date upon which the final module is
issued], the HKMA expects an AI to have:
(a) Developed its operational resilience framework; and (b) Determined the timeline by which it will have implemented the operational resilience framework, and become operationally resilient.
9.2.2 For the purposes of 9.2.1(a), AIs are expected to have
identified the operational resilience parameters and commenced a basic programme of mapping. The latter will be crucial to ensuring that an AI adequately understands the interconnections and interdependencies that underlie its critical operations, and in turn, is able to develop the other components of its operational resilience framework, including to identify the specific types of risks to critical operations delivery that need to be addressed, as well as how to most suitably conduct testing. The HKMA recognises that AIs may not be able to produce mapping that reaches the full level of sophistication at the initial stage, and instead, would expect AIs to make continual improvements as they obtain more experience in implementing their operational resilience frameworks.
9.2.3 Given the importance of operational resilience, the HKMA
expects AIs to become operationally resilient as soon as practicable. That said, the HKMA also recognises that becoming operationally resilient is a resource-intensive exercise (for reasons including that it involves mapping exercises which may be more complex for larger AIs, and could involve substantial system changes). Taking into consideration the need to accommodate AIs of different size and complexity, the HKMA has decided to allow AIs up to 2 years to become operationally resilient. In other words, the
timeline specified under Section 9.2.1(b) should not extend beyond 2 years from [1 year after the date upon which the final module is issued.] After this point in time, an AI will be expected to have fully implemented its operational resilience framework, including to have conducted scenario testing, and be able to satisfy the requirements in Section 1.3. Notwithstanding the 2-year time limit, AIs are encouraged to become operationally resilient as soon as their circumstances allow. The HKMA will engage in active discussions with AIs to review the suitability of their proposed timelines.
9.3 Supervisory approach
9.3.1 Following its risk-based supervisory approach, the HKMA will
assess the effectiveness of the operational resilience frameworks of AIs through a combination of risk-focused onsite examinations, off-site reviews and prudential meetings. Where needed, AIs may be required to submit selfassessments of their ability to remain operationally resilient. ————————— Contents Glossary Home Introduction
This module should be read in conjunction with the Introduction and with the Glossary, which contains an explanation of abbreviations and other terms used in this Manual. If reading on-line, click on blue underlined headings to activate hyperlinks to the relevant module. ————————— Purpose To set out the HKMA’s supervisory approach to business continuity planning and the sound practices which the HKMA expects AIs to take into consideration in this regard Classification A non-statutory guideline issued by the MA as a guidance note Previous guidelines superseded TM-G-2 “Business Continuity Planning” (V.1) dated 02.12.02 This is a new guideline. Application To all AIs Structure
1. Introduction
1.1 Terminology
1.2 Scope of business continuity planning
1.3 Supervisory approach
2. Board and senior management oversight
2.1 Establishment of policy, process and responsibility
2.2 Monitoring and reporting
3. Business impact analysis and recovery strategy
3.1 Business impact analysis
3.2 Recovery strategy formulation
4. Development of Business Continuity Plan
4.1 Overview
4.2 Crisis management process
4.3 Business resumption
4.4 Technology recovery
4.5 Business continuity models
4.6 Vital record management
4.7 Public relations and communication strategy
4.8 Other risk mitigating measures
5. Alternate sites for business and technology recovery
5.1 Selection criteria for alternate sites
5.2 Alternate sites for technology recovery
5.3 Alternate sites provided by vendors or other institutions
6. Implementation of Business Continuity Plan
6.1 Testing and rehearsal
6.2 Periodic maintenance
—————————
1. Introduction
1.1 Terminology
1.1.1 In this module:
“business continuity planning” refers to the advance planning and preparations which are necessary to identify the impact of potential losses arising from an emergency or a disaster; to formulate and implement viable recovery strategies; to develop recovery plans which ensure continuity of an AI’s operationsservices in that relation; and to administer a comprehensive testing and maintenance programme; “Business Continuity Plan (“BCP”)” refers to a collection of procedures and information which is
developed, compiled and maintained in readiness for use in the event of an emergency or disaster; “business impact analysis” refers to a management level analysis which identifies and assesses the impact of losing the various functions and services within an AI. The impact analysis tries to measure the potential loss and escalating losses over time in order to provide senior management with reliable data for the identification of critical operationsservices. Based on the results of the analysis, the AI should be able to identify the scope of the critical operationsservices to be provided and the time-frame in which the operationsservices should be resumed; “call-out tree” refers to a pre-defined sequence of points of contact of staff for dissemination of information; “crisis management team (“CMT”)” refers to a group of executives who would direct the recovery operations while taking responsibility for the survival and the reputation of the AI; “Critical operations” refers to: (i) activities, processes, and services performed by an AI, as well as (ii) the supporting assets (including people, technology, information and facilities) necessary for the delivery of such activities and services, which if disrupted, could pose material risks to the viability of the AI itself or impact the AI’s role within the Hong Kong financial system 1 . AIs should refer to “OR-2 Operational Resilience” for more details about this term. “crisis management” refers to the overall process designed to support the CMT when dealing with a specific emergency situation which might threaten the operations, staff, customers or reputation of an AI; and
1 These should include any “critical financial functions”, as defined in the Code of Practice “CI-1 Resolution Planning – Core Information Requirements”, that may be performed by the AI.
