Supervisory Policy Manual IR-1: Interest Rate Risk in the Banking Book

Supervisory Policy Manual
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 approach which the HKMA will adopt in the supervision of interest rate risk in the banking book (IRRBB) and in monitoring AIs’ level of IRRBB exposures Classification A non-statutory guideline issued by the MA as a guidance note Previous guidelines superseded IR-1 “Interest Rate Risk Management” (V.2) dated 14.12.2018 Circular “BCBS recalibration of shocks for IRRBB” dated 22.7.2024 Circular “Alternative Reference Rates: Interim Reporting Guidelines” dated 30.3.2021 Circular “Interest Rate Risk in the Banking Book: Scope of Application” dated 31.8.2018 Application To all AIs Structure

1. Introduction
   1.1 Terminology
   1.2 Background
   1.3 Scope
   1.4 Implementation
2. Sources of IRRBB
   2.1 Summary
   2.2 Gap risk
   2.3 Basis risk

Supervisory Policy Manual
2.4 Option risk
2.5 Credit spread risk
3. Effects of IRRBB
3.1 Summary
3.2 Earnings perspective
3.3 Economic value perspective
3.4 Embedded losses
4. Supervisory approach to IRRBB
4.1 Objectives and process
4.2 Basel Committee principles
4.3 Factors to be considered
4.4 Monitoring of IRRBB (earnings approach)
4.5 Review of capital adequacy (economic value approach)
4.6 Criteria for adequate internal systems
5. Local standardised framework for measuring IRRBB exposure
5.1 Standardised EVE risk measure
5.2 Slotting cash flows with optionality
5.3 Non-maturity deposits (NMDs)
5.4 Standardised interest rate shock scenarios
6. Oversight by AIs
6.1 Responsibilities of Board and senior management
6.2 Asset and Liability Management Committee
6.3 Independent risk management
7. Risk management policies, procedures and controls
7.1 Coverage
7.2 New services and strategies
7.3 Risk measurement, monitoring and control
7.4 Stress-testing
7.5 Limits
7.6 Internal controls and independent audits

Supervisory Policy Manual
Annex 1: Basel principles for the management of IRRBB
2: Factors influencing behavioural optionality —————————

1. Introduction
   1.1 Terminology
   1.1.1 In this module

• “IRRBB” means the risk to an AI’s financial
  condition resulting from adverse movements in interest rates that affect the AI’s banking book positions; and
• “OBS” means off-balance sheet.
  1.2 Background
  1.2.1 AIs’ normal activities of lending, taking deposits with
  differing maturities and interest rates and buying securities may expose them to IRRBB.
  1.2.2 While accepting some IRRBB is inherent in banking
  business, excessive IRRBB can pose a significant threat to AIs’ earnings and capital adequacy. AIs should therefore have a process to identify, measure, monitor and control IRRBB in a timely and comprehensive fashion.
  1.3 Scope
  1.3.1 The main focus of this module is on the management and
  measurement of IRRBB. It:
• provides guidance on the processes for effective
  IRRBB management;
• aims to help AIs evaluate the adequacy and
  effectiveness of their IRRBB management; and
• sets out how the HKMA monitors and supervises
  AIs’ level and management of IRRBB.
  1.3.2 Overseas incorporated AIs are exempted from the local
  IRRBB framework in cases where the parent group of the AI is not additionally represented in Hong Kong through a locally incorporated AI (referred to hereafter as “exempted

Supervisory Policy Manual
AIs”). The HKMA reserves the discretion to require any of these exempted AIs to implement the local IRRBB framework and to meet the corresponding reporting standards with a reasonable notice period for any reason that may be relevant for HKMA in exercising the key functions.
1.3.3 Exempted AIs are expected to manage their IRRBB
together with their parent groups based on the IRRBB standards of their home jurisdictions and in accordance with the BCBS standards.
1.3.4 For locally incorporated AIs that are exempted from the
market risk capital adequacy regime 1 and overseas incorporated AIs that are not exempted from the local IRRBB framework, this module applies to their positions in both the banking book and the trading book.
1.3.5 This module should be read in conjunction with IC-1 “Risk
Management Framework”. The criteria and sound practices for general risk management contained therein are also applicable to effective IRRBB management.
1.4 Implementation
1.4.1 This revised module will take effect from 1 January 2026.
2. Sources of IRRBB
2.1 Summary
2.1.1 The following subsections describe the primary forms of
IRRBB faced by AIs. They can be divided into three broad categories:

• gap risk;
• basis risk; and
• option risk.
  2.1.2 Gap risk and basis risk, in particular, are the major
  sources of risk underlying the IRRBB exposures of AIs that are active in retail banking activities. 1 Details of the market risk capital adequacy regime and the de-minimis exemption criteria as well as the requirements relevant to AIs which are exempted are set out in the Banking (Capital) Rules.

Supervisory Policy Manual
2.2 Gap risk
2.2.1 Gap risk is the risk arising from changes in the interest
rates on instruments of different maturities. The extent of gap risk depends on whether changes to the term structure of interest rates occur consistently across the yield curve (parallel risk) or differentially by period (non￾parallel risk).
2.2.2 Parallel risk is fundamental to banking business and some
AIs may take on this risk in their balance sheet as part of their strategy to improve earnings. It can, however, affect the income and economic value of an AI as interest rates fluctuate. For example, an AI that has funded a long-term fixed rate loan with a short-term deposit could face a decline in future income if interest rates increase. This is because the cash flows from the loan are fixed while interest payable on replacement funding will be higher after the short-term deposit matures.
2.2.3 Non-parallel risk materialises when unanticipated
changes in the shape of the yield curve have adverse effects on an AI’s income or economic value. As an example, the economic value of an AI’s long position in ten-year government bonds hedged by a short position in five-year government bonds could decline sharply if the yield curve steepens, even if the position is hedged against parallel movements in the yield curve.
2.3 Basis risk
2.3.1 Basis risk arises from imperfect correlation between
changes in the rates earned and paid on different instruments with otherwise similar repricing characteristics. As a result of these differences, the cash flows and earnings spread between assets, liabilities and OBS instruments of similar maturities or repricing frequencies will change.
2.3.2 Basis risk may also arise from the structural differences
between different interest rate benchmarks. For example, an AI may have mortgage loans priced at a different rate to that for its funding, e.g. priced at the prime rate and funded by HIBOR. HIBOR may rise while the prime rate remains unchanged. The AI has the option of increasing its prime rate but in practice its scope to do so may depend on whether other AIs will do the same.

Supervisory Policy Manual
2.3.3 This scenario affects the AI’s current net interest margin
through changes in the spread between earnings and payments on instruments that are being repriced. It will also affect future cash flows from these instruments, which will in turn affect the economic value of the AI.
2.4 Option risk
2.4.1 Option risk arises from interest rate option derivatives or
from optional elements embedded in an AI’s assets, liabilities and/or OBS instruments, where the AI or its customer can alter the level and timing of their cash flows. Option risk can be further characterised into automatic option risk and behavioural option risk.
2.4.2 Automatic option risk arises from standalone instruments,
such as exchange-traded and over-the-counter option contracts, or options explicitly embedded within an otherwise standard financial instrument, where the option will almost certainly be exercised if it is in the holder’s financial interest to do so.
2.4.3 Behavioural option risk arises from the flexibility
embedded implicitly or within the terms of financial contracts, such that changes in interest rates may affect the behaviour of the client. For example, AIs may experience a higher proportion of fixed rate loan commitments to be drawn down when the spread increases, and vice versa.
2.4.4 On the deposit side, customers can generally withdraw
early. Early withdrawal rights are equivalent to put options on deposits. If rates increase, the market value of customer deposits declines and customers may withdraw them and place them with the same AI, or a different one, at a higher rate. Another common product with behavioural optionality is non-maturity deposits (NMDs), which can be withdrawn at any time without notice, but a portion of which tend to remain with the AI in practice (i.e. core deposits).
2.5 Credit spread risk
2.5.1 While the three sources of risks listed above are directly
linked to IRRBB, credit spread risk in the banking book (CSRBB) is a related risk that AIs need to monitor and assess in their interest rate risk management framework.

Supervisory Policy Manual
CSRBB refers to any kind of asset/liability spread risk of credit-risky instruments that is not explained by IRRBB or by the expected credit risk or jump to default risk.
3. Effects of IRRBB
3.1 Summary
3.1.1 As described in section 2 above, changes in interest rates
can have adverse effects both on an AI’s earnings and economic value. Its IRRBB exposure therefore must be assessed from two separate but complementary perspectives, i.e. earnings and economic value.
3.2 Earnings perspective
3.2.1 In this traditional approach to IRRBB assessment, the
analysis focuses on the impact of changes in interest rates on accruing or reported earnings. Reduced earnings or outright losses can threaten the financial stability of an AI by undermining its capital adequacy and by reducing market confidence in it.
3.2.2 The component of earnings that usually receives most
attention is net interest income (NII), i.e. the difference between total interest income and total interest expense, taking account of hedging activity (e.g. via derivatives). This focus reflects both the importance of NII in AIs’ overall earnings and its direct link to changes in interest rates.
3.2.3 Market interest rate changes can also have an impact on
banking activities that generate fee-based and other non￾interest income. Non-interest income arising from many activities such as loan servicing and asset securitisation programmes can be highly sensitive to market interest rates.
3.3 Economic value perspective
3.3.1 Variations in market interest rates can affect the economic
value of an AI’s assets, liabilities and OBS positions. The economic value of an instrument represents an assessment of the present value of its expected net cash flows, discounted to reflect market rates. As fluctuations in interest rates will affect an AI’s earnings, they will also affect its net worth.

Supervisory Policy Manual
3.3.2 The economic value perspective reflects this sensitivity. It
provides a more comprehensive view of the potential long￾term effects of changes in interest rates than is offered by the earnings perspective. In contrast, changes in short￾term earnings, the typical focus of the earnings perspective, may not provide an accurate indication of the impact of interest rate movements on an AI’s overall positions.
3.4 Embedded losses
3.4.1 An AI should also consider the impact that past interest
rates may have on future performance. Instruments that are not marked-to-market may already contain embedded gains or losses due to past rate movements. These gains or losses may be reflected over time in the AI’s earnings. For example, a long-term fixed rate loan entered into when interest rates were low will result in an embedded loss when its funding is subsequently replaced by liabilities bearing higher interest rates over the remaining life of the loan. This embedded loss will be materialised over time until the loan is settled.
4. Supervisory approach to IRRBB
4.1 Objectives and process
4.1.1 The HKMA adopts a risk-based supervisory approach
which enables continuous supervision of AIs’ IRRBB through a combination of on-site examinations, off-site reviews and prudential meetings. The objective is to assess the adequacy and effectiveness of an AI’s IRRBB management process, the level and trend of the AI’s risk exposure and, in the case of a locally incorporated AI, the adequacy of its capital relative to the size of its exposure. See SA-1 “Risk-based Supervisory Approach” for details of the HKMA’s risk-based supervisory methodology.
4.1.2 AIs, unless exempted from the local IRRBB framework,
are required to submit timely and comprehensive information on their IRRBB exposures through the “Return of Interest Rate Risk in the Banking Book – MA(BS)12A” (“IRRBB Return”) on a quarterly basis. The HKMA uses this Return to evaluate AIs’ level of IRRBB based on both the earnings approach and the economic value approach (see subsections 4.4 and 4.5 below for more details). The

Supervisory Policy Manual information collected takes appropriate account of the range of maturities and currencies in each AI’s portfolio, including OBS items, as well as other relevant factors such as basis risk.
4.1.3 Exempted AIs are required to submit timely information on
their interest rate risk exposures through the “Interest Rate Risk Return – MA(BS)12” (“IRR Return”) on a quarterly basis.
4.1.4 Locally incorporated AIs that are exempted from the
market risk capital adequacy regime and overseas incorporated AIs that are not exempted from the local IRRBB framework are required to report in the IRRBB Return the aggregate of their interest rate risk exposures in the trading book and banking book. Where necessary, the HKMA may request individual overseas incorporated AIs that have material trading positions to comply with additional reporting requirements in order to distinguish between their trading and non-trading activities for monitoring purposes.
4.1.5 Locally incorporated AIs that are subject to the market risk
capital adequacy regime are only required to report their interest rate risk exposures in the banking book in the IRRBB Return as their trading positions in interest rate risk are monitored through the “Return of Capital Adequacy Ratio – MA(BS)3” and the “Return of Market Risk and CVA Risk Capital Charge – MA(BS)3A”.
4.1.6 The HKMA will discuss with an AI’s management to
identify the major sources of the AI’s IRRBB exposures and evaluate whether its measurement systems can identify and quantify adequately such risk exposures. The HKMA will also analyse the integrity and effectiveness of the AI’s IRRBB management process to ensure that its practices comply with the objectives and risk tolerance limits approved by the Board of Directors.
4.1.7 In considering whether an AI has appropriate systems for
managing IRRBB, the HKMA will have regard to the nature and complexity of the AI’s IRRBB exposures and its compliance with the standards and sound practices set out in IC-1 “Risk Management Framework” and this module.