“recovery strategy” refers to a strategy to resume the minimum set of critical operationsservices identified in the business impact analysis (e.g. use of another delivery channel to provide the same operationservice). “Severe but plausible scenarios” refers to situations that would result in significant disruptions, and while unlikely to occur, remain probable. AIs should refer to “OR-2 Operational Resilience” for more details about this term. “Tolerance for disruption” refers to the maximum level of disruption to a critical operation that an AI can accept. AIs should refer to “OR-2 Operational Resilience” for more details about this term.
1.2 Scope of business continuity planning
1.2.1 This module should be read in conjunction with “OR-2
Operational Resilience” with respect to the purposes of ensuring critical operations delivery and AIs should ensure that they are compliant with the relevant requirements specified therein.
1.2.11.2.2 The destruction of buildings, loss of life, and the
widespread dislocations to financial institutions’ operations resulting from the incident of 11 September, 2001 (“9/11”) have prompted many institutions to review their scope of business continuity planning. It is clear that the traditional scope of business continuity planning for inaccessibility of a single building for a short period is not adequate.
1.2.2 Efforts put into, and experiences gained from the
preparation of, the Y2K contingency plan were obviously not sufficient to cope with 9/11. Y2K was a known event and was essentially a software problem. Also, it did not raise the issue of destruction of people and property.
1.2.3 The HKMA recognises that BCPs involve a cost, and that
it may not be cost effective to have a fully developed and implemented plan for all the worst case scenarios. However, having regard to past events, it would seem sensible for AIs to plan on the basis that they may have to cope with the complete destruction of buildings and surrounding infrastructure in which their key offices,
installations, counterparties or service providers are located, the loss of key personnel, and the situation that back-up facilities might need to be used for an extended period of time.
1.2.4 AIs may find it useful to consider two-tier plans: one to
deal with near-term problems, which would be fully developed with the physical capacity to put it into immediate effect and the other, which might be in paper form, to deal with a longer-term scenario (e.g. how to lease additional premises and how to accommodate processes that might not be critical immediately but would become so over time).
1.2.5 Depending on the individual circumstances of AIs, the
longer-term plan may include plans on how to reconstruct the primary sites or to move to a new permanent work location. For example, this may require that duplicates of design documents, floor plans and cabling diagrams should be kept off-site.
1.3 Supervisory approach
1.3.1 The HKMA’s supervisory objective is to help ensure that
AIs have workable and well thought through BCPs to protect all the critical areas of their business and to cope with prolonged disruptions.
1.3.2 The HKMA will, in the course of its on-site examinations,
off-site reviews and prudential meetings with AIs, determine as appropriate the adequacy of their efforts being put into business continuity planning. In assessing the adequacy of AIs’ BCP, the HKMA will have regard to the practices set out in this module.
1.3.3 AIs should inform the HKMA promptly if their BCP is
activated. The HKMA should also receive periodic progress reports upon being notified until the final resolution of the crisis.