Supervisory Policy Manual
4.2 Basel Committee principles
4.2.1 The supervisory approach to IRRBB set out in this module
is based on the principles and practices expounded in the Basel Committee on Banking Supervision (“Basel Committee”) paper of April 2016, “Interest rate risk in the banking book”. The principles and practices have been integrated into the chapters SRP31 and SRP98 of the consolidated Basel Framework. The principles are listed in Annex 1.
4.3 Factors to be considered
4.3.1 In assessing the safety and soundness of an AI’s IRRBB
management and exposures, the HKMA will consider:

• the complexity and level of risk posed by its assets,
  liabilities and OBS activities;
• the adequacy and effectiveness of Board and
  senior management oversight;
• management’s knowledge and ability to identify
  and manage sources of IRRBB;
• the adequacy of internal validation of IRRBB
  measures, including sensitivity analysis and backtesting, in particular where changes in key parameters have occurred;
• the adequacy of internal measurement, monitoring
  and management information systems;
• the adequacy and effectiveness of risk limits on
  and controls over income and capital losses;
• the effectiveness of the AI’s IRRBB stress testing
  programme;
• the adequacy of the AI’s internal review and audit
  of its interest rate risk management process;
• the adequacy and effectiveness of the AI’s risk
  management practices and strategies, as evidenced from past and projected financial performance;
• the effectiveness of hedging strategies used by the
  AI to control IRRBB; and

Supervisory Policy Manual

• the appropriateness of the AI’s level of IRRBB in
  relation to its earnings, capital and risk management systems.
  4.3.2 These topics are discussed further in sections 6 and 7
  below.
  4.4 Monitoring of IRRBB (earnings approach)
  4.4.1 For AIs that are not exempted from the local IRRBB
  framework (“IRRBB-Reporting AIs”), the HKMA reviews the level and trend of these AIs’ IRRBB exposures using the quarterly IRRBB Return. The Return collects information on the following:
• the repricing positions of interest rate-sensitive
  assets, interest rate-sensitive liabilities and interest rate-sensitive OBS positions by different time bands and major currencies (i.e. Hong Kong dollar, US dollar and any other foreign currency that accounts for 5% or more of an AI’s total on-balance sheet interest rate-sensitive position in all currencies);
• a breakdown of interest rate-sensitive assets and
  liabilities into fixed rate, floating rate and managed rate items2 which have different repricing features and reference rates;
• the repricing positions of residential mortgage
  loans and deposits, which are the major components of AIs’ interest rate-sensitive assets and liabilities respectively;
• the weighted average yield and interest costs of
  interest rate-sensitive assets and liabilities, which provide more information for analysing AIs’ net interest income; 2 Fixed rate items are those assets and liabilities with interest rates fixed up to their final maturities. Floating rate items are those which will automatically be repriced at the next repricing date during the life of the items in accordance with movements in the relevant “reference rates” (such as HIBOR) and include those items for which the interest rates can be varied at the discretion of the counterparty. Managed rate items are those variable rate items (e.g. mortgage loans and non-maturity deposits) for which there are no fixed repricing dates and the interest rates can be adjusted at any time at the discretion of the reporting AI.

Supervisory Policy Manual

• a breakdown of the major types of interest rate￾sensitive OBS positions (e.g. interest rate swaps,
  cross currency swaps and options); and
• the impact of interest rate shock scenarios on
  economic value of equity (EVE) and earnings.
  4.4.2 IRRBB-Reporting AIs should follow the Completion
  Instructions for the purpose of reporting IRRBB in the IRRBB Return. The HKMA may review these AIs’ internal processes and assumptions for determining the behavioural maturity of interest rate risk positions in their portfolios.
  4.4.3 Based on the reported interest rate repricing positions in
  the IRRBB Return, the HKMA assesses the impact on an IRRBB-Reporting AI’s earnings over the next 12 months based on two standardised interest rate shock scenarios (parallel up and parallel down) as set out in subsection
  5.4. The HKMA will be particularly attentive to those
  IRRBB-Reporting AIs whose IRRBB exposures lead to a significant decline in earnings having regard to the nature and complexity of their activities.
  4.4.4 As basis risk is a major risk factor underlying AIs’ IRRBB
  exposures, the HKMA assesses the impact of changes in the relationships between key market rates on IRRBB￾Reporting AIs’ earnings using two hypothetical stress scenarios set out in the IRRBB Return. They are:
• all rates except for fixed and managed rates (e.g.
  the prime rate) on interest rate-sensitive assets are subject to the parallel up shock; and
• managed rates on interest rate-sensitive assets are
  subject to the parallel down shock while other rates remain unchanged. The changes are assumed to last for one month, three months, six months and 12 months, respectively. The HKMA will be particularly attentive to those IRRBB￾Reporting AIs whose basis risk leads to a significant decline in earnings having regard to the nature and complexity of their activities.
  4.4.5 Where an AI has significant exposures to gap risk or basis
  risk, the HKMA may review information from the AI’s internal management reports such as maturity/repricing

Supervisory Policy Manual gaps, earnings and economic value simulation estimates and the results of stress tests conducted. The HKMA will also discuss with the AI’s management to evaluate its strategy for managing those exposures and assess its capacity to absorb the risk of loss. Depending on the circumstances of each case, the AI may be asked to strengthen its capital position or reduce its IRRBB exposures (through, for example, hedging or restructuring existing positions) if necessary.
4.5 Review of capital adequacy (economic value approach)
4.5.1 Capital has an important role to play in mitigating and
absorbing the risk of loss from changes in interest rates. As part of sound management, AIs should incorporate the level of IRRBB they undertake into their overall evaluation of capital adequacy. Where AIs undertake significant IRRBB in the course of their business, an appropriate amount of capital should be allocated specifically to support this risk.
4.5.2 The HKMA expects locally incorporated AIs to maintain
adequate capital for IRRBB and to develop their own processes for internal assessment of capital adequacy. Specifically, locally incorporated AIs should consider their capital adequacy for IRRBB as part of the capital adequacy assessment process (CAAP, see CA-G-5 “Supervisory Review Process”), taking into account the following:

• the size and tenor of internal limits on IRRBB, and
  whether they have been reached;
• the effectiveness and expected cost of open
  hedging positions;
• the sensitivity of IRRBB measures to key
  assumptions;
• basis risk;
• the impact on economic value and earnings of
  mismatched positions in different currencies;
• the impact of embedded losses;
• the distribution of capital relative to risks across
  legal entities in the consolidation group;

Supervisory Policy Manual

• the drivers of the underlying risk; and
• the circumstances under which the risk might
  crystallise.
  4.5.3 The HKMA will evaluate whether a locally incorporated AI
  has adequate capital to support its level of IRRBB exposures and the risk those exposures may pose to its future financial performance. To facilitate the monitoring of a locally incorporated AI’s IRRBB and its capital adequacy, the HKMA prescribes six interest rate shock scenarios for the AI to measure its IRRBB exposures as reported in the IRRBB Return3 , and measure the EVE impact of the shocks using the standardised framework (see section 5).
  4.5.4 The HKMA will be particularly attentive to the capital
  sufficiency of “outlier AIs” – those locally incorporated AIs whose IRRBB exposures lead to an EVE decline of more than 15% of their Tier 1 capital as a result of applying one of the six standardised interest rate shocks.
  4.5.5 Where the HKMA is of the view that an AI’s level of IRRBB
  exposures is high in relation to its capital, the HKMA will discuss the concern with the AI’s management. Depending on the circumstances of each case, the AI may be asked to strengthen its capital position or reduce its IRRBB (through, for example, hedging or restructuring existing positions). The AI may also be subject to additional reporting requirements for its IRRBB exposures. However, the AI may not need to take any action if it can demonstrate that (i) its outlier status results from the AI’s specific IRRBB profile that is not adequately captured by the standardised framework; and (ii) the AI’s firm-specific internal model would be able to better reflect its actual IRRBB exposure.
  4.5.6 While overseas incorporated IRRBB-Reporting AIs are
  not subject to the capital adequacy regime in Hong Kong, the HKMA uses the standardised interest rate shocks to monitor their IRRBB in terms of economic value. In view 3 An IRRBB-Reporting AI’s exposures at quarter-ends reported in the IRRBB return should not be systematically lower than the average exposures during the quarter. 4 For exempted AIs, the IRR Return contains a model that applies a 200 bps parallel interest rate shock to the AIs’ interest rate risk exposures and measures such economic value impact. The AIs should follow the Completion Instructions for MA(BS)12 for reporting the interest rate risk exposures.

Supervisory Policy Manual of the limitations of the earnings approach, the economic value approach provides supplementary information about the impact of interest rate movements on an AI’s overall positions (see para. 3.3.2 above).
4.5.7 In monitoring the impact of the standardised interest rate
shock on the economic value of overseas incorporated IRRBB-Reporting AIs, the HKMA will have regard to the Tier 1 capital of their head office. Nevertheless, the 15% benchmark mentioned in para. 4.5.4 above will not apply.
4.6 Criteria for adequate internal systems
4.6.1 The HKMA will assess whether an AI’s internal
measurement system for IRRBB is adequate for managing risk in a safe and sound manner and for evaluation of its capital adequacy5 in the case of a locally incorporated AI.
4.6.2 An AI’s IRRBB management system should meet the
criteria set out in subsection 7.3 below. The system should be integrated into the AI’s daily risk management practices and its output should be used in reporting the level of IRRBB to the Board of Directors and senior management and, where appropriate, individual business line managers. The system should be capable of measuring risk under both the earnings approach and the economic value approach.
4.6.3 The HKMA will require AIs to bring their internal
measurement system up to standard if deficiencies are identified. Until the HKMA is satisfied that an AI’s measurement system is adequate, it may require the AI concerned to increase the frequency of reporting, to supply additional information and to keep its exposures within more prudent limits.
5. Local standardised framework for measuring IRRBB
exposure
This section applies to IRRBB-Reporting AIs only; i.e. it is not relevant for exempted AIs. 5 But it is the AI’s responsibility to ensure that its capital is adequate.

Supervisory Policy Manual
5.1 Standardised EVE risk measure
5.1.1 The calculation of the standardised EVE risk measure
involves the following key steps:

• For a given currency 𝑐 and time band 𝑘, calculate
  the net position 𝐶𝐹0,𝑐
  (𝑘) (and 𝐶𝐹𝑖,𝑐
  (𝑘) under interest rate shock scenario 𝑖 ) by slotting cash flows into time bands based on their earliest interest rate repricing dates, both under current conditions and under each of the six interest rate shock scenarios (see subsection 5.4). Both notional principals and coupon cash flows should be slotted. Note that the net position under interest rate shock scenarios may vary depending on the way cash flows with optionality and non-maturity deposits are slotted (see subsections 5.2 and 5.3).
• For each scenario 𝑖, calculate the impact on EVE
  (∆𝐸𝑖,𝑐
  (𝑘)) for a given currency and time band as ∆𝐸𝑖,𝑐 (𝑘) = 𝐶𝐹0,𝑐 (𝑘) ∙ exp(−𝑟0,𝑐 (𝑘) ∙ 𝑡𝑘) − 𝐶𝐹𝑖,𝑐 (𝑘) ∙ exp(−𝑟𝑖,𝑐 (𝑘) ∙ 𝑡𝑘), where 𝑟0,𝑐 (𝑘) denotes the current risk-free rate6 (at the midpoint of the time band), 𝑟𝑖,𝑐 (𝑘) denotes the risk-free rate under scenario 𝑖 (at the midpoint of the time band) and 𝑡𝑘 denotes the midpoint of each time band 𝑘 (measured in years).
• Calculate the interest rate option risk measure
  𝐾𝐴𝑂𝑖,𝑐 under each scenario 𝑖 as the current net value of interest rate options7 (𝑉𝐴𝑂0,𝑐 ) minus the net value of interest rate options under the interest rate shock scenario (𝑉𝐴𝑂𝑖,𝑐 ), i.e. 𝐾𝐴𝑂𝑖,𝑐 = 𝑉𝐴𝑂0,𝑐 − 𝑉𝐴𝑂𝑖,𝑐 . The net value of interest rate options under the interest rate shock scenario should be calculated using the new yield curve under scenario 𝑖, and assuming a relative increase in the implicit volatility of 25%. 6 This may be determined, for example, based on an interest rate swap curve. 7 This includes both bought and sold options and should be estimated according to the IRRBB-Reporting AI’s proprietary options pricing model.

Supervisory Policy Manual

• Calculate the impact on EVE across all time bands
  for a given currency and scenario as ∆𝐸𝑖,𝑐 = 𝑚𝑎𝑥(0, ∑ ∆𝐸𝑖,𝑐 (𝑘) 𝑘 + 𝐾𝐴𝑂𝑖,𝑐).
• Repeat the above for all applicable currencies. The
  aggregate EVE risk measure ( ∆𝐸 ) across all applicable currencies is calculated as the maximum loss across the six interest rate shock scenarios:
  ∆𝐸 = 𝑚𝑎𝑥
  𝑖∈{1,2,…,6}
  (∑∆𝐸𝑖,𝑐
  𝑐
  )
  5.1.2 Cash flows should generally be slotted into time bands
  according to the earliest interest repricing date. However, there are separate methodologies for slotting cash flows with optionality, including retail fixed rate loans subject to prepayment risk and retail term deposits subject to early redemption risk (see subsection 5.2 for details), as well as for non-maturity deposits (see subsection 5.3 for details).
  5.2 Slotting cash flows with optionality
  5.2.1 Retail fixed rate loans subject to prepayment risk
• Fixed rate loans subject to prepayment risk are
  fixed rate loan products where the full economic cost of prepayments cannot be charged, or charged only for prepayments above a certain threshold, to the borrower.
• This section applies to retail fixed rate loans only.
  Retail loans are defined as loans granted to customers satisfying the criteria in section 64(2) of the Banking (Capital) Rules. Non-retail fixed rate loans subject to prepayment risk should be considered as positions with embedded interest rate options.
• AIs should determine, using own estimates, the
  baseline conditional prepayment rate 𝐶𝑃𝑅0,𝑐,𝑝 for a given portfolio 𝑝 of homogeneous loan products subject to prepayment risk denominated in a given currency 𝑐, under the prevailing term structure of interest rates8 . The conditional prepayment rate 8 The baseline 𝐶𝑃𝑅 may vary over the life of the loan. In this case, it is denoted 𝐶𝑃𝑅0,𝑐,𝑝(𝑘).