2. Board and senior management oversight
2.1 Establishment of policy, process and responsibility
2.1.1 The Board of Directors 2 and senior management of AIs
have the ultimate responsibility for business continuity planning and the effectiveness of their BCP. The senior management should establish policies, standards and processes for business continuity planning 3 , which should be endorsed by the Board. The senior management should ensure that business continuity planning is taken seriously by all levels of staff and that sufficient resources are devoted to implementing the plan.
2.1.2 In this module, the suggested process for business
continuity planning consists of several key components as shown in the following diagram:
Business impact analysis
Business impact analysis
Implementation and Maintenance of BCP of BCP Development of BCP of BCP Establishment of alternate sites Recovery strategy formulation Recovery strategy formulation Board and senior management oversight Consideration of prolonged disruptions and catastrophic disasters The senior management should establish clearly which function in the institution has the responsibility for managing the entire process of business continuity planning (the BCP function).
2.2 Monitoring and reporting
2 For the purpose of this module, the responsibility of Board oversight of business continuity planning for overseas incorporated AIs in respect of Hong Kong operation should rest with the local senior management. In developing the policies, standards and processes for business continuity planning, AIs may find it useful to draw additional reference from some of the BCP resources, e.g. the Business Continuity Institute: “www.thebci.org” and the Disaster Recovery Institute:
“www.drii.org”.
2.2.1 The BCP function should submit regular reports to the
Board and senior management on the testing of its BCP. Any major changes to the BCP should also be reported to the senior management.
2.2.2 The internal audit function of an AI should conduct
periodic review of its BCP to determine whether the plan is realistic and remain relevant, and whether it adheres to the policies and standards established by the AI.
2.2.3 Given the importance of business continuity planning,
the Chief Executive of AIs should prepare and sign-off a formal annual statement submitted to the Board on whether the recovery strategies adopted are still valid and whether the documented BCPs are properly tested and maintained. The annual statement should be incorporated into the BCP and will be reviewed as part of the HKMA’s on-site examinations.
3. Business impact analysis and recovery strategy
3.1 Business impact analysis
3.1.1 The objective of the business impact analysis is to
identify different kinds of risks to business continuity and to quantify the impact of disruptions. The business impact analysis helps to identify those critical business activities, banking services and internal support functions which, in the event of a disaster, must be consistently and effectively delivered by an AI.
3.1.2 The business impact analysis normally comprises two
stages. The first stage is to identify critical operationsservices that must be maintained and continued in the event of a disaster. This usually entails an assessment of the overall exposure to the AI if the normal functions or services cannot be performed. The criteria for the assessment include the impact on customers, personnel, reputation and internal services as well as the financial and legal implications. The second stage is a time-frame assessment. It aims to determine how quickly the AI needs to resume the critical functions or servicesoperations identified, taking into account the tolerance for disruption set by the AI.
3.1.3 Based on the business impact analysis, the business
and support functions should be able to define the minimum level of critical operationsservices to be delivered in the event of a disaster.
3.2 Recovery strategy formulation
3.2.1 Individual critical business and support functions should
formulate their own recovery strategies on how to achieve the recovery time-frame and to deliver the minimum level of critical operationsservices derived from the business impact analysis. This involves determination of an alternate site, total number of recovery personnel and the related workspace, applications and technology requirements, office facilities and vital records required for the provision of such levels of operationsservices. AIs should take note that they might need to cater for processing volumes that exceed those under normal circumstances.
3.2.2 The result of the time-frame assessment in the business
impact analysis is the key determination factor for the recovery priority of individual services. The interdependency among critical operationsservices is another major consideration in determining the recovery strategies and priority. For example, the recovery of the front office operations is highly dependent on the recovery of the middle office and back office support functions.
3.2.3 Having performed the business impact analysis and
formulated the recovery strategies, individual critical business and support functions should have established the minimum BCP requirements for the provision of essential business and technology services levels. To avoid any unnecessary arguments and inappropriate BCP investment at a later stage, these BCP requirements should be approved by the senior management prior to proceeding to the development of the BCP. In addition, senior management should also ensure that the business continuity requirements should be considered at the planning and development stages of new business products and services.
4. Development of Business Continuity Plan
4.1 Overview
4.1.1 Once the recovery strategies for individual business and
support functions are determined and the BCP requirements are finalised, the development of the BCP should commence. The objective of the BCP is to provide detailed guidance and procedures to respond to and manage a crisis, to resume and continue critical operationsbusiness services and functions identified in business impact analysis, and to ultimately return to business as usual.
4.1.2 An effective BCP is forward-looking, and should be
validated for a range of severe but plausible scenarios which contain disruptive events and incidents. The BCP should identify critical operations as well as the key internal and external dependencies supporting these critical operations. It should incorporate business impact analysis, recovery strategies, testing programmes, training and awareness programmes, communication strategies and crisis management processes. An AI’s BCP for the delivery of critical operations, including those reliant on critical third-party services, should be consistent with its operational resilience framework. The same consistency requirement also applies to BCPs which may be contained within an AI’s recovery and resolution plans.
4.1.14.1.3 AIs should perform documentation of their BCP
along with the development of the recovery organisation, procedures and arrangements. The BCP documentation and development should not be treated as two unrelated and independent projects. Otherwise, it may be difficult to ensure that the content of the BCP is consistent with the actual recovery processes and arrangements. The BCP development process is summarised below.
4.2 Crisis management process
4.2.1 It is important to note that a disaster will evolve after
occurrence. AIs should establish a CMT to respond to and manage the various stages of a crisis. The CMT should comprise members of the senior management and heads of major support functions (e.g. building facilities, IT, corporate communications and human resources).
4.2.2 BCPs should set out a crisis management process that
serves as documented guidance to assist senior management in dealing with and containing an emergency to avoid spillover effects to the business as a whole. Senior management should identify potential crisis scenarios and where applicable develop specific crisis management procedures for managing these scenarios (i.e. a procedure to handle a bomb threat is different from that to handle a major power failure). The overall crisis management process, at a minimum, should contain the following:
the process for ensuring early detection of an emergency or a disaster and prompt notification to the CMT about the incident; the process for establishing the roles and responsibilities for managing operational disruptions and clear guidance regarding the succession of authority in the event of a disruption that impacts key personnel; the process for the CMT to assess the overall impact on the AI and to make quick decisions on the appropriate responses for action (i.e. staff safety, incident containment and specific crisis management procedures); the arrangements for safe evacuation from business locations (e.g. directing staff to prearranged emergency assembly area, taking attendance of all employees and visitors and tracking missing people through different means immediately after the disaster); clear internal decision-making process andclear criteria for activation of the BCP and/or alternate sites; the process for gathering updated status information for the CMT (e.g. ensuring that regular conference calls are held among key staff from relevant business and support functions to report on the status of the recovery process); the process for timely internal and external communications (see subsection 4.7 below); and
the process for overseeing the recovery and restoration efforts of the affected facilities and the business services.
4.2.3 If CMT members need to be evacuated from their
primary business locations, AIs should set up command centres to provide the necessary workspace and facilities for the CMT. Command centres should be sufficiently distanced from AIs’ primary business locations to avoid being affected by the same disaster.
4.3 Business resumption
4.3.1 Each relevant business and support function should
establish a business recovery team which may have subteams to carry out the business resumption process. Appropriate recovery personnel with the required knowledge and skills should be assigned to the teams. AIs should ensure that alternate recovery personnel are identified for all critical processes. Contact numbers for recovery and alternate personnel, including contact information after office hours, should be available for an emergency (e.g. as wallet cards). AIs should also consider a staff rotation plan in order to cover extended working hours for business recovery.
4.3.2 Generally, the business resumption process consists of
three major phases:
The mobilisation phase – This phase aims to notify the recovery teams (e.g. via a call-out tree) and to secure the resources (e.g. recovery services provided by vendors) required to resume business services. This phase might involve determination of the sequence for restoring business services if it needs to be different from the pre-determined sequence in the BCP; The alternate processing phase – This phase emphasises the resumption of the business and service delivery at the alternate site and/or in a different way than the normal process. This may entail record reconstruction and verification, establishment of new controls, alternate manual processes, and different ways of dealing with customers and counterparties; and
The full recovery phase – This phase refers to the process for moving back to a permanent site after a disaster. It is as difficult and critical to the business as the process to activate the BCP.