Supervisory Policy Manual
(CPR) under interest rate shock scenario 𝑖 is then given by:
𝐶𝑃𝑅𝑖,𝑐,𝑝 = 𝑚𝑖𝑛(1, 𝛾𝑖 ∙ 𝐶𝑃𝑅0,𝑐,𝑝) where 𝐶𝑃𝑅0,𝑐,𝑝 is the (constant) baseline CPR and 𝛾𝑖 is a multiplier for scenario 𝑖 as given in the following table:
Scenario number (𝑖)
Interest rate shock scenarios
𝛾𝑖 (scenario multiplier)
1 Parallel up 0.8
2 Parallel down 1.2
3 Steepener 0.8
4 Flattener 1.2
5 Short rate up 0.8
6 Short rate down 1.2

• The prepayments on the fixed rate loans can be
  broken down into scheduled payments adjusted for prepayment and uncompensated prepayments. The cash flows to be slotted into time band 𝑘 is given by:
  𝐶𝐹𝑖,𝑐,𝑝
  (𝑘) = 𝐶𝐹𝑖,𝑐,𝑆
  (𝑘) + 𝐶𝑃𝑅𝑖,𝑐,𝑝 ∙ 𝑁𝑂𝑖,𝑐,𝑝 (𝑘 − 1) where 𝐶𝐹𝑖,𝑐,𝑆 (𝑘) refers to the scheduled interest payment9 and principal repayment, and 𝑁𝑂𝑖,𝑐,𝑝 (𝑘 −

1. denotes the notional outstanding at the end of
   the previous time band10
   . AIs should repeat the above for each loan portfolio and applicable currency to calculate the aggregate position in each time band. 9 The scheduled interest payment should only be included when reporting net positions including coupon cash flows. 10 For simplicity, it is assumed that there is no annual limit on prepayments. If an AI has imposed an annual limit on uncompensated prepayments, this limit will apply.

Supervisory Policy Manual
5.2.2 Retail term deposits subject to early redemption risk

• Term deposits subject to early redemption risk are
  term deposits that can be withdrawn early at the discretion of the customer. Term deposits may only be treated as a standard fixed rate deposit and be slotted into the time band according to their contractual maturity dates if o the depositor has no legal right to withdraw the deposit; or o an early withdrawal results in a penalty that at least compensates for the loss of interest between the date of withdrawal and the contractual maturity date and the economic cost of breaking the contract11 . If neither of these conditions is met, the depositor holds an option to withdraw and the term deposits are deemed to be subject to early redemption risk.
• This section applies to retail term deposits subject
  to early redemption risk only. Retail deposits are defined as deposits placed with a bank by individual persons. Deposits made by small business customers, as defined in rule 39 of the Banking (Liquidity) Rules12 , can also be treated as retail deposits. Deposits from legal entities, sole proprietorships or partnerships should be categorised as non-retail deposits. Non-retail term deposits subject to early redemption risk should be considered as positions with embedded interest rate options.
• For each homogeneous portfolio 𝑝 of term deposits
  in a given currency 𝑐 , AIs must determine the baseline term deposit redemption ratio 𝑇𝐷𝑅𝑅0,𝑐,𝑝 and use it to slot the notional repricing cash flows. Term deposits which are expected to be redeemed 11 The economic cost of breaking the contract is the additional funding cost the AI would incur due to the early redemption, assuming the AI obtained term funding at the current interest rate, or the interest rate under the relevant interest rate shock scenario, to cover the remaining life of the original term deposit. 12 This definition, for the purpose of this Supervisory Policy Manual, applies to both Category 1 and Category 2 institutions.

Supervisory Policy Manual early should be slotted into the overnight time band (𝑘 = 1).

• The term deposit redemption ratio for time band 𝑘
  under scenario 𝑖 is obtained by multiplying 𝑇𝐷𝑅𝑅0,𝑐,𝑝 by a scalar 𝑢𝑖 as follows:
  𝑇𝐷𝑅𝑅𝑖,𝑐,𝑝 = 𝑚𝑖𝑛(1, 𝑢𝑖 ∙ 𝑇𝐷𝑅𝑅0,𝑐,𝑝) where the values of the scalars 𝑢𝑖 are set out in the following table:
  Scenario number (𝑖)
  Interest rate shock scenarios u𝑖 (scenario multiplier) 1 Parallel up 1.2 2 Parallel down 0.8 3 Steepener 0.8 4 Flattener 1.2 5 Short rate up 1.2 6 Short rate down 0.8
• The notional repricing cash flows which are
  expected to be withdrawn early under interest rate shock scenario 𝑖 are given by:
  𝐶𝐹𝑖,𝑐,𝑝
  (1) = 𝑇𝐷0,𝑐,𝑝 ∙ 𝑇𝐷𝑅𝑅𝑖,𝑐,𝑝 where 𝑇𝐷0,𝑐,𝑝 is the outstanding amount of term deposits in portfolio 𝑝. AIs should repeat the above for each deposit portfolio and applicable currency to calculate the aggregate position in each time band.
  5.3 Non-maturity deposits (NMDs)
  5.3.1 NMD slotting
• IRRBB-Reporting AIs have a choice to either slot
  NMDs into the appropriate time bands according to the earliest date on which their interest rates can be adjusted, or estimate their behavioural maturity using the methodology below.
• NMDs should first be segmented into retail and
  non-retail categories. Retail deposits should be

Supervisory Policy Manual considered as held in a transactional account when regular transactions are carried out in that account (e.g. when salaries are regularly credited) or when the deposit is non-interest bearing. Other retail deposits should be considered as held in a non￾transactional account.

• The AIs estimating behavioural maturities should
  first identify stable NMDs, i.e. those that are found to remain undrawn with a high probability, using observed volume changes over the past 10 years.13 The AIs should then identify core deposits, which are stable NMDs that are unlikely to reprice even under significant changes in the interest rate environment. All other NMDs are non-core deposits. The AIs are required to estimate the level of core deposits using this two-step procedure for each NMD category, subject to the caps in the following table:
  Cap on proportion of core deposits to total NMDs (%) Cap on average maturity of core deposits (years) Retail/transactional 90 5 Retail/non￾transactional 70 4.5 Non-retail 50 4
• NMDs should finally be slotted into the appropriate
  time bands. Non-core deposits should be considered as overnight deposits and accordingly should be placed into the overnight time band. Core deposits should be slotted according to their average behavioural maturity, which should be determined by the AIs using an appropriate procedure subject to the caps in the table above. 13 An AI should notify its supervisors if the relevant data available to it spans a period of less than 10 years but at least 7 years. In such cases, the AI should use all existing data to model NMDs. In cases where the available data spans less than 7 years, an AI should contact its supervisors for approval if it intends to use, and before using, such data to identify core deposits. In particular, the AI should suggest additional prudential cushions for the NMD model estimates to its supervisors. In either case, the AI should continuously extend its relevant database to cover at least 10 years.