4.3.3 For the first two phases, clear responsibilities should be
established and activities prioritised. A recovery tasks checklist should be developed and included in the BCP. It is recognised that certain tasks involved in the full recovery phase may depend on the nature of the disaster concerned and that it may be difficult to formulate detailed plans in advance. However, the BCP should at least identify and plan for activities which would be required in any events, for example, the verification of the safety and readiness of the permanent site.
4.4 Technology recovery
4.4.1 Business resumption very often relies on the recovery of
technology resources that include applications, hardware equipment and network infrastructure as well as electronic records. The technology requirements that are needed during recovery for individual business and support functions should be specified when the recovery strategies for the functions are determined.
4.4.2 AIs should pay attention to the resilience of critical
technology equipment and facilities such as the Uninterruptible Power Supply (“UPS”) and the cooling systems. Such equipment and facilities should be subject to continuous monitoring and periodic maintenance and testing. This would reduce the probability of having to activate the BCP and the inevitable disruptions to normal business.
4.4.3 Appropriate personnel should be assigned with the
responsibility for technology recovery. Alternate personnel needs to be identified for key technology recovery personnel in case of their unavailability to perform the recovery process.
4.5 Business continuity models
4.5.1 There are various business continuity models that could
be adopted by AIs to handle prolonged disruptions. The traditional model is an “active/back-up” model, which is widely used by many organisations. This traditional
model is based on an “active” operating site with a corresponding alternate site (back-up site), both for data processing and for business operations. This model may require significant investment if the “back-up site” needs to be equipped to support for prolonged disruptions of the “active site”.
4.5.2 An emerging split operations model, which has already
been used by some institutions, is a different business continuity model. This model is to operate with two or more widely separated active sites for the same critical operations, providing inherent back-up for each other (e.g. call centres for customer services). Each site has the capacity to take up some or all of the work of another site for an extended period of time. This strategy can provide nearly immediate resumption capacity and is normally able to handle the issue of prolonged disruptions.
4.5.3 The split operations model may incur higher operating
costs, in terms of maintaining excess capacity at each site and added operating complexity. It may be difficult to maintain appropriately trained staff and pose technological issues at multiple sites.
4.5.4 The question of what business continuity model to adopt
is for individual institutions’ judgement based on the risk assessment of their business environment and the characteristics of their own operations.
4.6 Vital record management
4.6.1 Each BCP should clearly identify information deemed
vital for recovery of critical business and support functions in the event of a disaster and the relevant protection measures. Vital information includes that stored on both electronic or non-electronic media (e.g. paper records).
4.6.2 Copies of vital records should be stored off-site as soon
as possible after creation. Back-up vital records must be readily accessible for emergency retrieval. Access to back-up vital records should be adequately controlled to ensure that they are reliable for business resumption purposes. For certain critical operationsservices, AIs should consider the need for instantaneous data back-up (e.g. adopting real-time data mirroring technology) to
ensure prompt system and data recovery. There should be clear procedures indicating how and in what priority vital records are to be retrieved or recreated in the event that they are lost, damaged or destroyed.
4.7 Public relations and communication strategy
4.7.1 AIs should formulate a formal strategy for communication
with key external parties (e.g. regulators, investors, customers, counterparties, business partners, service providers, the media and other stakeholders). The strategy needs to set out to which parties AIs should communicate in the event of a disaster. This will ensure that consistent and up-to-date messages are conveyed to the relevant parties. During a disaster, ongoing and clear communication is likely to assist in maintaining the confidence of customers and counterparties as well as the public in general.
4.7.2 The BCP should clearly indicate who can speak to the
media, and have pre-arrangements for redirecting external communications to designated staff during a disaster. AIs may find it helpful to prepare draft press releases as part of their BCP. This will save the CMTs’ time in determining the main messages to convey in a chaotic situation. Important conversations with external parties should be properly logged for future reference. Important contact numbers and e-mail addresses of key external parties should be kept in a readily accessible manner (e.g. in wallet cards or AIs’ intranet).
4.7.3 As regards internal communication, the BCP should set
out how the status of recovery can be promptly and consistently communicated to all staff, parent bank, head office, branches and subsidiaries, where appropriate. This may entail the use of various communication channels (e.g. broadcasting of messages to mobile phones of staff, AIs’ websites, e-mails, intranet and instant messaging).