Supervisory Policy Manual
5.4 Standardised interest rate shock scenarios
5.4.1 The change in the risk-free interest rate ∆𝑟𝑖,𝑐
(𝑘) for interest rate shock scenario 𝑖, currency 𝑐 and time band 𝑘 is calculated according to the equations below (see table below for values of 𝑅̅):
(i) parallel shock up:
∆𝑟1,𝑐
(𝑘) = 𝑅̅
𝑝𝑎𝑟𝑎𝑙𝑙𝑒𝑙,𝑐
(ii) parallel shock down:
∆𝑟2,𝑐
(𝑘) = −𝑅̅
𝑝𝑎𝑟𝑎𝑙𝑙𝑒𝑙,𝑐
(iii) steepener shock:
∆𝑟3,𝑐
(𝑘) =
−0.65 ∙ 𝑅̅
𝑠ℎ𝑜𝑟𝑡,𝑐
∙ 𝑒
−𝑡𝑘
4 + 0.9 ∙ 𝑅̅
𝑙𝑜𝑛𝑔,𝑐
∙ (1 − 𝑒
−𝑡𝑘
4 )
(iv) flattener shock:
∆𝑟4,𝑐
(𝑘) =
0.8 ∙ 𝑅̅
𝑠ℎ𝑜𝑟𝑡,𝑐
∙ 𝑒
−𝑡𝑘
4 − 0.6 ∙ 𝑅̅
𝑙𝑜𝑛𝑔,𝑐
∙ (1 − 𝑒
−𝑡𝑘
4 )
(v) short rates shock up:
∆𝑟5,𝑐
(𝑘) = 𝑅̅
𝑠ℎ𝑜𝑟𝑡,𝑐
∙ 𝑒
−𝑡𝑘
(vi) short rates shock down:
∆𝑟6,𝑐
(𝑘) = −𝑅̅
𝑠ℎ𝑜𝑟𝑡,𝑐
∙ 𝑒
−𝑡𝑘
5.4.2 The final post-shock interest rate, subject to a −2% floor,
is given by14
𝑟𝑖,𝑐
(𝑘) = 𝑚𝑎𝑥(𝑟0,𝑐
(𝑘) + ∆𝑟𝑖,𝑐
(𝑘), −2%)
14 For AIs taking spreads into account, the final post-shock interest rate is given by 𝑟𝑖,𝑐 𝑖𝑛𝑐𝑙.𝑠𝑝𝑟𝑒𝑎𝑑,𝑎𝑠𝑠𝑒𝑡𝑠(𝑘) = (𝑟𝑖,𝑐 (𝑘) + 𝑎𝑠𝑠𝑒𝑡 𝑠𝑝𝑟𝑒𝑎𝑑) for assets and 𝑟𝑖,𝑐 𝑖𝑛𝑐𝑙.𝑠𝑝𝑟𝑒𝑎𝑑,𝑙𝑖𝑎𝑏𝑖𝑙𝑖𝑡𝑖𝑒𝑠(𝑘) = 𝑚𝑎𝑥(𝑟𝑖,𝑐 (𝑘) − 𝑙𝑖𝑎𝑏𝑖𝑙𝑖𝑡𝑦 𝑠𝑝𝑟𝑒𝑎𝑑, −2%) for liabilities, where 𝑟𝑖,𝑐 (𝑘) is defined as in para. 5.4.2.

Supervisory Policy Manual
Specified size of interest rate shocks (𝑅̅ 𝑠ℎ𝑜𝑐𝑘𝑡𝑦𝑝𝑒,𝑐, in bps) ARS AUD BRL CAD CHF CNH CNY EUR GBP HKD IDR Parallel 400 350 400 200 175 225 275 225 400 Short 500 425 500 275 250 300 300 350 425 375 500 Long 300 175 200 150 150 200 250 200 300 INR JPY KRW MXN RUB SAR SEK SGD TRY USD ZAR Parallel 325 100 225 400 400 275 275 175 400 200 325 Short 475 100 350 500 500 375 425 250 500 300 500 Long 225 100 225 200 300 250 200 225 300 225 300
5.4.3 For MOP, the sizes of interest rate shocks are the same
as those for HKD. For all other currencies not explicitly mentioned, the applicable sizes of interest rate shocks to the risk-free rate for parallel, short and long are 400 bps, 500 bps, and 300 bps respectively.
5.4.4 The table above will be updated from time to time based
on parameter updates provided by the Basel Committee.
6. Oversight by AIs
6.1 Responsibilities of Board and senior management
6.1.1 An AI’s Board of Directors and senior management are
responsible for oversight of the IRRBB management framework and the AI’s risk appetite for IRRBB, which should be articulated in terms of the risk to both economic value and earnings. AIs must implement policy limits that are consistent with their risk appetite. See CG-1 “Corporate Governance of Locally Incorporated Authorized Institutions” and IC-1 “Risk Management Framework” for details of their risk management responsibilities. Many of the requirements and practices cited have a general application.
6.2 Asset and Liability Management Committee
6.2.1 The Board of Directors may delegate responsibility for
monitoring and management of IRRBB to the Asset and Liability Committee (“ALCO”), which is a designated committee usually composed of senior staff. Larger or more complex AIs should have such committees, responsible for the design and administration of IRRBB

Supervisory Policy Manual management. The ALCO should include members with clear lines of authority over the units responsible for establishing and managing positions. The Board should ensure that the AI’s organisational structure enables the ALCO to carry out its responsibilities.
6.2.2 The main role and functions of the ALCO are described in
CG-1 “Corporate Governance of Locally Incorporated Authorized Institutions”.
6.3 Independent risk management
6.3.1 The Board or senior management should assign
responsibility for managing IRRBB to individuals or units with appropriate experience and expertise. The responsible personnel should have an adequate understanding of all types of IRRBB faced throughout the AI.
6.3.2 There should be adequate segregation of duties in key
elements of the risk management process to avoid potential conflicts of interest. For example, the level of IRRBB is determined by how a particular transaction is evaluated based on current market rates. Such evaluation is normally conducted by the risk management or operations department of an AI while the actual transaction is performed by a risk-taking unit or front office. This is to ensure independent risk assessment of the transactions.
7. Risk management policies, procedures and controls
7.1 Coverage
7.1.1 Whatever the methodology chosen, an AI’s IRRBB
management procedures should be clearly defined and consistent with the nature and complexity of its activities.
7.1.2 The policies, procedures and limits (e.g. limits to fixed rate
deals, use of interest rate swaps, etc.) should be properly documented, drawn up after careful consideration of IRRBB associated with different types of lending, and approved and reviewed (at least annually) by management at the appropriate level. The policies and procedures should delineate delegated powers, lines of responsibility and accountability over IRRBB management decisions and should clearly define

Supervisory Policy Manual authorised instruments, hedging strategies and risk-taking opportunities.
7.1.3 There should also be an accurate, informative and timely
management information system for IRRBB. This is essential both to keep senior management and, where appropriate, individual business line managers in the picture and to facilitate compliance with Board policy.
7.1.4 AIs’ policies and procedures for IRRBB management
should cover the general criteria set out in IC-1 “Risk Management Framework” and other criteria specific to IRRBB as discussed in the following subsections.
7.2 New services and strategies
7.2.1 AIs should identify the interest rate risks inherent in new
services and activities and ensure that these are subject to adequate procedures and controls before being introduced or undertaken. For example, an AI specialising in prime-based mortgage loans that then engages in HIBOR-based mortgage loans with interest rate caps for customers should be aware of the volatility of HIBOR and the embedded option features.
7.2.2 AIs may be exposed to additional interest rate risk if they
develop products or services that enable greater access to customers who primarily seek the best rate. The introduction of e-banking services is an example of such services. This reinforces the need for AIs to react quickly to changing market conditions and to ensure that their pricing strategy has catered for an adequate interest spread to absorb any additional interest rate risk.
7.2.3 AIs should consider balancing cash flows and managing
the interest rate risk arising from new services or strategies through hedging, e.g. using swaps or other derivative instruments. Major hedging or risk management initiatives should be approved in advance by the Board or a committee such as the ALCO.
7.3 Risk measurement, monitoring and control
7.3.1 AIs should have IRRBB measurement systems that
encompass all significant causes of such risk. The systems should evaluate the effect of rate changes on both earnings and economic value meaningfully and accurately within the context and complexity of their