4.8 Other risk mitigating measures
4.8.1 AIs should have proper insurance coverage to reduce
the financial exposure that they may face during a disaster. AIs should regularly review the adequacy and coverage of their insurance policies in reducing any
foreseeable risks caused by disasters, such as loss of offices, critical IT facilities and equipment, and casualty. Insurance policies may also need to address AIs’ legal responsibilities for failing to deliver services to their customers and counterparties.
4.8.2 AIs should also incorporate the possible need to obtain
additional liquidity into their BCPs.
5. Alternate sites for business and technology recovery
5.1 Selection criteria for alternate sites
5.1.1 Most business continuity efforts are dependent on the
availability of an alternate site (i.e. recovery site) for successful execution. The alternate site may be either an external site available through an agreement with a commercial vendor or a site within the AI’s real estate portfolio. A useable, functional alternate site is an integral component of all BCPs.
5.1.2 AIs should examine the extent to which key business
functions are concentrated in the same or adjacent locations and the proximity of the alternate sites to primary sites. Alternate sites should be sufficiently distanced to avoid being affected by the same disaster (e.g. they should be on separate or alternative telecommunication networks and power grids).
5.1.3 AIs’ alternate sites should be readily accessible and
available for occupancy (i.e. 24 hours a day, 7 days a week) within the time requirement specified in their BCPs. Should the BCPs so require, the alternate sites should have pre-installed workstations, power, telephones and ventilation, and sufficient space. Appropriate physical access controls such as access control systems and security guards should be implemented in accordance with AIs’ security policy.
5.1.4 Other than the establishment of alternate sites, AIs
should also pay particular attention to the transportation logistics for relocation of operations to alternate sites. Consideration should be given to the impact a disaster may have on the transportation system (e.g. closures of roads or tunnels). Some staff may have difficulty in commuting from their homes to the alternate sites. Other
logistics, such as how to re-route internal and external mail to alternate sites should also be considered. Moreover, pre-arrangement with telecommunication companies for automated telephone call diversion from the primary work locations to the alternate sites should be considered.
5.2 Alternate sites for technology recovery
5.2.1 Alternate sites for technology recovery (i.e. back-up data
centres), which may be separate from the alternate business site, should have sufficient technical equipment (e.g. workstations, servers, printers, etc.) of appropriate model, size and capacity to meet recovery requirements as specified by AIs’ BCPs. The sites should also have adequate telecommunication (including bandwidth) facilities and pre-installed network connections as specified by their BCPs to handle the expected voice and data traffic volume.
5.2.2 AIs should consider arranging telecommunication links
from their alternate sites to the alternate sites of major customers, counterparties and service providers whose primary sites are close to AIs’ primary business locations and who may therefore be affected by the same disaster being catered for. Priority should be given to establishing telecommunication links to those parties upon which AIs’ critical operationsservices have a high dependency.
5.3 Alternate sites provided by vendors or other institutions
5.3.1 AIs should avoid placing excessive reliance on external
vendors in providing BCP support, particularly where a number of institutions are using the services of the same vendor (e.g. to provide back-up facilities or additional hardware). AIs should satisfy themselves that such vendors do actually have the capacity to provide the services when needed and the contractual responsibilities of the vendors should be clearly specified.
5.3.2 The contractual terms should include the lead-time and
capacity that vendors are committed to deliver in terms of back-up facilities, technical support or hardware. In some cases, a retainer agreement may be advisable to
ensure priority service from the vendors in the face of competing demands from other affected users. The vendor should be able to demonstrate its own recoverability including the specification of another recovery site in the event that the contracted site becomes unavailable.
5.3.3 Certain AIs may rely on a reciprocal recovery
arrangement with another institution to provide recovery capability. AIs should, however, note that such arrangement is often not appropriate for prolonged disruptions and an extended period of time. This arrangement could also make it difficult for AIs to adequately test their BCP. Any reciprocal recovery agreement should therefore be subject to proper risk assessment and documentation by AIs, and formal approval by the Board.