Supervisory Policy Manual activities. They should be able to flag any excessive exposures without delay.
7.3.2 Measurement systems should:

• evaluate all significant IRRBB arising from the full
  range of an AI’s assets, liabilities and OBS positions. If the same measurement systems and management methods are not used for all activities, an integrated view of IRRBB across products and business lines should be available to management;
• employ a variety of generally accepted financial
  models and ways of measuring risk, rather than relying on a single measure of risk;
• have accurate and timely data (in relation to rates,
  maturities, repricing, embedded options and other details) on current positions15 . Data inputs should be automated as much as possible to reduce administrative errors;
• document the assumptions, parameters and
  limitations on which they are based. Material changes to assumptions should be documented, justified and approved by senior management. Systems should also be sufficiently flexible to incorporate supervisory-imposed constraints on AIs’ internal risk parameter estimates;
• cover all significant sources of IRRBB (e.g. gap,
  basis and option risks). While all of an AI’s positions should be appropriately treated, its largest concentrations and positions should be assessed with special thoroughness, as should instruments which might have a material effect on an AI’s overall position (notwithstanding that they are not major concentrations) and instruments with significant embedded or explicit options; and
• assess exposures in different currencies (subject to
  para. 7.3.6 below).
  15 Any manual adjustments to underlying data should be clearly documented and the nature and reasons for the adjustments should be clearly understood.

Supervisory Policy Manual
7.3.3 Techniques to measure IRRBB exposure from an
earnings and economic value perspective comprise, in increasing degrees of complexity, simple calculations, static simulations using current holdings and highly sophisticated dynamic modelling techniques based on business forecasts and decisions. As a minimum, AIs should be able to use the standardised framework for measuring IRRBB exposure (see section 5). Where cash flows are slotted into different time bands (e.g. for gap analyses), the slotting criteria should be stable over time to allow for a meaningful comparison of risk figures over different periods.
7.3.4 AIs having complex risk profiles should employ more
sophisticated IRRBB measurement techniques such as simulation-based approaches. The assumptions underlying a simulation model can sometimes make it difficult to determine how much a variable contributes to changes in the simulation results. It is therefore necessary to supplement the simulation model by additional in-depth analysis or other simulation models to isolate the risk of each variable inherent in the existing balance sheet.
7.3.5 When assessing its IRRBB exposures, an AI should make
judgements and assumptions about how an instrument’s actual maturity or repricing behaviour may vary from the instrument’s contractual terms because of behavioural optionalities. The behavioural assumptions used should be conceptually sound and reasonable, and consistent with historical experience (see Annex 2 for a list of possible considerations). Such assumptions should be rigorously tested and aligned with the AI’s business strategies. The most significant assumptions should be documented, clearly understood by the Board or the relevant committee and subject to periodic review (at least annually).
7.3.6 AIs with positions in different currencies need to measure
their exposure to IRRBB in each currency. They may do so for each currency separately, on the grounds that yield curves for different currencies vary. AIs with material multi-currency exposures may, if they have the requisite skills and sophistication, decide to aggregate their exposures in certain currencies where there is assumed to be some correlation between interest rates for those

Supervisory Policy Manual currencies. Such AIs should review periodically whether these assumptions remain valid and assess their potential exposure if such correlations prove invalid.
7.3.7 Measurement outcomes of IRRBB and hedging strategies
should be reported to the Board or the relevant committee on a regular basis (at least semiannually), at relevant levels of aggregation (by consolidation level and currency). The reports should include at least the following:

• summaries of the AI’s aggregate IRRBB
  exposures, and explanatory text that highlights the assets, liabilities, cash flows, and strategies that are driving the level and direction of IRRBB;
• reports demonstrating the AI’s compliance with
  policies and limits;
• key assumptions such as NMD characteristics and
  prepayments on fixed rate loans;
• results of stress tests, including assessment of
  sensitivity to key assumptions and parameters, as well as sensitivity to changes in market conditions, with particular reference to portfolios that may be subject to significant mark-to-market movements; and
• summaries of the reviews of IRRBB policies,
  procedures and adequacy of the measurement systems, including any findings of internal and external auditors and/or other equivalent external parties (such as consultants).
  7.4 Stress-testing
  7.4.1 AIs should measure their vulnerability to loss in stressed
  market conditions, including the breakdown of key assumptions, and consider those results when establishing and reviewing their policies and limits for IRRBB, as well as in the CAAP.
  7.4.2 An AI’s stress testing framework for IRRBB should be
  commensurate with its nature, size and complexity as well as business activities and overall risk profile. The framework should have clearly defined objectives, well￾documented assumptions and sound methodologies, and

Supervisory Policy Manual should take into account the opinions of experts within the AI.
7.4.3 AIs’ IRRBB management systems should be able to
calculate, by currency, the impact on economic value and earnings of multiple scenarios, including the six standardised interest rate shock scenarios 16 set out in subsection 5.4, as well as internally selected interest rate shock scenarios addressing the AI’s risk profile according to its CAAP. Possible stress scenarios include:

• historical scenarios such as the Asian Financial
  Crisis in the late 1990s;
• changes in the relationships between key market
  rates (i.e. basis risk), e.g. (i) a surge in term and savings deposit rates and HIBOR but no change in the prime rate, and (ii) a drop in the prime rate but no change in term and savings deposit rates and HIBOR17;
• changes in interest rates in individual time bands to
  different relative levels (i.e. non-parallel gap risk);
• changes in the liquidity of key financial markets or
  changes in the volatility of market rates; and
• changes in key business assumptions and
  parameters such as the correlation between Hong Kong dollar and US dollar interest rates. In particular, changes in assumptions used for illiquid instruments and instruments with uncertain contractual maturities help understanding of an AI’s risk profile.
  7.4.4 When developing interest rate shock scenarios, AIs
  should consider the following:
• the scenarios should be severe and plausible, and
  sufficiently wide-ranging to identify parallel and non-parallel gap, basis and option risks;
• special consideration should be given to
  instruments or markets where concentrations exist; 16 AIs are only required to calculate the impact of two standardised interest shock scenarios (parallel up and parallel down) on earnings (see para. 4.4.3). 17 These scenarios for basis risk are incorporated in the IRRBB Return.

Supervisory Policy Manual

• AIs should assess the possible interaction of
  IRRBB with other risks, e.g. credit risk and liquidity risk;
• AIs should assess the effect of adverse changes in
  the spreads of new assets or liabilities replacing those positions maturing over the horizon of the forecast on their NII;
• AIs with significant option risk should include
  scenarios that capture the exercise of such options. For example, AIs that have products with sold caps or floors should include scenarios that assess how the risk positions would change should those caps or floors move into the money. AIs should also develop interest rate assumptions to measure their IRRBB exposures given changes in interest rate volatilities;
• AIs should specify the term structure of interest
  rates that will be incorporated and the basis relationship between yield curves when building interest rate shock scenarios;
• AIs should estimate how interest rates that are
  administered or managed by management might change;
• AIs should consider the time it would take to reduce
  or unwind unfavourable IRRBB exposures, and their capability or willingness to withstand accounting losses in order to reposition their risk profile;
• Forward-looking scenarios should incorporate
  changes in portfolio composition, new products, new market information and emerging risks; and
• AIs should perform qualitative and quantitative
  reverse stress tests to identify key vulnerabilities.
  7.5 Limits
  7.5.1 AIs should establish and enforce operating limits and
  other practices that maintain exposures within levels consistent with their internal policies and that accord with their approach to measuring IRRBB. Such limits should be approved by the Board or a committee such as the ALCO.

Supervisory Policy Manual
7.5.2 In particular, AIs should set a limit on the extent to which
floating rate exposures are funded by fixed rate sources and vice versa to limit IRRBB. In floating rate lending, AIs should limit the extent to which they run any basis risk that may arise if lending and funding are not based on precisely the same market interest rate (e.g. HIBOR).
7.5.3 The limits should be consistent with AIs’ underlying
approach to IRRBB measurement and should be directed at how reported earnings and capital adequacy might be affected by changes in market interest rates. As regards earnings, AIs should consider limits on earnings volatility in both net income and net interest income under specified interest rate scenarios so as to quantify what portion of their IRRBB exposure arises from non-interest income.
7.5.4 The limits should be appropriate to the nature, size,
complexity and capital adequacy of the AI, as well as its ability to measure and manage its risks. Depending on the nature of an AI’s activities and business model, sub-limits may also be identified for individual business units, portfolios, instrument types or specific instruments. AIs with significant exposures to gap risk, basis risk or positions with explicit or embedded options should establish appropriate risk tolerances for these risks.
7.5.5 Limits on IRRBB should be related to explicit scenarios of
changes in market interest rates and/or term structures, e.g. movements up or down of specified ranges or a change in shape. These scenarios or changes should constitute genuine stress conditions and should be developed in the light of historic rate volatility and time needed to unwind, restructure or hedge an AI’s IRRBB position. They can also reflect measures from the underlying statistical distribution of interest rates, e.g. earnings at risk or economic value at risk techniques. The scenarios should cover all possible sources of IRRBB, e.g. gap, basis and option risks, and not just parallel shifts in interest rates or other simple scenarios.
7.5.6 There should be systems in place to ensure that positions
that exceed, or are likely to exceed, limits established by the AI should receive prompt management attention and be escalated without delay. There should be a clear policy on who will be informed, how the communication will take

Supervisory Policy Manual place and the actions which will be taken in response to an exception.
7.6 Internal controls and independent audits
7.6.1 AIs should have adequate internal controls over IRRBB as
an integral part of the overall internal control system. The effectiveness of such controls should be evaluated regularly by independent parties, e.g. internal or external auditors.
7.6.2 AIs should conduct periodic reviews of their risk
management process for IRRBB to ensure its integrity, accuracy and reasonableness. AIs with more complex profiles and measurement systems should have their internal models or calculations audited or validated by an independent internal or external reviewer. Reports written by independent reviewers should be made available to the HKMA.
7.6.3 In such independent reviews, the factors to be considered
include the quality of IRRBB management and the size of IRRBB, e.g.:

• the volume and price sensitivity of various
  products;
• how vulnerable earnings and economic value are
  to all forms of IRRBB, e.g. gap, basis and option risks;
• compliance with established policies and
  procedures and escalation procedures for any exceeded limits; and
• any significant changes that may affect the
  effectiveness of controls (including changes in market conditions, personnel, technology and structures of compliance with exposure limits).

Supervisory Policy Manual
Annex 1: Basel principles for the management of IRRBB
1.1 Background
1.1.1 The Basel Committee issued the paper “Principles for the
management and supervision of interest rate risk” (“IRRBB Principles”) in July 2004, setting out supervisory expectations for banks’ identification, measurement, monitoring and control of IRRBB as well as its supervision. In April 2016, the Basel Committee published standards for “Interest rate risk in the banking book” with revised IRRBB Principles for banks as summarised below. The standards have been integrated into the consolidated Basel Framework.
1.2 Principles for banks
1.2.1 IRRBB is an important risk for all banks that must be
specifically identified, measured, monitored and controlled. In addition, banks should monitor and assess credit spread risk in the banking book.
1.2.2 The governing body of each bank is responsible for
oversight of the IRRBB management framework, and the bank’s risk appetite for IRRBB. Monitoring and management of IRRBB may be delegated by the governing body to senior management, expert individuals or an asset and liability management committee (henceforth, its delegates). Banks must have an adequate IRRBB management framework, involving regular independent reviews and evaluations of the effectiveness of the system.
1.2.3 The banks’ risk appetite for IRRBB should be articulated
in terms of the risk to both economic value and earnings. Banks must implement policy limits that target maintaining IRRBB exposures consistent with their risk appetite.
1.2.4 Measurement of IRRBB should be based on outcomes of
both economic value and earnings-based measures, arising from a wide and appropriate range of interest rate shock and stress scenarios.
1.2.5 In measuring IRRBB, key behavioural assumptions
should be fully understood, conceptually sound and documented. Such assumptions should be rigorously tested and aligned with the bank’s business strategies.

Supervisory Policy Manual
1.2.6 Measurement systems and models used for IRRBB
should be based on accurate data, and subject to appropriate documentation, testing and controls to give assurance on the accuracy of calculations. Models used to measure IRRBB should be comprehensive and covered by governance processes for model risk management, including a validation function that is independent of the development process.
1.2.7 Measurement outcomes of IRRBB and hedging strategies
should be reported to the governing body or its delegates on a regular basis, at relevant levels of aggregation (by consolidation level and currency).
1.2.8 Information on the level of IRRBB exposure and practices
for measuring and controlling IRRBB must be disclosed to the public on a regular basis.
1.2.9 Capital adequacy for IRRBB must be specifically
considered as part of the Internal Capital Adequacy Assessment Process (ICAAP) approved by the governing body, in line with the bank’s risk appetite for IRRBB.

Supervisory Policy Manual
Annex 2: Factors influencing behavioural optionality
Fixed rate loans subject to prepayment risk Loan size, loan-to-value (LTV) ratio, borrower characteristics, contractual interest rates, seasoning, geographical location, original and remaining maturity, and other historical factors. Other macroeconomic variables such as stock indices, unemployment rates, GDP, inflation and housing price indices. Fixed rate loan commitments Borrower characteristics, geographical location (including competitive environment and local premium conventions), customers’ relationship with the AI as evidenced by cross￾products, remaining maturity of the commitment, seasoning and remaining term of the mortgage. Term deposits subject to early redemption risk Deposit size, depositor characteristics, funding channel (e.g. direct or brokered deposit), contractual interest rates, seasonal factors, geographical location and competitive environment, remaining maturity and other historical factors. Other macroeconomic variables such as stock indices, unemployment rates, GDP, inflation and housing price indices. NMDs Responsiveness of product rates to changes in market interest rates, current level of interest rates, spread between an AI’s offer rate and market rate, competition from other firms, the AI’s geographical location and demographic and other relevant characteristics of its customer base. ————————— Contents Glossary Home Introduction