6. Implementation of Business Continuity Plan
6.1 Testing and rehearsal
6.1.1 AIs should not consider their BCP as complete if the
plans have not been subject to proper periodic testing. Testing is needed to ensure that the BCP is operable. Testing of BCP should be conducted and validated for a range of severe but plausible scenarios that incorporate disruptive events and incidents. Testing entails verifying the awareness and preparedness of AIs’ personnel as well as determining how well the BCP really works.
6.1.2 AIs are expected to conduct testing of their BCP at least
annually. Senior management should participate in the annual testing and be aware of what they are personally required to do in the event of their BCP being invoked. In addition, both recovery and alternate personnel should participate in plan rehearsals to familiarise themselves with their recovery responsibilities.
6.1.3 All BCP related risks and assumptions must be reviewed
for relevancy and appropriateness as part of the annual planning of testing. The scope of testing should be comprehensive to cover the major components of the BCP as well as coordination and interfaces among important parties, and interdependencies with third
parties and intragroup entities. Depending on the testing objectives and parties involved, the types of BCP testing may include a desk-top structured walkthrough, a testing of particular components of the BCP or a fully integrated testing 4 . In particular:
staff evacuation and communication arrangements (e.g. call-out trees) should be validated; the alternate sites for business and technology recovery should be activated; important recovery services provided by vendors or counterparties should form part of the testing scope; AIs should consider testing the linkage of their back-up IT systems with the primary and back-up systems of key customers, counterparties and service providers; if back-up facilities are shared by other parties (e.g. subsidiaries of the institution), the AI needs to verify whether all parties can be accommodated concurrently; and recovery of vital records should be certified as
part of the testing.
6.1.4 Formal testing documentation (including testing plan,
testing scenarios, testing procedures and testing results) should be produced to ensure thoroughness and effectiveness of testing. Specifically, a post mortem review report should be prepared at the completion of the testing for formal sign-off by AIs’ senior management. If the testing results indicate a weakness or gap in the BCP, the plans and recovery strategies should be updated to remedy the situation.
6.2 Periodic maintenance
6.2.1 AIs should have formal change management procedures
to keep their BCPs updated in respect of any relevant
In a fully integrated testing, every facet of the BCP is tested together. This test is usually performed during non-business hours to avoid disrupting normal operations and customers.
changes with proper approval and documentation. In the event that a plan has been activated, a review should be carried out once normal operations are restored to identify areas for improvement. If vendors are needed to provide vital recovery services, there should be formal processes for regular (say, annual) reviews of the appropriateness of the relevant service level agreements.
6.2.2 Individual business and support functions, with the
assistance of the BCP function, should review their business impact analysis and recovery strategy, say on an annual basis. This aims to confirm the validity of, or whether updates are needed to, the BCP requirements (including the technical specifications of equipment of the alternate sites) for the changing business and operating environment.
6.2.3 The contact information for key staff, counterparties,
customers and service providers should be updated as soon as possible when notification of changes is received.
6.2.4 Significant internal changes (e.g. merger or acquisitions,
business re-organisation or departure of key personnel) should be reflected in the plan immediately and reported to senior management.
6.2.5 Copies of the BCP document should be stored at
locations separate from the primary sites. A summary of key steps to take in an emergency should be made available to senior management and other key personnel and kept by them in multiple locations (e.g. office, home, briefcase or AI’s website). ————————— Contents Glossary Home Introduction
Note from RegAlert. AI assistants can read this document in full, and search 70,000+ more, through the RegAlert MCP connector (https://mcp.regalert.today/mcp). Free with an account. How to connect ChatGPT, Claude or Cursor.
The documents this one cites, and the documents that cite it.
“Previous guidelines superseded TM-G-2 “Business Continuity Planning” (V.1) dated 02.12.02”
“These include but are not limited to: Supervisory Policy Manual (SPM) modules ... “TM-G-2 Business Continuity Planning”...”
“These include but are not limited to: Supervisory Policy Manual (SPM) modules “TM-G-1 General Principles for Technology Risk Management”...”
This document amends: Business Continuity Planning (Consultation)
More like this from HKMA
HKMA published 5 documents in the last 30 days. We email you each new one the day it's published.
Source: Hong Kong Monetary Authority — original document · Summary generated with machine assistance and checked by an AI reviewer before publication; the authoritative text is the regulator's original document. How RegAlert works
In RegAlert: added · updated